key: cord-0865242-ikysbbc0
authors: Ritchey, Jerry W; Levy, Julie K; Bliss, Susan K; Tompkins, Wayne A.F; Tompkins, Mary B
title: Constitutive expression of types 1 and 2 cytokines by alveolar macrophages from feline immunodeficiency virus-infected cats
date: 2001-05-10
journal: Veterinary Immunology and Immunopathology
DOI: 10.1016/s0165-2427(01)00250-1
sha: abd20e33c3e08a2d1beed89375fcae1e510dd99a
doc_id: 865242
cord_uid: ikysbbc0

Abstract Evidence suggests that feline immunodeficiency virus (FIV), causes pulmonary immunodeficiency. The overall objective of this study was to explore FIV-induced alterations in cell counts and cytokine gene expression in the pulmonary compartment during the acute stage infection. Bronchoalveolar lavage (BAL) cells were collected from FIV-infected and control cats at 0, 4, 10, and 16 weeks post-FIV infection for phenotype and cytokine analysis. The major change in BAL cellular populations following FIV-infection was the development of a neutrophilia. Total BAL cell counts and relative numbers of alveolar macrophages (AM), eosinophils, and lymphocytes remained similar in both groups. The RT-qcPCR analyses of AM purified from BAL showed constitutive expression of TNFα, IL6 and IL10 mRNAs that peaked during the acute stage of infection then declined. The TNFα and IL6 bioactive protein secretion showed a similar response. In contrast, IFNγ expression increased progressively with time after infection and paralleled a progressive increase in FIV-gag mRNA in AM. The IL12 p40 expression also differed from the other cytokines in that there was a progressive decrease in the number of cats with AM IL12 expression following FIV infection. Infection of AM in vitro with FIV also caused an increase in TNFα and IL6 mRNA and bioactive protein suggesting that the increased cytokine response by AM following infection of cats with FIV is an intrinsic characteristic of FIV-infected AM. In summary, pulmonary immune changes seen in FIV-infected cats are similar to those seen in HIV-infected human patients.

The lung has been recognized as one of the most common targets of secondary infections, including Mycobacterium tuberculosis, Cryptococcus neoformans, Histoplasma capsulatum, and Toxoplasma gondii, in human immunode®ciency virus (HIV) infected patients. The primary line of defense against such pathogens is the activated alveolar macrophage (Beck and Shellito, 1989) . Activation of alveolar macrophages (AM) is a biphasic process regulated by a number of cytokines generated by antigen speci®c T cells (IFNg, TNFa) and the AM itself (TNFa, IL6, IL12). Early activation occurs as a result of IFNg and TNFa production from NK cells and macrophages. A more sustained activation of AM, which is required for successful control of intracellular pathogens, is dependent upon development of a Th1 immune response (IFNg and IL12).

Alveolar macrophage (AM) from HIV-infected patients constitutively produce cytokines, such as TNFa, IL10, IL8, IL10, and IL12, indicative of a state of activation (Agostini et al., 1991a; Denis and Ghadirian, 1994; Trentin et al., 1992; Twigg et al., 1992 ), yet they are not able to successfully control intracellular pathogens. As both Th1 and Th2 cytokines are produced by activated macrophages, it may be the ratio of the particular cytokines that is important in determining whether macrophages will function properly or not. For example, although low levels of IL12 are constitutively produced by the AM of HIV-infected patients, levels could not be increased following stimulation by Staphylococcus aureus (Denis and Ghadirian, 1994) . These data suggest that although AM from HIV-infected patients constitutively secrete cytokines associated with activation, when faced with a secondary infectious agent, the AM cannot reach the level of activation necessary for successful clearance of the organism, possibly due to a failure in mounting a protective Th1 immune response.

Feline immunode®ciency virus (FIV), a lentivirus morphologically and biochemically similar to HIV, has a pathogenesis and clinical disease pattern that parallels that seen in HIV infection (Bendinelli et al., 1995; English et al., 1994) . The acute infection is followed by a variably long asymptomatic period characterized by a progressive decrease of CD4 lymphocytes that results in the hallmark inverted CD4 :CD8 ratio in the peripheral blood (Bendinelli et al., 1995; English et al., 1994) . The asymptomatic period is followed by the development of AIDS-associated disorders, many of which mimic AIDS in humans (Bendinelli et al., 1995; English et al., 1994) . Davidson et al. (1993) and Yang et al. (1996) reported that FIV-infected cats challenged with an avirulent strain of T. gondii developed severe respiratory disease resulting in 60±80% mortality/euthanasia. The FIV-infected cats developed systemic toxoplasmosis characterized by necrotic/ in¯ammatory foci in the liver, spleen, lymph nodes, and most severely in the lung. The lesions were accompanied by abundant replication of Toxoplasma tachyzoites readily noticeable in the sites of in¯ammation. In contrast, FIVÀ cats developed only a transient, mild clinical disease characterized by anorexia and lethargy without clinical signs of pulmonary involvement. Levy et al. (1998) examined the peripheral and bronchial lymph nodes from normal and FIV-infected cats following challenge with T. gondii and reported a marked difference in the cytokine pro®le in the two groups of cats. As compared to control cats challenged with T. gondii, cells harvested from the lymph nodes of FIV-T. gondii co-infected cats produced high levels of IL10 and IFNg but were de®cient in IL2 and IL12 (p40) mRNA production. Dean and Pedersen (1998) also reported elevated levels of IL10 and IFNg mRNA in lymph nodes of FIV-infected cats.

As the pulmonary macrophage microbicidal activation process and resistance to T. gondii is dependent upon an effective IL12-dependent Th1 immune response to T. gondii antigens, we were interested in determining what effect FIV-infection may have on AM in the lung. In this study, we characterized changes occurring in bronchoalveolar lavage (BAL) cell counts, differentials, and cytokine mRNA and protein expression from the AM of cats infected with FIV. To verify that any alterations in cytokine message and protein expression were associated with a productive viral infection, FIV-gag mRNA expression was quantitated in AM infected with FIV in vivo and in vitro.

Twenty-two, 6-month-old, speci®c pathogen free (SPF), neutered male cats were obtained from Liberty Laboratories (Liberty Corners, NJ). The SPF cats were seronegative for feline panleukopenia, feline herpesvirus, feline calicivirus, coronaviruses, feline leukemia virus, feline immunode®ciency virus, feline syncytia-forming virus, Chlamydia, and Toxoplasma.

The NCSU 1 strain of FIV was originally isolated from a naturally infected cat at the North Carolina State University College of Veterinary Medicine and has been described in detail elsewhere (English et al., , 1994 Davidson et al., 1993) . Twelve cats were inoculated with FIV intravenously with 5 Â 10 6 TCID 50 of cell-free virus culture supernatant as described by Davidson et al. (1993) . The remaining 10 cats were sham inoculated with medium. All FIV-infected cats were antibody positive by commercial ELISA (IDEXX, Westbrook, ME) and provirus positive by PCR by 4 weeks post-FIV infection.

Bronchoalveolar lavage (BAL) was performed as previously described (Hawkins and DeNicola, 1989; Ma et al., 1995) . Brie¯y, cats were anesthetized with ketamine and intubated so that the end of an endotracheal tube was positioned within the distal trachea. Lavage was performed in three aliquots through the tube with the cat in left lateral recumbency using a volume of 5 ml/kg body weight of sterile phosphate-buffered saline (PBS) in each aliquot. The aliquots were pooled, total viable BAL cell counts determined by trypan blue dye exclusion, and differential cell analyses performed on Diff±Quik stained cytospins. Using the differential to determine the numbers of AM per milliliter of BAL¯uid, duplicate cultures of AM (along with other contaminating BAL cells) were seeded in RPMI supplemented with 10% fetal bovine serum and antibiotics containing 1 Â 10 6 AM/ml each. Following overnight incubation, nonadherent cells were removed by washing with HBSS. From one culture, adherent cells were dislodged and an aliquot used to verify AM purity, which was always >95% as determined by differential analyses. Remaining cells were lysed and total cellular RNA prepared for quantitation of cytokine gene expression (see below). In the duplicate culture, AM were cultured with fresh medium for an additional 24 h, when culture supernatants were harvested and frozen at À858C until analyzed for TNFa and IL6 bioactivity.

Puri®ed AM (prepared as described above) from normal SPF cats were infected in vitro with FIV-NCSU 1 at an MOI of 1 for 24 h. Total cytoplasmic RNA and culture supernatant were collected on days 1, 3, and 7 post-infection and frozen until analyzed for FIV-gag mRNA, cytokine mRNA, and TNFa and IL6 bioactivity.

The IL6 bioactivity was determined using an IL6-dependent murine plasmacytoma cell line, T1165, as previously described (Ma et al., 1995; Nordan and Potter, 1986) . Brie¯y, serial two-fold dilutions of the supernatants to be assayed were made and 50 ml added to 50 ml of T1165 cells in 96-well culture plates. Controls included murine IL6 containing P388D cell supernatant (positive control) and medium alone (negative control). Each dilution was run in triplicate. After 2 days of incubation, 20 ml of medium containing 1 mCi 3 H-thymidine (ICN, Costa Mesa, CA) was added to each well and incubated overnight. Cells were then harvested and incorporation of 3 H-thymidine (in counts per minute) measured using a scintillation counter. Units of IL6 per milliliter (U/ml) were de®ned as the reciprocal of the dilution of the supernatant which stimulated 50% of the maximal counts per minute in each assay, multiplied by 10 since the assay volume was 0.1 ml.

The TNFa bioactivity was measured as previously described (Ma et al., 1995; Laster et al., 1988) with minor modi®cations. Brie¯y, LM cells, a mouse ®broblast cell line that is sensitive to TNF, were plated in 96 well¯at-bottom plates at a concentration of 4 Â 10 4 cells in 100 ml and cultured overnight. Fifty microliters of serial two-fold dilutions of AM cell supernatants were then added in triplicate along with 20 ml of culture medium containing 8.5 mg/ml actinomycin D. Serial dilutions of human recombinant TNFa (UBI, Lake Placid, NY) of known bioactivity were used as a standard and positive control. Following overnight incubation, 25 ml of 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide at 5 mg/ml in PBS was added to each well. Following a 4 h incubation, all supernatant was removed and 100 ml of 0.01N HCl in isopropanol was added to each well. After a 30 min incubation, 100 ml of ddH 2 O was added to each well and absorbency measured at 570/630 nm in an ELISA plate reader. Units of TNFa were determined from a standard curve derived from dilutions of the human recombinant TNFa standard. The murine ®broblasts may be either more or less sensitive to feline TNFa when compared with the human recombinant standard. Therefore, the values reported are meant for relative comparisons of feline TNFa levels between cats and timepoints and are not intended to convey an absolute quantitative measurement.

Total cellular RNA was isolated from puri®ed AM using the guanidinium thiocyanate technique as described by Rottman et al. (1995) . Brie¯y, RNAstat-60 (Tel Test B, Friendswood, TX) was added to monolayers of puri®ed AM (5 Â 10 6 cells) and the cells lysed by pipetting up and down multiple times. The RNA was subsequently isolated according to kit instructions.

Cytokine mRNA levels were quantitated by a RT-qcPCR method shown previously to be sensitive, accurate, and reproducible (Levy et al., 1998; Rottman et al., 1996) . The competitive DNA fragment consists of a linear array of upstream and downstream primer sequences for IFNg, IL2, IL4, IL6, IL10, IL12 (p40), TNFa, and G3PDH separated by a DNA spacer (Rottman et al., 1996) . Ampli®cation of each primer pair yields a competitor product of 573 base pairs, which is different than the cDNA products. A second 489 base pair competitive DNA fragment was constructed to allow quanti®cation of FIV-gag mRNA by RT-qcPCR. The FIV-gag primers were chosen from sequences of the NCSU 1 isolate molecular clone (Yang et al., 1996) .

Following reverse transcription of an RNA sample, equal amounts of cDNA were ampli®ed using cytokine speci®c primer pairs and serial ®ve-fold dilutions of a DNA competitive fragment (500 Â 10 À5 to 1:02 Â 10 À5 amol). The PCR products were separated by electrophoresis on 2% agarose gels, stained with ethidium bromide, and photographed under 300 nm UV transillumination with a Polaroid camera. An image was acquired using Adobe Photoshop and analyzed using the NIH Image Software Program. The log of the cytokine product¯uorescence was plotted against the log of the number of copies of competitor in the sample, resulting in the cytokine copy number at the x intercept. Constitutive mRNA expression for the measured cytokines is presented as a ratio of the cytokine to G3PDH to allow for quantitative comparisons between cytokines and animals.

Comparisons of results between FIV-infected and uninfected cats were performed using nonparametric analysis (Wilcoxon's ranking test). Differences between infected and uninfected cats were considered signi®cant at P < 0:05. Comparisons of FIV-p26 gag expression between cats within the FIV-infected group were made using the Student's ttest with an alpha level of 0.05.

As expected, all of the SPF cats infected with FIV-NCSU 1 became PCR for provirus in peripheral blood mononuclear cells (PBMC) and developed antibody to the gag proteins by 4 weeks after infection (data not shown). The FIV-infected cats demonstrated a progressive decline in the peripheral blood CD4 :CD8 ratios secondary to a decrease in CD4 cells and an increase in CD8 cells (Fig. 1) . These ®ndings are characteristic for infection by FIV and are similar to those previously reported for experimental infection with the NCSU 1 virus isolate (English et al., 1994; Davidson et al., 1993; Yang et al., 1996; Tompkins et al., 1991) .

Examination of cell populations in BAL¯uids clearly showed that the total cell counts derived from infected and control groups did not vary appreciably throughout the course of the experiment ( Fig. 2A) . This indicates that FIV infection of SPF cats alone does not result in measurable changes in pulmonary cell numbers. Differential counts on the recovered BAL populations from control and FIV-infected cats demonstrated that, with 

The HIV infection induces AM activation accompanied by the elaboration of a number of cytokines, such as TNFa, IL6, IL8, IL10, and IL12 (Agostini et al., 1991a; Denis and Ghadirian, 1994; Trentin et al., 1992; Twigg et al., 1992) . To determine if FIV infection caused a similar activation of AM, BAL cells were collected from the FIV-infected and control cats, and constitutive cytokine production was measured. Alveolar macrophages (AM) from FIV-infected cats produced signi®cantly greater levels of bioactive TNFa at 4 (P 0:001) and 10 (P < 0:001) weeks post-FIV infection than did AM from control cats (Fig. 3A) . Similarly, the FIV-infected cats had signi®cant elevations in bioactive IL6 at 4 (P 0:002) and 10 (P 0:003) weeks post-FIV infection (Fig. 3B ). The increased constitutive TNFa and IL6 secretion by AM coincided with changes in the CD4 and CD8 subsets in the circulation (Fig. 1) . At 16 weeks post-infection, when the CD4 :CD8 ratios had stabilized, there still remained signi®cant levels of constitutive TNFa (Fig. 3A , P 0:001) and IL6 (Fig. 3B , P 0:035) production by AM.

Because of the limited availability of bioassays or ELISAs for feline cytokines, alterations in other cytokines were assessed by quantitating changes in mRNA levels using RT-qcPCR. Interleukin-6, IL10, IL12 (p40), and TNFa were evaluated as monokines that are important for the initiation and regulation of innate and acquired immunity. For the most part, TNFa and IL6 mRNA expression correlated with the bioactive protein production ( Fig. 4A and B) . Alveolar macrophages (AM) from FIVinfected cats had elevated TNFa mRNA levels that bordered on statistical signi®cance at 4 weeks (P 0:056) and 10 weeks (P 0:045) post-FIV infection. At 16 weeks post-FIV infection, the TNFa mRNA levels were similar to that of control animals. The The IL10 mRNA expression in AM from FIV-infected cats peaked at 4 weeks post-FIV infection (Fig. 4C , P 0:007) and declined to control levels at 16 weeks post-FIV infection. The control cats consistently expressed lower levels of IL10 mRNA than FIV-infected cats throughout the experiment (Fig. 4C) .

Although generally considered a product of activated lymphocytes, IFNg expression (mRNA and bioactive protein) from macrophages has been experimentally documented in other model systems (Munder et al., 1998; Nugent et al., 1985) . Therefore, we investigated the ability of AM from FIV-infected cats to produce IFNg. In contrast to AM from control cats, AM from FIV-infected cats produced high levels of IFNg mRNA. Unlike the other cytokines, IFNg mRNA levels increased progressively over the course of the infection (Fig. 4D , P 0:003 at 10 weeks; P 0:002 at 16 weeks). Table 1 shows that the numbers of FIV-infected cats expressing IFNg mRNA also increased with time, whereas IFNg mRNA was detected in only one control cat at one time point (16 weeks post-FIV infection). Thus, FIV-induced changes in IFNg mRNA expression by AM was manifested by a time-related increase in the number of cats expressing IFNg mRNA, as well as an increase in the level of expression for individual cats. Table 2 shows that the AM collected from control cats consistently maintained low but measurable constitutive IL12 p40 mRNA expression at all timepoints. In contrast, the AM Table 1 ). from the FIV cats demonstrated a progressive decrease in IL12 mRNA expression with time after infection, such that there was a progressive decrease in the number of cats with detectable IL12 mRNA. When compared as groups, the AM from FIV-infected cats expressed signi®cantly less (P 0:032) IL12 mRNA that control cats at week 16 (data not shown). However, the FIV-infected cats that maintained IL12 mRNA expression (3/ 12) had levels similar to that seen in the control cats (data not shown).

The FIV-gag mRNA was not detected in either the control or FIV-infected groups cats before FIV infection. Following FIV infection, there was a steady increase of FIV-gag expression over the 16-week course of infection; with peak gag mRNA expression occurring at 16 weeks post-FIV infection (Fig. 5) . The increase in gag mRNA was due to Table 2 Numbers of cats with constitutive IL12 mRNA expression in puri®ed AM

Weeks post-FIV infection 0 4 10 16 FIV 11/12 9/12 6/12 3/12 FIVÀ 9/12 10/10 10/10 10/10 Fig. 5 . Quantitation of constitutive FIV-gag mRNA expression from AM collected from FIV-infected cats. Alveolar macrophages (AM) were puri®ed by plastic adherence and total cytoplasmic RNA was extracted and analyzed by RT-qcPCR utilizing oligo-primers speci®c for p26 of FIV. The FIV-gag mRNA was not detected in the control cats at anytime or in FIV-infected cats before FIV infection. Data represents the mean AE S:E: of the FIV-gag:G3PDH ratio from 12 FIV-infected cats. There was a signi®cant increase in FIV-gag mRNA expression throughout the experiment (4 < 10 weeks, P 0:001; 10 < 16 weeks, P 0:002). an absolute increase in RNA expression in each individual cat, as all of the cats expressed detectable p26 mRNA throughout the experiment.

The above data suggests that the constitutive cytokine expression by AM recovered from FIV-infected cats is a direct response to FIV infection. To test this directly, AM from normal cats were infected with FIV in vitro, and TNFa and IL6 mRNA and protein were assayed. Fig. 6 indicates that AM cultured in the absence of virus produced little to no TNFa ( Fig. 6A and B) or IL6 (Fig. 6C and D) mRNA or protein even after 7 days in culture, whereas those infected with FIV produced signi®cant levels of TNFa and IL6 message ( Fig. 6A and C) at 1 day after infection and protein (Fig. 6B and D) at 3 and 7 days after infection. Addition of irradiated FIV to AM cultures failed to induce a cytokine response (data not shown), suggesting a virus infection-dependent process. In support to this conjecture, FIV-gag mRNA was easily detectable in AM infected with live FIV but not irradiated virus at either 3 or 7 days after infection (data not shown).

There is considerable evidence that cellular in®ltrates and pro-in¯ammatory cytokines may play a role in the clinical manifestation of HIV-associated pulmonary disease. However, reports of speci®c cellular and cytokine changes in HIV-infected patients have been controversial. While it is clear that altered expression of IL1, IL6, IL8, IL10 and TNFa, in the pulmonary compartment is associated with HIV infection (Agostini et al., 1991a; Denis and Ghadirian, 1994; Trentin et al., 1992; Twigg et al., 1992) , it is not known whether this is an intrinsic characteristic of viral infection of pulmonary cells (AM) or due to cofactors, such as secondary opportunistic infections. In this study, we addressed the question of cellular changes and cytokine expression in the pulmonary compartment of FIV-infected SPF cats, a lentivirus animal model in which interpretation of results is not confounded by potential exposure to secondary pathogens.

Total cell counts and absolute numbers of AM, lymphocytes, and eosinophils were similar in both FIV-infected and control groups. Although most studies on HIV-infected patients report no differences in total cell counts or AM numbers (Rankin et al., 1988; Meltzer et al., 1993) , others have reported increases in both (Beck and Shellito, 1989; Agostini et al., 1991b) . Agostini et al. (1991b) reported increased AM numbers in approximately one-third of HIV-infected patients whether or not they had concurrent opportunistic infections. Failure to demonstrate a decrease in AM numbers from HIVinfected patients and FIV-infected cats suggests that the high incidence of opportunistic pulmonary diseases in these lentiviral infections is not due to a numerical de®ciency of macrophages in the lungs.

The most consistent cytologic feature of the BAL from FIV-infected cats was a neutrophilia at 16 weeks post-FIV infection. Similarly, increases in distinct cell populations including neutrophils, lymphocytes, and macrophages have also been reported in HIV-infected patients, in some cases, without altering the total cell counts (Agostini et al., 1988; Davis et al., 1993) . Because the pulmonary neutrophilia was not accompanied by either a systemic neutrophilia, clinical symptoms of illness, or bacterial contamination of the BAL¯uid, the neutrophilia is related to FIV infection and not a result of a secondary bacterial pneumonia. Also, as a pulmonary neutrophilia did not develop in the control cats, the neutrophilia observed in FIV-infected cats is not a result of the sequential BAL. A neutrophilia in the BAL population has been reported in cats during the asymptomatic stage (range 8 months to 4 years) of FIV infection (Hawkins et al., 1996) that is similar to the neutrophilia reported in later stages of HIV infection (Agostini et al., 1991b) . Cadore et al. (Cadore et al., 1997) also reported a neutrophilic alveolitis in cats naturally infected with FIV and suffering from a number of secondary complications. The mechanism that regulates the recruitment of neutrophils to the lung is most likely the release of macrophage-derived cytokines, perhaps in response to secondary infections. In humans, IL8 and GM±CSF have been found to be increased in patients infected with HIV-1 Denis and Ghadirian, 1994) . In this study, we did not speci®cally evaluate chemokine expression from AM; however, it would not be unreasonable to expect that the constitutive cytokine release exhibited by AM from FIV-infected cats (TNFa, IL6, IL10, IFNg) would also include IL8, GM±CSF, and other chemokines.

An important difference in the BAL cell populations of FIV-and HIV-infected subjects is the failure of FIV-infected cats to develop the pulmonary lymphocytosis that is characteristic of HIV infection (Agostini et al., 1988) . There was no difference in the absolute numbers of lymphocytes between FIV-infected and control cats, and lymphocytes never contributed to more than 4% to the total BAL cell population. Based upon absolute lymphocyte counts, 25% of patients with early HIV infection and 50% of patients with advanced disease show increased numbers of lymphocytes, particularly CD8 lymphocytes, in the lung (Agostini et al., 1991b) . These CD8 cells are thought to be involved in the host defense mechanism against HIV (Beck and Shellito, 1989) . While this study and previous studies have shown that FIV-infected cats develop a population of activated CD8 cells in the peripheral circulation (Jeng et al., 1996; Bucci et al., 1998 ), Hawkins et al. (1996 and the present study show that signi®cant numbers of lymphocytes are not recoverable from the pulmonary immune compartment by BAL.

As the absence of a``lymphocytic alveolitis'' in FIV-infected cats is not due to a failure in the development of activated CD8 lymphocytes in the circulation, one possible explanation is that there is a difference between HIV and FIV in recruitment of these cells to the lung. This could be related to the level of virus expression in the lung of cats versus the expression of viral antigens in the lung of HIV-infected patients. Although this study quantitated FIV-gag mRNA expression in puri®ed AM, determination of actual antigen expression and location within the lung was not performed. While Ma et al. (1995) identi®ed FIV mRNA by in situ hybridization in only a small fraction (1 in 10,000) of the total BAL cell population, studies to determine the extent of FIV antigens present in interstitial lung tissue have not been reported. An alternative possibility is that the CD8 pulmonary lymphocytosis characteristic of HIV-infected patients is a response to subclinical secondary pulmonary infections that would not occur in FIV-infected SPF cats maintained in isolation facilities. In support of this, Cadore et al. (1997) described an interstitial pneumonitis in cats naturally infected with FIV and suffering from a variety of other infectious disease complications. The pneumonitis was characterized by a neutrophilic and lymphocytic alveolitis. We also have recently observed that FIVinfected cats do have large numbers of activated CD8 cells in the circulation that are not observed in the lungs unless they are exposed to a secondary pulmonary pathogen. Following exposure of FIV-infected cats to T. gondii, there is a rapid in¯ux (1±3 days) of activated CD8 cells into the pulmonary compartment that is not seen in normal cats challenged with T. gondii (unpublished observations).

Constitutive elaboration of cytokines by AM from FIV-infected cats showed many similarities to their HIV-infected counterparts. Expression of IL6, IL10, and TNFa mRNA was initially upregulated during the acute stage infection and then decreased to near control levels by 16 weeks post-FIV infection. Elaboration of bioactive TNFa and IL6 proteins was still evident at 16 weeks post-infection. As this experiment was interrupted at 16 weeks post-FIV infection, it is not known if the low level of constitutive TNFa and IL6 activity would remain or eventually return to normal levels. However, Levy et al. (1998) reported elevated expression of TNFa mRNA in lymph node cells from FIV-infected cats as long as 12 months post-infection. Alveolar macrophages (AM) from clinically healthy (asymptomatic) or diseased (ARC, AIDS) HIV-infected patients produced constitutive TNFa bioactive protein throughout the course of HIV infection (Israel-Biet et al., 1991; Krishnan et al., 1990; Millar et al., 1991) . In contrast to the rise and fall of TNFa, IL6 and IL10 mRNA levels, IFNg mRNA expression by AM from FIVinfected cats continued to increase throughout the experiment.

For these experiments, plastic adherence was utilized for puri®cation of AM from other BAL cell populations (eosinophils, neutrophils, and lymphocytes). Whereas AM purity was con®rmed by differential analysis (at least 95%), there remains a possibility that a cell type other than the AM is responsible for the IFNg expression. Contaminating lymphocytes would seem to be the leading candidate; however, lymphocytes were only rarely, if ever, demonstrated on post-plastic adherence differentials. The predominant contaminating cell type were eosinophils. Eosinophils have been demonstrated to elaborate IFNg mRNA in some model systems (Woerly et al., 1999) ; however, the levels of IFNg mRNA demonstrated in the puri®ed AM culture were independent of the level (5% or less) of eosinophilic contamination. For this reason coupled with the precedent of macrophage derived IFNg in the literature (Munder et al., 1998; Nugent et al., 1985) , AM were considered the most-likely source of the IFNg mRNA. The increase in IFNg mRNA in AM is similar to that reported in the peripheral blood and lymph nodes of HIV-infected patients (Graziosi et al., 1994; Graziosi et al., 1996; Than et al., 1997) and FIV-infected cats (Levy et al., 1998; Dean and Pedersen, 1998) . It has yet to be determined if AM from HIV-infected patients constitutively produce IFNg. If subsequent studies show that the AM is in fact not the source of the IFNg mRNA, this should not diminish the fact that signi®cant IFNg mRNA levels were only detected in the FIV-infected cats. It is also possible that the increased cytokine production by AM from the FIV-infected cat was due to in vitro manipulation of previously sensitized cells. However, AM from normal cats were collected and cultured in vitro in parallel to those from the FIV-infected cats, supporting our conclusions that the elevated cytokine production in the latter group is due to FIV infection and not an artifact of culture. Alveolar macrophages (AM)from most FIV-infected cats lost their ability to produce IL12 mRNA by 16 weeks post-FIV infection. This is somewhat different from HIV infection, where AM exhibit low level constitutive production of IL12 (Denis and Ghadirian, 1994) . Despite the constitutive production of IL12, HIV-infected AM stimulated with S. aureus do not increase IL12 production as do control AM (Denis and Ghadirian, 1994) . Similarly, Levy et al. (1998) reported that FIV-infected cats cochallenged with T. gondii did not respond with increased IL12 mRNA expression in the lymph nodes, as did T. gondii challenged control cats. Therefore, quantitation of cytokine mRNA expression in the lung and lymph nodes of FIV-infected cats indicate that susceptibility to T. gondii may be due (at least in part) to an inability to generate an IL12dependent Th1 immune response in the lung. Interestingly, the number of cats in this study without detectable IL12 mRNA expression (9/12 or 75%) correlates well with the expected mortality rate (60±80%) previously reported for the FIV-NCSU 1 -T. gondii cochallenge model (Davidson et al., 1993; Yang et al., 1996) .

In vitro infection of AM with FIV also induced expression of TNFa and IL6. The expression of HIV in macrophages and lymphocytes is closely related to cellular activation and cytokine signaling (Kazazi et al., 1992; Poli et al., 1990a,b; Serpente et al., 1992) . In this study, FIV-gag expression that increased throughout the experiment did not correlate with TNFa or IL6 expression that peaked at 4 weeks post-FIV infection. Interestingly, IFNg was the only cytokine that was expressed with kinetics similar to the FIV-gag expression, suggesting that mechanisms regulating expression of TNFa and IL6 in AM are different from those regulating IFNg. Upregulation of cytokines was dependent upon virus replication in AM, as FIV infection of AM in vitro induced cytokine mRNA and protein synthesis, whereas exposure of AM to irradiated virus failed to induce either cytokine mRNA or protein. The progressive increase in virus burden in AM also differs markedly with transient acute stage (2±4 weeks) high virus burden in plasma and PBMC (Jeng et al., 1996) . This observation supports the conjecture that macrophages may be a major virus reservoir during the asymptomatic stage of infection.

The IL10 mRNA synthesis was also elevated in AM during the acute stage infection. However, similar to TNFa and IL6, peak IL10 expression did not correlate with AM virus burden. The continued increase in FIV-gag and IFNg mRNA in the presence of high levels of IL10 is somewhat unexpected, as IL10 is a potent inhibitor of HIV replication (Akridge et al., 1994; Weissman et al., 1994) , as well as macrophage cytokine expression (Gazzinelli et al., 1992; Murray et al., 1997) . However, when IL10 is present at levels that do not inhibit TNFa and IL6 synthesis, IL10 will actually augment TNFa and IL6 mediated HIV expression (Weissman et al., 1995) . The FIV replication may be similarly regulated; however, additional studies will be needed to resolve this question.

Concurrent upregulation of IFNg and IL10 is consistent with recent observations that both FIV and HIV causes upregulation of some Th1 (IFNg, TNFa), as well as Th2 (IL10) cytokines and a suppression of other Th1 (IL2, IL12) cytokines (Levy et al., 1998; Dean and Pedersen, 1998; Graziosi et al., 1994 Graziosi et al., , 1996 Than et al., 1997) . The data reported herein indicate that AM from FIV-infected cats have a similar constitutive cytokine pro®le as has been reported from lymph node cells and PBMC from FIV-infected cats and HIV-infected patients.

From the standpoint of immunoprotection from intracellular pathogens, levels of both protective (TNFa, IFNg) and nonprotective (IL10, IL6) cytokines were elevated. As both HIV patients and FIV-infected cats are susceptible to intracellular pathogens, this suggests that it is not simply the appearance of protective cytokines exclusive of nonprotective cytokines, but the relative levels of the cytokines, in both quantity and timing of appearance. In support of this, the concurrent increased IL10 and decreased IL12 mRNA expression in FIV-infected cats has been shown to be predictive of susceptibility of FIV-infected cats to T. gondii (Levy et al., 1998) . The FIV-T. gondii coinfection, thus, should provide a powerful model to dissect the role of cytokines in the pulmonary immunode®ciency in FIV-and HIV-infected patients.

Phenotypical and functional analysis of bronchoalveolar lavage lymphocytes in patients with HIV infection

Alveolar macrophages from patients with AIDS and AIDS-related complex constitutively synthesize and release tumor necrosis factor alpha

Prognostic signi®cance of the evaluation of bronchoalveolar lavage cell populations in patients with HIV-1 infection and pulmonary involvement

Release of granulocyte-macrophage colony-stimulating factor by alveolar macrophages in the lung of HIV-1-infected patients. A mechanism accounting for macrophage and neutrophil accumulation

IL-10 is induced during HIV-1 infection and is capable of decreasing viral replication in human macrophages

Effects of human immunode®ciency virus on pulmonary host defenses

Feline immunode®ciency virus: an interesting model for AIDS studies and an important cat pathogen

The CD8 phenotype mediating antiviral activity in FIV-infected cats is characterized by reduced surface expression of the CD8 beta chain

Interstitial lung disease in feline immunode®ciency virus (FIV) infected cats

Feline immunode®ciency virus predisposes cats to acute generalized toxoplasmosis

Update: HIV infection and pulmonary host defenses

Cytokine response in multiple lymphoid tissues during the primary phase of feline immunode®ciency virus infection

Dysregulation of interleukin 8, interleukin 10, and interleukin 12 release by alveolar macrophages from HIV type 1-infected subjects

In vivo lymphocyte tropism of feline immunode®ciency virus

Development of clinical disease in cats experimentally infected with feline immunode®ciency virus

IL-10 inhibits parasite killing and nitrogen oxide production by IFN-gamma-activated macrophages

Lack of evidence for the dichotomy of TH1 and TH2 predominance in HIV-infected individuals

Kinetics of cytokine expression during primary human immunode®ciency virus type 1 infection

Collection of bronchoalveolar lavage¯uid in cats using an endotracheal tube

Effect of FIV infection on lung in¯ammatory cell populations recovered by bronchoalveolar lavage

Tumor necrosis factor production in HIV-seropositive subjects. Relationship with lung opportunistic infections and HIV expression in alveolar macrophages

Evidence for CD8 antiviral activity in cats infected with feline immunode®ciency virus

Recombinant interleukin 4 stimulates human immunode®ciency virus production by infected monocytes and macrophages

Alveolar macrophages in AIDS patients: increased spontaneous tumour necrosis factor-alpha production in Pneumocystis carinii pneumonia

Tumor necrosis factor can induce both apoptic and necrotic forms of cell lysis

Elevated interleukin-10-to-interleukin-12 ratio in feline immunode®ciency virus-infected cats predicts loss of type 1 immunity to Toxoplasma gondii

TNF-alpha responses are depressed and IL-6 responses unaltered in feline immunode®ciency virus infected cats

HIV infection of the lung. Role of virus-infected macrophages in the pathophysiology of pulmonary disease

Production of tumor necrosis factor-alpha by blood and lung mononuclear phagocytes from patients with human immunode-®ciency virus-related lung disease

Murine macrophages secrete interferon gamma upon combined stimulation with interleukin (IL)-12 and IL-18: a novel pathway of autocrine macrophage activation

T cell-derived IL-10 antagonizes macrophage function in mycobacterial infection

A macrophage-derived factor required by plasmacytomas for survival and proliferation in vitro

Stimulated human alveolar macrophages secrete interferon

Interleukin 6 induces human immunode®ciency virus expression in infected monocytic cells alone and in synergy with tumor necrosis factor alpha by transcriptional and post-transcriptional mechanisms

Tumor necrosis factor alpha functions in an autocrine manner in the induction of human immunode®ciency virus expression

Acquired immune de®ciency syndrome and the lung

A reverse transcription-polymerase chain reaction technique to detect feline cytokine genes

A reverse transcription-quantitative competitive polymerase chain reaction (RT-qcPCR) technique to measure cytokine gene expression in domestic mammals

Suboptimal and optimal activation signals modulate differently the expression of HIV-1 and cytokine genes

Cytokine pattern in relation to disease progression in human immunode®ciency virus-infected children

Early events in the immunopathogenesis of feline retrovirus

Spontaneous production of interleukin-6 by alveolar macrophages from human immunode®ciency virus type 1-infected patients

Role of cytokines in alveolar macrophage accessory cell function in HIV-infected individuals

Interleukin 10 blocks HIV replication in macrophages by inhibiting the autocrine loop of tumor necrosis factor alpha and interleukin 6 induction of the virus

IL-10 synergizes with multiple cytokines in enhancing HIV cells of monocytic lineage

Expression of CD28 and CD86 by human eosinophils and role in the secretion of type 1 cytokines (interleukin 2 and interferon gamma) inhibition by immunoglobulin complexes

Molecularly cloned feline immunode®ciency virus NCSU1 JSY3 induces immunode®ciency in speci®c-pathogen-free cats

This work was supported by Public Health Service Grants AI-38177, AI-01458, AI-01227, and RR-0124. We thank Debra Anderson, Janet Dow, Dale Brown, Joe Bucci, Doug Gebhard, and Tedd Childers for excellent technical assistance.