key: cord-023216-avn8f2w3
authors: nan
title: Symposium summaries
date: 2004-10-18
journal: Pediatr Pulmonol
DOI: 10.1002/ppul.20142
sha: 
doc_id: 23216
cord_uid: avn8f2w3

nan

The creation of mouse models of CF has provided new opportunities to elucidate disease pathogenesis, correlate genotype with phenotype, and evaluate the safety and efficacy of novel CF therapies. Epithelial tissues from CF mice exhibit the loss of cAMP-stimulated Clsecretion characteristic of the human disease, and CF mouse models display a range of abnormal intestinal and pulmonary phenotypes dependent on the CFTR genotype, independently segregating modifier genes, and environmental factors (diet, pathogens) (1) .

Studies in CF mice may also be helpful in answering the question: How much CFTR is enough? Quantifying the level of gene and protein expression necessary for restoring lung and intestinal function is of paramount importance for the design of somatic gene therapy or protein therapy protocols.

One strategy is to reduce the expression levels or activity of Cftr in wild type mice by genetic or pharmacological approaches (e.g. insertional mutagenesis; siRNA; specific CFTR inhibitors) and determine the threshold level at which the disease symptoms become manifest. A typical example is the creation of the exon 10 insertional CF mouse ( Cftr tm1Hgu ) . Unlike the CF mice generated by gene replacement (knockout null mice), the original mutant mouse with a mixed background showed residual low levels of normal Cftr mRNA (<10% in the lungs) as a result of exon skipping and aberrant splicing (2) , associated with a 30-50% residual intestinal and nasal cAMP-mediated chloride secretion and only mild intestinal obstruction (95% preand post-weaning survival on solid diet); outbred MF1/129 Cftr tm1Hgu mice raised outside the isolator displayed signs of inflammation, reduced mucociliary clearance , and impaired airway clearance of aerosolized, CF associated bacteria (1) , suggesting that a low residual expression of normal Cftr may ameliorate intestinal obstruction but not CF lung pathophysiology. Recently, the phenotypic evaluation of two Cftr tm1Hgu inbred mutant strains (named CF/1-Cftr TgH(neoim)Hgu and CF/3-Cftr TgH(neoim)Hgu ) generated in Hannover showed similarly low Cftr transcript level in the intestine, a low residual amount (10-20%) of mature Cftr protein in the apical membrane of the enterocytes, but close to normal (60-70%) chloride secretory responses to cAMP in the Ussing chamber (3) . These findings suggest that ~20% of wild-type Cftr protein is sufficient to normalize transepithelial intestinal chloride secretion, survival and weight gain.

In contrast to intestine, the level of Cftr mRNA, immunostained Cftr protein, and cAMP-stimulated chloride secretion in the nasal epithelum of these mice was almost normal, but the amiloride-sensitive sodium absorption, previously found to be 8 fold increased in Cftr-null mice (4), remained ~2-fold above normal values. This finding suggests that a complete correction of the hyperabsorption of sodium in the airways, a known hallmark of CF, is difficult to reach even at virtually normal Cftr protein levels. As yet the consequence of this persistent abnormality in sodium transport for lung pathophysiology remains to be explored.

An alternative strategy is to express human CFTR (hCFTR) in Cftr-/-mice and to examine rescue of the gut defect by the transgene. For example, a transgenic mouse expressing hCFTR cDNA under control of the intestinal fatty-acid binding protein (FABP) promoter showed normal survival and functional correction of ileal goblet cell and crypt hyperplasia, and a ~25% rescue of transepithelial chloride current (5) ; assuming that 100% chloride secretion corresponds with ~20% mature Cftr (cf. CF/1 mice), it follows that ~5% Cftr protein is sufficient to eliminate intestinal pathology. The notion that only partial correction of CFTR channel activity may have a significant clinical impact was confirmed by replicating the G480C trafficking mutant in the murine Cftr gene (6): 8% of mature Cftr protein corresponded with ~40% residual cAMP-stimulated jejunal chloride secretion and the complete absence of intestinal pathology.

Insight into the question: How much CFTR is enough, is also relevant for attempts to ameriolate CF disease pharmacologically, e.g. by promoting the correct processing and trafficking of F508del CFTR by small molecules emerging from high-throughput screens, by altering chaperone function (7) or by suppressing premature stop mutations (8) .

One approach is to treat CF mutant mice in vivo, or isolated tissues ex vivo with the rescue compounds/

approaches and compare the gain in Cftr protein expression with the extent of functional rescue and the amelioration of disease symptoms. For example, ex vivo incubation of ileal mucosa from congenic FVB Cftr tm1Eur mice homozygous for the F508del mutation at low temperature (12h, 26 0 C) resulted in a gain in mature Cftr protein (band C on a western blot) from a starting value of ~4% to a maximal level of 25% of the amount in Cftr +/+ ileum, concomittant with the full restoration of cAMP/genistein-stimulated transepithelial chloride current (9) . Assuming that the potentiator genistein fully corrects the ~3-fold lower open probability of F508del CFTR channels, this outcome confirms that ~20% of mutant Cftr protein, in the presence of a potentiator, is sufficient to fully restore intestinal chloride transport. Complete functional rescue could also be reached by exposing the ileal mucosa for 6 h at 37 0 C to the proteasome/calpain inhibitor ALLN, likewise resulting in a gain in steady state levels of apical Cftr protein up tõ 20% of normal levels (9) .

Importantly, even untreated Rotterdam FVB Cftr tm1Eur mice maintained on solid food (Hope farms SRM-A) did not show signs of intestinal or lung pathology despite a low level of mature F508del Cftr protein (~4%) in the intestine and a reduced level in the nasal epithelium (~30%). However a shift to a different solid diet (RM3, Special diet Services) caused 100% mortality by intestinal obstruction after 3-4 days, whereas changing the background from FVB to C57BL/6 provoked clear symptoms of lung pathology, without altering the nasal bioelectrics. These findings not only define optimal conditions for testing the efficacy of F508del Cftr rescue approaches in these mice, but stress the importance of diet composition and genetic modifiers in CF pathology irrespective of the level of transepithelial chloride and sodium transport.

what is termed a "genotype-driven" design. The second approach is to select patients with a similar level of disease severity and then determine their genotypes in a "phenotype-driven" study. Both approaches have been utilized extensively in CF research, and each approach has yielded a number of important insights:

1) A small amount of wild-type CFTR RNA transcript ameliorates disease severity. "Leaky" splice-site mutations that permit production of some normally-spliced CFTR transcript can moderate disease severity in the pancreas, lungs and vas deferens. Correlating estimates of the amount of wild-type transcript associated with "leaky" splice-site mutations demonstrates that approximately 3-4% wild-type transcript is associated with a less severe pancreatic disease, and approximately 8-10% allows individual to escape from CF lung disease (1;2).

2) Partially functional CFTR can reduce disease severity. Phenotype-driven studies searching for mutations occurring in patients with mild forms of CF revealed a disproportionate number of amino acid substitutions (3;4) . Analysis of CFTR bearing these substitutions in cell-based systems has revealed that many retain chloride channel properties of the wild-type protein (5) . In general, the degree of residual chloride channel function of CFTR correlates with the severity of the CF phenotype, in particular pancreatic status and degree of sweat chloride abnormality (6;7). Correlation of chloride channel function and the severity of lung disease is less clear since there is a paucity of mutations that are convincingly associated with an improved pulmonary outcome. One exception is the A455E mutation that confers a milder lung phenotype (8) , CFTR bearing this mutation has altered processing and chloride channel properties (9) .

3) Increasing the amount of a partially functioning protein can avoid the life-limiting pulmonary complications of cystic fibrosis. The best example of this phenomenon is the R117H mutation. This mutation occurred at least twice in human evolution in different CFTR genes, one bearing an efficient splice site (7T) in intron 8, and the other bearing an inefficient splice site (5T) in intron 8 (10) . CFTR bearing R117H is properly folded and targeted to apical membranes however conduction of chloride via the CFTR channel is reduced from wild-type CFTR (5) . Genes with R117H and 5T have reduced levels of full-length transcript leading to reduced production of partially functional CFTR bearing R117H. Individuals bearing R117H-5T and a severe CF mutation present with cystic fibrosis manifesting as progressive lung disease, pancreatic sufficiency, absence of the vas deferens in males and elevated sweat chloride concentration. On the other hand, R117H with the efficient 7T variant results in normal levels of full-length transcript bearing the R117H mutation leading to normal levels of CFTR protein bearing this mutation. Males with R117H-7T and a severe CF mutation generally have congenital absence of the vas deferens, a form of male infertility seen in CF males, and females generally have no clinical phenotype. Importantly, lung disease is usually absent in these individuals. Thus, increasing the amount of a partially functional CFTR mutant can alleviate the lifelimiting complication of CF. 4) Tissues affected in CF require different levels of CFTR function to operate normally. An early revelation of genotype/phenotype studies was that the presence of one or two "mild" mutations was associated with a milder form of pancreatic disease but not milder lung disease or sweat gland abnormalities (11) . From a large number of studies, it appears that the sinuses are most sensitive to reduction in CFTR function, followed by the vas deferens, lungs, sweat gland, and pancreas. Thus, the level of correction needed to reverse CF pathology will depend upon the target organ.

Studies of genotype/phenotype conducted in CF strongly suggest that correction of most features of the disease will occur when 10% of normal CFTR function is achieved. However, individuals bearing genotypes that are predicted to generate this level of CFTR function have minimal differences in biochemical measurements such as sweat chloride levels and nasal potential difference measurements. Improvement in measurable CFTR function in vivo may require 25% function based on what is observed in CBAVD males. Males who have mild mutations in one CFTR gene and a severe mutation in the other CFTR gene can present with normal or near normal lung function, borderline to elevated sweat chloride levels and nasal potential difference measurements that are abnormal, but can be distinguished from those who have life limiting forms of CF (12) or are normal (13) .

The CF respiratory disease phenotype includes thick mucus secretion and colonization of the lung with bacteria, such as Pseudomonas aeruginosa (PsA). Presumably, repeated bacterial infections compounded by an inability to clear the infecting organism results in a profound neutrophil migration into the airways and the secretion of a variety of inflammatory mediators, including neutrophil elastase, reactive oxygen species, and cytokines, such as the chemokine IL-8. The existence of an underlying defect in the airway immune response of CF patients remains controversial; however, increasing evidence suggests that the CF lung exhibits an exaggerated immune response even in the absence of bacterial infection (1;2). Moreover, increasing evidence suggests that the cystic fibrosis transmembrane conductance regulator (CFTR) may directly or indirectly regulate CFassociated airway inflammation and bacterial clearance.

In the CF community, the advance of gene therapy has precipitated the question of how much CFTR expression and function is required for the correction of CFTR-related functions, including the resolution of inflammation and bacterial clearance within the CF lung. Studies examining the effect of CFTR on these responses have been performed utilizing both in vitro and in vivo model systems. Specifically, in vitro studies have focused upon the effect of CFTR expression and / or function in airway epithelial cell models on inflammatory mediator expression, NF-B activation, and bacterial adherence. In these studies, a variety of non-CF, CF, and CFTR corrected-CF primary and immortalized airway epithelial cell lines have been utilized; however, little consensus has been reached on the role that CFTR plays in regulating inflammatory and bacterialrelated responses. For example, several reports suggest that mutant CFTR expression and / or function correlates directly with enhanced IL-8 and IL-6 (3;4) protein secretion, reduced RANTES (5), IL-10 (6), and iNOS (7) production, exaggerated NF-B activation (8) , and altered bacterial adherence (9;10). In contrast, additional studies indicate that little or no correlation exists between mutant CFTR expression and IL-8 (5;11-13) or IL- 6 (11;14) and IL-6 (14) production and NF-B activation (14;15) .

With regard to in vivo studies examining the role of CFTR in the pathophysiology of the CF lung, there is a general consensus that, when challenged with PsAembedded agarose beads, mice lacking functional CFTR exhibit heightened pulmonary inflammatory responses as compared with wild-type controls (16;17) ; such responses include increased production of TNF␣ and IL-1␤. There is less agreement, however, between in vivo studies that examine the role of CFTR in the bacterial clearance of an infected CF lung. A collection of recent studies indicate that correction of mutant CFTR in the lungs of murine CF models enhances the clearance of Burkolderia cepacia (18) , but not of PsA (16;17) . Interestingly, some studies have proposed that nutritional deficits, which result from CF-related intestinal defects, contribute to the exaggerated immune response in both CF patients and CF mouse models (19) . In contrast, a recent report demonstrates that CFTR knock-out (KO) and gut-corrected CFTR KO mice displayed similar indices of inflammation when challenged with PsA (20) . This study also showed that CFTR KO mice and wildtype control littermates exhibited similar lung inflammatory responses in the presence of a PsA infection (20) .

Such discrepancies between these collective in vitro and in vivo studies may be due to differences in the airway epithelial cell and mouse models utilized, tissue culture conditions, exogenous stimuli, bacterial strain, and type of bacterial infection. Importantly, these discrepancies highlight the on-going controversies regarding the role of CFTR in regulating inflammatory responses within the CF lung and, therefore, make it difficult to project the amount of CFTR required to attenuate such responses. In conclusion, the relationship between the level of CFTR expression and excessive inflammatory responses warrants further study.

The earliest concept of CF pathophysiology was that disease arises because affected organs are blocked by abnormally thick secretions 1, 2 . Thick mucus was initially attributed to a defect in mucin molecules, but the discovery of electrolyte disturbances in sweat gradually led to a radically different concept: the basic defect is in ion transport, which leads to thickened secretions because of deficient water within mucus. It was long known that the first step in water secretion is often mediated by anion secretion. With the discovery that the defective CF gene codes for CFTR, a channel for chloride and bicarbonate, it was immediately apparent that defective secretion of electrolytes and water could account for observed pathology in many CF-affected organs such as the vas deferens, pancreas, and intestines. An exception is the sweat duct, where CFTR anion channels play an important role in electrolyte absorption, leaving water behind 3 .

What about airways? People with CF typically die from unremitting lung infections. Both ciliated surface epithelia and submucosal gland serous cells express CFTR, and decreased anion-driven fluid secretion from both of these cell types contributes to mucus stasis. In addition, CF airways have an additional defect: lack of CFTR disinhibits ENaC, leading to increased sodium-dri-ven fluid absorption: if extreme enough, hyper-absorption alone can produce mucus stasis and a CF-like disease 4 . We propose still a third defect: gland mucus contains a rich complement of proteins that act as anti-microbials, anti-oxidants, and inhibitors of serine proteases. We hypothesise that the bio-availability of these compounds is reduced in CF because of the altered mucus properties(a precedent for this concept has been established in CF mouse intestines 5 . Furthermore, the ability of mucus to inhibit bacterial growth wanes with time, so static mucus eventually becomes a bacterial growth medium 6 .

Our laboratory has recently focused our efforts on gland secretion and serous cell models. Glands are the major secretory organs of healthy lungs and gland serous cells express CFTR . When provoked, glands produce copious amounts of mucus that traps pathogens and inhibits their growth while they are swept from the airways 7 . Mucus clearance normally keeps the airways sterile, and it has been proposed that altered gland function in CF airways impairs the mucus shield e.g. 8, 9 . Early studies of airway glands established many of their properties, but were carried out before it was understood that CFTR is expressed in serous cells. Since then, studies by Ballard and colleagues, and more recently by our lab and the Verkman lab, have expanded our understanding of glands and have established that CF glands have defective secretion.

We developed an optical method for rapid and accurate assessment of single gland secretion rates 10 and are using it to study detailed features of single gland secretion in sheep, pigs and humans. Glands secrete to both VIP and ACh. Gland secretions are hypotonic 11 , and secretions to both mediators are slightly acidic relative to bath 12 . We found that in humans with CF, the glands are completely refractory to VIP or forskolin 13 , and produce about half normal amounts of fluid in response to acetylcholine. Secretions to ACh have increased viscosity 11 .

To look more closely at gland secretion, we have developed methods to isolate single glands or portions of glands, so that responses of single tubules and single cells can be monitored with Nomarski differential interference contrast microscopy. These studies reveal that mucus is a highly heterogeneous substance, and provide evidence for exocrine, apocrine and holocrine secretion within a single gland. Surprisingly, VIP, which produces zero mucus outflow from CF glands, produces clear evidence of secretion when individual CF gland tubules are monitored, suggesting that proteins are being released, but without sufficient water to exit from the gland duct. These findings accord well with earlier predictions made by Ballard and colleagues using pharmacological block of fluid secretion 14 .

A complementary approach is to use cell culture models of serous cells. To help bridge the gap between secretion studies of glands and electrophysiological studies of cell sheets, we designed a 'virtual gland' that allows us to establish secretion rates (Jv) and collect secreted fluid from secretory cells such as Calu-3 cells 15 . With that link established, we can more easily interpret electrophysiological studies of Calu-3 cells now being carried out in many laboratories. Recent work from our laboratory indicates that Calu-3 cells express Kir 4.2 and possess an apical K + channel with properties similar to Kir 4.2. They also possess an apical H-K ATPase, resulting in the co-ordinate secretion of H + (via the H-K ATPase) and HCO 3 -(via CFTR). This system provides one method whereby an initially isotonic secretion can be rendered hypotonic by the subsequent conversion of H 2 CO 3 into CO 2 and H 2 O.

The hallmark of cystic fibrosis (CF) lung disease is chronic colonization/infection of the airways by a variety of bacteria, including H. influenzae, S. aureus, and P. aeruginosa. In CF, bacteria grow in regions of the lung that are normally sterile. The clinical course of CF lung disease correlates with the acquisition of bacterial infection and its progression. While the pathogenic mechanisms linking CFTR mutations to CF lung disease continue to be a subject of research and debate, it is accepted that the normal host defense barriers in the lung are fundamentally altered. Whatever the cause of the underlying propensity for infection, it is lung specific. The result is a characteristic susceptibility to bacterial infection. These features indicate that CF impairs the innate defenses of the lung. Prominent components of innate immunity include the peptides and proteins secreted into airway surface liquid (ASL).

The airway epithelium directly interacts with the environment in a dynamic fashion. Because of this interplay between the host and environment, systems evolved to clear or inactivate the pathogens encountered (see Fig) . These defense systems include components of both innate and adaptive immunity. Innate immunity complements adaptive immunity by its "ever ready" nature, as many of its components are continuously present and require no previous encounter or memory for their activity (4) . In the airways, the main cellular components of innate immunity are the epithelium and its products, and macrophages, dendritic cells, NK cells, cytotoxic T cells, and neutrophils. Innate responses may be triggered via pathogens interacting with pattern recognition receptors such as the Toll-like receptor (TLR) family (26, 41). The airway epithelium senses and responds to microbial challenges and can vectorially secrete its specialized products in the apical or basolateral direction to mount the appropriate response.

In addition to serving as an important physical barrier, two major functions of the epithelia lining the conducting airways and submucosal glands are the production and modification of ASL and the secretion of factors that contribute to host defenses. These agents act by a variety of mechanisms including disruption of microbial cell walls, sequestration of nutrients, and acting as decoys of microbial attachment. Presumably, ASL composition provides an optimal microenvironment for the function of the microbicidal compounds. ASL contains several products with bacteriostatic or bactericidal activities, forming an important line of defense (8, 31, 40) . The specialized cells of submucosal glands (serous cells) also contribute to the secretion of airway surface fluid and are a major source of the production of macromolecules including antimicrobials and mucins (25, 43). The antimicrobial peptides and proteins produced by surface and submucosal gland epithelia act in a broad-spectrum fashion, in some cases exerting antimicrobial effects against bacteria, fungi and viruses. Major ASL protein components in terms of concentration include lysozyme, lactoferrin, and secretory leukocyte proteinase inhibitor (SLPI) (8, 9, 20, 39, 42) . Additional components are surfactant proteins A and D (SP-A, SP-D), collectin family members that are opsonins for bacteria and viruses (11, 22, 23) , the alpha and betadefensins (32), cathelicidins (2) , NGAL (lipocalin 2) (10), and members of the lipid transfer/lipopolysaccharide binding protein family (3) . While the activities of many ASL components have been defined, it is clearly a complex milieu, containing >1000 proteins (24), and the identity and function of many constituents are unknown.

The pulmonary host defense defect in CF is complex. It is likely that the nature of the defect changes as the disease progresses. Early lung disease is characterized by intermittent bacterial infection and the onset of inflammation. Hallmarks of established lung disease include established bacterial infection with organisms living in biofilms, chronic neutrophilic inflammation, and progressive bronchiectasis. In the ~15 yrs since the CFTR gene was discovered, many ideas emerged to explain the link between mutations and lung disease. One hypothesis is that CFTR mutations alter the processing of cell surface asialoglycoproteins such that the CF epithelia are more suitable for bacterial attachment (17) . A second is that CFTR at the cell surface functions as a bacterial receptor for clearance by epithelial phagocytosis; mutant CFTR is unable to perform this function (28). A third hypothesis is that CFTR mutations cause increased NaCl and liquid absorption across epithelia, diminishing the ASL volume and impairing normal mucociliary clearance. As a secondary effect of impaired clearance, bacterial adherence and colonization of the airways begins (5, 6) . A fourth theory is that mutations in CFTR impair the normal ability of epithelia to modify the electrolyte composition of airway surface liquid (ASL) by reabsorbing NaCl (36). As a result, ASL ionic strength increases and this secondarily impairs the function of many endogenous antimicrobials (35, 36, 40) . Alterations in the composition of secretions from submucosal glands may contribute to dehydration of the ASL layer and increased viscosity of CF secretions (18, 19) . Furthermore, the trapped mucus may create an environment that is well suited for bacteria such as Pseudomonas to live as biofilms (33, 34) . In addition to primary or secondary CFTR-associated changes that lead to pulmonary disease manifestations, polymorphisms in other genetic loci may influence the CF phenotype (12, 29, 30) . Several candidate modifier genes have been proposed for CF including mannose binding protein (14) , HBD-1 and HBD-2 (30), alpha 1-antitrypsin enhancer (16) , and HLA class II (1) . These disease pathogenesis hypotheses are not necessarily mutually exclusive, and underscore controversies in our current understanding of CF lung disease. A theme common to all these hypotheses is that altered innate defenses predispose to an increased susceptibility to colonization with bacteria.

The chronic infection and inflammation associated with CF causes epithelial injury and repair responses (21) . This may also cause the epithelium to be impaired. For example, the production of the collectins SP-A and D progressively declines with disease progression (27). Proteolytic enzymes, arising from CF airway secretions or resident bacteria, also cleave and inactivate many proteins and receptors involved in innate and adaptive immune responses, further impairing host defense. For example, Pseudomonas proteases cleave IgG and destroy its opsonic activity (13) . There is also evidence for host or pathogen derived enzymatic cleavage of transferrin, lactoferrin, and SLPI (7, 38) . Recently, Taggart et al demonstrated that the cysteine protease cathepsins B, L, and S, in concentrations that are present in CF secretions, degrade and inactivate HBD-2 (37). In addition, the protease rich environment of the CF airways has been recently shown to cleave TLRs-2 and -4 that may further impair innate immune signaling (15) . The exuberant production of mucus and DNA released from host cells and bacteria creates a complex anionic matrix that may also bind cationic antimicrobial proteins and further impair their activity. In summary, the single gene defect in CF causes complex primary and secondary host defense defects that initiate and then perpetuate progressive chronic lung disease. Understanding these defects may lead to improved treatment for CF lung disease. Emerging evidence from in vitro and in vivo studies suggests that airway surface liquid (ASL) volume depletion initiates the cascade of events that results in CF lung disease. Although regional differences may exist in the relative contributions of the superficial epithelium and glands to ASL volume depletion, a common sequence follows that is manifest by periciliary liquid volume depletion, concentration of mucins in the mucus layer, adhesion of the mucus layer to airway surfaces, delayed ciliary and cough-dependent clearance, inflammation perhaps without overt bacterial infection, and ultimately acquisition of bacterial infection and promotion of bacterial biofilms. Four topics that pertain to this scenario will be reviewed.

• Mechanisms of autoregulation of ASL volume by normal and CF airway epithelia: The focus will be on the signals that sense airway surface liquid volume and the mechanisms by which these signals coordinate the deactivation/activation of epithelial ion channels (ENaC, CFTR, CaCC) to control ASL volume and hence surface mucus clearance. • Regional differences in airway surface liquid volume regulation in normal and CF: Data will be reviewed on the expression in the normal airway of CFTR in the proximal versus distal regions of the lung and the superficial versus gland epithelia. In addition, studies on the volume regulatory capacity of the normal and CF bronchiolar region will be reviewed. • Functional consequences of depleted ASL volume:

The focus will be on mucus adhesion as being an initiating lesion in the progression of the obstructive and infectious components of CF lung disease. Data demonstrating adhesion both in vitro and in vivo will be reviewed, as well as inferences as to the nature of the adhesive interactions. • Methods of restoring ASL volume therapeutically:

The actions of hypertonic saline in restoring ASL volume in normal and CF airway epithelia in vitro will be reviewed, and correlates to recent in vivo studies of the effects of hypertonic saline on mucus clearance and lung function made. In addition, the actions of selected novel small molecule pharmaceutical compounds will be reviewed. In sum, it appears that studies focused on the hypothesis that ASL volume depletion is the initiating event in CF lung disease have been fruitful in regard to generating follow-on studies investigating the details of the pathogenetic sequence of CF disease, including the importance of mucus adhesion to airway surfaces and the importance of concentrated mucins in promoting not only the acquisition of infection but the biofilm mode of bacterial growth that characterizes CF. The charge now is to assign in different lung regions the relative roles of superficial epithelia versus glands in order to design effective therapies for early and, it is hoped, established CF lung disease. Mucociliary clearance (MC) represents the first line of innate defense against the environmental contaminants and diverse exogenous sources of injuries to which the airways are exposed (1) . The ability of the airways to clear the mucus is depending on several factors including the number of ciliated cells and ciliary beat frequency (CBF), the volume of the airway surface liquid, the degree of hydration and the rheological properties of the periciliary (sol) and gel airway mucus. The airway secretory glandular cells may also markedly contribute to the hydration of the sol and gel mucus and therefore may directly influence the airway mucus transport (2) .

There is now increasing evidence that the airway surface liquid (ASL) covering the surface of the normal airways is an isotonic liquid, characterized by an optimal height (near the length of the cilia), and ideal viscosity and surface properties. The ASL volume must be adjusted continuously in relation to the environment of the airway epithelium. Any mechanical stimulation such as excess mucus load or external injury will first stimulate CBF, activate surface effectors and regulate the volume of ASL (3) .

Different studies using freshly excised tissues and primary cell cultures have shown that the steady-state periciliary layer volume is maintained by a balance between Na + absorption and Clsecretion (1). Tarran (3) has demonstrated that normal airway epithelium senses and autoregulates ASL height by adjusting the rates of Na + absorption and Clsecretion to maintain mucus transport. It is well accepted that excess liquid at the periciliary layer level may be removed by a transepithelial ENaC transport. On the opposite, ENaC inhibition will induce Clsecretion controlled by CFTR activity. In these conditions, it can be easily speculated that in CF airways, the upregulation of Na + and downregulation of Clrelated to CFTR defect will lead to a depleted ASL associated with abnormal mucus transport.

In CF patients with mild disease, compared with age-matched normal subjects, a reduced mucociliary clearance has been reported (4). A progressive decrease in MC is generally associated with increasing disease severity (5) . In patients with advanced CF disease, inflammation, infection and remodeling of the airway may be directly responsible of the decreased MC. Whether MC is delayed early in CF infants before any infection is still unknown. Most of the MC studies in CF mice have reported a large variability of the data (6) . In the normal murine lower airways according to the mouse strain, the mucociliary clearance varies from 0 to 4.5 mm/min, the lowest values being related to the low number of ciliated cells rather than to a decreased ciliary activity. In CF mice, the CBF is not decreased. Although we and others (7) have reported a decreased rate of MC in CF mice, other investigators failed to demonstrate a difference (6) . Interestingly, overexperession of ENaC in a recently published mouse model showed that the decreased volume of ASL in the lower airways was associated with a large decrease (~ 60 %) in MC (8) .

Although the specific contribution of abnormal biochemical and rheological properties to the impairment of mucus transport at an advanced stage of the disease is accepted, the impairment of MC in early CF disease is still not clearly defined. Secretory proteins, mucins and lipids are major components of airway mucus that contribute to the antibacterial defense of the airway epithelium but also contribute to the physical (adhesion), rheological (visco-elasticity) and transport capacity of the airway mucus. Secreted mucins such as MUC 5AC are reported to be enriched in the sol phase of CF mucus and mucins sulphation and sialylation are increased in CF secretions (9) . We have also earlier shown in secretions from adult CF patients that the phospholipid content moves towards a profile of poorly lubricant and therefore more adhesive mucus characterized by a decreased content in surface-active molecules such as phosphatidylglycerol and phosphatidylcholine (10) . Whether these biochemical and rheological changes are constitutive or related to a possible defective intracellular acidification due to CFTR mutation is unknown.

In human fetal tracheal xenografts in the SCID mice (11), we could not demonstrate any dehydration nor increased viscosity of the airway liquid collected in the lumen of the CF xenografts despite the fact that we observed an increased number of amiloride-sensitive Na + channel and an inefficient cAMP-dependent Clchannel. In fetal airway mucosa, volume activated Clchannels may also contribute to the fluid secretion as previously shown in the late stages of lung development and therefore could compensate the defective mucus fluid hydration related to CFTR mutation.

One important cause of impaired mucus transport in CF may be related to an abnormal glandular secretion. When the secretion response to glandular agonists is inhibited in pig airways, the mucus liquid becomes less hydrated and more viscous (12, 13) and MC is markedly decreased, likely due to an uncoupling between mucin and liquid secretion. Abnormal secretion by CF glands is expected on the basis that CFTR is expressed in airway glandular serous cells not only at the apical membrane

Edith Puchelle, Sonia Baconnais, Jean-Marie Zahm INSERM UMR-S 514, Reims, France surface but also at the level of membrane secretory granules (14) . Moreover, fluid secretions are reduced and viscosity is increased in CF gland cells and in pig tissue treated by CFTR inhibitors (15) . The implication of abnormal airway submucosal gland function in CF has been suggested. Wine and Joo (2) hypothesized that exocytosed mucus may remain as condensed packets for a long time after secretion and that, as in CF intestinal crypts, the granule contents may be trapped as undispersed granules within the lumen. Using phase contrast dynamic videomicroscopy, we could confirm in CF glandular cells that under basal conditions, mucus exocytosis and expansion of secretory granules is markedly delayed. In parallel, we could demonstrate by quantitative dark field imaging and X-ray microanalysis on freeze dried airway glandular cell cryosections, that the mucus hydration of the CF secretory granules is significantly decreased and the ion content (Na + , Mg 2+ , P, S and Cl -) is significantly higher compared with the non-CF secretory granule mucus content (Baconnais et al., unpublished data). Our data provide strong evidence that in CF cells, the alterations in ion composition and water content within the serous granules may partly explain the formation of thick mucus plugs lying over a dehydrated periciliary layer on airway surfaces. As proposed by J. Wine (2) "too little water, to late" could explain that the secretory granules of glandular cells cannot correctly expand and may remain anchored to the surface of the cells. In such conditions, the glandular mucus cannot correctly disperse in the airway surface liquid.

Taken altogether, these results suggest that the abnormal exocytosis associated with abnormal ion and water content of secretory granules inside the glandular cells may accentuate the dehydration of the periciliary layer and contribute to the deficient mucus transport in CF airways. Other alternative pathways may partly compensate this glandular mucus and ASL dehydration and more or less prevent a major early failure of MC. Nevertheless, the inadequate hydration of epithelial fluid at the apical surface and in intracellular secretory granules level favour the hypothesis that in CF, decreased mucus transport is a reality. Lung disease in patients with cystic fibrosis (CF) is associated with neutrophil-dominant lower airway inflammation and increased lower airway concentrations of pro-inflammatory cytokines and neutrophil elastase even among infants and young children with CF (1, 2) . Several studies report a robust inflammatory response in the airways of both bacterial culture-positive and culturenegative children; some studies show a greater inflammatory response in those patients with at least 5x10 4 colony forming units/milliliter of bacteria in their bronchoalveolar lavage (BAL) fluid (1) . Lung disease in CF can have regional heterogeneity that complicates understanding the relationship between initial airway infection and the inflammatory response (1). There is emerging computed tomography evidence for destructive regional inflammation/infection, i.e. bronchiectasis, in stable young children with CF (4) (5) (6) . Such data emphasize the need for more research on the pathogenesis and treatment of early lung disease in CF, and the further development of outcome measures to assess global and regional progression of early lung disease. Currently, lower airway inflammation and infection in non-expectorating patients can only be assayed by BAL. There is a growing interest in proteomic analysis of blood and exhaled breath condensate to identify non-invasive markers of lower airway inflammation and infection.

Patients with CF acquire a unique set of respiratory tract pathogens in an age-dependent sequence. Early airway infections are most frequently caused by S. aureus and non-typeable H. influenzae (1) . There is controversy on the role of S. aureus and H. influenzae in the progression of early lung disease. Pseudomonas aeruginosa (Pa) is a ubiquitous environmental organism and the most significant pathogen in cystic fibrosis (CF) lung disease. Pa prevalence increases with age, with positive respiratory tract cultures in 15-30% of infants, 30-40% of children 2-10 y.o.,~60% of adolescents, and ~80% of adults with CF (2). Reported risk factors for early Pa infection include prior S. aureus respiratory tract infection, chronic anti-staphylococcal antibiotic treatment, female gender, delta F508 homozygous genotype, and attendance at CF clinics without cohorting (2) .

There is currently no optimal method to assess lower airway Pa infection in young children. Sources of respiratory tract secretions for culture include expectorated or induced sputum, oropharyngeal (OP) swabs, endolaryngeal suction, and BAL fluid. Sputum expectoration is the least invasive and most widely used method in older children, but children < 6 years of age can rarely expectorate. Sputum induction with hypertonic saline is generally not successful in children < 8 years of age. OP cultures are widely used as a surrogate for lower airway cultures in non-expectorating CF patients, but have poor sensitivity and positive predictive value for lower airway Pa infection; OP Pa isolates can have different genotypes than BAL isolates from the same patient (2) . Despite the limitations, the current U.S. standard of care is to collect quarterly OP cultures.

The host immune response to Pa antigens appears to be the earliest marker of Pa infection; this is controversial and likely depends on the assays performed (2, 3) . Recent U.S. studies found serologic responses to Pa exotoxin A and whole-cell membrane proteins occurred prior to cultureevidence of Pa infection in young children (2) . The Danish CF center uses precipitins against Pa antigens as a marker of chronic infection (3); they have reported reduced serologic responses to Pa with intensive anti-Pa treatment. Limitations of serological markers include lack of commercially available standardized assays, and lack of specificity to the site of Pa infection (i.e. upper or lower airway). Generally, patients with CF are initially infected with unique environmental isolates of Pa (1,2). Regardless of treatment, patients commonly have a period of intermittent isolation of Pa from the respiratory tract prior to becoming chronically infected (1) . The duration of intermittent Pa + cultures can vary widely, and this complicates the decision regarding aggressive early intervention. Early Pa isolates are generally non-mucoid in phenotype, highly antibiotic sensitive, and present at relatively low density (2). These features suggest a "window of opportunity" for early intervention with antibiotics to possibly delay or prevent chronic Pa infection and lung disease progression. Early age at Pa acquisition adversely affects lung disease and survival; yet, acquisition of Pa does not appear to cause an immediate and rapid decline in lung function (2) . The early clinical features of Pa infection are subtle and variable, including slightly reduced lung function, and poorer chest radiograph and clinical scores (2). By school age, the impact of early Pa isolation is established. Among patients from whom Pa was isolated during the first 5 years of life, there was a 2.6 times higher risk of death, on average a 14% predicted lower FEV 1 , and 7% lower weight percentile at 8 year follow-up compared to patients without Pa isolated early in life (2) .

Danish investigators were the first advocates for aggressive treatment of early Pa infection in CF (3).

Subsequently, most published studies have shown a microbiologic effect of antibiotics for early Pa infection based on transient eradication of Pa from upper and lower respiratory cultures (2) . There has been no randomized, placebo-controlled trial to evaluate clinical efficacy and safety, including drug toxicities and emergence of resistant Pa or new pathogens. Most studies suggest that recurrent Pa infection is inevitable regardless of the initial treatment regimen (2) . The potential sources for recurrent or persistent infection include the sinuses, undetected residual lower airway infection, or re-infection with an environmental isolate.

Current options for early anti-Pa treatment include oral quinolones, inhaled tobramycin and colistin, and IV antibiotics (1-3). Macrolides have not been investigated for this indication. Attractive features of oral quinolones include ease of administration, evidence for efficacy in older patients, and accumulating safety data in pediatrics. Quinolone monotherapy may result in rapid emergence of resistance. Limited safety data exist for inhaled tobramycin in young children, but it has a well-established safety profile in patients > 6 years of age (1) . Inhaled colistin at the time of Pa acquisition is used extensively in some European centers, primarily in combination with oral ciprofloxacin (3) . IV antibiotics are established treatment for pulmonary exacerbation in chronic Pa infection (1); their role in the therapy of initial Pa infection deserves further investigation. An ultimate goal for treatment of early Pa infection is to pre-scribe the least invasive and safest treatment for the shortest duration necessary to achieve both microbiologic and clinical benefit. The pathogenesis of CF lung disease is due to dysfunction of CFTR in the airway epithelium and results in impaired chloride secretion and increased sodium absorption, leading to dehydrated hyperviscous mucus secretions and impaired mucus clearance. The hallmarks of CF lung disease are chronic bacterial infection, mostly notably Pseudomonas aeruginosa, and exaggerated neutrophil-mediated inflammation, eventually leading to airway obstruction and death. Although the role of bacteria in CF lung pathology is well established the consequences of respiratory viral infection in CF infants are less well understood. Although viral infection of the lung can result in serious illness, viral infection of the CF lung may have more severe consequences. There is also the likelihood that viral infection can establish an environment that is favorable for subsequent bacterial infection, a process termed "bacterial superinfection".

The upper respiratory tracts of healthy individuals often contain commensal bacteria such as Haemophilus influenzae and Streptococcus pneumoniae but are asymptomatic for infection. However, these pathogens can become a threat when concurrent with a respiratory viral infection. Of the common viruses infecting the respiratory tract of humans, influenza, respiratory syncytial virus (RSV) and parainfluenza virus (PIV) result in the highest morbidity and mortality especially amongst children. In particular, RSV is the most common respiratory virus associated with the development of otitis media in children and is often associated with the respiratory bacterial pathogen nontypable Haemophilus influenzae (NTHi). NTHi is responsible for a significant proportion of bacterial conjunctivitis, otitis media, sinusitis, bronchitis and pneumonia all of which may also be associated with concurrent viral infection. Many mechanisms have been proposed to result in bacterial superinfection including disruption of innate immune mechanisms, e.g., impaired cough reflex; reduced efficiency of mucociliary clearance; the accumulation of excess or altered airway secretions; and, reduced macrophage-mediated phagocytosis. Substantial evidence also exists for viral-induced up-regulation of bacterial adherence receptors although whether these changes occur at the lumenal surface of the respiratory epithelium or as altered secretory products is unknown.

Although viral-induced pathology and bacterial superinfection can occur in otherwise normal individuals, patients with CF or other underlying lung diseases are considered 'high-risk' for developing a greater severity of disease. More rapid pulmonary deterioration in CF patients has been associated with respiratory viral infections (1) and RSV infection produced more severe and prolonged disease in CF infants compared to infection of non-CF infants (2) . However, respiratory viral infections were found to just as likely occur in CF infants as in non-CF infants but more likely to result in lower respiratory tract infection, impaired pulmonary function, and hospitalization for CF infants (3) . In combination with these clinical data there also exists ample anecdotal evidence that CF patient lung disease gets much worse after viral infection. In terms of specific interactions with respiratory pathogens there is good clinical and experimental evidence that H. influenzae superinfection is associated with a concurrent airway viral infection and H. flu is recognized as an early bacterial pathogen in CF airways. However, data for an association of P. aeruginosa superinfection with airway viral infection is lacking.

In the context of CF lung disease and the consequences of viral infection in CF individuals a number of important questions remain to be addressed: Another rarely encountered species, B. ubonensis, should also be considered a member of the Bcc based on phylogenetic assessment. All of these 'new' species have been identified among banked CF sputum isolates previously identified merely as 'B. cepacia,' indicating that all have caused infection in CF patients for many years. However, the distribution of these species in CF is quite disproportionate. Studies from several countries indicate that B. cenocepacia (formerly 'B. cepacia genomovar III') is the most prevalent Bcc species recovered from CF patients. B. multivorans is the next most common species, with the remaining Bcc species generally being recovered from relatively small numbers of patients. Careful assessment of CF sputum isolates has also identified strains that, although clearly members of the Bcc, can not be placed definitively into one of the ten currently described species; these 'species indeterminate' strains may represent still additional Bcc species that await definition.

Genotyping (DNA fingerprinting) analyses have demonstrated that in some CF care centers multiple patients are infected with the same Bcc strain. Some strains are even more broadly distributed. Among these are strain ET12, prevalent in eastern Canada and the UK, and strain PHDC, which has been recovered from CF patients in 24 US states and, most recently, in Europe (1). ET12 and PHDC, as well as several other common strains reported to date, reside in the species B. cenocepacia. Shared strains belonging to other Bcc species, including B. cepacia, B. multivorans, and B. dolosa, have been less frequently identified (2) . The extent to which such common strains are inherently transmissible between patients is unclear and must be assessed in the context of the infection control measures employed to prevent their spread. Further, available data suggest that Bcc species differ with respect to their preferred niche(s) in the natural environment. A better understanding of the ecology of Bcc species (and specific strains) is needed to assess the risk posed by natural reservoirs and to develop strategies to prevent acquisition from the environment.

Limited clinical outcomes data, including studies in lung transplant recipients, have shown an association between infection with B. cenocepacia and greater rates of morbidity and mortality (3, 4) . These observations have led some to conclude that B. cenocepacia is the "bad cepacia," implying that all strains in this species are easily transmissible and cause particularly severe infection, while other Bcc species present little risk. However, most studies have involved small numbers of patients infected with a limited set of common strain types, with results being extrapolated broadly to the entire species. In fact, many, if not most, CF patients infected with B. cenocepacia harbor seemingly 'unique' strains (i.e., not shared by other patients), and many patients have remained chronically infected with this species for prolonged periods with little apparent impact on progression of lung disease. Furthermore, infection with Bcc species other than B. cenocepacia has been associated with sepsis and death (so-called cepacia syndrome). Rather than ascribing degrees of relative virulence to entire Bcc species, it seems more likely that virulence and transmissibility vary among specific Bcc strains. Of course, as yet undefined human host factors almost certainly contribute to clinical outcome, and so broad conclusions regarding relative Bcc virulence will be difficult.

A number of recent investigations employing improved animal and in vitro models of infection have yielded important insight into potential Bcc virulence factors and pathogenic mechanisms. The molecular events underlying the ability of Bcc to adhere to, invade and elicit a robust pro-inflammatory response from human airway epithelia have been studied by using immortalized cell lines as well as primary well differentiated human cell cultures (5, 6) . Similar to P. aeruginosa, Bcc are capable of expressing homoserine lactones likely involved in quorum sensing and biofilm formation (7) . A number of genes required for bacterial survival in vivo have been recently identified (8) , and provide an exciting opportunity to define potentially completely novel virulence factors. The availability of the complete genome sequence of B. cenocepacia strain J2315 (a representative of the ET12 lineage) provides a critical asset to investigations of Bcc pathogenesis. The sequencing of several other Bcc strains and the development of a Burkholderia DNA microarray are underway and will surely further enhance ongoing research efforts.

In summary, during the past several years, a great deal has been learned regarding Bcc taxonomy, clinical microbiology and epidemiology. Clearly, there is now a need for a more complete understanding of Bcc ecology and for large scale outcomes studies to better assess the relative risk of infection with specific Bcc strains. Ultimately, better definition of the bacterial virulence factors and human host factors that determine diseases progression is necessary to devise novel therapeutic and preventative strategies. Increasingly, the emergence of CF respiratory pathogens, including bacteria with intrinsic or acquired antimicrobial resistance, non-tuberculous Mycobacterium (NTM) species and fungi, challenge clinicians to better characterize their prevalence and epidemiology, pathogenecity in CF lung disease, optimize treatment of infection, and prevent acquisition in both healthcare and non-healthcare settings.

Staphylococcus aureus binds avidly to respiratory mucin and is often is the first pathogen to colonize the respiratory tract, with a prevelance of ~45% by one year of age. S. aureus may cause chronic inflammation of the CF respiratory tract by up-regulation of proinflammatory cytokines exotoxin-mediated immune dysregulation and tissue damage. (ref) Currently, 9.2% of CF patient S. aureus strains are resistant to methicillin (MRSA). MRSA strains may be transmitted via direct contact in and outside of healthcare settings, may cause transient or chronic respiratory colonization, and have been associated with severity of underlying disease and adverse outcomes in both CF and non-CF patients (Cosgrove, others) Although the impact among CF patients is limited to date, there is growing concern about genetically distinct community-acquired MRSA (CA-MRSA) strains that posses a unique resistance cassette and cause virulent soft tissue infections and necrotizing pneumonia mediated by the cytotoxin Panton-Valentine leukocidin (ref).

Stenotrophomonas maltophilia and Achromobacter xylosoxidans are intrinsically multidrug-resistant gramnegative pathogens whose acquisition is promoted by use of broad-spectrum antimicrobials. They can cause healthcare-associated infections, including pulmonary exacerbations among CF patients (refs). Among CF patients, current prevalence rates of 9.4% for S. maltophilia and 5.2% for A. xylosoxidans likely are underestimates because of the failure to use appropriate selective media and misidentification. Although many patients have only transient colonization with these organisms and their pathogenecity remains poorly defined, longterm colonization S. maltophilia was associated with decreased lung function 2 to 7 years after acquisition in one study (ref) . Studies have suggested that acquisition may occur both from environmental sources or from person to person. Saphrophytic fungi, including Aspergillus spp. and Scedosporium apiospermum, are ubiquitous and may be found in soil and in the hospital environment associated with construction activities or water intrusion. Aspergillus spp., predominantly Aspergillus fumigatus, and can colonize the lungs of CF patients and can cause allergic bronchopulmonary aspergillosis (APBA) in up to 7.8% of patients and rarely aspergilloma and invasive aspergillosis (285). Registry data likely underestimates the prevalence of Aspergillus. In the aerolized tobramycin trial, up to 25% of CF patients were colonized with Aspergillus spp ( ) Person-to-person transmission remains a rare but potential risk among patients with wound or tracheal infection (ref) .

Further studies are needed to assess the impact of use of selective media and infection control precautions on the prevalence of these emerging pathogens and to better define risk factors for acquisition in and outside of healthcare settings, the clinical impact and treatment strategies. It is pertinent to look at our evolving understanding of the clinical manifestations of CF in the context of key milestones that have occurred since the recognition of CF as a distinct clinical entity in 1938 (1) . From initial recognition/description through 1953, most patients with CF died in infancy or early childhood of pneumonia and severe malnutrition. The diagnosis was suggested by clinical features along with evidence of exocrine pancreatic deficiency on duodenal fluid analysis, and then confirmed at autopsy. Recognition of the sweat gland defect in 1953 (2) and development of the quantitative pilocarpine iontophoresis sweat test in 1959 (3) led to the inclusion of a positive sweat test result as an essential criterion for the diagnosis of CF. Recognition of physiologic abnormalities of sweat and serous-secreting glands led to a change in terminology from mucoviscidosis to cystic fibrosis.

From 1960 to 1989, the CF phenotype was better defined and it became clear that CF was characterized by significant phenotypic heterogeneity (4). During this time it was recognized that 1) there is an important subset of patients with chronic sinopulmonary disease typical of CF, pancreatic sufficiency and borderline or normal sweat electrolyte concentrations (variant, atypical or non-classic CF) (5) and 2) almost all males with CF are infertile secondary to bilateral absence of the vas deferens (6) . There were two highly significant laboratory observations during this period; raised immunoreactive trypsin levels in dried blood spots could be used for newborn screening and a raised bioelectric potential difference across CF respiratory epithelia could be used as a diagnostic tool.

The identification of the CFTR gene in 1989 led to significant advances in our understanding of genotypephenotype correlations and also introduced the possibility of using mutation analysis for diagnostic (including prenatal) testing and carrier screening. It became recognized that there is a subset of "mild" mutations which are associated with pancreatic sufficiency and, in some instances (missense mutation A455E) with a less severe pulmonary phenotype. Also, it was recognized that there are a number of disorders (CBAVD, idiopathic chronic pancreatitis, chronic rhinosinusitis, ABPA, diffuse panbronchiolitis) in which CFTR mutations (often associated with residual CFTR function) are frequently identified, but in which a diagnosis of CF may not be appropriate. This led to the introduction of diagnostic terms such as CFTR or CFTR-related disease or CF-like disease. Further complicating the issue of "what is a CF diagnosis?" were the observations that CF mutations cannot be identified in a subset of patients who have clinical features typical of CF (7) , and that some individuals who carry two known CF mutations may remain asymptomatic over extended periods of time.

Also, it has become clear that differences in endorgan sensitivity to CFTR protein deficiency and the genetic context in which a mutation exists can influence the CF phenotype. Intron 8 variants (8) and modifier genes (9) may play an important role in determining the severity of the pulmonary and gastrointestinal manifestations seen in patients with CF.

In 2004, there is still controversy regarding the diagnosis of CF. The overwhelming majority of patients fall into the category of "classic CF" in which there are typical clinical features in association with elevated sweat chloride concentrations and, in the majority of cases, two identified CF mutations. However, there remains a small but increasing subset of patients, often adults, who have atypical clinical features, a normal or borderline sweat test result and identification of one or no CF mutations, and a subset of asymptomatic newborns identified by newborn screening (two CF mutations and/or a positive sweat test result) but who lack clinical criteria to support a CF diagnosis. This has led to the somewhat controversial proposal (10) that it might be appropriate to have a diagnostic category of "genetic pre-CF" for some individuals and raises the issue of how a disease such as CF is defined in the genomic era in which gene alterations are likely to be found in the absence of clinical features (11) . It is clear that a genetic abnormality by itself does not equal a disease and that both clinical criteria and genetic abnormalities will need to be used to define a disease state. 

Peadar G Noone, M.D.

The diagnosis of cystic fibrosis is straightforward in the majority of patients, particularly pancreatic insufficient (PI) patients. Diagnostic criteria (set forth by a Consensus Statement, drawn up by a panel of experts in 1998) include clinical / phenotypic data, as well as data on CFTR genetics and function; these can be applied in most clinical situations (1). The criteria were designed to encompass as many of the potential clinical scenarios as possible; the potential pitfalls that may occur with the laboratory tests of CFTR were discussed. An emphasis was placed that the guidelines are unlikely to cover every possible scenario (practicing clinicians will quickly agree). Difficulties may be encountered in patients with atypical disease that do not conform to traditional clinical definitions, in conjunction with CFTR data that may be difficult to interpret -establishing the line between proven CF and "CF-associated" diseases thus may not be easy for the clinician. As new knowledge is gained through clinical research and clinical practice, and with the potential for increasing presentation of atypical patients the guidelines will need to be refined and updated.

The traditional laboratory investigation of CFTR function is the sweat test. It is strongly recommended that this test be carried out in a center experienced in its performance, and that the only acceptable test is the pilocarpine iontophoresis sweat test. Furthermore, Clconcentrations rather than Na + concentrations are regarded as more accurate for discriminating between CF and non-CF, though sweat Na + may still be used in some laboratories. The literature in regard to sweat Clconcentrations has varied somewhat. The older literature recommends a Clconcentration of 70 mmol/L as diagnostic of CF; these data are based on reference values obtained from PI CF patients during the 1960's, and do not include pancreatic sufficient (PS) patients, or any recently described phenotypes (2) . In view of the increasing awareness of atypical and "variant" CF, a level > 60 mmol/L is now regarded as diagnostic for CF (1) . What is more difficult to interpret is a level less than 60 mmol/L, although in the context of a compatible phenotype, the higher the Clconcentration, the more suspicious of CF the clinician is likely to be. Somewhat arbitrarily, a level of between 40 and 60 mmol/L is regarded as a gray zone (previously 50-60 mmol/L), often including PS / atypical CF patients. Complementation with nasal potential difference (NPD) measures can be useful. NPD is now more widely available, has been standardized, and can be performed safely in all age groups, including neonates (3). NPD requires significant technical experience and validation for each center / operator, and across groups of sufficient patients and controls. The best evidence of CF is a raised basal PD consistent with increased Na + absorption across airway epithelia. Across 8 different sites in the US, Standaert et al found that patients with CF (n = 31) had a mean basal PD (mV) of -45.3+11.4 (vs -18.2+8.3 for non-CF). Similarly, a reduced or absent CFTR-mediated Clconductance (mV) was present in patients with CF (+3.2+3.5 as compared with -23.7+10.2 for non-CF). Despite these advances in NPD measurement techniques and standardization, difficulties arise in atypical CF patients who may have normal or high normal Na + transport (up to -30 mV), although usually such patients have reductions in CFTR-mediated Clconductance, intermediate between that of CF and non-CF (4). As a technical note, inflammation of the nasal mucosa renders the test very difficult to perform and interpret (it may cause a falsely low basal PD and apparently absent Clconductance because of loss of epithelial integrity). Finally, molecular testing demonstrating two "CF-causing" mutations (i.e. associated with altered CFTR synthesis from a premature termination signal, alternations in intron splice sites, or novel amino acid sequences) is strong evidence of CF (1) . Difficulties arise when faced with one CF-causing mutation, after completion of currently available commercial tests. Then the issue arises as to how aggressive to be in searching for another CF-causing mutation on the opposite allele. Exactly defining a CF-causing mutation can be a difficult process, even after consultation with a CF molecular geneticist or other experts in the field of "atypical CF" is necessary

An example of clinical phenotypes described since the Consensus Statement include patients with chronic pancreatitis of unknown etiology, some of whom have what is now termed CFTR-related pancreatitis (4) . Initial CFTR genetic tests showed that some of these patients were carriers of a severe CF mutation, and only after extensive tests or gene sequencing did a subset turn out to have a mild or "variable" mutation in CFTR on the opposite allele. Sweat chloride values were less than 60 mmol/L in some, with a mean value for all CFTR compound heterozygotes of 54 mmol/L (n = 6). Nasal PD tests showed values intermediate between that of normal and "classic" CF, supporting the concept of a subset of CFTR mutations resulting in some residual CFTR function and intermediate CFTR physiology (5) . The flip side of the coin are patients with a phenotype strongly resembling CF but with normal tests of CFTR structure / function (6) . Such patients have some phenotypic features of CF, for example sino pulmonary disease, congenital absence of the vas deferens, gastrointestinal disease, or elevated sweat chlorides (>40 mmol/L). Proportions of patients referred with these criteria (n = 74) had two (n = 29), one (n = 15), or no (n = 30) mutations in CFTR. As with the patients with idiopathic pancreatitis, those patients referred already known to have one CF mutation were most likely to have another mutation (associated with reduced CFTR function) on the opposite allele. The hypothesis was raised that a non-classic CF like phenotype may result from causes outside the CFTR gene.

Finally, it used to be thought that isolation of mucoid Pseudomonas aeruginosa in the airway secretions of someone with bronchiectasis was strongly suspicious of CF. However, other diseases such as primary ciliary dyskinesia may also be associated with such respiratory tract microbiology. Thus, interpretation of the "clinical phenotype" portion of the current definition of a CF diagnosis is difficult when one considers these groups of patients.

Studies are underway looking at the issue of defining more precisely the diagnostic criteria of sweat / NPD / molecular tests in the context of different phenotypes ranging from classic to mild CF to heterozygote and non-CF individuals. 

Defects in the CFTR gene cause a wide spectrum of manifestations, from classic cystic fibrosis (CF) diagnosed in childhood with multi-system involvement including sino-pulmonary disease and pancreatic insufficiency, to monosymptomatic conditions not diagnosed until much later in life (e.g. congenital absence of vas deferens (CBAVD), recurrent pancreatitis). There is now confusion between what is considered cystic fibrosis and how to label other disorders associated with CFTR mutations. There is a growing population of patients in whom a diagnosis of CF can neither be made nor ruled out. The lack of a diagnosis can be stressful for the individual and health care professionals currently have difficulty providing prognostic guidelines.

The discovery of the CF gene has not been as helpful as hoped in these situations. The common CF-causing mutations detected on standard screening panels are usually seen in patients with classic disease, in whom the diagnosis can be made by a positive sweat test. Patients who present with atypical manifestations are more likely to have less common mutations; however extensive genetic testing of the remaining 1,000 mutations or variants in CFTR is not financially feasible or practical.

The standard diagnostic sweat test has not been validated across this new clinical spectrum and so its role in assessing patients presenting with unusual manifestations is unclear. We know that patients presenting with atypical features may have negative or "borderline" sweat test results. Newer investigations examining the function of CFTR, such as measurement of nasal transepithelial potential difference, have not been standardized or validated at all.

In an effort to refine the diagnostic criteria for CF, we have initiated the systematic recruitment of subjects in well-defined cohorts defined as follows: 1. typical CF 2. atypical CF (i.e. pancreatic sufficiency) 3. other conditions related to CFTR(e.g. CBAVD, idiopathic recurrent pancreatitis) 4. obligate CF heterogyzgotes 5. unaffected controls

The CF patients have been clinically diagnosed and identified from the Toronto CF Clinics (pediatric and adult), with 550 current patients who have been extensively genotyped, classified as pancreatic sufficient or insufficient and who are followed regularly with clinical information prospectively collected in an extensive database. Patients with unusual manifestations or disorders associated with CFTR mutations are recruited from three sources: a cohort of men with CBAVD from a male infertility clinic; patients with idiopathic, recurrent pancreatitis after extensive investigations to rule out other etiologies of pancreatitis; and patients referred to the Toronto CF Clinics with bronchiectasis of unknown etiology. Identification of subjects in the latter category has been enhanced by mailings to pulmonologists and general practitioners in the province and CF clinics in Canada. Presentations to the medical personnel in the Toronto and regional CF Clinics have been made to alert them to the opportunity to refer patients in whom CF has been queried but not established or excluded. Obligate heterozygotes are the siblings or parents of patients with classic CF. Control subjects included siblings of known CF patients who do not carry CF gene mutations, as well as unaffected volunteers.

All subjects are fully investigated with clinical assessment; complete gene scanning of CFTR with multiplex heteroduplex shift analysis and characterization of DNA fragments by direct sequencing analysis; sweat chloride testing, using the Gibson and Cooke method, and nasal transepithelial potential difference measurement performed (using method of Knowles) on the same day as the sweat test.

These measurements are prospectively performed in these well-defined patient and control populations, and will form the basis for classification and regression tree analysis as described later in this Symposium. This novel, multi-variable approach to defining the reference parameters of these diagnostic tests will help us to refine the diagnostic process. The ultimate goal is a redefinition of the diagnostic criteria for cystic fibrosis, and a more precise definition of related syndromes. According to the 1998 CF Consensus Statement [1] , a diagnosis of CF is established when a CFTR-associated phenotype is present in conjunction with either two abnormal sweat chloride measurements, or two disease causing CFTR mutations, or an abnormal nasal potential difference (PD) result. Although the sensitivity and specificity of these guidelines are high when applied to pancreatic insufficient patients and healthy controls, the growing number of patients identified with pancreatic sufficiency as well as CFTR-associated clinical phenotypes has resulted in numerous diagnostic dilemmas for which current guidelines are frequently inadequate. The reference range for sweat chloride, established in 1972, was based on assessments of primarily severe CF patients and healthy controls (who were not genotyped). Similarly, the CF-causing mutations included in the consensus statement do not reflect current knowledge. Furthermore, objectively defined guidelines and reference values for nasal PD as a diagnostic test have not been established. Our group is currently evaluating a cohort of patients presenting with diverse phenotypes associated with mutations of the CFTR gene to redefine the laboratory parameters that confirm or exclude a CF diagnosis. Study subjects are recruited from among conventionally diagnosed CF patients, healthy controls, obligate heterozygotes, and query CF patients presenting with infertility, pancreatitis, or sino-pulmonary disease. Patient evaluation includes comprehensive medical history, sweat chloride, nasal PD measurements on the same day, pulmonary function testing, serum trypsinogen, and extensive CFTR mutation testing by multiplex hereroduplex analysis and DNA sequencing.

Methods for characterizing this increasingly heterogeneous population are necessary to refine diagnostic criteria used for clinical and research purposes. In order to identify structure and "natural" grouping within the spectrum of CF disease observed in our data, cluster analysis [2] , an exploratory technique, is used. The most commonly used hierarchical procedures, agglomerative methods, successively fuse similar individuals or groups together in clusters of increasing size and heterogeneity, ending with a single group containing all subjects. The resulting sequence of clusters is represented graphically using a dendogram, a tree diagram where individuals correspond to the 'leaves' whose 'stems' are grouped together at 'nodes'. The distances between clusters, a measure of the difference between the groups, is represented by the lengths of the stems. The dendogram thus provides a useful visual representation of the clustering algorithm that can assist in making decisions about the number of clusters present in the data. To gain further insight into the underlying factors defining the individual clusters, a figure is derived that borrows from the concept of microarray data analysis. The variables used in the cluster analysis are depicted in separate columns across the X-axis, with individual subjects each represented as a single row across the figure. Shades of gray, ranging from white for values in the normal range to black for values representing greatest disease severity, are used to represent each data value. Using this figure, researchers can simultaneously visualize all the variables included in the cluster analysis, revealing the underlying structure of the data and unusual observations within the clusters.

Classification analysis is then used to define a rule that can be used to assign new subjects to one of the clusters. Unlike cluster analysis where no assumptions are made about the number of groups or their structure, in classification analysis the groups are known and are used as the basis for developing a diagnostic rule. The classification and regression tree (CART) method of Breiman et al. [3] is ideally suited for developing diagnostic rules ( Figure 1 ). This method uses recursive partitioning to segregate patients into one of two or more known groups. The decision tree representation incorporates complex interactions in a transparent and intuitive rule that can easily be applied to new cases. Moreover, the method does not rely on normal distribution assumptions and is particularly effective for developing rules in the presence of a large number of variables. Prior distributions and cost of misclassification can be incorporated into the model to optimize the diagnostic rule for use in a clinical or screening situation. Cutoff values selected by CART can be further adjusted to optimize sensitivity and specificity evaluated by normal distribution assumptions, where appropriate.

Statistical analysis of the cystic fibrosis diagnosis cohort will identify homogeneous subgroups within the spectrum of CFTR disease, providing alternative labels for subjects with an uncertain diagnosis of CF. Longitudinal studies of these patient groups will refine models for prognosis and survival, assisting in frequency of follow-up and treatment decisions and providing patients and their families with more precise information.

Supported by NIH DK9096-10 and CCFF. Figure 1 : 60 controls, 35 obligate heterozygotes, and 89 patients with CF were used as a training sample for CART analysis to establish a diagnostic rule based on sweat chloride and change in nasal PD from infusion to the final value after superfusion with isoproterenol (⌬Am+Clfr+Iso). The first split based on a value of ⌬Am+Clfr+Iso of 6.5mV has perfect sensitivity and specificity for discriminating CF from non-CF. Further splits based on sweat chloride at 25.5 and 82.5 mmol/L segregate the majority of controls and CFPI patients at the extremes, with overlapping groups in the intermediate range.

Finally, a split based on a ⌬Am+Clfr+Iso of -6.5mV further discriminates between controls and an intermediate group of controls and heterozygotes. By basing the initial split on ⌬Am+Clfr+Iso rather than sweat chloride, the CART rule correctly classifies CF-PS patients with sweat chloride values in the normal range.

Jane L. Burns, M.D.

Antibiotic therapy is a mainstay of treatment for CF patients hospitalized with a pulmonary exacerbation. In order to use antimicrobial agents effectively, it is important to understand the individual patient's microbiology, the patterns of antibiotic resistance within your CF center, the activity of the drugs in CF, and the optimum mode of administration.

Using each patient's culture and susceptibility data as a guide, it is usually possible to identify a combination of two or three agents from different antimicrobial classes (e.g., β-lactams, aminoglycosides, quinolones) that demonstrates in vitro activity against the predominant pathogens. The use of agents from different classes with different mechanisms of action offers the potential for synergistic activity. It also decreases the likelihood of development of antibiotic resistance. Even if susceptibility data are unavailable, knowing the organisms most recently cultured from a respiratory sample may help direct therapy. While Pseudomonas aeruginosa is often susceptible to a combination of a β-lactam and aminoglycoside, Burkholderia cepacia complex is uniformly resistant to the aminoglycosides and Stenotrophomonas maltophilia is never susceptible to the carbapenem antibiotics.

The patterns of antibiotic resistance within a given CF center may also influence the choice of antimicrobial therapy. For example, methicillin-resistant Staphylococcus aureus is being seen increasingly at some CF centers (1) and may be susceptible to clindamycin and trimethoprim/sulfamethoxazole, or perhaps only to vancomycin.

Knowledge of the local resistance patterns can help determine the best choice of agents.

The pharmacokinetics of antimicrobial agents in CF patients may be distinct from those of other individuals. Because many CF pathogens are relatively antibioticresistant, optimizing drug levels at the site of infection in the lung is critical. For drugs where serum monitoring is performed, such as the aminoglycosides, the use of a previously determined dose is optimum, unless renal function has changed. For other agents, where monitoring is not available, the use of CF-specific dosing is recommended (2) .

The route of administration is the final decision that must be made in identifying antibiotics for inpatient treatment. Questions that often arise include the necessity for parenteral rather than oral administration of quinolones and the advisability of combined inhaled and parenteral tobramycin. These decisions may need to be individualized for each patient.

Once therapy has been initiated, clinical response is often adequate to determine hospital course and the advisability of continuing therapy as an outpatient. If a patient is not improving as expected, additional culture and susceptibility testing may be indicated. 

Physiotherapy is one of the cornerstones of treatment in CF and is part of the multi-disciplinary treatment approach. P. aeruginosa may survive longer in airways abundantly filled with sputum, and by removing infected lung secretions the rate of proteolytic tissue damage can be decreased 1 . Combining intravenous anti-biotic therapy with intensified airway clearance physiotherapy has been shown to improve lung function more than IV anti-biotic treatment alone 2 . Higher sputum volumes are often associated with increased inflammation, lung obstruction and destruction. Inflammatory processes increase metabolism and in turn minute ventilation and work of breathing 3 .

There are many reasons why patients with CF may be admitted to hospital. Some of the most common reasons include acute flare up of CF lung disease, intestinal obstruction, preparation for elective surgery and for management of CF complications such and hemoptysis and pneumothorax. The objectives of hospital admissions vary depending on the reason for the admission. In the case of acute exacerbation, the main objective will be to restore optimal physical function as quickly as possible. Broadly speaking, this is achieved through a multi-disciplinary approach including inhalation therapy, clearance of sputum and mucus plugs and pharmacotherapy to reduce infection and inflammation which in turn may decrease the work of breathing. Further, a graduated exercise program to assist in sputum mobilization 4 and increase physical working capacity and musculo-skeletal well-being (including attention to posture, muscle strength and mobility) should commence as soon as appropriate. Nutritional and psychosocial support are also part of the care package. Ongoing age appropriate education about effective and efficient airway clearance, inhalation therapy and the need to adhere to treatment regimens should be incorporated in treatment sessions. Each treatment session should be seized as an opportunity for education. Establishing good rapport, offering age appropriate treatment and gaining the patient and family's trust and confidence are crucial in the delivery of effective treatment. An inpatient admission should include:

• relevant past history • recently recommended home physiotherapy program including inhalation therapy (agents, order and timing), airway clearance therapy (ACT) and physical exercise program and adherence • the possibility of gastroesophageal reflux 5 in relation to physiotherapy • clinical status including subjective and objective measures of the following -amount, color, consistency and ease of expectoration of sputum -oximetry/pulmonary function tests/peak expiratory flow rate -breath sounds on auscultation, respiratory rate and pattern of breathing -exercise tolerance (current activity & incidental exercise/ exercise tolerance tests) -musculo-skeletal problems (posture, pain, muscle tightness/weakness, oedema) -urinary incontinence during coughing and forced expirations

Establishment of a treatment program incorporated in the multi-disciplinary care plan: 1. Inhalation therapy: 6-8 broncho-dilator, mucolytic (pulmozyme, isotonic or hypertonic saline etc) inhaled anti-microbial and anti-inflammatory agents. The timing and order of these treatments in relation to ACT and exercise are important and need to be established. 2. Choice of airway clearance therapy: [9] [10] [11] [12] [13] Positioning Manual techniques Effective forced expirations

The Active Cycle (huffing) of Breathing Positive Expiratory Pressure Autogenic Therapy Drainage

Oscillating Positive Expiratory Non-Invasive Pressure Therapy Ventilation 3. Establishment of an exercise program to:

-Increase aerobic capacity, improve muscle strength and joint mobility -O 2 therapy if needed -Promote normal postural alignment and physical function -Motivate increased exercise participation in the short and long term 4. Discharge planning occurs once the objectives of the admission have been achieved. A home treatment program needs to be negotiated with each patient individually. An early review of progress is usually organized in the CF outpatient department. The dosage and number of treatments per day need to be individualized to each patient's needs. Acutely ill patients may need shorter treatments spread over the 24 hour period which should be available 7 days/week 14 . Treatments should be timed before meals or at least 2 hours after to allow gastric emptying (which is delayed in some patients) 15 in order to minimize gastroesophageal reflux during ACT. Appropriate information needs to be recorded in each patient's unit record to meet communication and medico-legal requirements. The quality of treatment and adherence are crucial in achieving optimal outcomes short and longterm. 

Noreen Roth Henig, M. D.

The majority of ICU admissions for CF occur in the setting of respiratory complications of CF including pulmonary exacerbations from chronic infection leading to respiratory failure, hemoptysis and pneumothoraces. Admission to an intensive care unit (ICU) for progressive respiratory failure has different implications than admission for hemoptysis or a pneumothorax, an extrapulmonary manifestation of CF, or for post-surgical care. Critical care for patients with cystic fibrosis encompasses a large number and wide range of important issues. The critically ill CF patient has needs specific to the underlying disease that distinguishes them from other ICU patients. The CF team is well positioned to guide the intensivists in these issues as well as in endof-life discussions. Factors such as local CF center practices, ICU norms, patient preferences, and the availability of lung transplantation predictably influence admissions and outcomes in the ICU.

In the largest published American series of adult CF patients in the ICU, 74% of the 135 admissions were for a CF specific respiratory event. [1] Extrapulmonary manifestations of CF account for additional ICU admissions, although the exact number is not known. Two such extrapulmonary indications for ICU admission are heat stroke and hematemesis from esophageal varices secondary to cirrhosis. Renal failure due to drug toxicity, narcotic overdoses, and anaphylaxis to antibiotics are unfortunate complications of usual CF care that may result in ICU admissions. In the largest European experience published, post-surgical admissions make up the majority of ICU admissions (18 of 31 admissions). [2] In that series, 16 of the 18 admissions were for pleurodesis. In other centers, abdominal and pelvic surgeries are more common indications for ICU admission. Although published experience in this area is limited, it appears that CF patients admitted to an ICU for any indication other than progressive respiratory failure are likely to survive to discharge from the hospital.

Although most physicians feel comfortable utilizing ICU resources for potentially reversible problems in CF patients, progressive respiratory failure presents a unique challenge. An early multicenter study of patients who received endotracheal intubation with mechanical ventilation for respiratory failure showed abysmal outcomes: 85% of patients died within 6 weeks of intubation, either on or off the ventilator. Heavy sedation and immobilization of the patient often accompany endotracheal intubation and mechanical ventilation and compromise airway clearance. Suctioning of secretions through an endotracheal tube, either blindly or with a bronchoscope, is less effective than a cough and expectoration in an awake patient. In addition, patients who cough while intubated are perceived as "bucking" the vent, uncomfortable, and at risk of self-extubation, and are commonly treated with heavier sedation and/or neuromuscular blockers. The combination of poor airway clearance and the potential for the development of myopathies from immobilization and drugs may actually contribute to poor outcomes in CF patients.

Non-invasive mechanical ventilation (NIV) provides ventilatory support to patients with respiratory failure without the limitations of sedation and immobility imposed by endotracheal intubation. The advantage is that patients can cough and expectorate while using NIV. It is increasingly used in patients with CF for a variety of indications and in both inpatient and outpatient settings. The uses of NIV are beyond the scope of this discussion, but its increasing use and acceptance has changed how

End of life care in cystic fibrosis has undergone a revolution in the past decade. Prior to the 1990s, end of life care for patients with advanced CF lung disease was based on the assumption that aggressive interventions were inappropriate; the advent of lung transplantation unraveled this consensus about the use of aggressive care at the end of life. There is anecdotal evidence that patients with end-stage CF lung disease may now see their choices as a stark dichotomy: either aggressive technologic intervention in the hopes of life extension, or acceptance of limited lifespan in order to receive high quality palliative care. This "either transplantation or palliation" model fails to recognize that the goals of pursuing restorative therapy and palliation are not by their nature in conflict, and that a mixed model of end of life care, incorporating the new realities of lung transplantation, is possible.

Although patients and physicians may not explicitly acknowledge it, opting for lung transplantation is certain to alter care in the final months of life. Patients on the transplant list are more likely to opt for aggressive measures to sustain life in order to increase the chance of receiving a suitable organ. As a result, transplantation requires a reordering of priorities for patients and families, and a re-envisioning of the good death. It is unrealistic to assume that patients and their families have given no thought to the matter; autobiographical accounts of life with CF contain frank discussions about the circumstances of dying, both before and after transplantation. It is clear that the topic, while difficult, is important to patients, and it is clear that patients and families can both consider transplantation and make end of life plans.

Consideration of transplantation requires that the patient, family, and physician directly confront the lethal and where CF patients are treated. For instance, in the American study of CF patients in the ICU, almost a third with infective exacerbations were treated with NIV. In some hospitals, NIV is used on the general medical ward and patients who fail NIV do not opt for endotracheal intubation.

All patients with CF in the ICU require skilled attention to CF specific care, including attention to nutritional needs, enzyme supplementation, bowel care, drug clearance, hydration, and diabetes management. The nutritional needs of CF patients include overall high caloric needs, the decision between parenteral and enteral feeding, and the use of powdered enzyme supplements through feeding tubes vs. elemental formulas. Distal intestinal obstructive syndrome is a potentially devastating complication of ICU care and results from immobilization, dehydration, narcotic analgesics, and the difficulty in determining appropriate enzyme doses in enterally fed patients. Hydration is a difficult management issue in CF patients who are prone to dehydration but who may have pulmonary edema as a contributor to the respiratory failure. Drug metabolism, especially of aminoglycosides, presents unique challenges and often results in the need for dose alterations. A final issue is diabetes management. Critical illness may worsen glucose control in CF, even in patients previous not recognized to have impaired glucose tolerance or CF-related diabetes. Critically ill CF patients likely benefit from the same aggressive glucose control as other critically ill patients. [3] 

Inevitably, discussions of end-of-life occur in the ICU. Ideally, the CF center team is aware of a patient's preferences regarding end-of-life care including aggressive interventions such as ICU admission, mechanical ventilation and lung transplantation, although when faced with imminent death, a patient's perspective on end-of-life issues may change. Additionally, the availability of transplantation, both lung and liver, may change the decision making paradigm, and even the local practices of the transplant center may influence end-of-life care. Ultimately, the guiding principle that there is no "one-sizefits-all" approach to CF and critical care allows flexibility and individualized care at a most challenging time. nature of cystic fibrosis. Physicians, eager to provide hope for recovery, may downplay the struggle that transplantation can be for patients and families, yet this does patients and families a great disservice. Discussion of transplantation needs to be initiated at a point that allows adequate time for the family and patient to come to terms with the threat of approaching death and the uncertain prospect of transplantation.

Since administration of the usual therapies for CF only rarely causes distressing symptoms for patients, caregivers may not face difficult decisions about discontinuing therapies. At the end of life, intravenous or inhaled antibiotics, pancreatic supplements, oxygen, and the rest of the usual daily therapies for CF can be continued in most cases without a great burden to patients both on and off the transplant waiting list. There is one notable example: if BiPAP has been instituted as a measure of symptom relief, its removal may be both psychologically and symbolically difficult for the patient, family and clinician; the decision to remove assisted ventilation may be a visible sentinel of the approaching end of life or of the "failure" of transplant listing.

Pain control. Initiation of pain control therapies such as opiates may have also have powerful symbolic importance for the patient family and clinician. Yet skilled pain control is simply the duty of compassionate care, and concerns over the symbolic nature of opiate use should be so great that the result is a patient in avoidable pain. The common fear of respiratory suppression following use of opiates in chronic lung disease is unfounded, and concerns over addiction in end of life care have been dispelled by years of experience with their use in a wide variety of palliative situations. Retrospective studies suggest that the frequency of headache and chest pain in CF patients increases steeply in the final year of life, and opiates should be used in sufficient amounts to relieve both acute and chronic pain.

Hemoptysis. Patients with a history of marked hemoptysis may have a major bleeding episode as the final event in life. This possibility should be discussed and plans made as to what actions will be taken. Simple measures such as a supply of dark towels and plans for rapid control of the symptoms of dyspnea and choking can lessen the suffering of both the patient and the family in such an event.

Respiratory Failure. For most patients, death will follow a gradual decline in ventilation, with the accompanying "narcosis" of elevated CO2 levels. Yet for many, there can be waxing and waning consciousness, as well as changes in work of breathing, and pharmacologic therapies for control of dyspnea will be needed. Although there is no prospective research on the use of opiates to control the dyspnea of CF respiratory failure, the use of opiates in other deaths due to COPD strongly suggests that moderate doses of opiate can provide excellent control of end-stage dyspnea.

There may well be other family members with CF, or a previous death from CF in the family. These deaths may have formed a "script" for the family as an example of either a good or bad death, and an open exploration of the family's view of the previous death can be invaluable in uncovering unspoken expectations. In addition, there is often a tight circle of friendship among patients with CF, and the death of one member of the circle will have powerful and long lasting effects on the other patients, particularly if death occurs suddenly.

Clinicians too may have a "script" of how they expect a patient's dying to proceed, and should be aware that their expectations may conflict with those of the patient or family. Clinicians should openly discuss their approaches to end of life care and transplantation with one another. Consultation with palliative care teams within the hospital can bring a valuable perspective. 

Until recently, most CF clinicians have operated on the assumption that the care they provided was uniformly excellent. However, it has become increasingly clear that there are differences in outcomes among different CF clinics, suggesting that there are more and less effective approaches for providing CF care. Thus, while we await the expected advances in care that will be provided in the long run by breakthroughs in biomedical research, we have the opportunity to take better advantage of existing knowledge to rapidly improve the outcomes for people with CF. There are several key strategies that healthcare providers can use in order to make positive changes in the care they deliver. These include:

1. Appreciate that changes must be made to the system of healthcare delivery.

The first step, and one that is often the most difficult for physicians, is to understand that simply working harder within a non-supportive system will not yield the results desired. The model of the individual physician who by force of intellect and will establishes the correct diagnosis and prescribes the appropriate therapy to cure a patient is anachronistic and does not apply to the care of patients with CF. In caring for a chronic disease like CF, multiple caregivers must communicate and integrate a complex set of data and then prescribe therapy based on the appropriate use of that data. While it is incumbent on the system to ensure that providers are knowledgeable regarding best, or ideal practices, it further needs to support consistent application of those interventions that the provider knows to be optimal. Variation in outcomes (when adjusted for variation in risk) is then due to variation in the system's ability to provide this support in a consistent manner.

In general, CF care providers have successfully adopted a longitudinal approach to dealing with this chronic disease. In addition, it is useful to conceptualize and work towards instituting an idealized system of chronic care delivery that consists of various interdependent components inside and outside the local clinic setting. Wagner's chronic illness care model provides a useful framework for consideration and may be summarized as follows (see also http://www.improvingchroniccare.org): a. Community resources -Medical center-based subspecialists should form partnerships with community organizations and primary care providers to support and develop interventions and fill gaps in needed services. Furthermore, providers should publicly advocate for policies that improve patient care. b. Overall health delivery system -Organizations should create a culture that promotes safe, high quality care. All levels of the organization should visibly support improvement and provide incentives based on quality of care. There should be open and systematic handling of errors and quality problems and care should be coordinated within and across organizations. c. Patient and family self-management -Individual resources should be exploited by training and empowering patients and their families in self-assessment, goal-setting, action planning, problem-solving and follow-up. Furthermore, patients' input should be sought in planning all aspects of care as well as delivery system design. d. Delivery system design includes the structure and function of the clinic, from the telephone to the reception area to the examination room. It is essential to define roles and distribute tasks among team members, use planned interactions, ensure regular followup, and give care that patients understand. e. Decision support comprises the promotion of evidence-based care at the provider-patient interface. This is accomplished via the use of clinical tools, guidelines and algorithms to ensure that intended care is actually prescribed, and reliance on rote memory is minimized. f. Clinical information systems -At the individual patient level, the system might provide timely reminders for providers and patients, facilitate individual patient care planning, and allow data to be easily shared in order to coordinate care. At the clinicwide population level, the system can help to identify relevant subpopulations for proactive care, and allow providers to monitor performance of the practice team and care system. It is the lack of the latter data that keeps many providers in the dark regarding the true effectiveness of their care.

An effective organizational change strategy is an essential component of improvement work. Without a disciplined approach, practitioners who are newly aware of the extent of their system's deficiencies will attempt immediate, large dramatic changes which fail either in their planning stage because they get bogged down in endless preparatory meetings or self-destruct in their implementation phase because of the number of unanticipated problems encountered. One effective approach, is the Langley et al's Plan/Do/Study/Act (PDSA) cycle. To implement the process, the first step is to plan the details of a small test of change [plan] . The planned change is then carried out [do] . Once the change is attempted on a small scale, data on its effectiveness is gathered [study]. Following discussion of what was learned by the initial endeavor, the change strategy is then modified and reattempted [act] . Through repeated use of this cycle, small hunches, theories, and ideas gradually coalesce to into significant changes that consummate in significant improvement. The essential key to success is the use of small changes that are easily accomplished, followed by the analysis of data to evaluate the impact of the intervention.

The use of data on performance is essential to recognize where opportunities for improvement exist, and also in order to garner feedback on what changes truly result in improved outcome. Once an organization determines specific actions that must take place in order to improve outcomes, it needs to track the consistency with which those actions are taken. Improved performance on these "process" measures can be measured as a preliminary step to improvement in the outcome measures that are the true goal of the work. Feedback must be received promptly and on a regular basis, and data should be reported visually in a way that can be understood and used by members of the care team as well as interested outsiders.

The synergy that derives from collaboration among workers investigating the same problem is well known to scientific researchers, the most successful of whom are typically embedded in networks of cooperating laboratories within and outside of their home institutions. This strategy is equally effective for the development and spread of innovations for improvement in the delivery of health services. The most commonly used cooperative model is one that recognizes "best practices" and attempts to copy and spread approaches used by providers whose outcomes are the best within their field. However, novel and effective ideas for how to accomplish certain specific goals exist even at centers whose overall performance is average, especially if they are actively striving to improve their outcomes. Thus, collaboration among various centers and healthcare workers who are trying to accomplish the same or similar goals is an important and effective strategy to accelerate change.

Supported in part by CFF SCHECH01C0QI.

The steady increase in median survival for those with cystic fibrosis (CF) is, in large part attributed to CF center-based care and the introduction of new therapies. However, variation in practice patterns and utilization of existing therapies, as well as patient outcomes, suggests less than optimal application of existing knowledge in caring for those with CF. Epidemiologic data identifies "average" CF care patterns and outcomes, but does not inform CF care providers on strategies for benchmarking above average, or "going beyond good enough".

Improving CF center care using existing therapies and scientific knowledge requires thoughtful examination of current care patterns to identify potential areas for improvement. The National Initiative for Children's Healthcare Quality (NICHQ) recruited 16 CF care centers of varying demographics across the United States to create a collaborative group. This group focused on two goals to improve CF center care: Improving nutrition and reducing environmental tobacco smoke (ETS) exposure. This collaborative group of centers spent eighteen months working towards these goals utilizing the Care Model for Child Health.

The experience of one pediatric CF Center within the collaborative is presented demonstrating successful strategies for improving outcomes in nutritional status and reducing ETS exposure, through appropriate process changes, partnership with families, consistent aggressive nutritional interventions, and application of existing knowledge. Practical application of elements within the Care Model for Child Health incorporated within the collaborative are shared including changes in clinical information systems, decision support, delivery system design, family education, and a prepared, proactive CF care team.

The results of the collaborative efforts resulted in a steady increase in median weight percentile for all children, and an even greater increase in those children in nutritional failure at the first recorded encounter. Key factors in the achievement of the collaborative goals included: nutrition focus and process changes, and engaging families and patients to teach about enteral supplementation options. 4. Preview of patients prior to clinic, in addition to postclinic review. 5. Improving self-management by providing specific nutrition education (monthly newsletters) and written self-management plan at each visit.

Targeted efforts to reduce environmental tobacco smoke (ETS) exposure resulted in a 75% documented "no smoking policy" for all environments for children within the center, and a 25% "quit-rate" among parental smokers. A simple three-pronged strategy of: a) systematic assessment of ETS, b) clinician education in tobacco cessation counseling, and c) identification and provision of appropriate support materials to families, allowed for significant success in the overall reduction of ETS exposure.

The experience of this center demonstrates the potential to improve the care and outcomes of children with CF through systematic application of existing knowledge and attention to basic tenants of the CF center-based interdisciplinary care model. Identification of variation in practice allows for focused interventions in areas where changes can, and will drive CF care beyond "good enough".

It is the goal of all health care practitioners to render optimal care to those who seek their assistance. To this end, the Learning Leadership Collaborative of the Cystic Fibrosis Foundation has begun the promotion of the education of our community in the processes of Quality Improvement. As the project for the mini-ATP course at Intermountain Health Care we developed a standardized approach to relate to our patients the importance of their nutritional status with the goal of having no patients in need of urgent nutritional rehabilitation by June of 2004.

Noting the great variance in the rate of nutritional compromise among the many CF centers, and accepting the concept that a calorie will function similarly regardless of geography, it was our hypothesis that raising awareness of nutritional status in our patient population, when coupled with specific recommendations for nutritional intervention, would result in a significant improvement in the nutritional state of our patient population.

In September of 2003 we stratified our pediatric and adult population into three categories (green, yellow, and red) based on weight/height percentile for patients less than 2 years of age, BMI percentile for pediatric patients greater than 2 years of age, and BMI in adult patients equal to or greater than 20 years of age. The three categories of green, yellow, and red rendered visual reminders to institute category and age determined protocols for nutritional maintenance or rehabilitation. A flowchart was created to assist in accurately stratifying each patient at each visit.

In October of 2003, we began universal distribution of growth or BMI charts at all regular visits. We also began institution of a "calorie contest" designed to raise the awareness of patients and their families regarding the caloric content of varied food combinations. Those found in need of nutritional rehabilitation were treated in accordance with established CF guidelines. We tracked our adherence to our own program by recording all of the above interventions in a flow sheet placed at the front of each of our patient's paper charts. Our team rapidly adapted to our new formalized process and we were able to achieve greater than 80% adherence to all of our interventions by 12/1/03.

To monitor the effect of our interventions we created graphical displays of the monthly total of patients within each category as % of total monthly attendees. These data were graphed within control limits calculated from retrospective analysis of monthly data for 6 months prior to our interventions. Upon reviewing our statistical process charts of our outcomes up until June 2004, we have not achieved the lofty goal of remitting our entire patient population from nutritional compromise. However, there has been notable improvement in both the pediatric and adult program outcome data as manifested by more visits per month with patients classified as "green zone". The data indicate a more prominent effect in the adult population, which has prompted us to readdress our process to further augment our positive outcomes.

This process has served to enhance the nutritional status of our CF patient population and has provided a mechanism by which we may continue to improve many aspects of patient care.

In our lives, sickness comes to loved ones, accidents leave their cruel marks of remembrance, and tiny legs that once ran are imprisoned in a wheelchair. Mothers and fathers who anxiously await the arrival of a precious child sometimes learn that all is not well with this tiny infant. A missing limb, sightless eyes, a damaged brain, or the term "Cystic Fibrosis" greets the parents, leaving them baffled and filled with sorrow. Sixteen years ago I had visions of what parenting a child would be like. My visions were based on hopes and dreams, which didn't include a life threatening illness. I remember waiting in a tiny room with staunch white walls that seemed to enclose us with its smells of alcohol. My husband and I anxiously waited the arrival of the pulmonary doctor as we nervously passed our two month old back and forth between us. I remember the doctor's red, white and blue tie, his graying hair and the subtle way he cleared his throat as he pronounced, "Your son has Cystic Fibrosis." I felt isolated, overwhelmed and numb as we left the hospital corridor. There followed the inevitable blaming, the condemnation, and the perennial questions: "Why such a tragedy in our family?", "How did this happen?", "Where was God?". Why, where, how-those recurring words-did not bring back our lost dreams, our son's perfect body, and the plans we had as parents. I have learned that self-pity, personal withdrawal, or deep despair will not bring peace, assurance, or help, which I needed to courageously, continue on my parenting journey. Rather, I had to find a way to go forward, look upward, and move onward.

Sixteen years later The Cystic Fibrosis Parent-to-Parent Support Network is proving to be just this type of help by offering comfort, hope and encouragement from parents who are reliable allies that immediately understand the situation because they have been through it themselves.

Imagine the atmosphere of a room filled with ninetyeight mothers of children with Cystic Fibrosis. Each of these mothers was experiencing a different place on the Cystic Fibrosis journey; some were newly diagnosed, and others were more experienced dealing with various struggles of CF. Each mother was given the opportunity to ask a question or give advice structured around the challenge of Cystic Fibrosis; "How do I get my threeyear-old to sit still while taking a breathing treatment?", "I feel like all I do is give breathing treatments, chest physical therapy and medicine all day. What can I do to manage my time better?", "How can I get my teenager to take his enzymes when he goes out with friends?". How did they get support that day? A connection was made between mothers; the loneliness and despair were replaced with friendship and companionship.

The Robert Wood Johnson Pursuing Perfection Grant supporting healthcare quality improvement at Cincinnati Children's Hospital encouraged focus on the viewpoint of patients and parents and perspective of their needs in an effort to improve care. A qualitative study of parent to parent self-help conducted by parents and researchers and published in the Journal of Pediatric Psychology (1) determined that an overwhelming majority of parents were positively helped by parent to parent support. This type of support was shown to increase parents' acceptance of their situations and their sense of coping. Parents were also able to define their roles in the healthcare process, perfect their self-management skills and in turn feel more confident in caring for their children. With support from the Cincinnati Children's Hospital and the Pursuing Perfection Grant, and knowledge and experience from studies (2,3), a core group pf patients, parents and CF Center staff was formed. The core group identified a void in support that the parents could fill. The next step was to promote a Parent-to-Parent support network with the goals of providing strategies for improved patient care, tips from other parents, coping skills, and hope and comfort.

I don't want to make it sound like things just fell into place, because in actuality, it took a lot of hard work. As with any change, there were concerns and conflicts. We experienced challenges but we were able to work through them. The legal department wondered whether parents would be compliant with HIPPA. There were concerns with training and how well the staff and parents would work together. Any of these concerns could have been used as an excuse to quit. But, as John F. Kennedy said, "The problems of the world cannot possibly be solved by skeptics or cynics whose horizons are limited by the obvious realities. We need…to dream of things that never were." We have a great team of people who, despite the difficulties, are committed to succeeding and reaching a common goal. In fact look at what we have been able to accomplish so far.

One-to-one support is a personalized type of support that is made available to parents particularly during periods of increased stress, such as at the time of diagnosis. Parents are carefully matched with a mentoring parent who understands their situation. Our team consists of

Tracey Blackwelder Cincinnati Children's Hospital, Cincinnati, Ohio twelve diverse mentoring parents who were selected by the healthcare team. These selected parents were trained in communication and listening skills.

Quarterly Newsletter is a quarterly parent-to-parent support newsletter that is mailed to all families of patients of the Cincinnati Cystic Fibrosis Clinic. This newsletter is coordinated and edited by parents, and is written with the parent's perspective. It is an effective way to stay in touch with parents and it provides continuous uplifting stories, updates, self-management skills, and influential hints and tips that help parents feel motivated and successful.

Social functions offer an opportunity to share experiences, strengths and hope. These functions enable parents to connect with others and find their way out of isolation, frustration and hopelessness and turn their challenges into positive forces. Socializing also provides a venue for information regarding successful care strategies.

Promoting awareness is an important component of parent-to-parent support to reach families. Stewart Henderson Britt said, "Doing business without advertising is like winking at a girl in the dark. You know what you're doing, but nobody else does." We connect with parents through brochures, voice mail, list-serve web service and e-mail.

In summary, the parents and staff are working together as a team. Just like two oxen yoked together can accomplish more than they can separately when working for a common goal, our combined team of parents and staff are accomplishing amazing things as we work together to "change the outcome" in our pursuit of perfect care.

Ainbinder,J., Blanchard, L., Singer, G.H.S., Sullivan 

The US Cystic Fibrosis Foundation Patient Registry has been collecting annual data from over 20,000 patients cared for at accredited CF care centers since 1982. This data has been used to document the steady improvement in the survival of people with CF, and improvements in lung function and in nutritional status. How can we now use it to identify opportunities for change and to monitor progress in the care of people with CF?

For the past 5 years we have examined the variation in treatments and outcomes between care centers and given feedback reports from the Registry to the centers on their processes and outcomes. The amount of variation at the center level is completely invisible to providers who care for patients one at a time in a busy clinic, so these reports are an important first step towards understanding each center's care and outcomes in the context of CF care in the US. Clinicians want to know how they are doing; eventually patients and families are going to demand to know as well.

There has been a steady improvement in those measures that have been highlighted in the reports. In the past five years, the rate of screening for CF related diabetes nationally has gone from 60% to 82%, and in 2002 six centers reported screening 100% of appropriate patients. The proportion of patients adhering annually to standards for numbers of pulmonary function tests, sputum cultures, and outpatient visits continues to rise. However, the center reports are limited to working within the framework of the Clinical Practice Guidelines for Cystic Fibrosis. The Guidelines summarize the best evidence for different aspects of CF management, but they also define those areas of clinical care that experts consider important.

There is center level variation in every aspect of patient care: Antibiotics are used chronically in over 60% of patients at some centers, zero percent at others (see figure on next page); TOBI is used in 23 to 100% of Pseudomonas aeruginosa positive patients; Pulmozyme is used in 16 to 100% of patients; annual hospitalization rates vary from 0 to 80% of adults, with median lengths of stay from 2 to 15 days; and home IV antibiotic use rates range from 0 to 71%, with median courses between 5.5 and 29 days. Nutritional processes and outcomes are similarly variable.

Does all this variation matter? We have started the process of answering the question. "Are some centers the 'Best' and what does that mean for patient care?" By examining lung function and nutrition outcomes for children and adults annually from 1998 to 2001, ranking centers and selecting those that were reliably in the top 10 in several categories we identified 7 centers with consistently good outcomes across patient groups. Since 1997 the median age of death is 28.8 years for the "Best" centers compared to a mean of 23.2 years for the rest of the country, and median survival is almost 7 years longer (p<0.0001).

This analysis is the basis for the CF Foundation's vision statement: "We believe that the life expectancy of CF patients can be extended by 5 to10 years through the consistent implementation of existing evidence-based care."

What are these centers doing to be so successful? The "Best" centers are diverse; small and large, academic, for profit, rural and urban. The Registry can suggest areas in which to focus further study but it does not collect the level of information that is required to completely describe exemplary CF care. How could a clinically useful data instrument describe a center where there are fail safe systems in place, where every patient gets everything they need at every visit, where handoffs between care givers are smooth, where culture results are never lost, where patients and families are well informed, activated, responsible, and supported? The Registry allows us to report what is currently achievable by some clinical centers and sets the performance possibilities for the present. The highest performing centers don't necessarily use more healthcare resources than other centers. Benchmarking visits by clinic teams may help us to understand the fine detail of care that produces the greatest success.

What can be done right now to accelerate the improvement in CF care that we have seen over the past decades? Historically, the Registry has been used for epidemiologic purposes; the conversion to a patient management and quality improvement tool requires that the Clinical Practice Guidelines be written in such a way that adherence is quantifiable. We must keep the Guidelines up to date for capturing and measuring exemplary CF care. Already PortCF helps track individual patient results, and has some important patient management and reminder tools. PortCF is being enhanced so that center level reports will be available much more quickly, and eventually, the reports will be available as the data is collected. But data that is not examined is like a tree falling in the woods.

The Patient Registry gives a broad view of CF care in the US and allows us to track progress and focus our attention on achievable goals. Variation is a treasure, but it does not yield its secrets easily and to learn what creates success is going to require the concerted efforts of all the players -clinicians, patients and families, basic scientists, and the Foundation.

Cystic Fibrosis Foundation Patient Registry 2002 Annual Report to the Center Directors, Bethesda, Maryland.

The year 2003 was a one of dramatic change for the CF Foundation Patient Registry. First, the implementation of new regulations pertaining to patient privacy necessitated that all care centers gain approval from their local Institutional Review Boards (IRBs) for submission of data to the Registry. A formal consenting/ assenting process was required at all but 4 care centers and 3 affiliate centers. Despite these logistical challenges, 2003 Registry data was submitted from 195 sites including 118 care centers, 24 adult programs with distinct registry center numbers, and 53 affiliate centers. Patient retention in the Registry was excellent. We observed a modest (approximately 6%) drop in patient numbers as compared to 2002, but fully expect the numbers to rebound in 2004. The care center teams deserve accolades for their work on behalf of the Patient Registry.

The transformation of the Patient Registry from the collection of year-end summary data to the web-enabled, encounter-based Port CF application was the other major change in 2003. The web-based application has opened up new challenges and opportunities. A number of resources can now be placed at the fingertips of our care teams including consensus documents and patient education materials. Ongoing data entry throughout the year provides the potential for real time feedback of information to the care team. For example, the current iteration of Port CF includes reporting features such as the Patient Roster, Patients Due for a Visit, and the Patient Summary Reports. The summary report displays complications, respiratory microbiology, and graphical trends of pulmonary function and nutritional status. The query tool provides the capability to identify subsets of patients, e.g., those with CF-related diabetes, which require special monitoring. Our intent with these reports and query capabilities is to provide access to patient and center level information that will facilitate health care delivery.

The future for Port CF holds even greater promise. Our quality improvement activities are focused on three core care areas in which suboptimal outcomes are associated with morbidity and mortality: 1) growth and nutrition, 2) pulmonary, maintenance and exacerbations, and 3) CFrelated diabetes mellitus. The significant variability in these outcomes throughout the care center network suggests that they are high leverage areas for accelerating the rate of improvement in CF care. We are currently collecting data in Port CF on the pertinent medical outcomes in these areas, but many of the associated practice patterns are not being captured. Collecting this data throughout the year and making individual and aggregate information about these practices available to the care centers will be instrumental in supporting further improvement activities. Decision support in the form of reminders and algorithms can also be incorporated for key care decisions.

Further enhancements to Port CF will occur in concert with an updating of our clinical practice guidelines. Our goal is to create a dynamic process with better integration of Port CF, the practice guidelines and clinical care (see Figure) .

Standing committees will be formed in the three core care areas. The Growth and Nutrition Committee has recently been constituted and will be the first area for enhancements to Port CF. One of the first tasks for the committees will be to develop a list of strong recommendations for care based on the scientific evidence and clinical experience. Work done in our quality improvement collaboratives will also inform the process. The strong recommendations of the guidelines committees will serve as a basis for new questions and reports derived from the new data for Port CF. While striving to incorporate value-added data fields into Port CF, an important consideration will be the impact of data collection and data entry on care centers.

In summary, Port CF is a critical tool for supporting improvement activities and for tracking the outcomes of those efforts. Further evolution of Port CF will occur in concert with ongoing quality improvement activities and an updating of our clinical practice guidelines. The continued success of the Registry is dependent upon the commitment and hard work of the care center teams and the willingness of patients and families to share their data.

The response to recurrent injury, in liver and in other organs is one of wound healing. Many different types of injury (i.e., chronic hepatitis, ethanol, biliary tract disease, cystic fibrosis, iron or copper overload, etc.) lead to injury and ultimately to hepatic fibrosis and cirrhosis.

Cirrhosis in turn leads to hepatocellular dysfunction and portal hypertension, each of which have grave clinical sequelae. Cirrhosis in cystic fibrosis results in the same clinical complications as is found in patients with other forms of liver disease.

Extensive investigation over the past 20 years has established that the effector in the liver wound healing process is the hepatic stellate cell (also known as a lipocyte or Ito cell). A central feature of the wounding response to liver injury is the transformation of resident stellate cells from a "quiescent" (normal) to an "activated" (injured liver) state ( Figure 1 ). Characteristics of this transition include morphologic and functional changes. Morphologic changes include loss of vitamin A, acquisition of stress bundles, and development of prominent rough endoplasmic reticulum (1) (2) (3) (4) . One of the earliest described and most prominent effects of stellate cell activation is production of increased quantities of extracellular matrix, including types I, III and IV collagens, fibronectin, laminin and proteoglycans, some of which are increased by greater than 50-fold (3). Further, the available evidence now indicates that the overall increase in extracellular matrix protein deposition typical of cirrhosis can largely be ascribed to excess production by stellate cells (5) . Thus, a great deal of emphasis has been placed on elucidating mechanisms underlying stimulation of fibrogenesis by stellate cells. A number of events, typically acting in concert, play a role in stimulating stellate cell fibrogenesis (6) (7) (8) (9) (10) (11) (12) (13) . For example, stellate cells produce chemotactic factors (14) , appear to be important in matrix remodeling (15) , and produce a number of "vasoactive peptides" (8, 9) , including the potent vasoconstrictor, endothelin-1 (ET-1) (9, 16). Indeed, ET-1 and vasoactive peptides have emerged as important components of the fibrogenic response.

An important emerging theme is that current methods for precise diagnosis and quantitation of the effects of therapy are required. Liver biopsy remains the gold standard for assessment of liver fibrosis. However, it is not without potential problems. Therefore, a number of methods have been used to non-invasively assess fibrosis. These include imaging techniques, non-specific or indirect markers of liver fibrosis, and tests that more directly assess the fibrogenic biology present in the liver. Such non-invasive tests are likely to emerge as critical components of the clinical armamentarium in management of hepatic fibrosis.

A number of therapies (i.e., colchicines, ursodeoxycholic acid, PTU, vitamin E, etc…) have been tested in patients with fibrosing liver disease. For the most part, these approaches have proved to be ineffective. However, newer preclinical studies have highlighted a number of putative therapies to abrogate fibrogenesis. Such therapies are targeted at reducing or removing the primary injury, modulation of stellate cell activation, stimulation of matrix degradation or stimulation of stellate cell death. For example, interferon gamma, a cytokine with potent immunomodulatory activity, has profound effects on multiple aspects of stellate cell activation. These data have led to clinical trials in humans in which interferon gamma is being examined in patients with fibrosis. Other specific, mechanism based, "anti-fibrotic" therapies are rapidly emerging. Liver disease is a major complication in many patients with cystic fibrosis (CF). The precise prevalence of CFassociated liver disease (CFLD) is difficult to determine as no reliable method of diagnosis currently exists. Traditional measures of liver function do not correlate with the severity of liver disease and there is no specific CF genetic association. Thus the diagnosis is not usually made until cirrhosis or complications such as portal hypertension are established. While clinical and imaging studies may detect these late changes, and invasive liver biopsy, which is not always reliable, due to the focal nature of fibrotic changes might aid diagnosis, the challenge exists to develop a sensitive, reliable non-invasive assay to identify CF patients who are at risk of developing severe liver disease, which might also serve as a guide to monitoring progression or response to therapy. Although the basic mechanisms remain unclear, abnormal bile transport and biliary fibrosis implicate abnormal biliary physiology in the pathogenesis of CFLD. We have documented a definitive role for Hepatic Stellate Cells in the fibrogenesis associated with CFLD, and suggested a potential mechanism for the induction of HSC collagen gene expression, through the production of the cytokine TGF-beta(1) by bile duct epithelial cells (1) . Accumulated bile acids have been proposed as potential mediators of cytokine production, and , indeed, we recently found a correlation between both cholic acid and cholic acid/chenodeoxycholic acid levels with histological liver injury and fibrosis progression (2) .

A potential monitoring role for serum bile acids and serum measures of components of hepatic matrix remodelling, which may serve as markers of early fibrosis is suggested by these studies. We thus have evaluated, by preliminary cross-sectional study, the utility of selected serum markers ie bile acids, hyaluronic acid (HA), collagen type IV (CL-IV), prolyl hydroxylase (PH), matrix metalloproteinases (MMPs), and tissue inhibitors of MMPs (TIMPs), by their correlation with the degree of hepatic fibrosis in patients with CFLD. We compared with serum levels from patients with CF and no liver disease (CFnoLD), and age matched control data. Methods: Bile acid composition was determined by gas-liquid chromatography/mass spectrometry in bile, urine, and serum samples from 30 children with CFLD, 15 children with CF but without liver disease (CFnoLD), and 43 controls. Sera from 36 CFLD, 37 CFnoLD and 38 control subjects were assessed for TIMP-1, CLIV, MMP-2, HA and PH by enzyme immunoassay. Serum markers were correlated with blinded hepatic fibrosis score on liver biopsies from 30 CFLD subjects.Results: There were significant correlations between serum cholic acid levels and the ratio of serum cholic acid/chenodeoxycholic acid with hepatic fibrosis score, the degree of inflammation, and limiting plate disruption. TIMP-1, PH, and CL-IV In CFLD, there was a negative correlation between fibrosis score and both TIMP-1 (r = -0.39, p = 0.06) and PH (r = -0.48, p = 0.008). There were no differences in TIMP-1, CL-IV, HA or PH between CFnoLD and controls, nor in MMP-2 between the three groups.

The correlation between both cholic acid and cholic acid/ chenodeoxycholic acid levels with histological liver injury and fibrosis progression suggests a potential monitoring role for these bile acids in CFLD. Elevated levels of serum TIMP-1, CL-IV and PH in CFLD suggests that these may be indicators of fibrogenesis in CF.

Higher levels of TIMP-1 and PH were associated with mild fibrosis, and these, in concert with serum bile acid measurements may be useful markers for the early detection of CFLD. We suggest that these serum markers be subject to longitudinal study in CF as markers of progression of CFLD, and/or response to future antifibrogenic therapies. 

Clinical and biochemical parameters have been shown to be very insensitive in the identification of cystic fibrosis associated liver disease (CFLD) (1). Thus, many centers have begun to investigate the use of radiologic screening for CFLD. This is challenging in that the histologic findings with CFLD can range from steatosis through early fibrosis to biliary cirrhosis (2) . The primary therapy that is currently available, ursodeoxycholic acid, has been suggested to be most effective in early disease (3, 4) , but may improve advanced CFLD (5) . Thus, the challenge for treatment remains the identification of "early" CFLD for potential treatment and the identification of "silent" advanced CFLD to institute appropriate anticipatory management. To that end, radiologic studies offer noninvasive imaging to assess the liver. This summary will discuss ultrasonography, CT scanning, MR scanning and radioisotope excretion studies.

Ultrasonography (US) utilizes short pulses of high frequency sound that are reflected back at the interfaces of tissues of different acoustic properties. US is low in cost, portable and involves no radiation exposure. It is limited by the inability to penetrate bone or air which can prevent a complete examination of abdominal organs and occasionally by the depth of penetration required.

US is the imaging modality of choice to screen for biliary disease in CF. Cholelithiasis and biliary dilation can easily be determined. US has also been suggested as a tool to detect early changes in liver architecture in CFLD (6) (7) (8) . US has an excellent sensitivity and specificity for end stage cirrhosis with portal hypertension (9) as the findings of a nodular liver, splenomegaly and varices are generally due to cirrhosis. However, the goal of screening for CFLD would be to identify early liver disease prior to the progression to cirrhosis. Two ultrasound classifications for CFLD have been proposed (6, 8) :

In Montreal, they have demonstrated minimal interobserver variability in the scoring and interpretation of US (8, 9) . Abnormal US findings were more common in patients with abnormal AST, ALT or GGT. However, only about 50% of patients with abnormal AST, ALT or GGT had an abnormal US and 10-20% with of patients normal values had an abnormal US (9) . No study has investigated inter-center variability using either scoring system or the correlation between early US findings and liver biopsy evidence of fibrosis in CF.

US can be used to follow patients for the development of US abnormalities. In a longitudinal US study by Leanerts, serial US was performed for an average of 6 US per patient (8) . Of the 106 normal patients at the start of the study, 18 (17%) had any abnormalities identified during longitudinal screening. Of these, 5 developed and continued with a heterogeneous liver, 6 developed a nodular liver and 8 progressed to portal hypertension. Kaplan Meier risk analysis predicted that approximately 15% of subjects would develop a nodular liver on US by 18 years of age. In this study, Thirty subjects received ursodeoxycholic acid during the course of the study, clearly not all due to US abnormalities as 10-30% had biochemical abnormalities despite a normal US.

It is difficult to make a firm recommendation about the use of US in screening for CFLD. The sensitivity and specificity of early US findings for fibrotic CFLD have never been studied, making interpretation of the meaning of these findings difficult at present. It seems likely that US will increase the identification of "silent" CF associated liver imaging abnormalities, but the timing, frequency and cost effectiveness of the routine use of US in screening for CFLD remains to be determined. A prospective study of the sensitivity of this modality would be helpful in our interpretation of these studies.

Computed Tomography (CT) detects small differences in the attenuation of X-rays to identify structures not visible on standard radiographs (10) . The addition of oral and IV contrast improves the identification of tissues and blood vessels and can detect differences in perfusion in organs such as the liver.

Much like US, CT is very effective at identifying end stage cirrhosis with portal hypertension. The CT findings in end stage cirrhosis from a variety of causes have been published (11) . However, accurate CT imaging of the liver requires IV contrast and radiation exposure. In biliary cirrhosis, the primary end stage lesion of CF, nodularity of the liver is common and easily seen on CT. However, there are no studies of CT findings in early CFLD. Similar to US, there have been no studies correlating CT and early hepatic fibrosis in CF or other liver diseases. At present, CT cannot be recommended for screening for CFLD given the need for contrast and the radiation exposure compared to US. However, CT is likely the most sensitive imaging modality for screening for hepatic cancer in patients with cirrhotic CFLD.

Magnetic Resonance Imaging (MRI) detects the density of protons in tissue water and lipids and their relaxation times (10) . In MRI, the magnetic environments of protons in fat, intracellular water and extracellular water have different relaxation properties leading to sharp contrast differentiation between tissues with differing contents of fat or water.

MRI is very valuable for non-invasive imaging the biliary and pancreatic tree (magnetic resonance cholangiopancreatography). MRI findings in CF have been presented with excellent demonstration of biliary and pancreatic abnormalities (12) . It is the imaging modality of choice for hemagiomas. The value of MRI in diffuse parenchymal liver disease is still investigational (13) . Similar to US and CT, MRI can detect the gross abnormalities of cirrhosis and portal hypertension. At present, the use of MRI for screening for early CFLD would require further investigation.

Hepatobiliary Scintigraphy. After IV injection, several radioisotopes are extracted by the liver and excreted into the bile and eventually into the intestine. These are detected by a scintillation camera. In CF, abnormalities detected have included dilated intrahepatic and extrahepatic bile ducts and delayed excretion and intestinal appearance of the radioisotope (14) (15) (16) . Delayed intestinal appearance seems to predict a better response to ursodeoxycholic acid (3) . However, the specificity and sensitivity of this imaging modality for CFLD has never been studied. Given the technical difficulties with the study, it has been superceded by MRI, CT and US for routine hepatic imaging and is not suitable as a screening test for early CFLD. Liver disease is increasingly recognized as a major cause of morbidity and mortality in patients with Cystic Fibrosis (CF). The typical hepatic manifestation of CF is focal biliary cirrhosis and results from biliary obstruction and progressive periportal fibrosis; it is the most clinically relevant CF-associated hepatic problem, since extension of the initially focal lesions may lead to multilobular biliary cirrhosis, portal hypertension and related complications (1, 2) .

Surveys on CF patients with liver disease have reported rapid progression of cirrhosis to portal hypertension and high mortality rates due to complications of cirrhosis and lung disease (3, 4) .

In a retrospective study spanning a 26-year period, of 44 children with CF investigated for cirrhosis, esophageal varices developed in 86% and in half of these patients esophageal bleeding occurred early during the second decade of life; liver failure occurred in 36 % of them at a mean age of 15 years (3). In another retrospective study, over a period of 28-year, 16/31 CF patients with multilobular biliary cirrhosis died as a result of respiratory (10 cases) and hepatic (6 cases) complications (4).

More recently information regarding natural history of liver disease in CF patients have been obtained through prospective studies in carefully monitored CF patients (5, 6) .

In a cohort of 177 consecutive CF patients identified over a ten-year period and followed for a median time of 14 years at the CF Center of the University of Milan, there were 48 incident cases with no incidence peak in any age group (5). We estimated that crude incidence of liver disease in our whole CF population was 27%, corresponding to 2.4 cases per 100 patient-years. Median age at time of diagnosis of liver disease was 7 years (range: 2 months to 18 years) and in 14 of the 48 patients with liver disease (29%) this complication was diagnosed during the first five years of life. In our cohort, after the age of 10 years, incidence declined and, by the age of 12, liver disease had developed in 45 out of 48 patient. Liver disease should be therefore considered a relatively early complication of CF, which may be susceptible to prophylactic strategies.

The 48 patients with liver disease (treated with ursodeoxycholic acid since 1988) were followed for a median time of 8.5 years (range: 1 month to 20 years), corresponding to 407 patient-years. Cirrhosis was already present at time of diagnosis of liver disease in 5 of these patients (10%) and developed in 12 additional patients during follow-up after a median period of 5.1 years from time of diagnosis of liver disease (range: 6 months to 15 years). Incidence rate of cirrhosis was 4.5 cases per 100 liver disease patient-years (95% CI: 2.3 to 7.8). Among the 17 patients with liver cirrhosis who were altogether detected, 13 (76%) developed portal hypertension (incidence: 28.8 cases per 100 cirrhotic patient-years, 95% CI: 15.4 to 49.3), 4 (24%) esophageal varices (incidence: 3.2 cases per 100 cirrhotic patient-years, 95% CI: 0.9 to 8.3), 1 (6%) developed liver decompensation (incidence: 0.4 cases per 100 cirrhotic patient-years, 95% CI: 0 to 2.0) and underwent liver transplantation at the age of 10 yrs. At the end of follow-up, of the 45 patients with liver disease who were still alive 31 had not developed liver cirrhosis after a median time of 7.9 years (range: 2 to 15 years) from time of diagnosis of liver disease. Present evidence concerning efficacy of ursodeoxycholic acid on patient survival in CF is still inconclusive (7), but we cannot exclude that treatment with this bile acid also contributed to the favourable clinical course we have observed.

In order to establish the impact of liver disease on the outcome of CF, we also compared the main clinical outcomes of patients who developed liver disease with those observed in the group of patients who did not (5) . No significant differences between patients with or without liver disease were observed in the occurrence rates of clinically relevant end-points, like respiratory failure, need of oxygen therapy or number of hospitalisations during the last two years of follow-up. These data indicate that liver disease does not expose CF patients to higher risk of developing respiratory failure or other major outcome events. Mortality from any cause also did not differ significantly among patients with different liver status (death rate ratio 0.44), indicating lack of association between liver disease and mortality in CF.

Other studies addressing CF-related liver disease prospectively (6) have also provided evidence that this is a relatively early complication of CF, and that its clinical course may be milder than previously reported by retrospective studies (3, 4) . In a cohort of carefully managed Swedish CF patients who had been followed-up for more than a decade, liver disease developed before adolescence in 35 % of cases and in none of the patients this occurred in adulthood. Rate of progression was slow, as also indicated by histologic changes at repeat liver biopsy during the follow-up (6) .

In summary, liver disease associated with CF develops in a significant percentage of CF patients, usually before or at puberty, displays a slowly progressive course and until the most advanced stages are attained, it does not seem to influence nutritional status and severity of pulmonary involvement. Only in a minority of patients, often in the pediatric age, liver disease may represent the main clinical problem and its progression may be unusually rapid. The role of genetic modifiers in determining this variability in terms of severity is presently under study (8) . The most prevalent CF-causing mutation is a three base pair deletion that results in the loss of phenylalanine 508 (F508) in the first nucleotide binding domain (NBD1) of CFTR. Over ninety percent of all CF patients carry at least one delta F508 allele (1) . Not surprisingly, the molecular and cellular phenotypes of this mutation have been the focus of intensive study over the past fifteen years. This work has established that the mutant CFTR does not efficiently traffic to, nor accumulate at, the plasma membrane (2) due to a disruption of NBD1 folding (3; 4) and destabilization and rapid turnover of the minor fraction that does reach the membrane (5).

The inefficiently folded delta F508 CFTR is recognized by cellular quality control systems and marked for degradation by the proteasome (6; 7) . The lack of properly folded, native CFTR at the apical membrane underlies the aberrant fluid and electrolyte homeostasis that define the CF pathology.

Delta F508 CFTR that does fold and reach the plasma membrane retains some function as a chloride channel (8; 9). Thus, promoting the folding of delta F508 CFTR would be predicted to at least partially ameliorate the pathology (3; 10). In this regard, conditions that promote folding and stabilize native structures, such as reduced temperature (11) and addition of osmolytes (12) (13) (14) have been shown to improve mutant CFTR folding and partially restore function. However, treatments such as these, as well as alterations of cellular chaperones that assist in folding, lack specificity and the consequences of chronic alterations of these parameters on cellular function remain unknown. More specificity could theoretically be achieved if ligands that bind tightly to the native, but not unfolded CFTR could be identified (3; 10; 15). The favourable binding energy for these ligands would thereby be coupled to the otherwise unfavourable folding of mutant CFTR (15) . Ideal compounds would promote folding and potentiate CFTR function although weak anatagonists could also have utility.

Detailed structural knowledge of CFTR and its domains provides a context for understanding not only the functional mechanics of CFTR, but also the role of F508 in folding and further suggest sites which could be targeted for ligand binding. The recent solution of the murine NBD1 at high resolution using X-ray diffraction methods provides the first such structural information (16) . The structure reveals that F508 is exposed on the surface of the NBD in an alpha-helical sub-domain. Replacement of the phenylalanine with any other amino acid has no measurable effect on the folding or final structure of the isolated NBD, indicating that it is the loss of the backbone that leads to inefficient NBD folding (17) . In contrast, charged or bulky side chains at the 508 position interfere with the the later steps in the folding of the full-length CFTR. Together with the surface exposure of the 508 position, these results indicates that F508 may be at an interface between the NBD and other domains of CFTR, most likely the TMDs. These results have implications for understanding the function of CFTR and the development of ligands to promote mutant CFTR folding for therapeutic benefit. Structural predictions suggest that CFTR, a member of the ABC superfamily of transport ATPases, consists of two homologous halves, each comprised of six transmembrane (TM) helices and a nucleotide binding domain (NBD1 and NBD2) that are connected by the regulatory (R) domain (1). This complex, multidomain structure, conceivably, renders the post-translational folding of CFTR sensitive to point mutations, while ensures its regulation by phosphorylation and ATP-binding/hydrolysis. Depending on the expression system examined, 20-80% of the newly synthesized wt CFTR degrades at the endoplasmic reticulum (ER), as coreglycosylated folding intermediate (2) (3) (4) . The remaining 20-80% of the CFTR undergoes an ATP-dependent posttranslational conformational maturation (3) (4) (5) and attains an export competent, folded conformation that is incorporated into transport vesicles at the endoplasmic reticulum (ER). During traversing the cis/medial Golgi, the high mannose-type N-linked oligosaccharides of CFTR (core-glycosylated form) are converted to complex-type oligosaccharides (2) .

It was proposed that the ATP-dependent conformational maturation of the nascent, core-glycosylated CFTR is a prerequisite for entering the distal stages of the secretory pathway from the ER (3). This conclusion was verified by examining the in vivo and in vitro protease susceptibility of the early folding intermediates and the fully mature CFTR either in its core-or complex-glycosylated form (3, 6) . The conformational maturation of CFTR is facilitated by an array of chaperone interactions. Accordingly, mutations perturbing the folding process of CFTR alter the association/dissociation dynamics of these chaperones and vice versa; modulation of chaperone activity has an effect on CFTR folding (7) .

In the CF gene more than 1000 mutations have been identified, leading to impaired biosynthesis, processing, activation and stability of CFTR or a combination of these (8) . Missense mutations, including the most frequent one, deletion of phenylalanine at position 508 (∆F508) in the NBD1, are believed to interrupt the post-translational folding of CFTR (2) (3) (4) 6) and target the core-glycosylated intermediate for degradation via the ubiquitin-proteasome pathway at the ER (9,10). While exposure of ER-retention signals, may also contribute to the inability of folding intermediate(s) to exit the ER (11), the consequence of ∆F508 at the molecular/structural level remains elusive.

Numerous model systems, represented by synthetic peptides and isolated nucleotide binding domains, harboring the wild-type sequence or the ∆F508, have been applied to unravel the structural impact of the mutation. In our approach we used systematic amino acid substitution of the F508 residue in the context of the full-length CFTR to investigate the role of the side-chain and back-bone structure of F508 on the biogenesis, stability and interdomain interaction of the channel. Hydrophobic amino acids with side chain volumes comparable to, but not larger than phenylalanine supported the post-translational folding and stability of CFTR. On the other hand, charged amino acids, as well as proline and glycine substitutions were non-permissive for both folding and stabilization of the channel. While deletion of F508 caused limited conformational defect in the NBD1, similar to non-permissive amino acid substitutions, it substantially compromised the conformation of NBD2 and its interaction with NBD1 (unpublished data). These and other results, discussed in the symposium, strongly suggest that hydrophobic side chain interactions of F508 are indispensable to ensure normal folding NBD2 and thus to achieve global stabilization of the native conformation of CFTR.

Reduced temperature or chemical chaperones can partially rescue the folding defect of ∆F508 CFTR (12) The majority of patients with cystic fibrosis (CF) express a mutant cystic fibrosis transmembrane conductance regulator (CFTR) protein that is defective in folding (1). The misfolded protein is retained in the endoplasmic reticulum (ER) and is not trafficked to the cell surface (misprocessed mutants) (2). These patients would benefit from therapy that can cause the mutant CFTR protein to fold properly and be delivered to the cell surface in a functional form.

Recently, we discovered a strategy for correcting misfolded defects in P-glycoprotein (P-gp) that has direct application to CF. P-gp and CFTR are both members of the ATP-Binding Cassette (ABC) family of proteins. The ABC proteins generally have two transmembrane domains (TMDs) and two nucleotide-binding domains (NBDs). Mutations in P-gp at equivalent positions to those in CFTR that cause it to be misprocessed also caused the mutant P-gp to misfold and be retained in the ER. The misprocessed P-gp mutants, however, could be induced to fold properly when they were expressed in the presence of a specific drug substrate (3). The protein was then trafficked to the cell surface where it was fully active. Therefore, P-gp is a useful model system for studying how processing mutations interfere with folding of ABC proteins.

The next objective was to understand how the misprocessed CFTR and P-gp mutants were different from the mature wild-type proteins and how drug substrates the mutant channel to the cell surface. Biochemical assays revealed that the ∆F508 mutation not only impairs the gating kinetics, but also decreases the channel stability by tenfold at the cell surface (13) . To unravel the cellular machinery responsible for the recognition and elimination of the mutant from the plasma membrane, the internalization and recycling rates of CFTR were determined. Whereas native CFTR recycled from sorting endosomes back to the cell surface, misfolding by ∆F508 and ∆70 mutations prevented recycling and facilitated lysosomal targeting of CFTR by promoting its ubiquitination. Rescuing the peripheral folding defect or down-regulating the E1 ubiquitin-activating enzyme stabilized the mutant CFTR without interfering with its internalization. These and other observations, in concert with the preferential association of the mutant CFTR with Ub-binding proteins (e.g. Hrs and STAM-2) and other components of the Ub-dependent endosomal sorting machinery (e.g. TSG101), establish a functional link between Ub-modification and lysosomal degradation of misfolded CFTR from the cell surface (13) . Our data provide evidence for a novel cellular mechanism of CF pathogenesis in case of ∆70 CFTR. The results also imply that efficient relocation of the ∆F508 CFTR from the ER to the cell surface will require the conformational stabilization of the mutant. Finally, these observations suggest a paradigm for the quality control of membrane proteins in general, involving the coordinated function of the ubiquitination and Ub-dependent endosomal sorting machinery in post-Golgi compartments.

act as specific chemical chaperones to rescue the misprocessed mutants. Initial studies suggested that processing mutations affect the structure of P-gp (4, 5) or CFTR (6) by trapping the mutant proteins in the ER in "loosely-folded" conformations that resemble the immature core-glycosylated form of the wild-type protein.

These trapped proteins are inactive.

One folding step that may be sensitive to the presence of processing mutations is the establishment of correct domain-domain interactions between the two TMDs and the two NBDs. To test whether processing mutations affect domain-domain interactions, we constructed a Pgp/CFTR(⌬F508) chimera in which the predicted third ␣-helix in NBD1 of P-gp was replaced with the equivalent region from CFTR containing the ⌬F508 mutation. It was found that the ⌬F508 mutation caused misprocessing of the P-gp/CFTR chimera and loss of activity by disrupting packing of the TM segments in the TMDs (7) . It also appeared that the ⌬F508 mutation disrupted interactions between NBD1 and the first cytoplasmic loop connecting TM segments 2 and 3. More recently, we showed that processing mutations could also disrupt interactions between NBD1 and NBD2 (Loo et al., 2004, submitted) .

These results indicate that processing mutations may interfere with folding of ABC proteins by hindering the proper interactions between the domains. The drug substrates (specific chemical chaperones) may act by stabi-lizing proper domain-domain interactions during the folding process. Mutations in CFTR profoundly influence the function of respiratory epithelial cells, with resultant mucous accumulation, inflammation and infection. CFTR plays numerous roles in cellular function, interacting directly and indirectly with cellular proteins and processes that regulate many aspects of epithelial cell biology. In order to identify pathways by which CFTR influences cell function, and to explore the effects of CFTR on cellular processes per se, we have generated a series of transgenic mice (1) lacking CFTR (FABP-CFTR, CFTR-/-), (2) overexpressing human CFTR in lung epithelial cells (SP-C-hCFTR), and (3) expressing the human ∆508 CFTR (SP-C-∆508 CFTR, CFTR-/-, FABP-hCFTR). All of the mice are viable and without overt pulmonary pathology. RNA microarray analysis of the lung tissue was compared in lungs from adult mice of each genotype using the

Jeffrey A. Whitsett, M.D. and Yan Xu, Ph.D.

Affymetrix system. RNAs differentially expressed were cross-compared. Comparison of SP-C-∆508 CFTR, CFTR-/-vs. CFTR-/-demonstrated a strong, positive correlation (R = 0.938), and a weak, negative correlation (R = -0.372) with the SP-C hCFTR overexpression, indicating human CFTR ∆508 did not correct the abnormalities in gene expression characteristic of CFTR deficiency. Genes regulating inflammation, neutrophil chemotaxis, as well as fluid and electrolyte transport were induced in CFTR-/-and SP-C-∆508 CFTR, CFTR-/-mice, while those regulating cell growth, cell matrix and adhesion, and angiogenesis were significantly decreased. Increased expression of human CFTR (wild type) in the ∆508 CFTR induced genes in the heat shock protein family, indicating that this subset of genes was induced in response to CFTR or misfolded ∆508 CFTR. Dnajb1 and Hsp105 RNAs were increased by human CFTR and human ∆508 CFTR, perhaps indicating a response to excess or misfolded CFTR protein. Genes involved in the endoplasmic reticulum response pathway (ERAD) were not induced by overexpression of CFTR or ∆508 CFTR in vivo. While RNAs in the ∆508 and CFTR deficient mice were quite similar, several genes involved in host defense including S100a8, S100a9, H2-Eb1, H2-Dmb1, H2-T10, and c-fos were induced in CFTR-/-, but decreased in the ∆508 mutant mice. Several genes regulating lipid homeostasis were uniquely influenced by the ∆508 CFTR (Scd-1, lipocalin, and carboxyesterase-1). In summary, increased expression of CFTR and ∆508 CFTR similarly induced genes in the heat shock protein (HSP) pathway, but did not induce ERAD transcriptional responses. Genomic responses to CFTR deficiency were, in general, not corrected by the ∆508 CFTR. CFTR influences genes regulating host defense, inflammation, and fluid and electrolyte transport by processes likely mediated by its direct interaction with other cellular proteins and by influencing adaptive responses to the lack of CFTR. 

The absence of CFTR activity from the CF airway epithelium results in abnormal regulation of ion transport mechanisms leading to a reduced volume of airway surface liquid (ASL) producing mucus dehydration and decreased mucociliary clearance, both hallmarks of CF lung disease. The consequences of these physiological alterations in the lung eventually result in chronic bacterial infection, mostly notably Pseudomonas aeruginosa, and exaggerated neutrophil-mediated inflammation, eventually leading to airway obstruction and death. A logical therapeutic strategy for CF lung disease would restore CFTR function to the airway epithelium and gene transfer strategies remain one approach towards this goal.

At present, the regions of the CF lung that would require CFTR expression for restoration of normal physiological function and reduction of disease symptoms are not well established. In normal human lung, the ciliated epithelial cells of the surface epithelium are considered to express CFTR. However, restoration of normal ion transport functions to CF airway epithelium after delivery of CFTR exclusively to ciliated cells has not yet been achieved to test this hypothesis. The possible involvement of the submucosal glands in the pathogenesis of CF lung disease suggests that cell-types other than ciliated cells may also require CFTR expression for normal function (1). However, the predominance of ciliated cells in the upper and lower airways especially in CF infants who have not developed chronic airway inflammation, and the accessibility of the ciliated surface epithelium to intralumenal-delivered therapeutics identifies ciliated cell-types as attractive targets for CF lung therapeutic strategies.

A critical factor for CF gene transfer strategies has been overcoming the low transduction efficiency for intralumenal delivered vectors. There are several extracellular barriers to gene transfer vectors in human airway that may result in inefficiency of gene transfer, including: the mucociliary clearance system; the glycocalyceal barrier; the absence of most viral receptors from the airway lumen; and, the slow rate of lumenal endocytosis in airway epithelial cells. Although strategies to circumvent these barriers with commonly available vectors are under investigation, i.e., retargeting vectors to new apical surface receptors, transiently opening tight junctions or using new serotypes of the common vectors, the results so far published have not shown significant improvement. Novel vector-types are also being assessed for airway gene transfer that may be more efficient at breaching epithelial cell apical barriers. For example, Sendai virus, human coronavirus 229E, and lentiviral vectors pseudotyped with Ebola virus or Sendai virus envelope proteins have shown improved gene transfer efficiency to airway epithelial cells in vitro and/or in vivo (2) (3) (4) (5) . However, the usefulness of these vectors remains to be determined.

We have chosen to investigate the utility of the human parainfluenza viruses (PIV) as gene transfer vectors since these viruses exclusively infect ciliated cells of the human airway epithelium after intralumenal delivery. To test these vectors, we used an in vitro model of human pseudostratified, mucociliary airway epithelium (HAE) that recapitulates the morphologic and phenotypic characteristics of the in vivo human airway epithelium. Recent studies have revealed that this model system displays the phenotypic differences that occur between CF and non-CF airway epithelium, i.e., reduced chloride ion transport, hyperabsorption of sodium ions, the failure to regulate the depth of the ASL, and, the dehydration of secreted mucus that results in cilial dysfunction and mucostasis.

We have previously shown that human recombinant parainfluenza virus (hPIV) efficiently infects human ciliated airway epithelial cells from the lumenal surface. To test our hypothesis that CFTR expression in ciliated cells may restore normal physiological function to CF epithelia, we generated a hPIV that expressed CFTR (hPIV-CFTR) and systematically tested the effects of CFTR expression in CF HAE. Western analyses performed 48 hrs after apical surface inoculation of CF HAE with hPIV-CFTR, hPIV expressing GFP (hPIV3-GFP) or vehicle control alone showed that only hPIV-CFTR produced expression of fully glycosylated "Band C" CFTR. Immunodetection of CFTR protein in histological sections of CF HAE inoculated with hPIV-CFTR localized CFTR to the apical surface of ciliated cells in regions corresponding to the microvilli of the ciliated cells. To determine whether CFTR expression in CF ciliated cells resulted in chloride ion channel activity, CF HAE were assayed in Ussing chambers after viral inoculation. Compared to control CF HAE (vehicle or hPIV-GFP treated), inoculation with hPIV-CFTR resulted in a significantly increased forskolin-mediated activation of CFTR chloride channel activity [1.5 ± 0.1 _A/cm 2 for control (n = 18) and 23 ± 2.5 _A/cm 2 for hPIV-CFTR (n = 9)]. To determine whether this level of CFTR expression in ciliated cells was sufficient to restore the altered fluid transport of CF HAE, we assessed the regulation of ASL depth by measuring the rate of absorption of liquid across the apical surfaces of CF HAE 48 hrs after inoculation with vehicle control alone or hPIV-GFP or hPIV-CFTR. In control CF HAE, the ASL stabilized at 3.1 ± 0.3 _m (n = 12), a depth associated with decreased mucus transport. However, in CF HAE inoculated with hPIV-CFTR, liquid absorption ceased at a depth of 7.3 ± 0.3 _m (n = 12), a depth similar to that maintained by normal HAE (7.4 ± 0.6 _m, n = 7). A hallmark of CF lung disease is mucostasis, a parameter that is recapitulated in CF HAE. Therefore, to determine if the restoration of fluid transport to CF HAE by CFTR expression in ciliated cells was sufficient to allow the transport of mucus secretions, we measured the capacity of CF HAE to transport mucus after inoculation with hPIV-CFTR. For control CF HAE, rotational velocity of mucus transport was low with little or no mucus transport (1 ± 1 _m/sec, n = 5) compared to significant mucus transport in CF HAE inoculated with hPIV-CFTR (21 ± 2 _m/sec, n = 5).

Therefore, hPIV as a gene delivery vehicle has circumvented a major barrier to the effective use of gene transfer vectors, i.e., the restricted uptake of vectors from the apical surface. We have demonstrated that expression of CFTR in ciliated cells of CF HAE resulted in restoration of the ability of the epithelial cells to regulate the depth of the ASL to levels that were sufficient for effective mucociliary transport. Whether, expression of CFTR in ciliated cells in the airway epithelium in vivo will be sufficient to restore these physiological indicators of CF lung disease remains to be determined. Since hPIV-CFTR is the most efficient and efficacious vector that we have so far tested for CFTR gene transfer and results in correction of pathophysiological consequences of the CFTR defect, efforts will be directed at generating derivatives of hPIV vectors that can be safely delivered to the human airway in vivo.

The genetic etiology of human respiratory diseases is being increasingly emphasized as a means of better understanding disease pathogenesis, with the ultimate goal of improving preventive strategies, diagnostic tools and therapies. Considerable effort and expense is currently being expended in attempts to detect genetic loci contributing to both 'simple' (single-gene) and 'complex' human diseases. Association and linkage studies comprise the two dominant strategies, the former aiming to find disease predisposing alleles at the population level and the latter focusing on familial segregation. Although both strategies have compelling strengths, association analyses are currently more widely conducted and likely to spread even further in the future, particularly in the pharmacogenetics domain. Facilitating these studies are technical developments in molecular genetics and in the use of gene-specific variants derived from the human genome. In addition, extensive catalogues of anonymous DNA sequence variants across the human genome have begun to be compiled. Some large-scale, population-based human samples have been or are expected to be collected (e.g., EPIC, ISIS, Million Women Study, MRC/Wellcome Trust Biobank UK), and the use of DNA variants in drug development is expanding. The coupling of high-throughput molecular technology, large numbers of genetic variants, and large samples offers some new opportunities for understanding the etiological basis of many diseases, including Cystic fibrosis (CF).

Modifier genes are classically defined as genes that have small quantitative effects on the level of expression of another gene. Studies of modifier genes therefore attempt to detect the sometimes subtle modulation of the main effects of a major locus regulating a disease (the CF transmembrane conductance regulator [CFTR] gene in the case of CF). Such effects may modify disease severity and progression. A broad spectrum of disease severity exists in CF, and it has become increasingly clear that CFTR genotype does not account for the wide variance in the clinical spectrum of CF phenotype. Many family-based and population-based study designs are possible to attempt detection of modifier genes in CF. This symposium will review the statistical concerns related to genetic studies of modifier genes and will discuss study design issues related to the investigation of modifier effects in gene discovery projects.

University of Western Australia. Phenotypic heterogeneity of CF lung disease is not explained by intragenic mutations in CFTR, and likely reflects genetic heterogeneity at other loci, plus environmental effects. A multi-center study of gene modifiers for CF lung disease is underway to identify non-CFTR genetic contributions to the variability of CF lung disease. The study is designed as a case-control association study in unrelated patients with the same CFTR genetic background (homozygous ∆F508). The analysis focuses on comparing the prevalence of alleles and genotypes in patients who have "severe" (lowest 25 th percentile for age) versus "mild" (highest 25 th percentile) lung disease. We have enrolled 864 subjects, who are segregated into three groups: 1) 320 young subjects with "mild" lung disease (age 15-28 yrs), 2) 255 older subjects with "mild" disease (>age 29 yrs), and 3) 289 subjects with "severe" disease (age 8-25 yrs). Longitudinal analysis of spirometric measures covering five years up to enrollment (average (20 measures per patient) has been performed by a mixed model analysis with empirical Bayes estimates of individual slopes and intercepts. The intercepts (at birth) for the FEV 1 (% Pred.) for these three groups are 116, 112, and 108 (% Pred.), respectively. The rates of decline (%/year) for the two groups of mild subjects were similar (-1.2 and -1.0), whereas the severe cohort had a more rapid rate of decline (-3.7). The mild versus severe groups also differed in body mass index (BMI). The BMI Z-score was nearnormal (-0.11 and -0.4) for the younger and older mild lung disease cohorts, respectively, but much worse (-1.28) for severe patients. The prevalence of Pseudomonas aeruginosa during the 3 years up to enrollment was similar for all three groups (-86-88%). Thus, CF patients in this protocol are characterized as "mild" and "severe", and other key clinical variables are defined. Genotypes in 815 subjects have been completed for selected SNPs in genes previously studied as modifiers of CF lung disease, including alpha-antiprotease, ACE, β-adrenergic receptor 2, GSTM1, GSTM3, GSTP1, IL-10, MBL2, NOS3, TNFβ and TGFβ1. Results show a striking association of SNPs in TGFβ1 with severe lung disease (P = 0.0004; odds ratio >2). SNPs flanking TGFβ1 show no association, which implies that TGFβ1 is the causative modifier gene, as has been reported for asthma and COPD. Ongoing studies are testing other genetic variants in several pathogenic cascades in CF related to infection and inflammation. *Reporting for the Gene Modifier Study Group. Supported by CFF KNOWLE00A0, NIH R01 HL68890, and NIH R00046.

Garry R. Cutting, MD Institute of Genetic Medicine, Johns Hopkins University, Baltimore, MD USA Genetic, environmental, and stochastic factors are the three major factors contributing to phenotype variation. For single gene disorders the nature of the defect in the disease-causing gene is a major contributor to variation. However, in most single gene disorders, including CF, the disease-causing gene is not the sole determinant of variation indicating an important role for the other major factors. This point is highlighted by the observation that cystic fibrosis patients with identical CFTR genotypes have considerable variability in the severity of lung disease (1). Study of related affected individuals provides a method to dissect the contributions of genetic background, environmental and stochastic factors to phenotype variability (2). This is achieved by comparing the degree to which monozygous (MZ) twins, dizygous (DZ) twins and siblings are similar to each other for selected traits. If pairs of MZ twins are found to be highly correlated (e.g. if one member of the twin pair has a trait; the other twin also has the trait) then one could attribute this similarity to genes (since they share 100% of their genes) or the environment that they share (since they are together in utero and grow up together in the same household). If DZ twins are examined for the same trait and found to have lower rates of correlation than MZ twins, then genetic factors can be implicated since DZ twins share only 50% of genes but have the same degree (or nearly so) of shared environment as MZ twins. On the other hand, similar high levels of correlation among MZ and DZ twins implicates the environment that they share. Finally, siblings share 50% of genes like DZ twins but have a lower degree of shared environment since they are born at different times (although still raised in the same household). Thus, siblings provide an opportunity to test the role of environment on genetically similar individuals. A trait that shows similarly high levels of correlation among twins but a lower level of correlation among siblings would implicate environmental rather than genetic factors for the trait under question.

The European Twin and Sibling Study has demonstrated the utility of this approach. This group reported that genetic factors independent of CFTR can influence phenotype severity by altering chloride secretion (3;4). To increase our understanding of the role of these major factors, a CF Twin and Sibling Study is being conducted in the United States. Over 100 participating centers located throughout the United States have contributed patient information to the study. At the current time, 50 pairs of monozygous twins, 23 pairs of dizygous twins, and 239 sibships have been enrolled into the study.

Several themes are emerging from the U.S. Twin and Sibling Study. First, variation in pulmonary disease, as measured by FEV 1 % predicted is minimal among twins. In fact, MZ and DZ twins show similarly high levels of correlation for this parameter suggesting that genes other than CFTR have a limited role in differences in lung disease severity. Furthermore, lower levels of correlation for FEV 1 % predicted among siblings both by cross-sectional and longitudinal analysis supports the concept that genes are not a major factor, but that sharing of environment plays an important role. This conclusion is consistent with observations from CFTR genotype-CF phenotype studies over the past decade. However, the small number of enrolled DZ twins in both studies reveals a limitation in this approach. Analysis of a larger collection of siblings will allow rigorous evaluation of the role of environment, and the specific factors therein, that influence pulmonary outcome.

Second, evaluation of GI manifestations of CF in twins and siblings has revealed an important role for genetic factors. Previous studies have documented familial clustering of meconium ileus (MI) and a locus for this trait has been mapped on chromosome 19 (5) . MZ twins display a high rate of concordance for MI (92%) while DZ twins have a lower rate (22%) strongly implicating genetic factors, as has already been noted. Liver disease and CF-related diabetes mellitus follow the same pattern suggesting that genetic factors also underlie these traits.

Third, some twins and siblings exhibit considerable difference in the age that they acquire bacterial infections. Most impressive is the observation that MZ twins can differ by several years in the initial age of acquisition of Pseudomonas aeruginosa or Burkholderia cepacia. Since MZ twins share 100% of genes and environment, the difference in acquisition of these bacteria suggest a substantial role for stochastic factors.

In summary, study of affected twins and siblings permits an assessment of the contribution of genetic factors to trait variation. This is an important step prior to the investment of considerable effort to identify modifier genes for particular traits. Genomic analysis is now widely used to probe the basis for disease and how disease may be expressed in individual patients. The attempt to identify specific disease-associated mutations, polymorphisms, or transcript patterns currently underpins much of our thinking about disease mechanisms, but the overall complexity of genebased data and the limited amount of functional annotation continue to present challenges in the interpretation of these experimental results. Going forward, it will become increasingly important to look beyond single genes to genetic networks and ultimately to the integration of these data with quantitative measures of biological function at the system level to discover key molecules that can explain and classify clinically-relevant phenotypes.

We have taken a systems biology approach to discovery that is based on the concurrent analysis of gene expression in combination with two other large-scale data streams: metabolomics and quantitative histomorphometry. Metabolomics uses mass spectrometry to measure changes in the relative abundance of endogenous small molecular weight biochemicals (e.g. amino acids, sugars, lipids) present in cells, tissue and biofluids. Quantitative histomorphometry uses computationally intensive image analysis of normal and pathological histology slides to quantify tissue structures and identify phenotypic changes in response to toxicity or progress of disease. The assessment of the integrated cellular state using biochemical profiles and quantitative tissue analysis is an important complement to the large-scale molecular state of a biological system. We are currently applying this approach to the study of drug-induced liver injury and to the discovery of novel drug targets and biomarkers for Type II diabetes.

Hepatotoxicity: Acetaminophen (APAP) overdose remains one of the most common causes of hospital admissions for acute liver toxicity. The initiating mechanism is believed to involve hepatic metabolism of APAP to a reactive oxidative species, NAPQI, which subjects the liver to oxidative stress, resulting in depletion of glutathione. Oxidative stress is also believed to play a pivotal role in the hepatotoxicities of a wide range of chemical and pharmaceutical agents. Therefore, APAP may serve as a useful model toxicant to elucidate the mechanism of liver injury. Single oral doses of APAP producing gross liver pathologies ranging from no histopathological change to frank necrosis were administered to rats. Tissue and biofluid samples were analyzed by LC/MS (ToF) with an ESI source in either positive or negative mode. Mass spectra at each retention time were matched to a library of known standards using proprietary software and linked to metabolic pathways. Analysis of individual metabolites at the tissue level revealed major perturbations in pathways associated with known injury and repair mechanisms -e.g. depletion of glutathione and cystathionine (oxidative stress), decreases in NAD and various nucleotides (nucleic acid repair), and a decrease in CDPcholine (phospholipid turnover). Several of these decreases in liver metabolites were also reflected in urine, suggesting that they might serve as useful biomarkers for early detection of liver disease. Time and dose-related changes were also observed in various other biochemicals that were not predicted by current knowledge.

Type II Diabetes: As a disease model, obesity, insulin resistance and hyperglycemia were induced by feeding C57Bl/6 mice high-fat diets. Responses were monitored over 16 weeks. High-fat fed obese mice were chosen that demonstrate either hyperinsulinemia or symptoms of Type II diabetes and were then compared to control mice fed a normal diet and demonstrating normal weight gain, fasting plasma insulin levels and fasting blood glucose levels. The mice selected at each time point were sacrificed and 14 tissues (liver, skeletal muscle, pancreas, white adipose tissue, heart, kidney, brain, testes, skin, tongue, stomach, intestine, spleen, and tongue) were harvested and frozen in liquid nitrogen until processed for RNA isolation and metabolic profiling with adjacent sections fixed in formalin for histological processing.

By combining data sets, we are able to identify novel and important drug targets and biomarkers that would be overlooked in a conventional target discovery processes. This unique approach to the discovery of novel drug targets and biomarkers is based upon the hypothesis that tissue structure, which is at the core of most diagnosis, holds the key to unlocking the link between genes, biochemicals, function, and disease. Through correlations, differential analysis, and knowledge of gene function and pathways, we are able to identify biologically relevant drug targets and biomarkers that support this hypothesis. Co-variant analysis of gene expression and tissue feature data identified different type of genes than were found using gene expression analysis alone. Comparing the ontological classifications of the 303 genes that were up or down regulated greater than 2-fold, and 306 genes were closely correlated to liver features, fasting glucose, or insulin at sacrifice revealed, but there were only 56 genes in common. Comparing the molecular function and biological processes ontological classifications associated with these two sets also shows distinct differences in the classes of gene families represented in the individual groups. Many of the genes identified through co-variant analysis were altered <2-fold and therefore would have been lost in conventional analyses.

In conclusion, these results indicate that a systems biology approach combining gene expression analysis with biochemical profiling and quantitative tissue analysis can help identify disease mechanisms, drug targets and biomarkers that other gene-only approaches often overlook.

We are indebted to DiAthegen (Athens, OH) who is a strategic research partner in the diabetes target discovery program. This research was funded in part by the National Institute of Standards and Technology -Advanced Technology Program (NIST ATP), award #70NANB2H3009.

Pseudomonas aeruginosa is an interesting organism since it provides a prototype for a nutritionally diverse, free living organism that is found in many microenvironments such as soil, water, plants, and animals, and because it is a common cause of hospital-acquired and cystic fibrosis (CF) patient infections, in part because of its high intrinsic resistance to many antibiotics.

The most important recent development in P. aeruginosa research was the publication of the genome sequence in August 2000 (Stover, KC, et al. Nature 406:959-964). However a genome sequence is only as good as the usage we put it to. To permit us to capitalize on and reach beyond the genome sequence with maximum efficiency, we and other laboratories have been developing genomic tools. These include bioinformatics, whole genome knockout and gene fusion libraries, and microarrays.

A major bioinformatics resource has been provided by the Pseudomonas genome web site (www.pseudomonas. com), which is an attractive, information rich resource, maintained with funding from the Cystic Fibrosis Foundation. This site includes all known information about each of the genes in the prototype strain PAO1, permitting investigators to find genes of interest through key word searching or browsing through the genome. Other resources are being continually developed, e.g. the recent development of PSortB has permitted a prediction of the location of every gene in the genome and especially key secreted or surface genes that may be involved in pathogenesis. Using the tactic of fusing fragments of the P. aeruginosa genome to an N-terminal signal sequence lacking alkaline phosphatase gene, we have confirmed several hundred of these predictions.

Whole genome knockout libraries are of tremendous value in pathogenesis. For example one can use such libraries to determine if a particular mutation attenuates virulence. In addition to the established library, we have made a knockout library that consists of random gene fusions to the promoter-less luxABCDE cassette. This library is not as comprehensive as the library published by Jacobs et al (PNAS USA 100:14339-44; , but creates in addition to knockouts, a reporter system for expression of the particular gene into which the lux cassette, encoding light production, has been inserted. Thus one can also study regulatory events related to pathogenesis.

By far the most utilized genomic tool has been microarrays, which permit the analysis of the expression of every gene in the genome simultaneously. Many of the known regulatory events in Pseudomonas that are connected to pathogenesis are being probed through the use of microarrays. As the tool-kit for post-genomic studies grows, we are making rapid in-roads into understanding how this complex organism responds to life outside and in the host.

Funding through the Functional Pathogenomics of Mucosal Immunity program grant, provided through Genome Prairie and Inimex Pharmaceuticals is gratefully acknowledged. In the chronic airway infections that afflict cystic fibrosis patients P. aeruginosa live in matrix-encased groups known as biofilms (2, 6, 8) . Biofilms are often thought of as communities of bacteria because the cells occupy a defined territory, they can exhibit coordinated behavior facilitated by communication, and because key biofilm characteristics (like their marked resistance to killing) cannot be generated by the bacteria living as individuals (9) . The functioning of many types of biological communities is enhanced by diversity. Diversity is beneficial because the presence of diverse subpopulations increases the range of environmental conditions in which some community members will survive, or even thrive (1, 7, 10) .

We have observed that growing wild-type P. aeruginosa in biofilms for 2-7 days produced considerable colony morphology variation. Planktonic (free-swimming) batch cultures grown to log, stationary, or late stationary phase produced no variants. The variant phenotypes we studied were heritable, suggesting genetic changes produced them. A prime candidate for mediating such variation is RecA. This protein is involved in phase variation and adaptive mutation, both which can be triggered by specific environmental cues (3, 5) . Inactivation of recA dramatically reduced biofilm-induced colony variation and this defect was complemented with chromosomally inserted recA.

The involvement of RecA, which could mediate genetic change anywhere in the chromosome, led us to hypothesize that biofilm-induced diversity could extend to other functions. To test this, we examined the effect of biofilm growth on additional phenotypes, and found that biofilm growth produced heritable variation in several other traits. Furthermore, some variants exhibit specialized functions in biofilms. Isolates from CF airway infections have also been observed to undergo phenotypic diversification (4) . Our data suggest that shortterm biofilm growth can rapidly produce genetic diversification of P. aeruginosa. This diversity may increase the stability of biofilm communities and speed the evolution of the resident bacteria. 

Despite advances in the assessment and management of pain in children, there are still subpopulations of pediatric patients for which pain and its management has not been well studied. Children with cystic fibrosis (CF) are such a population. Pain and discomfort associated with CF may be secondary to direct involvement of the organs (e.g., abdominal cramping and pleuritic chest pain), physiologic changes that may occur related to organ involvement (e.g., headaches from hypercarbia and sinusitis, musculoskeletal pain), or due to treatment such as procedural pain. Historically, pain management has not been included in an overall treatment plan, and commonly used analgesics such as opioids have long been avoided in the CF population because of the fear they may compound respiratory difficulties.

Only a few studies have focused on assessment or management of acute or chronic pain symptoms in children and adolescents with CF. The first published report of chronic pain in a relatively older and sicker population with CF was conducted via a retrospective chart review of patients who died and patients who were referred to a pain treatment service. This study demonstrated that patients with CF primarily complained of chest pain and headaches but back pain, abdominal pain, and limb pain also occurred regularly. 1 In the patients referred for treatment of chronic pain, safe and effective management with opioid analgesics was achieved. Another study, focused on musculoskeletal complications experienced by pediatric patients with CF, found through clinical assessment, serology, and bone scan that 21 of 125 (17%) patients had joint pain. 2 Finally, pain and dyspnea have been shown to be a common symptom at the end of life in children with CF. 3 Opioids were effective in managing these symptoms in terminal patients. Although these studies provide evidence of the type and nature of pain associated with CF, to our knowledge, there has not yet been any published report of the multidimensional experience of acute and chronic pain in a broad, un-referred population of children and adolescents with CF. Unrecognized and untreated pain in children with CF may be associated with increased morbidity, poorer prognosis, and decrements in their overall quality of life. Therefore, research to characterize the full spectrum of the pain experience from the patient's point of view may help to identify areas where pain management intervention is needed.

Our research group initiated a study, the DCH Pain in CF Study, to characterize the experience of acute and chronic pain and its impact on the quality of life in children with CF. Forty-six children ages 8-17 (M = 13.0 years; 52% male) with CF were recruited at one university-based and two community-based pediatric CF clinics. A previously validated survey instrument collected standardized data regarding the location, frequency, intensity, duration, associated emotional distress, and perceived functional limitations of chronic pain, and the intensity of procedural pain.

Findings demonstrated that chronic pain was a regular experience for children and adolescents with CF, occurring at least weekly in 46% of children. Commonly reported locations of pain were: abdominal/pelvic region pain (50%); chest pain (37%); head pain (33%); leg pain (15%); and arm/hand pain (13%). Most children (71%) reported the intensity of their pain as being mild and lasting for "less than 1 hour" (65%). However, a small subgroup of patients (15%) reported pain lasting onehalf day or longer and 11% reported experiencing moderately intense pain. There were no significant differences in pain reports by age or gender.

Pain location and frequency were related to pulmonary function and perceived limitations. Patients with chest pain had lower FEV 1 % (p<.05) than those patients with no chest pain. These patients also had more perceived functional limitations due to pain (p<.001) than patients without chest pain. Patients with weekly pain had more perceived functional limitations (p<.0001) than patients with less frequent pain.

The majority of patients reported experiencing mild procedural pain. A small number of patients reported experiencing severe pain from blood draws (7%), PICC line placement (9%), gastrostomy tube placement (8%), PFTs (2%), throat cultures (8%), and chest physiotherapy (3%). Patients who experienced higher intensity of pain from PICC line placement were more likely to also experience higher intensity of chronic pain (p<.05). Patients who experienced higher intensity of pain from

Patrick Hubbard, R.N., B.S.N

The purpose of this study was to develop a web-based education program tailored to patients with CF who may be experiencing pain, and to investigate, via the website, the pain experiences of those patients by studying their pain reports, disability, and coping strategies.

This study was a cross-sectional survey of CF patients. Recruitment efforts focused on informing patients of the website and research study. A sample of 18 participants was recruited. Three separate questionnaires were employed: a demographic questionnaire, the Pain Disability Index, and the Pain Response Inventory.

Approximately half of the sample reported experiencing pain episodes on a daily basis lasting two hours or less. The average intensity of a pain episode was reported to be in the moderate range. Participants reported pain disability highest in areas of recreation, occupation, and social activities. The most commonly used coping strategies included active and accommodative coping techniques such as problem solving, acceptance, and self-encouragement.

Pain management is clearly a problem for some young adults with CF. This study provided information about the areas of disability caused by pain, as well as common coping strategies utilized by patients. The findings of this study provide direction for the future of pain management in CF patients. PFTs and chest physiotherapy were more likely to experience chronic chest pain (p<.05). The majority of children reported that medication (41%), rest (63%), relaxation (41%), heat or cold (39%), distracting activities (36%), and family/friends (39%) provided some pain relief. Twenty-four percent of patients reported taking no medication for pain. Of those taking medication to treat pain, acetaminophen, NSAIDs, or a combination of the two was used. None of the patients reported taking opioids for pain management. It was interesting to note that all the patients who reported head pain were taking some medication, mostly NSAIDs.

Children with CF appear to experience pain, especially abdominal, chest, and head pain, on a regular basis. The finding of frequent chest and head pain is consistent with the work of Ravilly et al. 1 The majority of patients in our study described their pain as mild and of relatively short duration, however a significant number of patients seem to have pain of long enough duration to potentially have a negative impact on daily functioning. A good example of this is that children with chest pain have reduced FEV 1 % and more perceived functional limitations as a result of the pain.

Our study provides further insight into the experience of pain for patients with CF with mild to moderate disease who were seen as outpatients. It is important to note that even in a less sick population, problems with chronic pain were very common. Evidence suggests that pain assessment should be a routine part of the clinical care of children and adolescents with CF. Further research is clearly needed to better understand the sources of pain and how best to provide relief. Pharmacological interventions were minimal in this population and possibly contribute to the under treatment of pain. Effective pain management may serve an integral role in reducing morbidity and improving quality of life in children with CF. Figure 1 . Relationships between pain, coping, and disability. 

Assessment of health related quality of life (HRQOL) in children and adolescents with cystic fibrosis (CF) is important to better understand disease and treatment-related factors that impact function and well-being, and to evaluate the effectiveness of therapies and methods of drug delivery. Limited data exist concerning the HRQOL of youth with CF, although the recent development and validation of CF specific measures of HRQOL for children and adolescents 1 will likely increase the use of QOL measurement in clinical trials. Measurement of how symptoms, such as pain, impact on HRQOL is essential, but these data have not yet been reported in children and adolescents with CF.

There are preliminary data that patients with CF do in fact experience and/or suffer from pain related to their disease or complications of their disease. In the only published study of chronic pain in CF, headaches, chest pain, back pain, abdominal pain, limb pain, arthritis, and neuropathic pain were experienced by the majority (84%) of an older sicker pediatric population with CF. 2 It is well documented that chronic pain can have a significant influence on HRQOL of youth by impacting a variety of normal daily activities, such as school attendance and social interactions as well as psychological functioning . 3 In addition, pain may limit a child's ability to take deep breaths, cough, and tolerate pulmonary toilet exercises that are so impor-tant for patients with CF. Therefore, our research group initiated a study, the DCH Pain in CF study, to examine the impact of pain on the HRQOL of children and adolescents with CF.

A convenience sample was recruited from the CF center at Doernbecher Children's Hospital (DCH). The study sample consisted of 46 children and adolescents with CF (24 males, 22 females) and their parents; n = 30 child subjects (ages 8-13 years) and n = 16 teen subjects (ages 14-17 years). Participants completed two survey instruments to measure HRQOL and pain characteristics. The Cystic Fibrosis Questionnaire (CFQ) 1,4 is a disease-specific quality of life measure, with different versions for children, adolescents, and parents. The CFQ Teen Version includes 12 subscales (the child version includes 8 subscales) to assess CF specific symptoms and perceptions of HRQOL. Higher scores indicate better quality of life. Pain characteristics were measured using a self-report questionnaire assessing frequency, location, duration, intensity, and emotional upset due to pain. Adequate reliability and validity have been reported in the investigators' previous research with this pain questionnaire. 5 In addition, to assess disease severity, forced expiratory volume in 1 second (FEV 1 percent) was obtained from the most recent clinic visit.

As expected, findings demonstrated that having frequent pain (at least once/week) produced decrements in the HRQOL of children and adolescents with CF. Specifically, children with frequent pain reported increased treatment constraints compared to children with less frequent pain. Adolescents with frequent pain, on the other hand, reported negative impact on their HRQOL in many areas including physical and role functioning, vitality, eating disturbances, respiratory and digestive symptoms, as well as overall perception of their health, compared to adolescents with less frequent pain. CF specific symptoms, in particular, respiratory and digestive symptoms, were associated with more frequent, longer lasting, and more intense pain that is often in more than one location for both children and adolescents. Several specific pain locations were related to decreased HRQOL. In particular, having chest pain was related to decrements in physical, emotional, and role functioning, increased treatment constraints, respiratory symptoms, and poorer overall health perception. The presence of abdominal/pelvic pain was associated with more digestive symptoms. Disease severity was also associated with HRQOL; a significant positive association was found between FEV 1 percent and respiratory symptoms.

Children and adolescents with CF, who are seen in an outpatient setting, experience frequent pain that is associated with decrements in their HRQOL. In the DCH Pain in CF study, adolescents experienced the greatest negative impact on their HRQOL from frequent pain. Moreover, patients experiencing chest or abdominal pain were particularly at risk for having more CF specific symptoms and reduced HRQOL.

Our study highlights the need for future research to identify methods to reduce pain and improve quality of life, particularly in adolescent patients with CF. Effective pain management is likely to include both pharmacological and psychological interventions. The role of biobehavioral or psychological interventions for CF has received limited research attention. A recent systematic review 6 identified only eight studies of psychological interventions for CF, and none were focused on pain relief. However, the three studies that used biofeedback, music, and massage therapy to assist with physiotherapy found some effectiveness. 6 Such interventions may also play a potential role in providing pain relief to these children and adolescents. Pain relief should constitute an important treatment goal for youth with CF, as adequate pain treatment is expected to lead to broad improvements in health-related quality of life. 

Pain in patients with cystic fibrosis (CF) results from discomfort arising from the disease (e.g., chest pain, gastrointestinal complaints, headaches, renal stones, arthritis, and end-stage disease), and caused by therapies (e.g., chest physiotherapy, intravascular access, and surgery.) Characterization of an individual's pain is necessary in order to treat it effectively. Types of pain include acute, recurrent, and chronic. Patients' experience of pain is dependent not only on its physical nature, but also on their emotional, psychological, and spiritual states. Social factors usually contribute to the impact of pain in patients' lives.

Evaluation of pain involves review of physical and psychological factors. Patients should be asked regarding the onset, duration, and frequency of the pain. The nature of the pain can be described by its location, distribution, quality, and intensity. Psychological factors that impact the experience of pain include anxiety, depression, substance abuse, and certain behaviors, e.g., reinforcement by caregivers who attend to the patient when pain is reported. In order to gain insight into how patients perceive their pain, they should be queried about their beliefs regarding the reason for the pain, its meaning, when it is most bothersome, what would change when the pain improves, and their expectations of the therapy for pain.

Hypnosis is useful in the treatment of pain because it can affect perception of pain, and psychological interactions with the experience of pain. Hypnosis can be defined as an altered state of consciousness, characterized by a heightened state of suggestibility, which allows instruction and alteration of perception based on patients' specific needs. Hypnosis may block the pain signal by allowing the triggering of a substitute signal through focus on an alternative thought, which usually is pleasant. Hypnosis can reduce acute pain, chronic pain, depression, and anxiety, and can decrease dependence on pain medications. It is more effective in controlling pain than placebo, relaxation, distraction, or alpha-feedback.

Self-hypnosis for the control of pain can be taught by a qualified hypnotherapist or members of the CF care team who have been trained in hypnotherapy, e.g. physicians, nurses, social workers, or psychologists. Many patients can be instructed in 1 or 2 sessions, which require 5-50 minutes. Some patients need several additional sessions, especially if there is a large psychological component to their presentation. Steps for instruction in hypnosis can include a pre-hypnotic interview, induction, deepening, hypnotic suggestions, and trance termination. Techniques that can be used in hypnosis to help improve patients' mastery of their pain experience include egostrengthening, dissociation, time distortion, amnesia, positive visualization, and validation. Self-hypnosis can be used to reduce the intensity or develop complete anesthesia at the site of the pain. The pain sensation can be altered, e.g., changing a sharp pain to a dull, warm, or tingling sensation. Patients also can learn to relocate pain to a part of the body where it is not as bothersome.

The patient was an 11-year-old with CF who became agitated whenever he required phlebotomy, which typically required 30 minutes, as he became very upset, cried, and struggled. The patient explained that he was frightened by the thought of a needle in his arm. Use of a topical anesthetic was not helpful, probably because it did not alleviate his anxiety about the procedure.

The patient was interested in learning how to use selfhypnosis to help himself. A single 10-minute instruction session was offered by his cystic fibrosis physician. Hypnosis induction was achieved by rolling his eyes back as far as possible, inhaling deeply, and closing his eyes as he exhaled. Deepening was achieved with the suggestion that the patient imagine being in a favorite place. The patient was coached to imagine what he might see, hear, smell, feel, and taste there. In order to promote dissociation, it was explained that the more he paid attention to his different senses the more relaxed he could become. To help with his comfort control during phlebotomy, the patient was given the hypnotic suggestion that he imagine his antecubital fossa was covered by a magic sleeve that reduced his perception of pain at that site. Following hypnosis, the experience was validated when the patient was asked to pinch his antecubital fossa and reported no discomfort. He was congratulated for his success, thus providing ego-strengthening, and encouraged to use self-hypnosis as he saw fit, which reinforced his sense of self-mastery.

Subsequently, the patient was proud and almost eager to undergo phlebotomy with use of hypnoanalgesia alone. He reported feeling more confident. In later years the patient used self-hypnosis to reduce or eliminate headaches and abdominal pain, as well as to gain insight into how to cope better with stressful situations.

Hypnosis is best taught by those who have had specific training in hypnosis and hypnotherapy. Instruction to uninterested patients is likely to be of little benefit because the efficacy of hypnosis is affected by the patients' desires. Hypnosis sometimes can be used instead of medications for the treatment of pain, but oftentimes its optimal role is to augment the effectiveness of medical therapy.

In conclusion, self-hypnosis for patients with cystic fibrosis can improve mastery of their response to the disease, cooperation with therapy, and comfort for patients, family members and the health care team.

The primary defect in CFRD is insulin deficiency. Abnormal glucose tolerance is present in the majority of adult CF patients because of fibrosis-related partial islet destruction. Progression to overt diabetes, however, appears to require diminished function of the remaining beta cells, and this may be related to the same genetic defect that causes diminished beta cell function in type 2 diabetes. Insulin resistance must also play a role. In several different studies, insulin resistance has been reported to be decreased, normal, or increased. It is likely that these contrary observations have been related to different degrees of illness severity in the patient populations that were studied.

Why do we care if CF patients have diabetes? CFRD contributes to morbidity and mortality in CF. A prospective study followed pulmonary function for four years in 152 patients separated at baseline by OGTT testing into three groups: normal glucose tolerance, impaired glucose tolerance and diabetes without fasting hyperglycemia. At baseline, the groups did not differ in age, weight, pulmonary function or bacterial colonization. The rate of decline in pulmonary function over the fouryear observation period was directly related to the severity of glucose tolerance abnormalities at baseline. Pulmonary function deteriorated more rapidly in patients with IGT than those with NGT, and patients with diabetes without FH had the greatest decline. When patients were grouped at baseline based on the amount of insulin secreted during the OGTT, the degree of pulmonary decline over the subsequent four years was directly related to the degree of insulin deficiency at baseline, suggesting a cause and effect relationship between insulin deficiency and CF clinical deterioration. Insulin is a potent anabolic hormone that plays a pivotal role in carbohydrate, protein and lipid metabolism. Insulin deficiency may contribute to morbidity and mortality in CF by promoting an overall catabolic state. Mounting evidence supports the use of insulin to improve nutritional status and thus pulmonary function in CF.

A retrospective survival analysis at the University of Minnesota in 1988 found that while nearly 60% of CF patients without diabetes were still alive at 30 years of age, less than 25% of those with diabetes reached this age. Since 1988, we have been aggressively treating all patients with fasting hyperglycemia with insulin. A new survival analysis was performed on 1081 patients followed at the University of Minnesota over the last 25 years. Median survival time after a diagnosis of CFRD was 5.4 years, but after the first 5 years the survival curve was less precipitous, suggesting that the patients dying in the first few years after diagnosis were probably sicker than the long-term survivors. Median survival was 49.5 years for non-diabetic males, 47.4 years for diabetic males, 47.0 years for non-diabetic females, and 30.7 years for diabetic females. Thus, the diagnosis of diabetes had a dramatic negative impact on survival in women but not men with diabetes. This did not appear to be related to BMI, genotype, pulmonary pathogens, steroid use, or pregnancy. Women had worse PFTs at the time of diagnosis. We hypothesize that gender-related hormonal factors may modulate the impact of diabetes on CF prognosis, with either a deleterious effect of estrogen or a beneficial anabolic effect of testosterone.

Holley F. Allen, MD Department of Pediatrics, Baystate Medical Center Children's Hospital, Springfield, MA Cystic fibrosis related diabetes mellitus (CFRD) is different in etiology, course, and treatment than type 1 or type 2 diabetes. As life expectancy CF patients increases, so does the prevalence of CFRD. Evidence to support standardized recommendations for screening and diagnosis of CFRD is just emerging; such evidence to support specific treatment modalities is nearly non-existent.

CFRD occurs in CF subjects with pancreatic insufficiency, and increases with age, reaching a prevalence rate of over 30% in those over age 30 (1) . Estimates of the prevalence of CF vary from center to center, likely related to the strategies used for screening and diagnosing CFRD. CFRD is caused by insulinopenia and is exacerbated by insulin resistance. Initially it has an indolent course with absence or intermittent occurrence of the classic symptoms of hyperglycemia (polyuria and polydipsia) mostly during episodes of acute illness. Ketoacidosis is extremely rare.

Clinical status has been shown to decline prior to the development of overt symptoms of diabetes (2;3). To prevent the catabolic state which is present during insulinopenia, screening for abnormalities of glucose metabolism is advocated by most experts in the field.

OGTT is the gold standard for detecting abnormal glucose metabolism. The issues around recommendation of universal OGTT screening for CF patients arise primarily because: 1)OGTT is felt to be a cumbersome test to perform, 2)intermediate results such as impaired glucose tolerance (IGT) fasting<110, 2-hr impaired (140-200 mg/dl) may normalize in subsequent tests, and 3) no solid evidence shows that the asymptomatic CF patients who have diabetic glucose tolerance without fasting hyperglycemia (DGT), (fasting<110, 2-hr>200mg/dl) benefit from treatment.

Although other tests including HbA1c, FBG and actual BG are frequently used by physicians caring for CF patients (4), none have consistently demonstrated the sensitivity required of a screening test. HbA1c has been in the normal range in 16%(5) to 70%(6;7) of subjects at the time of CFRD diagnosis. Numerous groups have evaluated use of a combination of FBG and HbA1c and have reported that they are NOT reliable in identifying CFRD(8) (2) (9). Annual OGTT screening for CF subjects over the age of 10 has been advocated by several experts in the field (8) . Increasing numbers of U.S. centers are screening with OGTT. A CFRD Consensus Conference in 1998(10), however, did not recommend universal OGTT screening, but rather an algorithm based on random plasma glucose in asymptomatic subjects Ongoing research showing benefit from early detection of CFRD is needed to justify universal OGTT screening.

The goal of treatment of CFRD is to achieve optimal nutritional and clinical status by maintaining near normoglycemia, to avoid both short and long term complications. Because CFRD is a disease characterized by insulin insufficiency, insulin is generally the treatment of choice. Early in the development of CFRD, significant insulin secretory capacity remains. Most problems (hyperglycemia, decreased protein synthesis) occur in the post-prandial state, suggesting a possible role for short acting insulin secretagogues. Sulfonylureas have been used by some with success comparable to insulin early in the disease process (11) . Sulfonylureas bind with the sulfonylurea receptor (SUR), which like CFTR is a member of the ATP-binding cassette superfamily of proteins. Theoretical concern exists because of evidence that sulfonylurea like drugs bind to and inhibit the CFTR. Because of these concerns, sulfonlyureas should only be used by practitioners very aware of the risk benefit ratios.

Many insulin treatment regimes are available. The approach to insulin therapy needs to be tailored to the subject's preference, stage of disease, and ability to adhere to a particular regime. 1. Maximum flexibility with best post-prandial glycemic control is achieved using rapid acting insulin administered at the time of food consump- 

Dana S. Hardin, M.D.

CF related diabetes (CFRD) is a frequent clinical problem. The North American CF Foundation has greatly improved the knowledge and management of this disease by providing research funding, hosting the CFRD consensus conference and providing educational opportunities and written materials for CF caregivers. Although the current clinical initiatives are appropriate for many CF patients, particular patient subsets may have unique needs. Our group has had experience with both research and clinical management of two such groups, Hispanics with CF and pregnant women with CF, and will review what we have learned. A relatively unexplored method of treating patients with CFRD is the insulin pump. We have just completed a study of its use in CFRD and will share our findings as part of this seminar.

We studied 9 pregnant women and found that 88% developed gestational diabetes (GDM) by 12 weeks gestation, 99% by 24 weeks gestation and 100% by week 32 gestation. Our studies of the metabolic changes caused by pregnancy revealed that pregnant CF women have lower insulin levels, higher hepatic glucose production and more peripheral and hepatic insulin resistance than normal pregnant women. Insulin secretion did not increase during pregnancy, hepatic glucose production increased and insulin resistance worsened. We believe that these defects contribute to earlier presentation and greater incidence of GDM in CF. We also measured protein turnover and learned that pregnant women with CF are more catabolic than normal pregnant women and that catabolism worsened during pregnancy. Catabolism correlated with less gain in lean mass in the pregnant CF women. Furthermore, despite markedly greater caloric intake, pregnant women with CF had significantly less weight gain than normal pregnant women. Based on our findings, we recommend that women who are actively seeking pregnancy be screened for CFRD. During pregnancy, we recommend an oral glucose tolerance test (glucose load 100 grams) be administered each trimester to any woman not yet diabetic. Women who are diabetic should be treated with insulin to improve glycemic con-trol, improve weight gain and prevent muscle loss from on-going catabolism.

Although Cystic Fibrosis (CF) occurs less often in Hispanics than in Caucasians, clinicians have long recognized that the disease progresses more quickly, and the morbidity and mortality are higher, in this patient group. These observations have been supported by several research studies. We hypothesize that one major reason for increased morbidity and mortality in Hispanic CF patients is increased glucose intolerance occurring at a younger age. We believe the earlier development of CFRD and glucose intolerance is secondary to underlying insulin resistance, a hallmark of type 2 diabetes, which is worse in Hispanics. Our review of medical records supports our hypothesis and describes an incidence of normal random glucose levels in less than 5% of Hispanic children and 0% of Hispanic teens. Data from Moran et al collected at University of Minnesota indicate normal glucose tolerance occurs in 57% of children and 36% of teens.

To date we have performed measures of insulin secretion and insulin resistance using a frequently sampled IVGTT in 12 Hispanics with CF. We have found that insulin resistance is greater in these patients than in agematched Caucasians. Another possible factor affecting increased risk of CFRD in Hispanics may be advanced clinical disease. These patients tend to be sicker. Previous studies have identified a mechanistic role of cytokines, specifically TNF-_, in the development of insulin resistance, and increased hepatic glucose production. Elevated levels of IL-6 and TNF-_ have been associated with protein catabolism. We measured TNF-α in 19 CF children and found higher levels in the Hispanics (TNF-( = 98 ± 10 pg/ml) than in the Caucasian children (TNF-_ = 76 pg/ml). TNF-_ levels correlated with protein catabolism. Furthermore we have found poorer longitudinal growth and weight gain in Hispanic children, another indicator of worsened clinical status.

The general recommendation for screening for CFRD is to wait until age 13. We recommend that all Hispanic

Carol Brunzell R.D., L.D., C.D.E.

Maintenance of a healthy body weight and optimal nutritional status are critical to survival for the patient with CF. With the additional diagnosis of CFRD, nearnormal blood sugar control is essential to normalize metabolism of macronutrients and to facilitate weight gain and weight maintenance (1). These patients are at risk for diabetic microvascular disease, so optimal control of blood glucose is imperative for prevention (2, 3, 4) . The risk of macrovascular disease necessitating the typical lowfat, low-sodium diet restrictions for type 1 and type 2 diabetes do not apply to the patient with CFRD as macrovascular complications appears to be non-existent at the present time (5, 6) . However, one study showed that patients with CFRD had a 4% prevalence of elevated cholesterol and 16% prevalence of elevated triglycerides, suggesting that this recommendation may change as these patients live longer (7) . The typical diet for CF, high-calorie, high-fat, and high-sodium remains essential to optimize weight and nutritional status.

To meet the increased energy demands of CF many patients require some form of nutrition support in the form of oral or gastrostomy-delivered supplements. Fluctuations in appetite make a rigid meal plan unrealistic. Prescribing a complicated meal plan in addition to an already complicated medical regimen is not necessary today due to the dramatic improvements in insulin types and delivery systems. The need for simplicity is imperative. The CFRD consensus committee recommended matching insulin to carbohydrates for maximum flexibility as one approach to management (1) The use of carbohydrate counting is a simplified way to superimpose some diet organization onto the traditional CF diet and works well with the use of rapid-acting insulin to optimize blood sugar control.

The rationale for carbohydrate counting is based on the relative effects of carbohydrate, protein, and fat on blood glucose, with dietary carbohydrate having the most significant effect compared to protein and fat. (8, 9) The advent of insulin pump therapy and the insulin Lantus TM has made carbohydrate counting user-friendlier as many patients now administer rapid-acting insulin according to the amount of carbohydrates they plan on eating. One study showed excellent glycemic control in patients with type 1 diabetes using 1 unit of regular insulin for each 10 grams of carbohydrate consumed. Glycemic control was maintained over a wide range of carbohydrate ingested and was not affected by the glycemic index, fiber, caloric or lipid content of the meals (10) .

To confirm the approximate carbohydrate-to-insulin ratio, self-monitoring of blood glucose (SMBG) data and food records for at least 3 days are recommended, with pre-meal and frequent 2-hour post-prandial glucose values to verify the ratio. Most patients require approximately 0.5-2.0 units rapid-acting insulin per 15 grams carbohydrate (11) . While using fixed doses of insulin may not be the best approach, patients receiving fixed doses of insulin need to be consistent with their carbohydrate in conjunction with the time action of the insulin injected (1).

Currently, most patients who have CFRD without fasting hyperglycemia are not treated with insulin unless they are unable to maintain an appropriate weight or if pulmonary function is declining more rapidly than expected. At the University of Minnesota, the strategy in diabetes not children ages 5 and over receive an OGTT at least yearly, and random blood glucose levels should be checked during every acute illness. Further study may indicate a need for pre-emptive treatment for these children who are at high-risk for CFRD.

The insulin pump has been utilized for treatment of type1 diabetes, including pediatrics. The pump's benefits include adjustable basal coverage and the ability to give bolus insulin for all meals and snacks without giving an injection. CF patients require high caloric intake. Constant carbohydrate intake (a method for regulating blood sugar when traditional insulin is given) is not recommended for these patients. For this reason, the insulin pump may provide ideal therapy for CFRD. We have just completed a study on the use of the insulin pump in 9 CFRD patients. We found that as compared to multiple injections per day of subcutaneous insulin, the pump improved weight gain and lowered hemoglobin A1c. Reduced protein catabolism was also noted. All patients but one requested insulin pump therapy at the conclusion of the study. Based on our findings we believe the pump is a beneficial treatment for CFRD. As part of this seminar, we will share our experience regarding basal and bolus settings for the pump. treated with insulin is to spread carbohydrates throughout the day to minimize large carbohydrate loads without reductions in total calories. Patients obtain a blood glucose profile periodically at home to monitor status.

Patients with impaired glucose tolerance are at high risk of progressing to CFRD. The risks of IGT in the general population with regards to cardiovascular disease do not appear to be of concern for the patient with CFRD (5,6). The risk of microvascular disease with IGT is not known at present. Unlike diet strategies used in the general population, which recommended weight loss and a low fat diet for people with IGT, it is never appropriate to recommend weight loss or a restriction of fat and calories in CF patients. The only potential restriction may be to minimize the excessive consumption of regular soda or other sweetened beverages, and to try to maximize intake of more nutrient-dense foods to prevent weight loss. Spreading carbohydrates throughout the day may also be beneficial. It is internationally recognised that urinary incontinence (UI) is a common problem for women with CF. There is a reported prevalence of between 35-68% (1,2,3) . This compares to an approximate 8-13% prevalence of UI in healthy young women (4, 5) . UI is also reported as a problem for male CF patients (6) . There is an increased prevalence of UI described in children with CF (3). A wide-ranging age at onset is reported, from 5 to 45 years of age (1, 3) . Whilst acknowledging the symptoms of urinary leakage in children and adults with CF, this report will focus on the possible mechanisms of UI in women with CF.

The mechanism for preservation of continence is known to be a complex interaction of the muscles, fascia and ligaments of the abdomino-pelvic capsule and a competent urinary sphincter. In the general population risk factors for urinary leakage include female gender, obesity, genetic causes such as collagen weakness and conditions which increase intra-abdominal pressure (IAP) such as constipation, vaginal delivery and persistent cough (7) . Leakage may be related to features of urge and/ or stress incontinence. Urge incontinence occurs when leakage is accompanied or preceded by urgency, whilst stress incontinence occurs when the pressure within the bladder exceeds that of the maximum urethral closing pressure. It is necessary to perform urodynamic assessment in order to provide a diagnosis of genuine stress incontinence (8) . The aetiology of UI in CF has not been fully evaluated.

Cough is recognised as the major cause of leakage in CF, where long spells of intense coughing are seen (1, 3, 9) . Normally in healthy subjects the pelvic floor muscle (PFM) contracts prior to voluntary abdominal muscle activity suggesting a pre-programmed response to IAP (10) . However in the presence of increased stress to the pelvic floor (PF), for example with chronic cough, the continence threshold may be exceeded. Indeed chronic cough is recognised as a poor prognostic indicator for success of treatment of UI (11) . Coughing often occurs in a position of spinal flexion in which the urethral closing pressure is diminished.

Abnormal posture, which is increasingly reported in CF, may be implicated in the mechanism of urinary leakage (12, 13) . Increased kyphosis, over strengthened upper abdominals and a weaker transversus and multifidus are reported. It is recognised that the muscles of the abdomino-pelvic capsule play a role both as postural and respiratory muscles and therefore it is possible that dysfunction of any of these synergistic actions may contribute to the symptoms of urinary leakage (14) . All of these factors may reduce the ability of the PF musculature to provide timely, co-ordinated and sufficiently strong contraction to maintain adequate urethral closing pressure in response to increased IAP.

Invasive urodynamic studies are not routinely performed on CF women complaining of leakage as patients are surprisingly symptom tolerant and reluctant to be assessed (1, 3) . In the absence of a definitive diagnosis it is therefore important to recognise other potential causes of UI in relation to both the muscular and neural aspects of continence.

Altered muscle metabolism in female athletes with CF is described and this may suggest differences in the quantity and quality of muscle in CF patients (15) . Autonomic neuropathy in CF is also reported and may disturb the autonomic control of bladder filling and voiding (16) . The elevated inflammatory cytokines evident during pulmonary exacerbation may also have an effect on the muscle strength, as shown in-patients with COPD (17) .

In summary, UI in CF is a widely reported problem and it may be considered over simplistic to suggest that chronic cough is solely responsible for the mechanism of leakage. Multifactorial causes should be considered when posing the question of why UI occurs in CF.

Physiotherapist, spec. resp diseases, Bsc, Lund CF team, Lund University Hospital, Lund, Sweden Many children, adolescents and young adults with cystic fibrosis (CF) are reported as having a hyperinflated chest, stiff intervertebral and costovertebral joints, bad posture, thoracic kyphosis and back pain [1] [2] [3] [4] [5] [6] . The reported wedging or compression of vertebrae [3, 5] in adult CF populations may be caused by the ergonomic burden of the thoracic kyphosis in combination with osteoporosis [7] .

Pulmonary hyperinflation is a response to the clogging and collapse of airways in an attempt to keep airways open and ventilated, i.e. to keep the functional residual capacity (FRC) at the level of the closing vol-ume. Most often the degree of hyperinflation follows the degree of obstruction. With hyperinflation, expansion of the chest is accomplished by contracted inspiratory muscles. Expiratory flow and expired volume is controlled by eccentric contractions of inspiratory muscles and inspiration is accomplished by concentric contractions, but from an unfavourable point on the length/tension curve. The more hyperinflated the chest, the closer the inspiratory muscles get to active insufficiency. Atelectasis occurs if airways are not kept open, resulting in decreased ventilation distribution, and further increased breathing frequency, minute ventilation and work of breathing. Accessory muscles become involved in the process of hyperinflation. As hyperinflation increases and as ventilation becomes more insufficient these muscles are progressively more involved in the mechanics of breathing. Accessory muscles originate from the chest and insert on the upper limbs, cervical region and skull. Concentric contraction assists in the inspiratory movements required of the severely hyperinflated chest when upper limbs and head are fixed while sitting upright, leaning forward with elbows on a table and the head resting in the hands [8] .

All soft tissues shorten when not stretched to full range regularly. Shortening of the musculoskeletal tissues in the hyperinflated chest results in the elevated and protracted shoulders, the thoracic kyphosis, cervical lordosis and the decreased mobility of the chest [1, 9] that have been reported. A stiff, hyperinflated chest makes effective airway clearance therapy impossible. It has been shown that bad posture, thoracic kyphosis and back pain are partly reversible if properly treated [1, 6, 9, 10] . But trying to regain what has been lost can be difficult. It is time-consuming, uncomfortable for the patient and an added burden to the CF care package. Thoracic kyphosis due to wedged vertebrae cannot be rehabilitated, but rather increases the load on the adjacent vertebrae. Therefore preventing a stiff chest and bad posture is of great importance. Some simple physical exercises which maintains muscle length and strength and joint mobility as an incorporated part of the daily physiotherapy programme from the very beginning can make a big difference in the long run. Physical loading in upright positions stimulates bone accretion, which may reduce the risk of osteopenia/osteoporosis and the risk of spontaneous fractures including wedging and compression of vertebrae. If physical exercise aiming to preserve physical function is included in the treatment, patients can maintain good posture and chest mobility, even if lung disease should progress.

Nowadays most CF centres recommend physical activity to their patients. However, children and adolescents with CF, despite having good lung function, have been shown to be engaged in less vigorous spontaneous physical activity than their non-CF peers [11] . Whether this is due to less spare time caused by time consuming therapy, to protective parents or health care system, or to the disease itself can be discussed. Simply recommending patients to be physically active is obviously not enough. They probably need more active guidance and continuing encouragement to become and remain physically active.

All children need physical activity to develop motor maturity and body awareness. If children with CF are given the opportunity to experience pleasure and satisfaction during physical activity, much may be gained for future outcomes. The different types of exercises that should be included in a physical exercise pro-gramme from the very beginning or as soon as possible are [12, 13] : • chest mobility activities/exercises using movements around a vertical, sagittal and horizontal axis • shoulder mobility exercises, especially elevation and external rotation • muscle-strengthening activities/exercises, especially for postural muscles • working capacity training/activities/exercises Exercises must never be uncomfortable. For infants, toddlers, children and adolescents, the activities/ exercises must be stimulating, enjoyable and age-appropriate. Activities/exercises should always be individualized and provided at appropriate times in different settings. Team sports/activities offer parts of what is to be included in the CF care and provide the added benefits of normal social interaction. Good chest mobility allows effective airway clearance therapy. Good posture probably reduces the risk of back pain and spinal complications. Good posture contributes to positive body image and self-esteem.

There is definitive evidence that nutritional status is the single most important factor that determines aerobic and anaerobic exercise performance in subjects with cystic fibrosis (CF) 1,2,3 . However, diminished exercise capacity has been reported in subjects with normal nutritional status 4 , 5 . There are several factors that can limit skeletal muscle function in CF. This can be considered in the following categories:

Intrinsic abnormalities in the skeletal muscle cells of patients with CF have been demonstrated in the mitochondria of fibroblasts and leucocytes and include increased calcium concentration, lower nicotinamide adenosine dehydrogenase activity (respiratory chain enzyme complex) 6 and a higher pH optimum of nicotinamide adenosine dehydrogenase 7 . Moser and colleagues have suggested that there is a muscle related abnormality in oxygen metabolism in patients with CF 8 . Studies utilizing 31-P magnetic resonance spectroscopy 5, 9 during exercise have demonstrated inefficient and sub-optimal aerobic and anaerobic muscle metabolism in subjects with CF and good nutritional status. The mechanism of this remains unclear.

The class of the CFTR mutation may be a factor in determining muscle function 10 . Subjects with a CFTR mutation belonging to either class I or II have a lower peak aerobic capacity and anaerobic power compared to those with class III, IV or V CFTR mutations. In addition, compared to the DD angiotensin converting enzyme (ACE) gene polymorphism, the II polymorphism has been shown to be associated with less end organ damage in subjects with CF 11 and better anabolic response to exercise training programs in healthy subjects 12 .

Girls with CF have a lower peak aerobic capacity 13 and anaerobic power 10 . This gender difference is not unique to CF and have been attributed to differences in androgen levels, efficiency of fat metabolism, ratio of type I to type II muscle fibres and total muscle mass.

A significantly reduced breathing reserve is associated with reduced exercise capacity 14 and habitual activity 15 .

While nutritional status is a major determinant of habitual activity 15, 16 , increasing habitual activity as part of a training program can improve muscle 17 .

Arthropathy occurs in up to 8% of patients with CF and limit activity. This can lead to deconditioning 18 .

Although nutritional status is a primary determinant of muscle function, there are several other factors that can limit optimal performance.

Larry C. Lands, M.D., Ph.D.

With increased life expectancy for CF patients, we are required to pay attention to organ systems beyond the respiratory and gastrointestinal systems, whose dysfunction can cause significant long-term morbidity. Many authors have documented significant decreases in bone mineralization in CF patients 1-3 , with increased risk for fracture and pain.

There are two distinct phases with respect to the skeletal system. In the growing child, especially in the critical peripubertal and pubertal periods, the goal is to achieve maximal bone accretion 4 . Maximal bone mass is achieved by the early 20's and so a great emphasis must be made during the early growth phases. Once adulthood is achieved, the goal is to maintain this bone bank.

There are many influences on bone, including genetic, nutrition, body habitus, hormonal status, and exercise patterns 5 . One of the most significant factors affecting ultimate bone mass is genetics. This is also true for patients with CF, where there is a significant relationship between a mother and child's bone mineral density 6 .

The other major way in which parents can influence bone mineralization is through the encouragement of a healthy diet. A variety of population-based studies have demonstrated the positive influence of vegetable and fruit intake and the supply of vitamins K and C, and minerals, such as magnesium 5 . Certainly, calcium and vitamin D intake are very important. Sources of calcium include milk, tofu, and bok choy. These essential elements are particularly important during the phase of bone accretion in the first two decades of life 7 .

Other factors that play important roles in bone accretion and maintenance include sex hormones and inflammatory factors. CF patients are at risk for delayed puberty and thus delay in bone maturation. At the other end of the spectrum, premature menopause and andropause can lead to loss of bone mineral 8 . Production and circulation of pro-inflammatory cytokines, such as IL-1 and 6, and TNF-_, can affect the balance between osteoblastic bone formation and osteoclastic bone resorption in favor of bone resorption 9 .

Activity, both the amount and type, has a major impact on both bone accretion and retention 10 . A recent study from Australia highlighted these factors 11 . The Australian population is particular in that there is lots of sun exposure, promoting vitamin D formation, and the population is oriented towards being physically active. In this population, adolescent patients maintained good calcium intakes and vitamin D levels and were as physically active as their contemporaries. Yet their femoral neck bone densities were reduced, possibly due to the higher inflammatory level. In the adults, reductions in physical activity also played a significant role in reducing bone densities.

In a healthy pediatric population, differences in the amount of time spent being physically active affects bone mineralization 12 . In CF patients, both lung function and maximal exercise ability, which partially reflects the degree of physical activity, influence bone mineralization, such that higher lung function and exercise capacity are associated with greater bone density 13 . Since both lung function and peripheral skeletal muscle function influence exercise ability in both healthy individuals and CF patients 14 , greater muscular ability will thus positively influence bone mineralization.

Of the exercises which most promote bone growth and maintenance, weight bearing exercises are primordial 10;15-17 . This is applicable to both children and adults.

In summary, there are multiple influences on bone accretion and maintenance, with CF patients facing the additional challenges of chronic inflammation. It is important that a major effort be made to replete the bone reserves by early adulthood. This makes the peripubertal, pubertal, and post-pubertal periods the most critical. Physical activity, especially impact activities and weight-bearing, plays an important role in achieving optimal bone reserves and maintaining these throughout life.

Phosphorylation of the cystic fibrosis transmembrane conductance regulator (CFTR) by protein kinases is thought to be an absolute prerequisite for opening of CFTR channels. In addition, nucleoside triphosphates were shown to regulate opening of phosphorylated CFTR. CFTR contains multiple sites of phosphorylation by cAMP-dependent protein kinase (PKA) and protein kinase C (PKC) [1] . It was also shown that the membrane-associated cGMP-dependent protein kinase isoform II is able to phosphorylate CFTR [2, 3] .

Recently it was reported that in human sweat gland ducts endogenous CFTR is activated by heterotrimeric G proteins via a cAMP-independent pathway [4] . The question arose if alternative pathways exist to activate CFTR, apart from protein phosphorylation. Therefore we tested the effect of phosphatidylinositol 4,5-bisphosphate (PIP 2 ), a known regulator of ion channels and transporters [5] , on plasma membrane patches of oocytes, heterologously expressing human CFTR.

Our study demonstrates that phospholipids, like PIP 2 , enable kinase-independent activation of CFTR, resulting in ATP-responsiveness of PIP 2 -treated CFTR [6] . PIP 2 alone is not sufficient to open CFTR but ATP opens nonphosphorylated CFTR after application of PIP 2 . The effect of PIP 2 is independent of protein kinases, as PIP 2 activates CFTR in the complete absence of Mg. Phosphatidylinositol (PI) and phosphatidylinositol monophosphate (PIP) activate CFTR less efficiently than PIP 2 . PIP 2 application to phosphorylated CFTR may inhibit the CFTR chloride current. We suggest that regulation of CFTR by PIP 2 is a previously not recog-nized, alternative mechanism to control CFTR-mediated chloride conductance.

Supported by the Max-Planck-Society and the German Research Foundation(DFG). CFTR channels help determine the rates of chloride and fluid transport across the epithelia affected in cystic fibrosis. Early studies revealed two general modes of regulation, one mediated by secretagogues that elevate cAMP and activate cyclic-AMP dependent protein kinase (PKA), and another mediated by agonists that activate protein kinase C (PKC) and/or mobilize cell calcium. Maneuvers that elevate cAMP induce sustained secretion across dog trachea and other preparations whereas global stimulation of PKC by phorbol esters leads to robust secretion lasting 20-30 min followed by profound inhibition 1 . The stimulatory effect of PKC is due, in part, to activation of apical CFTR channels 2 whereas its antisecretory action probably reflects inhibition of basolateral potassium conductance, which is needed to maintain the driving force for apical chloride exit.

A direct role of CFTR in phorbol ester-stimulated secretion is implied by the presence of many consensus sequences for protein kinase C (PKC) phosphorylation on the R domain which, like the PKA sites, are highly conserved across species. The physiological significance of the PKC sites is less obvious than for PKA sites, however, since exposing inside-out membrane patches to PKC in combination with the diacylglycerol analog 1,2dioctanoyl sn-glycerol (DiC8) causes only a slight increase in channel activity (a few % compared to PKA). This weak response can not be attributed to the use of DiC8 as lipid activator because similar results are obtained when PKC is added in a cocktail containing phosphatidylserine and diacylglycerol under conditions that increase PKC activity by 100-fold. Rather than acting alone, PKC phosphorylation seems to regulate CFTR gating primarily by enhancing the rate and magnitude of its response to PKA, effects which are most noticeable in patches after channel activity has been allowed to run down. PKC phosphorylation of CFTR channels in vivo is undoubtedly regulated by hormones and/or transmitters but appears partially "constitutive" in cultured cells, perhaps a reflection of high PKC activity in cells soon after they are removed from medium containing serum and its growth factors.

CFTR regulation is often studied using potent activators (eg forskolin, cpt-cAMP, phorbol esters, etc), but to assess the physiological role of a pathway it is preferable to expose cells to a "first messenger" (eg VIP, acetylcholine, neurokinin) and block the pathway of interest. Pharmacologic inhibitors, antisense oligonucleotides, dominant-negative mutants and small interfering RNAs have all been used in various preparations, although specificity is always a concern. Chelerythrine and Gö6976 were both used to establish the role of PKC in stimulation of CFTR by PKA 3 , but it is now clear that chelerythrine is less potent and less specific than Gö6976 or the other bisindoylmaleimides (eg BIS1 or BIS10 4;5 ) and should be avoided. Interactions between signals could arise if an agonist activates more than one pathway or if multiple secretagogues stimulate the cell simultaneously. The role of such interactions in regulating CFTR has not been studied but may be important in vivo since several transmitters are often co-released and cells also receive inputs from circulating hormones and locally-generated agonists such as ATP, adenosine and NO. Sorting out the relative contributions and crosstalk between these signals remains a major challenge, however a key factor will probably lie in the subcellular localization of receptors and signaling molecules. The fact that phorbol esters have opposing actions at the apical and basolateral membranes itself implies that transport is controlled by multiple PKC isozymes having restricted distributions. PKC signaling involves activation by phospholipid and/or calcium, translocation to the site of action (eg plasma membrane), and binding to a receptor for activated C-kinase (RACK) 6 . RACKs are ideally suited to orchestrate localized signaling responses because they are located near PKC sub-strates and are specific for particular PKC isozymes. During secretion, basolateral uptake of chloride into human tracheal cells by sodium-potassium-chloride cotransporters is regulated by the PKC_ and PKC_ whereas apical chloride channels in the Calu3 cell line are regulated by PKC_ 7;8 . PKC_ binds to RACK1, which in turn binds the scaffolding protein EBP-50 (ezrin-radixin-moesin binding phosphoprotein of 50 kD); 9 ). Regulation of these associations appears complex since phosphorylation of a PKC site on the PDZ1 domain of EBP-50 can disrupt the PDZ1-CFTR interaction 10 . Together these findings raise the possibility that when activated, PKC could potentially regulate its own targeting to CFTR.

Finally, choosing the right preparation will be important for physiological studies of PKC signaling as cell lines may differ in their expression of receptors, anchoring proteins, PKC isozymes, and apical-basolateral polarization. Using first messengers rather than artificial activators such as phorbol esters should minimize artifacts caused by hyperstimulation, and isozyme-specific translocation inhibitor and activator peptides are available which act specifically on PKC_ or other PKC isozymes. These tools in combination with mutant channels that lack PKC sites but are fully competent for gating (eg the "6CA revertent" mutant 11 ) should provide new insights into the regulation of CFTR by PKC. CFTR is a chloride channel in the ATP-binding cassette (ABC) transporter protein family and contains its defining features, two membrane-spanning domains and two nucleotide-binding domains (NBDs) (1, 2) . Like other ABC transporters CFTR can bind and hydrolyze ATP. Earlier work has shown that enzymatic activity is required for normal channel gating (3, 4) . But this has long seemed puzzling because ion flow is passive without a fixed stoichiometric relationship to ATP hydrolysis. Furthermore, no other ion channel is known to require the energy of ATP hydrolysis for gating. Therefore we asked whether CFTR had another ATPdependent enzymatic activity that released less energy. In previous work we found that a recombinant NBD2 polypeptide had ATPase activity (ATP ( ADP + P i ) in the presence of ATP alone. But when AMP was also present, the NBD2 polypeptide functioned as an adenylate kinase (ATP + AMP ( ADP + ADP) and its ATPase activity was suppressed (5) . Under physiologic conditions an adenylate kinase reaction is reversible and releases very little energy (6) . We therefore tested the hypothesis that CFTR has adenylate kinase activity that gates the channel.

We did five sets of experiments. We first asked whether Ap 5 A (P 1 ,P 5 -di(adenosine-5') pentaphosphate), a well characterized inhibitor of adenylate kinases that does not inhibit ATPases, would inhibit CFTR. We found that Ap 5 A inhibited CFTR currents by reducing the channel opening rate. Our data with different di(nucleoside) polyphosphates indicate that the inhibition involved simultaneous binding to one of CFTR's two ATP binding-sites and to an AMP binding-site.

Second we asked whether AMP would influence CFTR gating. We found that AMP did not evoke currents in the absence of ATP but could increase currents at non-saturating ATP concentrations by increasing the channel opening rate. AMP altered the relationship between ATP concentration and current from one with no apparent cooperativity between the two ATP-sites to one with positive two-site cooperativity for ATP, indicating a different gating mechanism.

Third we asked whether the effect of AMP involved ATP:AMP phosphotransfer. To address this question we tested the effect of AMP-NH 2 (adenosine 5'monophosphoramidate), an AMP analog that cannot act as a phosphate acceptor. We found that AMP-NH 2 did not mimic the effect of AMP but inhibited gating produced by ATP alone (i.e., ATPase-dependent gating). AMP, but not ATP, reduced AMP-NH 2 inhibition, indicating that the inhibition occurred because AMP-NH 2 bound to the AMP-site, which is distinct from the ATP binding-site.

Fourth we asked whether ADP would alter CFTR gating through the readily reversible adenylate kinase reaction. We found that ADP inhibited ATP generated CFTR currents. In addition, like AMP, it induced positive twosite cooperativity for ATP. ADP-NH 2 (adenylyl 5´-phosphoramidate), an ADP analog that does not allow phosphotransfer, and GDP, which does not bind to an adenylate kinase AMP-site, failed to induce positive cooperativity. These results support the conclusion that phosphotransfer between two bound nucleotide diphosphates induced positive cooperativity.

Fifth we tested the NBD contribution to CFTR adenylate kinase activity. We found that the mutations K1250A and D1370N in NBD2 and the homologous NBD1 mutations K464A and D572N reduced the potency of ATP to stimulate CFTR currents, indicating a reduction in ATP binding and hydrolysis. The mutations K1250A and D1370N also abolished current stim-ulation by AMP and current inhibition by Ap 5 A while K464A and D572N did not. These differential effects suggest that NBD2, and not NBD1, contains the adenylate kinase activity that gates the channel. In addition, we studied the effects of the CF-associated mutation N1303K in NBD2. This mutation did not reduce the potency of ATP to stimulate currents, but abolished AMP-induced current stimulation and Ap 5 A inhibition, indicating that this mutation may be related to the AMP-site.

In summary, our data indicate that CFTR has adenylate kinase activity that regulates channel gating. When functioning as an adenylate kinase, CFTR showed positive cooperativity for ATP suggesting its two nucleotide-binding domains may dimerize. Thus, channel activity could be regulated by two different enzymatic reactions, ATPase and adenylate kinase. Both activities share a common ATP binding-site in the second nucleotide-binding domain. Our data indicate that it is the interaction of AMP with the AMP-site that induced adenylate kinase activity and suppressed ATPase activity. The AMP K m of 73±20 µM for CFTR gating suggests that CFTR will function as an adenylate kinase in vivo. Thus, at physiologic nucleotide concentrations adenylate kinase activity, rather than ATPase activity may control gating, and therefore involve little energy consumption. This work was supported by the NHLBI (HL29851-21 and HL1234-05), the Deutsche Forschungsgemeinschaft (Ra682/3-2 and Ra682/5-1) and the HHMI. Phosphorylation of CFTR in the R domain by protein kinase A (PKA) is essential for the channel to be active. However, after phosphorylation, ATP is required to open phosphorylated CFTR channels (Gadsby and Nairn, 1999; Zou and Hwang 2001) . ATP-dependent gating of CFTR is studied in inside-out membrane patches excised from NIH3T3 or CHO cells heterologously expressing wild-type or mutant CFTR. The opening rate of the channel follows a simple Michaelis-Menten function and the closed time histograms show a negative exponential component (Zeltwanger et al., 1999) . Both of these observations are consistent with the idea that CFTR gating is not in microscopic equilibrium (i.e., Injection of free energy from ATP hydrolysis drives the gating transition).

Since the open time is drastically prolonged when lysine 1250 (K1250) or glutamate 1371 (E1371), both critical for ATP hydrolysis, is mutated, it is reasoned that ATP hydrolysis at NBD2 closes the channel (Zou and Hwang 2001) . Several biochemical studies indicate that ATP binding at NBD1 is extremely tight (Szabo et al., 1999; Alexandrov et al., 2002; Basso et al., 2003) presumably because NBD1 does not hydrolyze ATP (Lewis et al., 2004) . If ATP is occluded in NBD1 (i.e., the offrate is extremely slow), association and dissociation of ATP at NBD1 cannot account for the electrophysiologically observed, millisecond-to-second gating transitions (Vegani et al., 2003) . By default, ATP opens the channel by binding to NBD2. Thus, NBD2 controls both opening and closing of CFTR.

Although ample experimental results have supported the role of NBD2 in channel closing, kinetic evidence for its role in channel opening is lacking. Since ADP competitively inhibits ATP-dependent opening of the channel, we reason that determining the site where ADP binds to affect the channel opening rate will identify the NBD that controls channel opening. Homology model of CFTR's NBD2 was built by using the framework of the crystal structure of NBD1 from mouse CFTR (Lewis et al., 2004) . According to our homology model, tyrosine 1219 (Y1219) was identified as the aromatic amino acid that interacts with the adenine ring of ATP at NBD2. Mutating Y1219 to glycine (Y1219G) shifts the ATP dose-response to the right with an apparent affinity >50 fold of that for wild-type channel. ADP inhibition was also dramatically reduced with this mutant. These results are consistent with the idea that ATP as well as ADP binds at NBD2 to modulate channel opening.

It thus appears that NBD2 can open and close the channel. Does NBD1 play any role in channel gating? Biochemical and crystallographic studies of NBDs in bacterial ABC transporters suggest that binding of two ATP molecules at the interface of two NBDs drives NBD dimerization Smith et al., 2002; Chen et al., 2003) . If an equivalent molecular motion happens in CFTR during opening-closing transition, the free energy of ATP binding at NBD1 should contribute to the overall energetics of association-dissociation of NBD dimer. We found that ADP can increase the closing rate of the channel (also see Weineich et al., 1999) , suggesting that ADP may act at another binding site other than NBD2. Mutation of lysine 464 at NBD1 also shortens the open time and the AMP-PNP-locked open time (Powe et al., 2002) , indicating that ATP binding at NBD1 affects the energetic stability of the open (or the locked-open) state. Based on these results, we propose a model that entails energetic coupling of ATP binding (NBD1) and hydrolysis (NBD2) in controlling CFTR gating.

Using automobile as an analogy, we picture that phosphorylation of the R domain serves as the ignition mechanism. ATP binding at NBD1 acts as the clutch that controls the speed of the energy-consumption engine at NBD2. 4 . CFTR null mice have more inflammation and higher mortality than normal mice after airway infection with Pseudomonas aeruginosa 5 . Because airway epithelial cells (AEC) express CFTR and orchestrate inflammation, they may mediate dysregulated inflammation in CF. Hyper-inflammation in CF could result from pro-inflammatory gene expression in AECs that is excessive, prolonged, and/or without stimulation.

Many in vitro models of AECs expressing a CF or non-CF phenotype demonstrate increased IL-8 release by CF cells either spontaneously or after exposure to inflammatory stimuli. The models include AEC lines from CF patients complemented with wildtype CFTR 6 and non-CF AEC lines with decreased CFTR function due to expression of the CFTR regulatory domain or antisense oligonucleotides 7, 8 . In one CF cell line, IL-8 release was prolonged after removal of the inflammatory stimulus 8 . Increased constitutive or induced IL-8 release has also been reported from primary CF AECs, including both surface cells cultured under submerged or airliquid interface conditions and gland cells [9] [10] [11] . The observation that multiple CF model systems from different laboratories display greater pro-inflammatory gene expression provides strong support for the intrinsic hyper-inflammatory hypothesis.

Altered expression of inflammatory mediators other than IL-8 has been demonstrated in CF AECs, including GM-CSF, IL-6, IL-10, RANTES, and nitric oxide synthase-2 8, [12] [13] [14] . Altered regulation of multiple mediators suggests modulation of common signaling pathways by mutant CFTR. The NF-B pathway appears to be an important target, likely via effects on upstream regulatory proteins [15] [16] [17] . Other signaling pathways may also function improperly in CF cells, including the TGF-␤1/Smad and JAK-STAT cascades 14, 18 . Thus, altered regulation of diverse signaling pathways have been found in CF epithelial cell models by several groups, and these may provide unique therapeutic targets for controlling airway inflammation in CF.

Multiple studies do not clearly support the hypothesis that inflammation in intrinsically dysregulated in CF. For example, greater IL-8 and neutrophils found in CF BAL samples could easily be explained by defective mucus clearance rather than a hyper-inflammatory response. Furthermore, many CFTR-expressing tissues in patients with CF are not inflamed, suggesting that any hyper-inflammatory phenotype must be airway-specific. Although multiple in vitro studies suggest hyper-inflammation in CF AECs, others show no change or even lower secretion of pro-inflammatory mediators 13, 19 . Most studies were done with immortalized cell lines, which are inherently unstable and acquire differences other than CFTR status over time, making them difficult to control for experimentally. Complementation with wildtype CFTR inconsistently affects inflammatory responses 10, 20 . Primary AECs freshly removed from inflamed CF airways release more IL-8, but this phenotype is lost in culture suggesting an acquired, rather than intrinsic, response 10, 12 . Recent studies of well-differentiated primary AECs show significant person-to-person variation, but few differences in CF versus non-CF cells 10, 11, 20 . Taken together, it is clear that hyper-inflammation has stimulus-and condition-specificity that is highly variable between cell models 10, 11 . Another concern is that potential bias in the literature against negative studies may over-represent positive outcomes. Finally, despite numerous investigations, one can argue that a solid mechanistic link between mutant CFTR expression and altered regulation of inflammation has not yet been established.

While CFTR mutation links to disease pathogenesis and effects on the host response are hotly debated, there is essentially universal agreement that severe and sustained inflammation impairs pulmonary function and ultimately destroys the CF lung. Anti-inflammatory therapy slows the progression of CF lung disease, confirming the importance of airway inflammation, but current therapeutic approaches are nonspecific and entail risks 21, 22 . Research directed towards understanding the regulation of ongoing airway inflammation in CF will suggest more selective treatments to mitigate lung damage without adversely affecting host defense and pathogen clearance.

The authors acknowledge colleagues whose work could not be cited due to page limitations.

CF airway inflammation is typified by the presence of neutrophils and their protein products, including elastase, myeloperoxidase and matrix metalloproteinases, in the airways. Neutrophils are attracted to the airway by interleukin (IL)-8, a chemokine which is released from epithelial cells and macrophages as part of the innate immune response to infection. Besides IL-8, epithelial cells and macrophages are capable of producing other pro-inflammatory cytokines such as tumor necrosis factor (TNF)-␣, IL-1␤, and IL-6, each of which are increased in the airways of patients with CF.

The gene expression of IL-8 and other pro-inflammatory proteins is regulated largely by transcription factors of the nuclear factor (NF)-B and activator protein (AP-1) families. In the context of the IL-8 promoter, the NF-B site serves as a cis-acting response element to diverse stimuli including TNF␣, IL-1␤, PKC␦ and rhinovirus (unpublished data), whereas the AP-1 site serves as a basal level enhancer (1). The basic NF-B complex is a dimer of two Rel family members, p50 (NF-B1) and p65 (Rel A). The AP-1 complex is a dimer of Fos and Jun family transcription factors. In unstimulated cells, NF-B is sequestered in the cytoplasm by IB proteins. Phosphorylation and degradation of IB allows NF-B translocation to the nucleus, where it regulates gene transcription by binding to specific sequences of DNA. The "classical" signaling pathway to NF-B activation includes successive phosphorylation and activation of NF-B activating kinase (NIK) and IB kinase (IKK). IKK consists of two catalytic subunits (IKK-␣ and -␤) and a regulatory subunit (IKK␥). While IKK␣ and IKK␤ contain similar kinase domains with essentially identical activation loops, they are functionally distinct. Recent studies suggest that IKK␤ serves as the target for pro-inflammatory signals, whereas IKK␣ plays a critical role in development (2, 3) .

The most potent IKK activator is the serine-threonine kinase NIK. NIK is required for activation of NF-B by non-typeable H. influenzae (4) and micrococci (5) . We have shown that successive activation of NIK and IKK␤ is required for IL-8 transcription induced by ligation of the asialoGM1 (6), the putative glycolipid receptor for P. aeruginosa. IKK␤ and the mitogen-activated protein (MAP) kinase kinases share structural elements, including the position of two activation loop serine phosphoaccepting sites. Thus, it was found that MAP kinase/extracellular signal regulated kinase (ERK) kinase kinase (MEKK) phosphorylates and activates IKK (7) . MEKK1 is activated by TNF␣ and IL-1, and dominant negative MEKK1 inhibits IKK␤ and NF-B activation (7, 8) .

The NF-B signaling pathway in CF could be activated either by cytokines, infection, or endoplasmic reticulum (ER) stress. TNF␣ and IL-1␤ are potent activators of NF-B-dependent gene expression. NF-B activation is required for maximal IL-8 expression in response to P. aeruginosa (9) (10) (11) . These studies also noted a requirement for the MAP kinases, which in turn function as upstream activators of AP-1 (1, 6) . Recent attention has focused on the mechanisms by which P. aeruginosa engagement of asialoGM1, a glycolipid which lacks transmembrane and intracellular domains, elicits host cell responses. AsialoGM1 ligation promotes ATP release from the host cell, which is followed by successive activation of a nucleotide receptor, phospholipase C, Ca 2+ flux and ERK, a MAP kinase family member (12) . Also, it has recently been shown that asialoGM1 is associated with Toll-like receptor (TLR)-2, and that stimulation with P. aeruginosa flagella mobilizes TLR2 into a lipid raft receptor complex containing the TLR downstream effectors myeloid differentiation (MyD)-88, IL-1 receptor-associated protein kinase (IRAK)-1 and TNF receptor associated factor (TRAF)-6 (13). The pathway from TRAF-6 to NF-B activation has not been identified, but may involve transforming growth factor-␤-activated kinase (TAK)-1 and NIK (5, 14) .

Infants with CF show airway inflammation without apparent infection, suggesting that inflammatory signaling pathways may be primarily upregulated in CF. Consistent with this, tracheas from CF fetuses grafted into severe combined immunodeficient mice show increased intraluminal IL-8 and leukocytic infiltration compared to those from normal fetuses (15) . On the cellular level, airway epithelial cell lines derived from CF patients show increased basal and stimulated NF-B activation and IL-8 expression vs. corrected cells (7, 16) . Similar observations have been made in CF airway gland cells (17) . Expression of ⌬F508 CFTR, an incorrectly folded protein which partially accumulates in the ER, but not G551D CFTR, which is trafficked normally to the epithelial cell surface, increases NF-B transcriptional activity in Chinese hamster ovary cells (16) , suggesting that CFTR misfolding, with subsequent ER stress, is responsible for dysregulated inflammatory signaling in CF.

Overload of the ER with mis-or unfolded proteins leads to the unfolded protein response (UPR), an integrated pathway leading to the expression of molecular

Marc B. Hershenson, M.D.

chaperones, attenuation of protein synthesis, and ERassociated degradation (18) . If these adaptive responses do not correct the protein folding defect, cells undergo apoptosis. In addition, a distinct but poorly-defined pathway leads to NF-B activation, perhaps via calcium mobilization or the generation of reactive oxygen species. The UPR and ER stress in CF cells have not been carefully studied. Acute inflammation is a normal, protective response to tissue injury that, in the ideal circumstance, leads to removal of injurious stimuli and damaged cells or matrix and results in restoration of normal structure and function. In most forms of acute inflammation, neutrophil influx is an early and essential component of the process. After reaching a peak the influx ceases and the emigrated inflammatory cells are removed as a key component of the resolution process. Emigration of monocytes into the inflamed lung is slower and appears critical to the neutrophil removal and resolution process. Ultimately the excess macrophages are also removed.

This simple scheme raises a number of important questions, including: 1) What initiates the decline in neutrophil accumulation?. 2) What causes the later influx of monocytes (1)? 3) By what mechanisms are the neutrophils recognized and designated for removal? 4) How are they removed? 5) How are the excess mononuclear phagocytes cleared from the lung?

In chronic, persistent inflammatory responses this normal, self-limited sequence is disrupted, perhaps because of persistent stimuli, initiation of new forms of stimulation, or abnormalities in the resolution. Persistent and recurrent neutrophil influx in CF may encompass all three of these abnormal effects.

Neutrophils are short-lived cells and usually undergo spontaneous apoptosis within a few hours of release from the bone marrow. Emigration into the lung can alter this apoptosis in complicated ways that can both delay or accelerate the process, but only by a few hours (2) . The apoptotic neutrophils are recognized and cleared locally due to their ingestion by macrophages and tissue cells, including epithelial cells. This is normally a very efficient process; so efficient that detection of significant numbers of apoptotic cells within a tissue may be considered to suggest some defect in the clearance process. For example, only a few percent of the neutrophils being cleared during resolution of bacterial pneumonia or ARDS are seen to be apoptotic. By contrast, examination of CF patients revealed up to 40% apoptotic neutrophils (3).

Apoptotic cell recognition and removal in vivo occurs quietly and unlike uptake by other phagocytic processes, is both non-inflammatory and actively anti-inflammatory (4) . Instillation of such cells into an ongoing inflammatory reaction in the lung enhances resolution (5) . While a large number of molecules have been implicated in this recognition, a specific receptor for phosphatidylserine (the phosphatidylserine receptor, PSR) appears to be critical for these anti-inflammatory and anti-immunogenic effects. In its absence, alternative receptors may in fact induce pro-inflammatory mediators and/or the lack of apoptotic cell clearance results in necrotic death with pro-inflammatory consequences.

These observations led to the possibility that in CF, apoptotic cell clearance was ineffective, leading to persistence of the apoptotic neutrophils, increased necrosis and liberation of pro-inflammatory cell contents, as well as abrogation of the normal anti-inflammatory consequences of apoptotic cell recognition. All of these would lead to persistence of the neutrophilic phase of inflammation. In seeking the mechanism for defective apoptotic neutrophil clearance, the known presence of high levels of active elastase in CF airways suggested a likely possibility. Apoptotic cell recognition receptors, particularly the PSR, are highly susceptible to proteolysis. Elastolytic cleavage of these in the airways would result in blockade of both apoptotic cell clearance as well as shutdown of the pro-inflammatory mediator generation (3) .

Extending this concept more generally, the possibility arises that during the induction of acute inflammation, active neutrophil elastase liberated during the early phase of neutrophil influx, can cleave and inactivate the PSR, thereby providing a window of time during which the acute inflammatory reaction can proceed. As vascular permeability increases, influx of protease inhibitors would inactive the elastase, now allowing incoming monocytes to mature into macrophages that are free to express their PSR and initiate the anti-inflammatory resolution phase.

In CF this puts a great deal of weight on the persistent presence of active elastase. To some extent, a positive feed-back loop from non-clearance of the apoptotic neutrophils, cytolytic release of their elastase content, with more suppression of apoptotic cell clearance, all in the presence of a high bacterial load, could drive a self-sustaining neutrophilic inflammation. This does not, how-ever, explain how it all got going in the first place, leading to questions about genetically driven alterations in apoptotic cell recognition and response as potential explanations for the very early post-natal infiltration of CF airways with neutrophils. 

Although there may still be uncertainty about whether misfolded protein or other intracellular consequences of CFTR defects initiate some inflammation in the lungs of CF patients, it seems clear that infection plays a major role in stimulating inflammation in the airways. There also seems to be a consensus that the response to infection in the CF airway is dysregulated and excessive (1) . Several studies in mice with CFTR mutations have shown excessive responses to the prototypic inducer of inflammation, lipopolysaccharide (LPS). Most investigators have reported that huge quantities of pro-inflammatory cytokines, chemokines, and low molecular chemoattractants are found in lung secretions of CF patients, mice with mutations in CFTR that have been challenged with LPS or bacteria, and the supernatants of cultured cells with defective or blocked CFTR. Decreased production of anti-inflammatory mediators such as lipoxins (2) and IL-10 (3) is also likely to be important in the dysregulation of the inflammatory response in the CF lung. There is considerable evidence that the NF-B dependent proteins, TNF, IL-1, IL-6 and IL-8 are over-produced in CF and contribute to both intrapulmonary and systemic morbidity (reviewed in 3). The major inflammatory effector cell in the CF airway is the neutrophil (PMN). Besides a role for IL-1 and TNF in priming PMN, they induce expression of adherence molecules which facilitate migration of PMN into the airway; and IL-8 is itself an important chemoattractant. Additional chemoattractants important in CF are the lipoxygenase product, LTB 4 , and the complement fragments C5a and C5a desarg (3, 4) . The PMN themselves contribute additional LTB 4 . More importantly, PMN products including active proteases, long stranded DNA, and oxidants all have deleterious effects on airway function and structure which are out of proportion to their necessary role in controlling infection. These PMN products contribute to airway obstruction, induce NF-B activation and additional pro-inflammatory cytokine production, and impair phagocytic defense mechanisms. Thus, the PMNs initiate and perpetuate a vicious cycle of infection, inflammation and airway destruction which eventually claims the life of the patient (3,4).

The paradigm above illustrates an array of targets for anti-inflammatory therapy. The role of infection as the major stimulus for the inflammatory process emphasizes the importance of controlling infection and reducing the burden of bacteria with antibiotics and airway clearance techniques/therapies. Numerous studies have documented decreases in inflammatory mediators after "cleanouts", confirming the importance of limiting this stimulus. While we all hope that correction of the basic defect will terminate the cycle of infection and inflammation, if chronic infection and airway damage are already established, this may not be possible. Inhibition of intracellular signaling pathways such as the I-B/NF-B pathway or correction of putative CF-related dysregulation of this pathway would reduce many of the proinflammatory cytokines and chemokines, and would thus hit many lung and systemic targets by reducing TNF and IL-8 (3). Aspirin, NSAIDS, and corticosteroids all inhibit NF-B activation (5, 6) , and ibuprofen and prednisone have shown beneficial effects in CF (7, 8) . The pleotropic effects of cytokines might suggest that inhibiting any individual pro-inflammatory cytokine may not be very effective. However, inhibition of TNF alone has been very effective in rheumatoid arthritis (RA), which has several features in common with CF, especially the role of PMN in destruction of the end organ (9) . Interestingly, methotrexate, the mainstay of RA treatment, has shown promising effects in small studies in CF (10) . Inhibiting production of individual pathways like 5-lipoxygenase is also possible, and inhibitors of individual chemoattractant receptors are being studied. Although preventing the PMN influx might seem to be a better strategy than attempting to neutralize their products one-by-one, the success of DNAase illustrates the efficacy of the latter approach. Similar efforts could be directed at elastase and PMNderived oxidants. In evaluating anti-inflammatory therapy in CF, it is important to differentiate short-term improvement in symptoms from long-term preservation of lung function (11) . A two-tiered approach is envisioned by the CFF TDN AIM program, in which preliminary 30 day studies looking for efficacy in reducing inflammatory mediators in induced sputum will be used as a screening tool to select the best agents for larger, long-term studies of efficacy in preventing loss of lung function. This strategy offers our best chance of fairly evaluating new agents. Any time one contemplates interfering with inflammatory mechanisms that also play important roles in host defenses, there is fear of allowing infection to get out of control. This has not occurred in trials of ibuprofen or corticosteroids in CF, but problems with infection, particularly activation of old TB in patients over 65, have been seen with the use of TNF inhibitors in RA (9) . Some of the concerns with these agents and other anti-inflammatories may be relevant to CF, but some may not. Adverse effects frequently occur in organs other than the lung and may or may not be related to that mechanism of action of the drug which is believed to be important in CF. Observation of the latter should lead to development of more specific agents in which the therapeutic effect is maintained while the action(s) causing the adverse effects is eliminated. Specifically modifying the drug and/or route of delivery to decrease access to sites of toxicity while preserving effects on the therapeutic targets is the principle behind the use of inhaled corticosteroids for asthma, but these drugs seem to have only limited efficacy in CF. Just as the long-term therapeutic effects must be considered, long-term adverse effects must also be weighed. It is particularly worrisome that "catch-up" growth has not occurred in CF patients who were given prednisone in earlier trials (12) . Although there are many reasons to believe that anti-inflammatory therapy should a part of every CF patient's regimen, recent CFF surveys suggest that use of these agents, particularly ibuprofen, is actually decreasing. This seems ironic in the face of the confirmation of the efficacy of ibuprofen presented by Schlucter et al at this meeting. Development of newer strategies, realistic assessment of the risk of adverse effects, development of better targeted agents, and better understanding of the role of inflammation in lung damage are all necessary for determining the appropriate place for anti-inflammatory therapy in CF. A small-molecule discovery program was established to identify inhibitors of wildtype CFTR and activators of ∆F508-CFTR (1). Primary high-throughput screening is done using automated instrumentation, a collection of >200,000 diverse drug-like small molecules, and a cellbased assay utilizing a halide-sensitive green fluorescent protein (2) .

Two interesting classes of small-molecule CFTR inhibitors were identified. The compound CFTR inh -172 contains a 2-thioxo-4-thiazolidinone core (3) . CFTR inh -172 reversibly inhibits CFTR Clchannel function with K I in the range 0.3-5 µM, depending on cell type/membrane potential. The mechanism of CFTR inhibition by CFTR inh -172 involves stabilization of the closed channel state with prolonged mean channel closed times by patch-clamp analysis (4), probably as a consequence of CFTR inh -172 binding to the first nucleotide binding domain of CFTR. CFTR inh -172 did not inhibit calciumor volume-activated chloride channels or the ATP-binding cassette protein MDR-1, and produced little toxicity in cell culture and mouse models. A single intraperitoneal dose of 0.25-1 mg/kg CFTR inh -172 inhibited cholera toxin-induced intestinal fluid secretion in closed-ileal loop models in mice and rats (3, 5) . CFTR inh -172 was also found in mice to reproduce the nasal epithelial ion transport defect in cystic fibrosis (6) , and in pig and human airways to reproduce cystic fibrosis defects in submucosal gland fluid secretion (7) . Analysis of rodent pharmacology indicated slow renal elimination without metabolism, and efficient liver uptake with enterohepatic recirculation (8) . The in vitro and in vivo results provide the rationale for further evaluation of thiazolidinone-type CFTR inhibitors as antidiarrheals and agents for pharmacological creation of cystic fibrosis animal models.

A second class of small-molecule CFTR inhibitors, glycine hydrazides, was identified using a screen designed to identify compound that might act at the external CFTR surface (9) . The compound N-(2-naphthalenyl)-[(3,5-dibromo-2,4-dihydroxyphenyl)methylene]glycine hydrazide (GlyH-101) reversibly inhibited CFTR Clconductance in <1 min. Whole-cell current measurements revealed voltage-dependent CFTR block by GlyH-101 with strong inward rectification, producing an increase in apparent inhibitory constant K i from 1.4 µM at +60 mV to 5.6 µM at -60 mV. Apparent potency was reduced by lowering extracellular Clconcentration. Patch-clamp experiments indicated fast channel closures within bursts of channel openings, reducing mean channel open time from 264 to 13 ms. GlyH-101 inhibitory potency was independent of pH from 6.5-8.0, where it exists predominantly as a monovalent anion with solubility ~1 mM in water. Topical GlyH-101 in mice rapidly and reversibly inhibited forskolin-induced hyperpolarization in nasal potential differences. In a closed-loop model of cholera, intralu-minal 2.5 µg GlyH-101 reduced by ~80% cholera toxin-induced intestinal fluid secretion. Compared to CFTR inh -172, GlyH-101 has much greater water solubility and rapidity of action, and a novel inhibition mechanism involving occlusion near the external pore entrance. Glycine hydrazides may be useful as nonabsorbable antidiarrheals in enterotoxic-mediated secretory diarrheas.

∆F508-CFTR 'potentiators' (correctors of defective gating) and 'correctors' (correctors of cellular misprocessing) were identified using FRT epithelial cells expressing human ∆F508-CFTR and an ultra-high halide sensitive green fluorescent protein. Initial screening revealed a tetrahydrobenzothiophene class of ∆F508-CFTR potentiators that corrected defective ∆F508-CFTR chloride channel gating at concentrations down to 100 nM (10) . Recent additional screening hasidentified 2 novel classes of potentiators with good efficacy and medicinal profiles. After structure-activity analysis and optimization, one class of compounds corrected defective ∆F508-CFTR gating at concentrations down to 10 nM. In cell attached patch-clamp experiments, the potentiators evoked increased channel open probability >5-fold by reduction of interburst closed time. Stimulation of chloride secretion was confirmed in low temperature-rescued human bronchial epithelial cells from ∆F508 CF subjects. Two novel features of the new class of ∆F508-CFTR potentiators include amplifed response (synergy) with cAMP agonists, and correction of other CFTR gating mutants including G551D-CFTR.

Screening of ∆F508-CFTR expressing cells cultured at 37 o C for correctors of defective cellular processing produced 3 classes of small-molecules with favorable medicinal properties that conferred greater chloride conductance to ∆F508-CFTR expressing cells than low temperature rescue. Complex glycosylation was confirmed as measured by immunoblot analysis, as was plasma membrane expression as measured using an external epitope-tagged CFTR. Correction was seen within 3-6 hours after compound addition and persisted for up to 18 hours after washout. A unique feature of some correctors was their ability to confer partial correction of defective ∆F508-CFTR gating, possibly by improved ∆F508-CFTR folding. The small-molecule potentiators and correctors maybe useful in therapy of CF caused by ∆F508 and possibly other mutations.

Cystic fibrosis (CF) is a fatal genetic disease caused by mutations in the CF transmembrane conductance regulator (CFTR) protein, a protein kinase A-regulated anion channel in the apical membrane of many epithelial tissues, including the airway. Approximately 70% of CF patients are homozygous for a single mutation that causes a deletion of phenylalanine at position 508 (∆F508-CFTR). This mutation causes improper folding of the protein, deceasing both channel cell surface density and gating. Either directly or indirectly, ∆F508-CFTR impairs ion transport, fluid secretion and mucociliary clearance, resulting in chronic infection, inflammation and eventual loss of lung function.

Our strategy for clinical treatment of CF is to discover drugs to increase cell surface expression and/or increase anion transport for mutant CFTR, including ∆F508-CFTR. The hypothesis is that such pharmacological "rescue" of CFTR would restore airway ion transport and ameliorate or slow the deterioration of lung function. Agents such as 4-phenylbutyrate and genistein have been demonstrated to increase the density and gating of ∆F508-CFTR in recombinant cells, respectively. However, these agents have other biological effects and have not been chemically optimized for activity on CFTR or for human pharmaco-kinetics. Furthermore, these agents show little effect on chloride transport in primary airway cultures from CF lungs, and therefore likely lack either the potency or the efficacy to be effective drugs in humans.

We have pursued a targeted approach to discover drugs that improve ∆F508-CFTR function. To identify starting points for chemistry optimization, we employed fluorescence-based assays of membrane potential in 3T3 cells expressing recombinant ∆F508-CFTR. Approximately 200,000 compounds were screened to identify small molecules that either increased ∆F508-CFTR gating in the presence of PKA stimulation (so-called "potentiators") or increased ∆F508-CFTR trafficking to the apical membrane (so-called "correctors"), or both. In addition to activity in the fluorescence assays, both potentiators and correctors were assayed for the ability to increase chloride transport in Ussing chamber measurements using monolayers of primary human bronchiolar epithelial cell from the airway of ∆F508-CFTR patients (∆F508-HBE). Potentiators were prioritized if they exhibited drug-like chemical features and efficacy and potency greater than or equal to genistein. Correctors were prioritized based on drug-like features and efficacy greater than or equal to temperature correction, which has been previously shown to partially reverse the defect in vitro. Based on these criteria, distinct chemical series have been identified for correctors and potentiators.

Chemical optimization is aimed at improving both in potency and efficacy measured in both the fluorescence assay and (F508-HBE chloride transport with the goal of achieving greater than 10% of ion transport observed in normal airway. Several thousand molecules have been synthesized in the program to date. The presentation will provide an update on recent progress in lead optimization and describe some of the key challenges moving toward compounds suitable for clinical testing. The most common mutation associated with CF is ⌬F508, which results in the in-frame deletion of the single codon encoding a phenylalanine residue at position 508. This mutation accounts for ~69% of all CF alleles, and ~90% of CF patients carry at least one copy of ∆F508 CFTR. The protein encoded by the (F508 CFTR allele has been classified as a processing mutant and as such, it is retained in the endoplasmic reticulum through interactions with elements of the ER's quality control chaperone machinery and tar-geted for subsequent degradation in the proteasome. However, investigators have identified a number of in vitro manipulations that allow cohorts of newly synthesized ∆F508 CFTR protein to depart the ER, undergo post-synthetic processing in the Golgi complex and be delivered to the cell surface. The resultant "rescued" surface population of ∆F508 CFTR proteins are able to function as chloride channels with properties that are similar, but not identical, to those of the wild type polypeptides.

In light of this behavior, numerous efforts aimed at developing small molecule pharmacotherapies for CF are focused on identifying compounds that can either stabilize the tertiary structure of the ∆F508 protein, or that can interfere with the interactions between the mutant protein and ER chaperones. By preventing these interactions, the newly synthesized mis-folded but functional ∆F508 CFTR protein might be allowed to escape recognition by the mechanisms that are responsible for its retention and ultimate degradation.

We have previously shown that compounds that inhibit the endoplasmic reticulum Ca-ATPase (SERCA) can induce the release of ER-retained _F508 CFTR, which is delivered to the cell surface and able to function. Many compounds that inhibit the Ca-ATPase are potentially toxic, and are thus not suitable as therapeutic candidates.

Curcumin is a non-toxic Ca-ATPase inhibitor that can be administered to humans safely in very large quantities. In preliminary studies the oral administration of curcumin to homozygous ∆F508 CFTR mice in mg/kg doses com-parable to those well tolerated by humans resulted in alterations of these animals' characteristic nasal potential difference (NPD) defect, such that there is a decrease in the baseline NPD and the appearance of an isoproterenol response. Furthermore, this treatment altered the electrical properties of gastrointestinal epithelia in these animals as demonstrated by the rectal potential difference assay. These effects are not observed in mice homozygous for a complete knockout of the CFTR gene. Curcumin also induces the functional appearance of _F508 CFTR protein in the plasma membranes of transfected BHK cells. Taken together, our initial studies suggest that curcumin is a nontoxic compound that can partially correct defects associated with the homozygous expression of ∆F508 CFTR. However, in a number of follow up studies using different in vitro cell models and in vivo mouse models have not resulted in similar changes in ion transport properties. The difference in these in vivo and in vitro studies will be discussed. Lastly, the findings from the first clinical study will be presented. An estimated one-third of mutations underlying human disorders result in premature termination of translation (nonsense or stop mutations) 1 . Nonsense mutations account for approximately 10% of the total mutant alleles in cystic fibrosis (CF) patients (Cystic Fibrosis Mutation database), however, in certain populations the incidence is relatively high. Among Ashkenazi Jews, nonsense mutations account for 64% of all CFTR alleles 2,3 . These mutations are associated with a severe form of the disease due to no CFTR chloride channel function. Aminoglycosides, in addition to their antimicrobial activity, can promote read-through of nonsense mutations in eukaryotic cells by allowing synthesis of full-length proteins 4, 5 . Aminoglycosides were shown to restore the CFTR function in cell lines and patients carrying the nonsense mutation W1282X 6-8 . In a recent randomized, double blind, placebo-controlled, crossover clinical trial 9 we showed that the aminoglycoside gentamicin can increase the expression of fulllength CFTR and correct the electrophysiological abnormalities, in CF patients carrying the W1282X nonsense mutation. However, in these studies several patients did not respond.

Here we studied the molecular basis for the variable response in CF patients carrying the W1282X mutation. A wide variability in the level of CFTR nonsense transcripts was found among these patients (n = 10). This variability correlated with the response of the patients to gentamicin treatment as demonstrated by nasal potential difference (NPD) measurements. All the patients with relatively high transcript levels showed a significant reduction in basal potential difference and/or a significant response to chloride-free isoproterenol solution, indicating functional restoration of CFTR following gentamicin treatment. In contrast, the patients with markedly reduced levels (n = 2, <26%) of CFTR transcripts did not correct the abnormal basal potential difference nor the chloride transport, indicating no response to the gentamicin treatment.

We then analyzed the effect of the CFTR transcript levels on the response to gentamicin treatment in two cell lines (CFP15a and CFP15b) from unrelated CF patients carrying the W1282X mutation. Analysis of the CFTR transcript levels revealed a significant difference between the cells, such that in CFP15a the level was 4.5 fold higher than in CFP15b. Chloride efflux measurements in CFP15a and CFP15b cells revealed no forskolinstimulated chloride efflux, indicating that the CFTR channels in both cell lines are inactive. Following treatment with 50 _g/ml gentamicin, restoration of the CFTR function was demonstrated only in CFP15a cells. Following treatment with 200 _g/ml gentamicin, chloride efflux was detected in both cell lines. These results indicate that the level of CFTR nonsense transcripts correlates with the response to gentamicin, similarly to the findings in the patients.

The level of nonsense transcripts is regulated by several factors, among which is the nonsense mediated decay (NMD) pathway. We hypothesized that inhibition of NMD would increase the level of nonsense transcripts available for read-through, and thus would improve the response to gentamicin. Hence we treated the CFP15a and CFP15b cells with cycloheximide (CHX), a potent inhibitor of NMD. This treatment significantly increased the level of nonsense transcripts. A direct inhibition of NMD by siRNA directed against the NMD factors hUpf1 and hUpf2, resulted in a similar increase in the level of CFTR transcripts. We further analyzed the effect of NMD inhibition on the CFTR chloride efflux. Fro this we studied the effect of siRNA against hUPf1 and hUpf2 on the CFTR function. The results showed a significant higher chloride efflux in response to gentamicin, indicating that inhibition of NMD can improve the response to gentamicin. Our results suggest that NMD plays a role as a genetic modifier of the response to aminoglycoside treatment in CF and many other inherited diseases. Lung transplant needs to be considered for all CF patients once they develop end-stage lung disease. Our goal should be that no CF patients should die without having had the option of considering lung transplantation. With still limited availability of lung grafts, up to 15 -40% of CF patients will die while awaiting transplant. Thus timing of referral for transplant is the issue -bal-ancing risks of death of CF lung disease with risks of transplantation; identifying those ill enough to benefit from transplant but well enough to survive the procedure.

Selection of candidates at the appropriate level of illness is therefore the challenge for the transplant team. For the CF team, it is obviously better to err on the side of a premature referral. As it may take time for the patient to prepare for the idea of transplant, this time has to be factored into the equation as well. Early referral to provide information for patient and family and to assess for transplant (but not necessarily listing) is likely the best option for patients. Early referral should not mean premature transplantation and the system should have the flexibility to move patients on and off an active waiting list to "fine tune" or optimize the timing of transplant.

The major difficulty is that CF is an unpredictable disease; the patients are young and can withstand severe dysfunction and treatment has a huge effect on disease stabilization. As patients become more ill, adherence to therapy can improve and intensification of treatment often occur.

Prediction of patients at high risk of dying has proven to be elusive -several models have shown we are better at predicting who will survive over who will die. Mayer-Hamblett et al used the CF Foundation National Patient Registry to find predictors of mortality (FEV 1 , Burkholderia cepacia, Pseudomonas aeruginosa, hospitalizations, use of IV antibiotics) which were incorporated into a multivariate logistic regression model. 2 This model did not significantly improve on the previous model developed in the Toronto CF population where FEV 1 < 30 % predicted alone was the most significant predictor of mortality (Kerem et al 3 ). Liou et al. also used data from the CF Foundation National Patient Registry and from UNOS (United Network for Organ Sharing) to develop a multivariate logistic regression model to predict 5-year survival and they showed that patients with a 5-year survival of < 30 % showed maximal survival benefit from transplantation. 4 However, this still does not help predict the timing of referral to transplant.

In addition to our inability to accurately predict at what stage a person should be referred and listed for transplant, we have to contend with other variables. Predicted long term survival depends on survival post transplantation plus the waiting time to transplant. Survival post transplant depends mainly on B. cepacia status (the species cenocepacia or genomar III appears to result in higher immediate post operative mortality)positive or negative). The waiting time to transplant depends on country and centre. In the USA, the previous organ allocation policy depended on body size and blood type and accrued active waiting time on the national waiting list. In Canada, waiting list time is affected by medical necessity giving much more flexibility for patients who suddenly deteriorate. A similar style of practice is present in many European countries. Waiting list of less than 1 year in Toronto allows for later referral when patients are more ill (which can be easier to predict) and patients feel more "ready" at time of referral as they have a lower functional status.

From a practical point of view, we must use the information and models that we have available, recognizing differences depending on our patient population, using predictors (FEV 1 , rate of decline in FEV 1 , bacteriology, CO 2 retention, hypoxia, clinical stability (hospitalizations, use of IV antibiotics), and functional status (6minute walk test)). All CF physicians will recognize older, clinically stable patients with slow decline in FEV 1 who may remain with FEV 1 below 30 % predicted for decades. Thus, it is imperiative for the CF clinician to communicate this information to the transplant center, because they have their finger on the pulse of each individual patient. The reality is that Afterall, the transplant pulmonologists focus more on post transplant after care than on pre transplant-care. Thenus, the goal of the CF centre is to maximize medical therapy and follow patient frequently to fine tune. Rehabilitation programs improve functional condition and provide an early warning system for clinical deterioration. Ongoing screening for CF complications which may impact the post transplant course (eg. CF related diabetes, osteoporosis) should not be overlooked. The concept of transplant is frightening and at the time of referral, patient may not feel that this is necessary for them. Emotional and psychological preparation of the patient and family is required. Linking the patient and family with another CF patient of similar age and life experience who has had a transplant is a helpful strategy.

Close links between CF centre and transplant centre lead to optimal care for patients with hopefully an enhanced ability to minimize deaths on the waiting list.

We performed 21 transplants in ventilator dependent patients from July 1988 to January 2000. Selected pretransplant characteristics of these patients are described in Table 1 . The pretransplant characteristics of the nonventilated patients and the stable ventilated patients were compared and no significant differences were found. The 5 unstable ventilated patients were all acute retransplantations for early graft failure and represented such unique and unpredictable situations that statistical comparisons with stable patients did not make sense. The single re-transplant in the stable group was a woman who developed bronchiolitis obliterans syndrome after an en-bloc double lung transplant for COPD and was relisted 17 months after her first transplant.

Living lobar lung transplantation was developed as a procedure for patients considered too ill to await cadaveric transplantation. The procedure involves using a lower lobe from each of two donors for each recipient.

138 living lobar lung transplants have been performed in 133 patients between 1993 through June 2004. 90 patients were adults (mean age 27) and 43 were pediatric (mean age 13.9). The primary indication for transplantation was cystic fibrosis (85%). At the time of transplantation, 72.4% of patients were hospitalized and 21.1% were intubated. 1, 3, and 5 year actuarial survival among living lobar recipients was 70, 54, and 45% respectively. There was no difference in actuarial survival between adult and pediatric living lobar recipients (p = 0.65). There were 66 deaths among living lobar recipients, with infection being the predominant cause (53.4%), followed by obliterative bronchiolitis (12.7%), and primary graft dysfunction (7.9%). The overall incidence of acute rejection was 0.8 episodes per patient. 78% of rejection episodes were unilateral. Age, gender, indication, donor relationship, preoperative hospitalization status, use of preoperative steroids, and HLA A, B, and DR typing did not influence survival or rejection. However, patients on ventilators preoperatively had significantly worse outcomes (odds ratio 3.06, p = 0.03; Kaplan-Meier, p = 0.002), while those undergoing retransplants had an elevated risk of death (odds ratio 2.50).

We examined the perioperative outcomes associated with the first 253 donor lobectomies performed.

There have been no perioperative or long-term deaths. 80.2% of donors (n = 203) had no perioperative complications, while fifty (19.8%) had one or more complication. The incidence of intraoperative complications was 3.6%. Complications requiring reoperation occurred in 3.2% of donors. 15.0% of donors had other perioperative complications; the most serious were two donors who developed pulmonary artery thrombosis, while the most common was the need for an additional thoracostomy tube or a thoracostomy tube for ≥ 14 days for persistent airleaks and/or drainage. Right sided donors were more likely to have a perioperative complication than left sided donors (odd ratio 2.02, p = 0.04), likely secondary to right lower and middle lobe anatomy.

These results support the continued use of living lobar lung transplantation in patients deemed unable to await a cadaveric transplant. We consider retransplants and intubated patients to be at significantly high risk due to the poor outcomes in these populations. Our experience has shown donor lobectomy to be associated with a relatively low morbidity and no mortality. This is important if this procedure is to be considered an option at more pulmonary transplant centers, given continued organ shortages and differences in philosophical and ethical acceptance of live organ donors.

The postoperative length of intubation, intensive care unit stay, and overall hospital stay were significantly longer for the stable ventilated patients as compared to non-ventilated patients.

When the 16 stable ventilated transplant recipients were compared with the 479 non-ventilated recipients, the difference between the survival curves failed to reach statistical significance with p = 0.15. Finally, the survival difference between stable and unstable ventilator dependent recipients was significant with p = .013.

Acute lung retransplantation for severe graft dysfunction is a high risk strategy that has led to no long term survivors in our limited experience in five patients. In each case the recipient fulfilled two criteria: acute lung dysfunction of such severity that the surgeon and pulmonologist agreed that survival was unlikely, and generally intact physiologic systems without systemic infection or multisystems organ dysfunction. Other strategies considered in this dire situation have been ECMO and profound hypothermia but the experience in these modalities is equally anecdotal and the results equally poor.

Lung transplantation for patients in stable respiratory failure is more promising. In our sixteen stable ventilated patients there were no postoperative hospital deaths and a long-term survival that is not statistically different from the general population of our program's transplant recipients. These patients do seem to be more prone to prolonged intubation and hospitalization after the transplant, but the results indicate that these problems are surmountable and that long term results are acceptable. Our practice continues to be the same as described in this paper: we will consider transplantation in stable patients who have declined on the waitlist but we are skeptical of and generally discourage transplantation for acute lung dysfuction, regardless of the etiology. The medical consequences of delayed diagnosis in CF can be thought of as short term or long term. In the short term, infants with cystic fibrosis (CF) are subject to a wide range of complications if diagnosis is delayed. The most striking complications are related to severe nutritional deficiencies and to hypoelectolytemia. Infants with CF can present with severe protein calorie malnutrition closely resembling Kwashiorkor. These infants have low albumin levels, edema, anemia and hepatomegaly from steatosis. They are also developmentally delayed at diagnosis. Trace metal deficiencies have also been described including severe zinc and selenium deficiencies. Zinc deficiency is associated with severe, weeping rash and immune dysfunction. Selenium deficiency can lead to congestive heart failure. Deficiency of alpha tocopherol, fat soluble vitamin E, is associated with severe anemia. Vitamin K deficiency can lead to intracerebral hemorrhage. (1) Linoleic acid deficiency likely coexists with other severe nutritional abnormalities and can cause rash. Hypoelectrolytemia can lead to lethargy, poor intake and seizures. A death related to hypoelectolytemia and seizures in an infant with unsuspected cystic fibrosis has been reported. (2) The causes of nutritional deficiency and hypoelectolytemia in infants with CF are well understood and help explain the short term consequences of delayed diagnosis. Because of exocrine pancreatic insufficiency, fat and protein malabsorption occur in most infants by two months of age. (3, 4) Trace element metabolism is also abnormal by two months of age. Albumin and fat soluble vitamin levels in young infants inversely correlate with the degree of nutrient malab-

Philip M. Farrell, M.D., Ph.D.

During the past 25 years, numerous investigations have identified medical and psychosocial benefits associated with early diagnosis of CF through neonatal screening. Research has included both observational studies and a randomized controlled trial. The following list of potential benefits provides a compelling rationale for universal newborn screening using the trypsinogen/DNA(CFTR) method. 1. Prevent deaths of undiagnosed patients-save lives! 2. Improve access-avoid geographic and fiscal barriers 3. Avoid disparities related to gender, race and ethnicity 4. Prevent protein-energy malnutrition and stunted growth 5. Prevent prolonged micronutrient deficiencies such as E and K 6. Reduce risk for cognitive dysfunction due to malnutrition 7. Preempt bronchopulmonary disease and Pa infection* 8. Provide genetic risk information ("counseling") to parents 9. Reduce costs for diagnosis and possibly treatment 10. Enhance clinical research opportunities while improving quality * Create the opportunity to initiate respiratory therapy before irreversibility develops i.e., the "point of no return," which will eventually facilitate prevention of lung disease sorption. We also know that impaired weight gain after birth is correlated with malabsorption. Hypoelectolytemia arises from sweat electrolyte loss. Sweat electrolytes are abnormally high in CF within a week after birth. Early diagnosis rather than delayed diagnosis along with treatment of pancreatic insufficiency and supplementation of salt intake might therefore avoid short term complications of CF in infancy. Two lines of evidence support this notion. The most important evidence that early diagnosis and treatment can avoid complications in infants with CF comes from the Wisconsin trial of newborn screening. This trial has clearly demonstrated that growth is improved in the short term (as well as the long term) in infants identified through newborn screening compared to those identified after symptoms appear. (5) The second line of evidence stems from observational studies of infants identified through newborn screening showing that nutrient levels and growth improve with treatment. Impaired growth is a long term consequence of that CF that is clearly related to delayed diagnosis. (5) Since improved pulmonary outcome in CF has been related to improved growth, (6, 7) it might be expected that delayed diagnosis could lead to worse pulmonary outcome. The one study that has looked at this carefully however showed worse pulmonary outcome in patients identified through newborn screening compared to those with delayed diagnosis. (8) There were special circumstances dealing with environmental Pseudomonas exposure that made this study difficult to interpret, however. Thus further studies of short term and long term medical consequences of delayed diagnosis are needed.

To further explore the short and long term medical consequences of delayed diagnosis of CF we examined data from the CF Foundation registry from 2000, 2001, and 2002. Short term consequences examined included presentation with hypoproteinemia or hypoelectrolytemia, as well as growth status Pseudomonas culture positivity, and hospitalization in the year of diagnosis. Long term medical consequences included growth status, Pseudomonas culture positivity, pulmonary function and hospitalization. We specifically examined four modes of diagnosis (delayed diagnosis, newborn screening, meconium ileus and prenatal diagnosis). More than 2500 new diagnoses were studied. We found that stunting, wasting and hospitalizations in the year of diagnosis were three times as common in infants with delayed diagnosis compared to infants identified through newborn screening. Occurrence of hypoproteinemia was significantly more common in infants with delayed diagnosis as was Pseudomonas culture positivity. We examined long term consequences by age group (2-5 years, 6-10 years, and 11-20 years). Delayed diagnosis was associated with stunting and wasting at essentially each age, higher percentage of mucoid Pseduomonas at each age, and decreased FEV1 (absolute and percent predicted) at ten years of age. Hospitalization rate was higher in the delayed diagnosis group than the newborn screened group at two age ranges.

Examination of the CF registry therefore supports the notion that delayed diagnosis of CF is associated with significant short term and long term medical consequences. These data strongly support the need for early diagnosis and treatment to improve outcome in CF.

Cystic fibrosis (CF) is not readily diagnosed without newborn screening. It has few unique features, is very rare, and varies in its presentation.(1) In the United States, half of all persons with CF were diagnosed after six months of age. (2) The median delay in diagnosis is well over one year in parts of the United States where newborn screening is not universal. The impact of delayed diagnosis of CF on families is poorly described. In the United States, most studies have been concerned with the clinical features of those with delayed diagnoses and have rarely explored the impact on the mother, father and siblings.(e.g., 3-6) A few bibliographic accounts of families with a child with CF have been written.(e.g., 7, 8) Outside the U.S., however, studies have found that misdiagnosis leads to increased anxiety, guilt and anger, and mistrust of the medical profession, and late diagnosis leads to more negative feelings about the pre-diagnostic period and less confidence in the medical profession.(e.g., 9, 10) We aimed to describe the impact of delayed diagnosis on U.S. families with CF at different life stages. We requested diagnostic stories from individuals on the CF Research, Inc. e-mail list in September 2003. We obtained 20 diagnostic stories and used an additional 20 stories found on the Cystic-L listserv dating back to 1997. We condensed the stories into that of a single family and qualitatively described the affected child's symptoms at different life stages and the medical and family responses to these symptoms. We provide likely impacts on health and family resulting from a diagnosis had it occurred at different ages.

We found that the age at which a person with CF was diagnosed can have large, irreversible negative impacts on that person's health status, quality of life, longevity, compliance with medical regimens, self-image, family structure, and major life decisions. Familial relationships were seriously impacted. The stress around not knowing what is wrong was damaging to the family. Economic losses can be expected. Strong views about incompetence in the medical care profession abound. These negative effects on families were likely to become compounded the later the diagnosis of CF was made.

We found that long-term parental guilt around not finding out sooner was difficult to avert altogether or remove later. Delayed diagnoses means that opportunities were lost to make informed decisions about health care, employment, housing, insurance, reproduction, and other matters. Even if untrue, it was common for parents to feel like appropriate steps could have been taken to prevent their child's lung disease, for example, if only they had known sooner about the diagnosis. A few parents who had a first child diagnosed on account of a second child being diagnosed early in life via newborn screening indicated that it was far better knowing the diagnosis than being tormented by not knowing it. It was extremely hard to plan for the future when it was unclear what was wrong with the child. There was parental regret, anger and pity for their older child who was not able to benefit in the way the younger child can through prevention and focused CF medical care, which works best for persons who present early without damaged lungs or nutritional deprivation.

In nearly all of the delayed diagnosis stories we reviewed, parents and persons with CF would end with the hope that in the future others could avoid the pain they suffered of not having known sooner.

1. How should we understand our collective use of "early" and "late" diagnosis of CF as we consider the ethics of newborn screening?

In most medical communities, often repeated case reports become part of our explanatory narrative, such that they express the shared understanding of our work. These explanatory narratives form a common language by which we express normative assumptions about the appropriate roles of patients and doctors, and the proper approach to the treatment and diagnosis of CF.

Case reports about "late" diagnosis of CF are one such common explanatory narrative. Among ourselves, we often tell the story of a child or young adult diagnosed with CF only after being seen in a specialist clinic. We all recognize this story immediately, and it has a kind of defining quality for those of us in the CF community: I doubt anyone who has come to a NACF conference could make it through the conference without either hearing or telling one version of this story. We tell it to each other even if we have just met, when the talk turns to our patients, and we commonly use this story, in one form or another, as a cautionary tale to our trainees.

Why do we tell each other this story so often? From a simple point of view, it's not so surprising, because the story makes us look smart-in particular, it confirms us as better clinicians than whoever saw the patient before us, and lets us comfortably and collectively shake our heads in acknowledging that another doctor could have missed the obvious diagnosis. Perhaps the frequent telling of the late diagnosis story is a just a harmless form of self-congratulation among colleagues, much like a group of carpenters telling stories about a non-carpenter neighbor's inept attempt at building a tree house for his child. But perhaps there is more to it.

In thinking about this story, we should first admit that the frequency of the story does not reflect the frequency of the event, that is, we are not simply telling the late diagnosis story so often because it is such a common one. The sort of late diagnosis reflected in the story happens far less often than, say, the first acquisition of pseudomonas, or the onset of CF related diabetes, or for that matter, the death of a patient from CF. Like many narratives, the moral of the story of "late" diagnosis is more complicated than it first appears. In this case, the "late" diagnosis story stands in for a complex set of assumptions about life with CF, CF clinical care, and the role of the CF physician. Our head-shaking reaction to the narrative of late diagnosis reflects our assumption that the patient's life would

Walter M. Robinson, MD, MPH Children's Hospital, Boston and Harvard Medical School analysis. This improves the detection sensitivity from 94% to approximately 99%. Additionally, the direct confirmation of CF from the newborn screen by identification of two alleles increased from 41% with ∆F508 analysis only to 64% with multiple mutation analysis. 7. Since the introduction of newborn screening for CF in Wisconsin, the total number of annual sweat tests has decreased from 1670 in the year 1991 to 804 in the year 2000. 4 Concomitant with this decrease in the number of sweat tests has been a decrease in the number of institutions performing sweat tests from 44 in the year 1991 to 15 institutions in the year 2000. The cost savings of a fewer number of sweat tests accounts for 77% of the cost to operate Wisconsin's CF neonatal screening program. 8. In order to achieve excellent nutritional outcomes, we exceed the CFF recommendation of quarterly clinic visits by utilizing every 6-8 week clinic visits in infants <1 year of age. Newborn screening provides the opportunity for better outcomes, and excellent follow-up care is essential to achieve outstanding results.

always be better if we had made the diagnosis sooner. That the story is much more complicated is evident in the outpatient clinic of every CF physician; the contemporary heterogeneity of the clinical course of CF, and the presence of a wide variety of clinical conditions at every age, simply belies the assumption that in any individual case a "late" diagnosis always means more suffering than an "early" diagnosis. And if the history of CF care has a cautionary tale for us, it is that our current therapies and models of care may well have unintended consequences, and that we cannot see ourselves as always wiser in our use of various therapies than our forebears. In short, it may be that more therapy at the youngest possible age is in every case a good idea, but then again it may not.

The late diagnosis narrative also supports an historically inaccurate view of the process of diagnosis; it is the state of diagnostic technology, rather the presence or absence of symptoms, which is the strongest determinant of the timing of diagnosis. Before the reliability of the sweat test was established, no child could be diagnosed with CF in the absence of characteristic symptoms; in this situation, there are no pre-clinical diagnoses, that is, there are no "early" diagnoses, and all "late" diagnoses are likely to be a reflection of unfamiliarity with the disease. As the diagnostic accuracy of the sweat test advances, we begin to see its use first to confirm a clinical suspicion based on signs and symptoms, then to establish the diagnosis in the face of a confusing set of signs and symptoms, and finally to screen for a diagnosis to detect the pre-symptomatic disease. In a very real sense, the screening aspect of the sweat test transforms the healthy into the diseased before the appearance of any symptoms, and in so doing redefines the boundary of "early" diagnosis.

If neonatal screening is adopted, we can only expect further reconfiguring of the boundaries of early and late diagnosis, so that a child diagnosed at six months of age would then be seen as a "late" diagnosis and the asymptomatic only child with a CF genotype will be unable to live the quiet life of the undiagnosed.

Considerations of newborn screening often center on parental anxiety and its function as either a benefit or burden of screening programs. A variety of types and causes of anxiety are put forth as amenable by (or caused by) newborn screening:

• Anxiety about the presence of disease in an apparently healthy infant • Anxiety about the lack of diagnosis for the symptoms seen in a sick infant • Anxiety about the detection of life threatening illness in the apparently healthy, asymptomatic infant • Anxiety about the detection of life threatening, as opposed to self-limited, illness in the sick infant • Anxiety about future reproduction • Anxiety about the blame for passing on a genetic illness • Anxiety about the implications of the carrier state of the newborn It is perhaps peculiarly modern that we are debating the role of a public health program in promoting or alleviating anxiety. Yet anxiety is a slippery diagnosis, and as in the late diagnosis story, our attribution of anxiety to others may reflect our own concerns rather than the concerns of others. How ought we as clinicians assess these claims and counterclaims about anxiety, and how ought we to weigh the relief or creation of anxiety when assessing the ethics of newborn screening?

Parts 

Although CFRD is different than type 1 or type 2 diabetes, the development of diabetes-induced complications is similar. These complications include eye, kidney, circulation and nerve problems, and are caused by high blood glucose levels. Thus the goals of CFRD management are similar to those of people with all forms of diabetes and help prevent development of complications from diabetes.

Insulin is an important anabolic hormone which prevents protein catabolism and studies have demonstrated catabolism in both adults and children with CF, and in both the fed and fasted state. Protein catabolism is asso- Appropriate nutrition is essential to management of patients with cystic fibrosis. Many cannot maintain appropriate weight or growth without supplemental nutrition support. Initiation of such support is facilitated by lifelong education of patients and their parents as to the importance of good nutrition in the management of CF. Dieticians play a role in every CF visit and families are exposed to information about gastrostomy-tubes long before they become necessary.

Nutrition failure is identified using the guidelines established by the consensus report on nutrition for pediatric patients with cystic fibrosis and the CF adult care consensus conference. Patients are initially evaluated for manageable causes of nutrition failure, including poor oral intake, uncontrolled malabsorption, and increased nutritional needs. Poor oral intake can result from many causes, including chronic sinusitis, chronic pulmonary infections, onset of CF-related diabetes, and gastrointestinal causes (e.g., gastroesophageal reflux or distal intestinal obstructive syndrome). Depression can also contribute to poor intake. Malabsorption may result from inadequate or inappropriate enzyme use, cholestasis, short bowel syndrome, or small bowel overgrowth. The presence of pulmonary or sinus infections may lead to increased nutritional needs which the patient may not be able to meet orally.

If oral supplementation and medical management fail, gastrostomy tube placement must be considered to restore normal nutritional status and normal growth. Percutaneous endoscopic gastrostomy placement is generally successful. Patients may then receive nocturnal drip feedings over 8-10 hours as supplement. Aggressive control of reflux and monitoring for abnormal glucose tolerance is important.

Preoperatively, we counsel patients and families that gastrostomy tube placement is usually permanent in CF patients and will allow normal activity. We review their pulmonary status and may request an NG tube feeding trial to demonstrate efficacy.

Patients are admitted after placement to allow pain management, provide classes for parents and caregivers, and maintain good pulmonary function after the procedure. G-tube feedings are usually an elemental or semielemental formula, with enzymes used at bedtime and available if the patient arouses at night. AM glucoses are monitored for a time after feeding introduction. Good follow-up is essential to monitor progress and reduce complications. Patients may experience pain or nausea during feedings, diminished daytime appetite, no weight gain despite gastrostomy tube feedings, onset of insulin requirement, or complications of the gastrostomy itself.

Malnutrition and Nutritional Growth failure in CF is a common and important issue in CF management because of its adverse effects on long term outcome. Optimal standard of care in most cases requires meeting the excess energy, protein and vitamin needs of CF patients via overcoming malabsorption with appropriate pancreatic enzyme replacement therapy (PERT), the use of oral supplements, sometimes helped by behavioral therapy, and /or, in some cases, enteral nutrition supplements (EN). There are, however, some CF -related complications , where conventional nutritional support is not enough, may fail, or is impossible to maintain. These include neonatal meconium ileus with consequent short bowel syndrome, CF related liver disease with liver synthetic dysfunction, severe lung disease with frequent exacerbations, chronic lung disease or pulmonary failure, CFrelated diabetes, and some cases of poor intake and/or compliance, particularly if these are manifest during critical phases of growth such as infancy or adolescence. Various adjuncts to nutritional therapy have been studied and may be considered in these circumstances (1) . Firstly, maximising PERT with acid suppression or increasing duodenal bicarbonate may enhance absorption of nutrients (2, 3) .However, there is insufficient evidence to indicate whether there is an improvement in nutritional status, lung function, quality of life, or survival (4). Secondly, enhancing EN with specific supplements, eg trophic factors such as glutamine for short bowel syndrome (5), and branched chain amino acids for liver disease may give added benefit (6) . Thirdly, improved protein turnover, and efficiency of energy/fuel use and consequent wight gain and growth may be achieved by adjunctive growth hormone therapy (7) . Other adjuncts, such as IGF-1, and megestrol acetate have been evaluated with more variable outcomes. .Finally, where EN is inadequate, or impossible, the use of parenteral nutrition may be necessary, and has documented benefits (10) . The latter is particularly useful in pre-lung transplant patients, where improving/ maintaining nutrition may have an important bearing on outcomes. Questions as to when and how these various adjuncts are best utilised are at present incompletely answered, but it does appear that any or all approaches are preferable to continuing to starve the patient in the face of progressive undernutrition.

Etiology and Pathogenesis

Some organic digestive disturbances in early life

The diagnosis of cystic fibrosis: A consensus statement

Cluster Analysis

Classification and regression trees

Chest physiotherapy -the mechanical approach to anti-effective therapy in cystic fibrosis

Effectiveness of home vs hospital care in the routine treatment of CF

Schwachman-Kulczcki score and resting energy expenditure in CF

Effect of exercise and physiotherapy in aiding sputum expectoration in adults with cystic fibrosis

Chest physiotherapy, gastro-oesophageal reflux, and arousal in infants with cystic fibrosis

European Respiratory Society Task Force on the use of nebulizers

Use of nebulised saline and nebulised terbutaline as an adjunct to chest physiotherapy

Inhaled hypertonic saline increases sputum expectoration in cystic fibrosis

Liver involvement in Cystic Fibrosis

Hepatobiliary complications of cystic fibrosis

Outcome of liver cirrhosis in cystic fibrosis. Management of portal hypertension

Liver cirrhosis in Cystic Fibrosis. Therapeutic implications and long-term followup

Giunta A and the Italian Group for the Study of Ursodeoxycholic acid in Cystic Fibrosis. Ursodeoxycholic acid for liver disease associated with Cystic Fibrosis: a double-blind multicenter trial

Natural history of liver disease in Cystic Fibrosis

Ursodeoxycholic acid for cystic fibrosis-related liver disease (Cochrane Review)

Genotype-phenotype correlation in cystic fibrosis: the role of modifier genes

Cystic Fibrosis Genotype/Phenotype Consortium. Correlation between genotype and phenotype in cystic fibrosis

Novel uses to study complex traits and genetic diseases

Residual chloride secretion in intestinal tissue of deltaF508 homozygous twins and siblings with cystic fibrosis. The European CF Twin and Sibling Study Consortium

Chloride conductance and genetic background modulate the cystic fibrosis phenotype of Delta F508 homozygous twins and siblings

Detection of a cystic fibrosis modifier locus for meconium ileus on human chromosome 19q13

Chronic pain in cystic fibrosis

The musculoskeletal complications of cystic fibrosis

End-of-Life Care in Cystic Fibrosis

Cystic Fibrosis Foundation. (2003) About CF. Retrieved

Methodological issues in webbased research

Ethical and legal aspects of human subjects to research on the internet

Graphics Visualization Usability Center tenth WWW user survey

Research and the web: Eyeballs or smiles?

Designing and implementing web-based surveys

Chronic pain in cystic fibrosis

Pain in infants, children, and adolescents

Self-hypnosis for patients with cystic fibrosis

Hypnosis, Theodore Roosevelt, and the patient with cystic fibrosis

Hypnosis in treatment of clinical pain: Understanding why hypnosis is useful

Adjunctive non-pharmacological analgesia for invasive medical procedures: a randomized trial

Hypnosis and hypnotherapy with children

Handbook of hypnosis for professionals

Abnormal glucose metabolism in cystic fibrosis

Diabetes mellitus associated with cystic fibrosis

Influence of the development of diabetes mellitus on clinical status in patients with cystic fibrosis

Identification and treatment of cystic fibrosis-related diabetes: a survey of current medical practice in the

Diagnosis of cystic fibrosis related diabetes: a selective approach in performing the oral glucose tolerance test based on combination of clinical and biochemical criteria

Glucose tolerance in patients with cystic fibrosis: five year prospective study

Diagnosis of diabetes in cystic fibrosis and thalassemia major

Diagnosis, screening and management of cystic fibrosis related diabetes mellitus

Diabetes mellitus in Danish CF patients: prevalence and late diabetic complications

Diabetic microangiopathy in patients with cystic fibrosis

The interaction of two diseases: diabetes mellitus and cystic fibrosis

Generalised atherosclerosis in an adult with CF and diabetes mellitus (Abstract)

Lipid metabolism in adults with cystic fibrosis

Abnormal lipid levels in cystic fibrosis

Effect of protein ingestion on the glucose and insulin response to a standardized oral glucose load

Ingested protein has little effect on glucose concentration or rate of glucose appearance in people with type 2 diabetes (Abstract #97)

Effects of meal carbohydrate content on insulin requirements in type 1 diabetic patients treated intensively with the basal-bolus (ultralenteregular) insulin regimen

Cystic Fibrosis Related Diabetes: an approach to diagnosis and management

Clinical Guidelines for Physio. Management of SUI

Back pain and spinal deformity in cystic fibrosis

Kyphosis and fractures in children and young adults with cystic fibrosis

Increased rate of fractures and severe kyphosis: sequelae of living into adulthood with cystic fibrosis

Musculoskeletal involvement in cystic fibrosis

Vertebral deformities and low bone mineral density in adults with cystic fibrosis: a cross-sectional study

Posture and cystic fibrosis

Impact of lung inflammation on bone metabolism in adolescents with cystic fibrosis

Postural relief of dyspnoe in severe chronic airflow limitation: relationship to respiratory muscle strength

Physiotherapy for Respiratory and Cardiac Problems, 3 rd edn

Chest mobilization and respiratory function

Habitual physical activity in children and adolescents with cystic fibrosis

Physiotherapy in infants and young children with cystic fibrosis: current practice and future developments

International Physiotherapy Group for Cystic Fibrosis (IPG/CF): Physiotherapy in the treatment of CF

Effects of nutritional status on exercise performance in advanced cystic fibrosis

Respiratory and peripheral muscle function in cystic fibrosis

Determinants of aerobic and anaerobic performance in cystic fibrosis

Exercise performance and breathing patterns in cystic fibrosis: male-female differences and influence of resting pulmonary function

Muscle function and resting energy expenditure in female athletes with cystic fibrosis

The Km of NADH dehyrogenase is decreased in the mitochondria of cystic fibrosis cells

Mitochondrial NADH dehyrogenase in cystic fibrosis

Muscle size and cardiorespiratory response to exercise in cystic fibrosis

Peripheral muscle weakness and exercise capacity in children with cystic fibrosis

The relationship between genotype and exercise tolerance in children with cystic fibrosis

End organ dysfunction in cystic fibrosis

Angiotensin converting enzyme gene insertion/ deletion polymorphism and response to physical training

Exercise performance and breathing patterns in cystic fibrosis: male-female differences and influence of resting pulmonary function

Multiple factors that limit exercise capacity in cystic fibrosis

Activity levels and relationship to lung function and nutritional status in children with cystic fibrosis

Habitual activity in children and adolescents with cystic fibrosis

A randomized controlled trial of a 3 year home exercise program in cystic fibrosis

Musculoskeletal manifestatons in cystic fibrosis

Low bone mineral density in adults with cystic fibrosis

Osteoporosis and osteopenia in adults and adolescents with cystic fibrosis: prevalence and associated factors

Histomorphometric analysis of bone biopsies from the iliac crest of adults with cystic fibrosis

The role of physiotherapy in the prevention and treatment of osteoporosis

Exercise, bone and nutrition

Bone Mineral Density and Bone Acquisition in Children and Young Adults with Cystic Fibrosis: A Follow-Up Study

Milk intake and bone mineral acquisition in adolescent girls: randomised, controlled intervention trial

Sex steroids and body composition in men with cystic fibrosis

Impact of lung inflammation on bone metabolism in adolescents with cystic fibrosis

Acquisition of optimal bone mass in childhood and adolescence

Bone mineral density in Australian children, adolescents and adults with cystic fibrosis: a controlled cross sectional study

A six-year longitudinal study of the relationship of physical activity to bone mineral accrual in growing children: the university of Saskatchewan bone mineral accrual study

Decreased bone mineral density in normal-growing patients with cystic fibrosis

Analysis of factors limiting maximal exercise performance in cystic fibrosis

Prospective ten-month exercise intervention in premenarcheal girls: positive effects on bone and lean mass

Highimpact exercise and bones of growing girls: a 9-month controlled trial

Former exercisers of an 18-month intervention display residual BMD benefits compared with control women 3.5 years post-intervention: a follow-up of a randomized controlled high-impact trial

Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA

Isotype-specific activation of cystic fibrosis transmembrane conductance regulator-chloride channels by cGMP-dependent protein kinase II

Apparent affinity of CFTR for ATP is increased by continuous kinase activity

cAMP-independent phosphorylation activation of CFTR by G proteins in native human sweat duct

The complex and intriguing lives of PIP2 with ion channels and transporters

Protein kinase-independent activation of CFTR by phosphatidylinositol phosphates

Phosphorylation-regulated Clchannel in CHO cells stably expressing the cystic fibrosis gene

Phosphorylation by protein kinase C is required for acute activation of cystic fibrosis transmembrane conductance regulator by protein kinase A

Specificity and mechanism of action of some commonly used protein kinase inhibitors

The bisindolylmaleimide GF 109203X is a potent and selective inhibitor of protein kinase C

Intracellular transport mechanisms of signal transducers

The first nucleotide binding domain of cystic fibrosis transmembrane conductance regulator is a site of stable nucleotide interaction, whereas the second is a site of rapid turnover

Prolonged nonhydrolytic interaction of nucleotide with CFTR's NH2-terminal nucleotide binding domain and its role in channel gating

A tweezers-like motion of the ATP-binding cassette dimer in an ABC transports cycle

Control of CFTR channel gating by phosphorylation and nucleotide hydrolysis

Structure of nucleotide-binding domain 1 of the cystic fibrosis transmembrane conductance regulator

Cooperative, ATP-dependent association of the nucleotide binding cassettes during the catalytic cycle of ATP-binding cassette transporters

Mutation of Walker-A lysine 464 in cystic fibrosis transmembrane conductance regulator reveals functional interaction between its nucleotide binding domains

ATP binding to the motor domain from an ABC transporter drives formation of a nucleotide sandwich dimer

Nucleotide occlusion in the human cystic fibrosis transmembrane conductance regulator. Different patterns in the two nucleotide binding domains

On the mechanism of MgATP-dependent gating of CFTR Cl -channels

Dual effects of ADP and AMP-PNP on CFTR channel kinetics show binding to two different nucleotide binding sites

Gating of CFTR by nucleoside triphosphates: quantitative analysis of a cyclic gating scheme

Gating of CFTR by ATP hydrolysis: Structure and Function

Early pulmonary inflammation in infants with cystic fibrosis

The relationship between infection and inflammation in the early stages of lung disease from cystic fibrosis

Quantitation of inflammatory responses to bacteria in young cystic fibrosis and control patients

Inflammation and infection in naive human cystic fibrosis airway grafts

Excessive inflammatory response of cystic fibrosis mice to bronchopulmonary infection with Pseudomonas aeruginosa

Diverse Pseudomonas aeruginosa gene products stimulate respiratory epithelial cells to produce interleukin-8

Overproduction of the CFTR R domain leads to increased levels of asialoGM1 and increased Pseudomonas aeruginosa binding by epithelial cells

Proinflammatory cytokine responses to P. aeruginosa infection in human airway epithelial cell lines

Altered cytokine production by cystic fibrosis tracheal gland serous cells

Inflammatory response in airway epithelial cells isolated from patients with cystic fibrosis

Cytokine secretion by cystic fibrosis airway epithelial cells

Altered respiratory epithelial cell cytokine production in cystic fibrosis

Chemokine expression in CF epithelia: implications for the role of CFTR in RANTES expression

In vivo alterations of IFN regulatory factor-1 and PIAS1 protein levels in cystic fibrosis epithelium

Activation of NF-B by adherent Pseudomonas aeruginosa in normal and cystic fibrosis respiratory epithelial cells

Exaggerated activation of NF-B and altered IB-␤ processing in cystic fibrosis bronchial epithelial cells

High susceptibility for cystic fibrosis human airway gland cells to produce IL-8 through the IB kinase ␣ pathway in response to extracellular NaCl content

Reduced Smad3 protein expression and altered transforming growth factor-␤1-mediated signaling in cystic fibrosis epithelial cells

Reduced interleukin-8 production by cystic fibrosis airway epithelial cells

Interleukin-8 production by cystic fibrosis nasal epithelial cells after tumor necrosis factor-␣ and respiratory syncycial virus stimulation

Alternate-day prednisone reduces morbidity and improves pulmonary function in cystic fibrosis

Effect of high-dose ibuprofen in patients with cystic fibrosis

Regulation of human airway epithelial cell IL-8 expression by MAP kinases

Abnormal morphogenesis but intact IKK activation in mice lacking the IKK␣ subunit of the IB kinase

IKK␣ provides an essential link between RANK signaling and cyclin D1 expression during mammary gland development

Activation of NF-B by nontypeable Hemophilus influenzae is mediated by TLR2-TAK1-dependent NIK-IKK␣/␤-IB␣ and MKK3/6-p38 MAP kinase signaling pathways in epithelial cells

Micrococci and peptidoglycan activate TLR2-MyD88-IRAK-TRAF-NIK-IKK-NF-B signaling pathway that induces IL-8 transcription

Hershenson MB. Signaling intermediates required for NF-B activation and IL-8 expression in CF bronchial epithelial cells

Coordinate regulation of IB kinases by MAP kinase kinase kinase 1 and NF-B inducing kinase

Ras and MEKK-1 coregulate AP-1 and NF-Bmediated gene expression in airway epithelial cells

Activation of NF-B via a Src-dependent Ras-MAPK-pp90 rsk pathway is required for Pseudomonas aeruginosa-induced mucin overproduction in epithelial cells

Pseudomonas pyocyanin increases interleukin-8 expression by human airway epithelial cells

CF pathogens activate Ca 2+ -dependent MAP kinase signaling pathways in airway epithelial cells

ATP transduces signals from ASGM1, a glycolipid that functions as a bacterial receptor

TLR2 is mobilized into an apical lipid raft receptor complex to signal infection in airway epithelial cells

The kinase TAK1 can activate the NIK-IB as well as the MAP kinase cascade in the IL-1 signalling pathway

Inflammation and infection in naive human cystic fibrosis airway grafts

Activation of NF-B in airway epithelial cells is dependent on CFTR trafficking and Cl (-) channel function

Selective upregulation of IL-8 expression in CF bronchial gland cells in vivo and in vitro

Signal transduction from the endoplasmic reticulum to the cell nucleus

Monocyte retention and migration in pulmonary inflammation. Requirement for neutrophils

Differential roles for alpha(M)beta(2) integrin clustering or activation in the control of apoptosis via regulation of akt and ERK survival mechanisms

Elastase-mediated phosphatidylserine receptor cleavage impairs apoptotic cell clearance in cystic fibrosis and bronchiectasis

Macrophages that have ingested apoptotic cells in vitro inhibit proinflammatory cytokine production through autocrine/ paracrine mechanisms involving TGF-beta, PGE2, and PAF

Phosphatidylserine-dependent ingestion of apoptotic cells promotes TGF-beta1 secretion and the resolution of inflammation

Cystic Fibrosis in Adults

Drug discovery in academia

High-affinity activators of CFTR chloride conductance identified by high-throughput screening

Thiazolidinone CFTR inhibitor identified by highthroughput screening blocks cholera-toxin induced intestinal fluid secretion

Altered channel gating mechanism for CFTR inhibition by a high-affinity thiazolidinone blocker

Prevention of toxin-induced intestinal ion and fluid secretion by a small-molecule CFTR inhibitor

CFTR involvement in nasal potential differences in mice and pigs studied using a thiazolidinone CFTR inhibitor

A small molecule CFTR inhibitor produces cystic fibrosis-like submucosal gland fluid secretions in normal airways

Pharmacology and antidiarrheal efficacy of a thiazolidinone CFTR inhibitor in rodents

Discovery of glycine hydrazide pore-occluding CFTR inhibitors: mechanism, structure-activity analysis and in vivo efficacy

Nanomolar affinity small-molecule correctors of defective ∆F508-CFTR chloride channel gating

Nonsense-mediated mRNA decay in health and disease

Highly variable incidence of cystic fibrosis and different mutation distribution among different Jewish ethnic groups in Israel

Cystic fibrosis in Jews: Frequency and mutation distribution

Suppression of a nonsense mutation in mammalian cells in vivo by the aminoglycoside antibiotics G-418 and paromomycin

Aminoglycoside suppression at UAG, UAA and UGA codons in Escherichia coli and human tissue culture cells

Suppression of a CFTR premature stop mutation in a bronchial epithelial cell line

A pilot study of the effect of gentamicin on nasal potential difference measurements in cystic fibrosis patients carrying stop mutations

Evidence that systemic gentamicin suppresses premature stop mutations in patients with cystic fibrosis

Gentamicin-induced correction of CFTR function in patients with cystic fibrosis and CFTR stop mutation

Developing cystic fibrosis lung transplant referral criteria using predictors of 2-year mortality

Prediction of mortality in patients with cystic fibrosis

Survival effect of lung transplant among patients with cystic fibrosis

Living Lobar Transplantation

Lung transplantation of ventilator dependent patients

Recipient Selection

International Guidelines for the Selection of Lung Transplant Candidates

Lung transplantation is warranted for stable, ventilator dependent recipients

Early diagnosis of cystic fibrosis through neonatal screening prevents severe malnutrition and improves long-term growth

Longitudinal evaluation of bronchopulmonary disease in children with cystic fibrosis

Bronchopulmonary disease in children with cystic fibrosis after early or delayed diagnosis

Cognitive function of children with cystic fibrosis: deleterious effect of malnutrition

References 1. Merelle ME, Griffioen RW, Dankert-Roelse JE. Cystic fibrosis presenting with intracerebral haemorrhage

Dehydration deaths in infants and young children

Pancreatic function in infants identified as having cystic fibrosis in a neonatal screening program

Pancreatic insufficiency, growth, and nutrition in infants identified by newborn screening as having cystic fibrosis

Early diagnosis of cystic fibrosis through neonatal screening prevents severe malnutrition and improves long-term growth. Wisconsin Cystic Fibrosis Neonatal Screening Study Group

Investigators and Coordinators of the Epidemiologic Study of Cystic Fibrosis. Growth and nutritional indexes in early life predict pulmonary function in cystic fibrosis

Stature as a prognostic factor in cystic fibrosis survival

Bronchopulmonary disease in children with cystic fibrosis after early or delayed diagnosis

The diagnosis of cystic fibrosis

Cystic Fibrosis Foundation

Cystic fibrosis diagnosed after age 13. Twenty-five teenage and adult patients including three asymptomatic men

Diagnosis of cystic fibrosis during adolescence

Clinical features associated with a delayed diagnosis of cystic fibrosis

Delayed diagnosis of US females with cystic fibrosis

The Spirit of Lo. An Ordinary Family's Extraordinary Journey

Dueling with cystic fibrosis: finding the chloride defect

Early versus late diagnosis: psychological impact on parents of children with cystic fibrosis

Effects of newborn screening of cystic fibrosis on reported maternal behaviour

Newborn Screening for Cystic Fibrosis in Wisconsin: Nine Years Experience with Routine Trypsinogen/DNA Testing

Genetic counseling and neonatal screening for cystic fibrosis: an assessment of the risk communication process

Sweat chloride concentrations in infants homozygous or heterozygous for F508 cystic fibrosis

Analysis of the costs of diagnosing cystic fibrosis with a newborn screening program

Comprehensive analysis of risk factors for acquistion of Pseudomonas aeruginosa in young children with cystic fibrosis

Pediatric Clinics of North America

Managing Cystic Fibrois Related Diabetes (CFRD). A teaching manual for cystic fibrosis related diabetes. The CF Foundation

Enteral tube feeding for cystic fibrosis

Consensus report on nutrition for pediatric patients with cystic fibrosis

Cystic fibrosis adult care: consensus conference report

Optimising nutrition in cystic fibrosis

Gastric acid suppression and treatment of severe exocrine pancreatic insufficiency

Omeprazole, a proton pump inhibitor, improves residual steatorrhoea in cystic fibrosis patients treated with high dose pancreatic enzymes

Drug therapies for reducing gastric acidity in people with cystic fibrosis.Cochrane Database Syst Rev

Trophic and cytoprotective nutrition for intestinal adaptation, mucosal repair, and barrier function

Nutritional Support in children with end-stage liver disease: a randomized crossover trial of a branched-chain amino acid supplement

Growth problems and growth hormone treatment in children with cystic fibrosis

Effects of glutamine and recombinant human growth hormone on protein metabolism in prepubertal children with cystic fibrosis

Improved growth and clinical, nutritional, and respiratory changes in response to nutritional therapy in cystic fibrosis

S20.4 WHEN ENTERAL NUTRITION IS NOT ENOUGH -ADJUNCTS TO NUTRITION SUPPORT IN CF

Newborn screening for cystic fibrosis (CF) has been performed as part of a randomized, controlled trial for patients born in Wisconsin between April 15, 1985 and June 30, 1994 . In the randomized, controlled trial, the study investigators were informed of a positive neonatal screen (Immunoreactive trypsinogen [IRT] in the first six years of the study and IRT/DNA in the last three years of the study) for patients in the early diagnosis group. The study investigators informed the primary care physician of the positive newborn screen and the infant was brought to Madison or Milwaukee (the two study centers) for sweat testing at six weeks of age. In July 1994, CF newborn screening was added to the panel of routine neonatal screening tests. Over the past decade, we have learned the following lessons from the routine CF newborn screening program. 1 1. Initially in the routine neonatal screening program, the newborn screen collection card had a check-off box for meconium ileus. If this box was checked, then DNA analysis for the ∆F508 mutation was performed regardless of the trypsinogen level. The rationale for this was to detect CF patients who theoretically could have a low trypsinogen level after surgery for meconium ileus. However, in the first 3 1/2 years of the routine CF newborn screening program, there were a total of 1266 infants in which meconium ileus was checked on the newborn screening specimen card. The staff at newborn nurseries were possibly checking this box erroneously for meconium stained amniotic fluid. In confirmed cases of CF with meconium ileus, there were no infants in which the trypsinogen level was less than the cut-off value. Therefore, the meconium ileus box on the newborn screening form was removed. 2. Initially in the routine newborn screening program, the written recommendation for an infant with a positive CF newborn screen was to have a sweat test performed at four weeks of age. Although the two CF Centers in the state of Wisconsin were listed on the report, it was not explicitly stated that sweat testing should occur at a CF Center. This led to two difficulties: a. Families were required to wait until the baby was four weeks of age for a sweat test. During this time, they encountered anxiety about the possibility that their child could have CF. b. A number of sweat tests were performed at hospitals that were not CF Centers. This led to questions about the accuracy of the sweat testing procedure. Additionally, families were not receiving genetic counseling in these non-CF Center settings.These issues were addressed by revising the positive CF neonatal screening report. The report no longer states that the infant should be four weeks of age for a sweat test and, therefore, there is now less of a delay from the time that the primary care physician notifies the family of the positive newborn screening test until the time that a sweat test is performed. Additionally, the newborn screening report now states that a quantitative pilocarpine iontophoresis sweat test should be performed at CF Centers or Affiliate Centers listed on the report. As a result of these changes, in the calendar year 2003, there was only one infant who had a sweat test outside of a CF Core Center or Affiliate Center. 3. We are aware that families are anxious when they learn that their infant has a positive neonatal screen (specifically, an elevated trypsinogen and one CF mutation).To help allay the parents' anxiety 2 , a letter has been created that accompanies the positive CF newborn screening report to the primary care physician. This letter is intended to be given to the parents and emphasizes that there is only a possibility of cystic fibrosis. The letter briefly describes the sweat test and that no special preparation is needed. Contact information for the CF core and affiliate centers and the state newborn screening coordinator is provided. Additionally, the letter directs the parents to a website that we have created that provides useful information about CF newborn screening <http://www.pediatrics.wisc.edu/childrenshosp/CF/nbs.html>. Within this website is a video entitled "Understanding Newborn Screening for Cystic Fibrosis". All of these educational efforts have significantly decreased the number of phone calls that we receive from panicked parents. 4. From our experience in the CF newborn screening program, it is clear that the traditional decision levels for sweat chloride (<40 mEq/L is normal, 40-60 mEq/L is borderline, and >60 mEq/L is diagnostic of cystic fibrosis) is not applicable to infants. We have demonstrated that sweat chloride levels of >30 mEq/L in an infant is suspicious for CF. Extended mutation analysis in infants with a sweat chloride of >30 mEq/L almost invariably reveals two CF mutations. This agrees with previous data. 3 5 . Infection control measures are of utmost importance, as clinic exposures in a small waiting room (110 sq. ft.) was a risk factor for acquisition of Pseudomonas aeruginosa. 5 6. CF newborn screening has evolved from an algorithm of IRT/∆F508 analysis to IRT/multiple (25) mutation

ciated with insulin deficiency and resistance to insulin's anti-catabolic effects on whole body protein turnover. Protein catabolism is an especially important consideration in the malnourished CF child and adult. Malnourishment, principally loss of muscle mass, has been associated with worsened morbidity and mortality.

Insulin lowers blood sugar levels and functions as an anabolic agent by increasing cellular uptake of ingested nutrients and conversion of the nutrients to energy. Insulin is the principal hormone preventing protein catabolism. The 1998 CF Consensus conference recommended that insulin be the only medical therapy prescribed for CFRD with fasting hyperglycemia. The recommendations were based on the review of existing medical literature. The use of insulin may be especially important in a malnourished patient with CF, and several studies have documented weight gain and improvement of catabolism with insulin. Insulin use can be made more convenient by using insulin injection devices and insulin pumps. There are many types of insulin and potential insulin regimens available for treating CFRD, and therapy should be individualized. However, there are several principles to guide the clinician in choice of therapy.

1. The regimen should as closely as possible fit the patient's life-style. 2. Meal coverage should be provided to prevent postprandial hyperglycemia. 3. The regimen should confer as low a risk as possible for the development of hypoglycemia. 4. In most cases, good glycemic control can be obtained using of multiple injections.

The broad classes of insulin are "rapid-acting," "short-acting", "longer-acting" and basal insulin. Insulin action (when it peaks in activity, and how long it lasts) may vary from person to person. There is also some variability from one day to the next in the same person. When deciding on an insulin regimen, three types of coverage should be considered.• Meal Coverage (Bolus insulin): A normal pancreas secretes insulin as a "bolus" to cover the meal (specifically the carbohydrates) ingested. Rapid-acting or short-acting insulin is given before meals to mimic this extra insulin "bolus." Generally, the best way to dose pre-meal insulin dose is to account for the carbohydrate content of the meal.

The normal pancreas makes a small amount on insulin at all times. These low levels of insulin are called "background" or "basal" insulin. Generally long-acting insulin such as Ultralente (lasts approximately 12-24 hours and has a small peak) or Lantus (lasts 24 hours and has no peak) are used. Most people with CF make some insulin, thus they often do not require as much basal insulin as people who have diabetes but do not have CF. However, protein catabolism is higher in malnourished patients; therefore basal insulin is an important part of the insulin regimen in these patients. In general a starting dose which minimizes the risk of hypoglycemia is 0.1 units per kilogram of weight per day. This dose should be increased as needed, but hypoglycemia should be avoided.• Correction: When the blood glucose level is too high, rapid-acting insulin is given in addition to the usual insulin dose to "correct" the blood glucose level. Correction is especially important during illness and to prevent hyperglycemia-related diabetes complications. In general, when a patient is normally insulin sensitive, one unit of rapid-acting insulin will lower the glucose level approximately 50 mg/dl.

Night-time enteral feeding are particularly useful in the malnourished CF patient, and some may only have high blood sugar levels during these feedings. Hyperglycemia at this time is best managed by a combination of longeracting (NPH or Lente) plus short-acting insulin (Regular), at the start of the feeding. A starting dose which minimizes the risk of hypoglycemia is 0.1 units per kilogram per dose of Regular and 0.1 units/kg/dose of NPH.

The CF consensus conference on diabetes did not recommend that oral agent be used to treat CFRD with fasting hyperglycemia (the only glucose tolerance category for they recommended mandatory treatment). There are multiple types of oral agents, and a few small studies have documented efficacy in CFRD. Research is needed to test potential oral therapies.