key: cord-0001677-iqflllk3
authors: Coon, Tiffany A.; McKelvey, Alison C.; Weathington, Nate M.; Birru, Rahel L.; Lear, Travis; Leikauf, George D.; Chen, Bill B.
title: Novel PDE4 Inhibitors Derived from Chinese Medicine Forsythia
date: 2014-12-30
journal: PLoS One
DOI: 10.1371/journal.pone.0115937
sha: 5d7ef1b6d6d20f3a2d0c1cad4daa1496ef8a195b
doc_id: 1677
cord_uid: iqflllk3

Cyclic adenosine monophosphate (cAMP) is a crucial intracellular second messenger molecule that converts extracellular molecules to intracellular signal transduction pathways generating cell- and stimulus-specific effects. Importantly, specific phosphodiesterase (PDE) subtypes control the amplitude and duration of cAMP-induced physiological processes and are therefore a prominent pharmacological target currently used in a variety of fields. Here we tested the extracts from traditional Chinese medicine, Forsythia suspense seeds, which have been used for more than 2000 years to relieve respiratory symptoms. Using structural-functional analysis we found its major lignin, Forsynthin, acted as an immunosuppressant by inhibiting PDE4 in inflammatory and immune cell. Moreover, several novel, selective small molecule derivatives of Forsythin were tested in vitro and in murine models of viral and bacterial pneumonia, sepsis and cytokine-driven systemic inflammation. Thus, pharmacological targeting of PDE4 may be a promising strategy for immune-related disorders characterized by amplified host inflammatory response.

Phosphodiesterase (PDE) is an enzyme that catalyzes the hydrolysis of a phosphodiester bond, most notably those of the second messenger cascade molecules cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP). cAMP and cGMP bind to the regulatory units of protein kinase A(PKA) allowing for phosphorylation thus transducing signal cascades in the cell [1] [2] [3] . Consequently, by severing these cA/GMP-dependent pathways, PDE has an enormous clinical significance monitoring cAMP and cGMP levels in cells. The PDE superfamily comprises 11 subfamilies, PDE1-11, respectively, each utilizing different substrate specificities. For example, PDE4, 7, and 8 strictly hydrolyze cAMP; PDE5, 6, and 9, on the other hand, only hydrolyze cGMP while the other family members (PDE1, 2, 3, 10, 11) can target both cAMP and cGMP for hydrolysation [4] . For many years, PDE enzymes have remained an interest within the pharmaceutical industry, as inhibition of PDE can increase the levels of cAMP or cGMP thus enhancing or prolonging their natural physiological effects [5] .

Currently, various small molecular compounds have been discovered to inhibit PDEs. For example, caffeine, aminophylline, and IBMX are nonselective PDE inhibitors [6] that increase intracellular cAMP, thereby activating PKA, thus inhibiting tumor necrosis factor (TNF) and other inflammatory cytokines, and reducing inflammation [7] . There are also selective PDE inhibitors. Specifically, PDE3 inhibitors such as inamrinone and milrinone, are used for short-term treatment of congestive heart failure [8] , and Cilostazol used in the treatment of intermittent claudication [9] . The PDE5 inhibitors Sildenafil, tadalafil and vardenafil boost cGMP levels in penile tissue [10] [11] [12] and are used primarily for erectile dysfunction, as well as having secondary indication in treatment of pulmonary hypertension [13, 14] . Recently , an opium alkaloid, Papaverine, was shown to inhibit PDE10 [15] . Among all of the PDE isoforms, PDE4 is the major cAMP-degrading enzyme found in inflammatory and immune cells. Selectively inhibiting PDE4 prevents the release of cytokines and other inflammatory factors and hinders the production of reactive oxygen species [16, 17] .

As the leading causes of infectious deaths in the US, sepsis and pneumonia are pathognomonically linked to a burst in cytokine release, i.e. cytokine storm, from pro-inflammatory cells including macrophages, lymphocytes, and polymorphonuclear leukocytes [18] [19] . Cytokine storm occurs in response to infection with virulent pathogens, host cell injury, or irritants that activate a multitude of receptors on immune effector cells. Under some conditions it is exaggerated (hypercytokinemia) and results in a fatal immune reaction with constant activation of immune effector cells that produce sustained or supraphysiologic levels of tumor necrosis factor (TNF), interleukin 1 (IL1), and interleukin-6 (IL6) that results in severe tissue injury and often death. Thus, selectively inhibiting PDE4 may ameliorate cytokine storm and prevent tissue injury.

The earliest Pharmacopoeia in the world, the ''Tang Ben Cao'', completed in 659 AD is the first known documentation of the Forsythia suspensa plant being used to treat fever, flu-like, and inflammatory symptoms. Since then, it has been habitually and widely used in Asia as one of the major traditional medicines. Nowadays, the extract of Forsythia suspensa seeds is used to treat numerous inflammatory diseases including but not limited to erysipelas, inflammation, pharyngitis, pyrexia, tonsillitis, and ulcers [20, 21] . Moreover, it has been shown in studies that the crude extract displays potential antibacterial, antiviral, choleretic and antipyretic effects [22] . Medicinal chemistry studies on Forsythia seeds have revealed phenolic compounds including lignans and flavonols which are suspected to be responsible for the various biological activities of the herb [22, 23] . In this study, we sought to better understand Forsythia lignan Forsythin structure-function relationships and it was used to design and test a new class of selective PDE4 inhibitors. These selective PDE4 inhibitors exert potent antiinflammatory activity in several murine models of inflammation.

Forsythin is one of the major bioactive compounds extracted from Forsythia suspensa seeds (Fig. 1A-B) (27) . An o-linked b-D-glucopyranosylated lignan, Forsythin readily can be hydrolyzed to remove a glucose moiety (Fig. 1C) . Using the structure of the predicted metabolite, we assessed the virtual docking with PDE4D (based on PyMol PDB 1 mkd.phg) using LibDock (Discovery Studio, Accelrys Software Inc./BIOVIA, San Diego, CA) (Fig. 1D ,E). Using ZINCPharmer [24] , we then generated a pharmacophore model ( Fig. 1E top left) that enabled the identification of over 100 high value targets from a 18.3 million purchasable compound library. We refined this search by performing a computer based docking analysis of the target compounds using LibDock (Discovery Studio 3.5) and selected 14 compounds (Fig. 2) for further testing using an in vitro phosphodiesterease activity assay.

Compounds 1-14 ( Fig. 2) were tested in vitro using PDE-glo phosphodiesterase assay (Promega, Madison, WI). Briefly, all the compounds were diluted in DMSO, and the final concentrations in each assay were 100, 10, 1, 0.1, 0.01, 0.001, 0.0001 mM. For the PDE4 activity assay, 10 mU of purified PDE4D (Millipore) was used per reaction. Compound 7, which is Forsythin, exhibited [IC 50 ] PDE458 mM. Other custom designed compounds such as 6, 9 and 13 exhibited much greater potency with an [IC 50 ] towards PDE4 ranging from 10-90 nM. Impressively, all of the test compounds exhibited selectively toward PDE4, in as much as the [IC 50 ].80 mM with PDE3, 5, 7, and 10 ( Fig. 3 ).

Tested compounds reduce TNF secretion in LPS stimulated mouse RAW264.7 and human PBMCs greatest potency towards PDE4 with [IC50]510 nM (Fig. 3) , it also exhibited comparable potency towards [IC 50 ] TNF ranging from 25-75 nM (Fig. 4C ).

To assess any in vivo anti-inflammatory activity, four compounds [C2, C6, C7 and C9, 1 mg/kg, intraperitoneal (i.p.) injection] were administered in a LPS-induced pneumonia model. Briefly, compounds were administered to mice at various doses through an i.p. injection 10 min after mice were given LPS (i.t. 3 mg/kg). Mice were then euthanized 16 h later using pentobarbital; BAL was collected and assayed for IL1, IL6 and TNF cytokine levels. These compounds varied in their ability to ameliorate lavage cytokines, protein concentration, and total cell counts from LPS-treated mice ( Fig. 5A-E ). In addition, these inhibitors diminished histological evidence of lung injury (Fig. 5F ).

As a widely used model of sepsis [25] , selective PDE4 inhibitors (C6, 7 and 9) were administered to mice at various doses through an i.p. injection, and 10 min later mice were given LPS (E. coli, 100 ug i.p.). Mice were euthanized 2 h later using pentobarbital where blood was collected and assayed for IL6 and TNF cytokine levels. All inhibitors exhibited high potency in vivo (inhibitory dose [ID 50 ] TNF50.01,1 mg/kg, ID 50 IL-650.1,1 mg/kg,) ( Fig. 6A-B) . In a separate 

To further test these compounds, mice were challenged with H1N1 (105 pfu/mose i.t.) without (vehicle control) or with C6, C7, C9, or C13 treatment (30 mg/ml added to drinking water containing 2% sucrose: estimated dose55 mg/kg/d). All four PDE4 inhibitors significantly improved the survival of mice infected with H1N1 compared with mice receiving diluent (Fig. 7A) . The PDE inhibitors also decreased lavage cell counts, protein concentrations, and lavage cytokine IL-6 and TNF (Fig. 7B-E) . Impressively, PDE4 inhibitors also significantly reduced serum cytokines induced from the H1N1-induced systematic inflammation (Fig. 7F ,G). The tested compounds were also able to lessen signs of tissue injury (Fig. 7H) . These results suggest that the selective PDE4 inhibitors suppress inflammation and preserve lung homeostasis after pulmonary H1N1 infection. 

It has been known for years that increased cAMP from PDE4 inhibition or gene knockout were able to markedly decrease inflammatory responses in immune cells [26] . cAMP controls inflammation through several distinct pathways, including cyclic nucleotide-gated ion channels, cAMP-activated protein kinases (PKA), or exchange proteins directly activated by cAMP (Epac) [27] . Specifically, cAMP actives PKA, negatively regulates NF-KB pathway, thus suppressing the release of proinflammatory mediators (eg, TNF-a, interleukin-17, Interferon-c) [28] . Moreover, cAMP actives transcriptions of anti-inflammatory gene such as interleukin-10 through transcription factor CREB (cAMP response elementbinding protein) [29] . Thus, pharmacological targeting of PDE4 may be a promising strategy for anti-inflammatory therapy.

Here we developed a novel class of selective PDE4 inhibitors that exert robust anti-inflammatory activity by impairing cytokine release in three preclinical models. The structural basis for our drug design emerged from Forsythin, the one of major compound extracted from Forsythia suspensa seeds. Among the 14 compounds tested in this study, synthetic compounds 6, 9 and 13 showed the potent IC 50 (10-90 nM) toward PDE4. Impressively, these compounds also exhibit high selectivity toward PDE4, with .80 mM IC 50 towards other PDE (Fig. 3 ). Compounds 6, 9 and 13 also showed potent IC 50 (20-450 nM) toward TNF inhibition in LPS stimulated RAW274.7 and PBMC cells. Many natural compounds, such as the flavonoid derivatives Hesperetin and prunetin, were also isolated from traditional Chinese medicines are selective inhibitors of PDE4 with an IC 50 ranging from 10-100 mM [30] . However, little is known about the structure-activity relationship between such lignans or their inhibitory effect on proteins in the inflammatory pathway. In this study, we determined that olinked b-D-glucopyranosided lignan Forsythin is a weak but selective inhibitor of PDE4. Structurally speaking, compound 6 and Forsythin are highly similar except that Forsythin is an o-linked b-D-glucopyranoside. In vitro, Forsythin is .80 times weaker in activity toward PDE4 compared with compound 6. However, our in vivo study suggested that Forsythin and compound 6 exhibit similar activities in several murine models (Fig. 5A, 6C, 7B ). This suggests that Forsythin may act as a prodrug that its glucopyranoside group will be potentially removed in vivo by b-glucosidase, and release the active lignin; which is a common of metabolic reaction in plant lignans and flavonols [31, 32] .

Pandemic H5N1 Influenza virus and SARS virus are known to cause dangerous cytokine storm resulting in severe damage to alveoli and lung tissue and thus Fig. 6 . PDE4 inhibitors lessen cytokine storm induced by LPS septic shock. C57/BL6 mice were administered i.p. nothing (CON), vehicle, 10 ug/kg, 100 ug/kg, 1 mg/kg or 10 mg/kg of compounds 6, 7 or 9. Mice were given LPS (E. coli, 100 mg) 10 min later through an i.p. injection. 2 h later mice were euthanized using pentobarbital and blood was collected for IL-6 and TNF measurements. Shown in panel A-B is the % inhibition of cytokine levels as a function of drug dose. The data represent n53 mice/group at each dose. C57/BL6 mice were also pretreated with compound 5, 6, 7 or 9 at 1 mg/kg. Mice were given LPS (E. coli, 100 mg) 18 h later through an i.p. injection. 2 h later the mice were euthanized using pentobarbital and blood was collected for IL-6 and TNF measurements (C-D). The data represent n54-6 mice/group, *P,0.05 versus vehicle. extremely high mortality rates [33, 34] . Today, viral pneumonia affects more than 100 million children and 100 million adults per year yet there are no effective treatments only supportive care [35] . Our H1N1 murine model recapitulates well with the deadly human viral pneumonia, as H5N1 Influenza infection is largely driven by an exuberant host response and amplification of cytokines and chemokines that contribute to influenza-induced morbidity and mortality [25, 36, 37] . Our pulmonary viral infection model was complemented with a murine sepsis model to provide assessment of PDE4 inhibitors on systemic inflammation by the endotoxin LPS [37, 38] . Importantly, in each of these models of inflammation, our PDE4 inhibitors (Compound 6, 7, 9, 13) were observed to lessen severity of tissue injury using varying modes of application (parenteral, oral), and none of the animals treated with the agents exhibited overt signs of distress. The dosing of these compounds also appeared to be excellent in vivo, with ,1 mg/kg efficacious dose in both LPS pneumonia and sepsis model. Specifically, these compounds also appeared to have long half-life in vivo. As shown in Fig. 6C -D, the septic shock model was carried out 18 h after the initial treatment (1 mg/kg, i.p.), yet we still observed a 40-60% decrease in serum cytokines. Overall, our pre-clinical studies demonstrate the biological efficacy of selective PDE4 inhibitors in murine models where both infectious and irritant factors trigger cytokine release. These selective PDE4 inhibitors could potentially have broad applications in acute and chronic inflammatory diseases. In fact the selective PDE4 inhibitor, Roflumilast, recently became the first FDA approved PDE4 inhibitor on the market [39, 40] . Roflumilast is a potent anti-inflammatory drug, particularly in pulmonary diseases such as asthma, COPD, and rhinitis. Thus, PDE4 is a potentially important drug target for the treatment of inflammatory diseases. However, further research and development clearly is needed to carefully ascertain the safety profile, distribution, elimination, and metabolism of these new chemical entities in larger models of inflammation. Successful results from these pharmacokinetic studies will set the stage for transition to clinical testing in subjects with acute and chronic immune-related illness.

RAW 264.7 cells were from ATCC. PBMCs were from Sanguine Life Sciences. PDE4 activity assay kit was from Promega. IL1b, TNFa, IL6 mouse ELISA kit, human TNFa were from R&D systems. LPS (E.coli) were from Sigma. Forsythin was from Stanford Chemicals; all other synthetic compounds were purchased from ChemDiv, Inc. and Vitas-M Laboratory, Ltd. All compounds are .98% pure by HPLC.

All animal experiments were approved by the University of Pittsburgh Institutional Animal Care and Use Committee (IACUC) under protocol (14023127). Mice were housed at University of Pittsburgh Animal Care Facility and maintained according to all federal and institutional animal care guidelines.

All mice were euthanized using i.p. injection of 100 mg/kg pentobarbital approved by University of Pittsburgh Institutional Animal Care and Use Committee (IACUC) under protocol (14023127).

Male C57LB/6 mice (purchased from Jackson Laboratories) were acclimated at the University of Pittsburgh Animal Care Facility and maintained according to all federal and institutional animal care guidelines and under a University of Pittsburgh Institutional Animal Care and Use Committee (IACUC)-approved protocol (14023127). Male C57LB/6 Mice were deeply anesthetized with ketamine (80 to 100 mg/kg of body weight, intraperitoneally [i.p.]) and xylazine (10 mg/kg, i.p.). Compounds were diluted in PBS and various amounts (0.01, 0.1, 1 and 10 mg/kg) of compounds were administered to mice though an i.p. injection. 10 min later, mice were given 100 mg of LPS (E. coli) through an i.p. injection. 2 h later, mice were euthanized. Plasma was collected and processed for cytokine assays.

Male C57LB/6 mice (purchased from Jackson Laboratories) were acclimated at the University of Pittsburgh Animal Care Facility and maintained according to all federal and institutional animal care guidelines and under a University of Pittsburgh Institutional Animal Care and Use Committee (IACUC)-approved protocol (14023127). Male C57LB/6 Mice were deeply anesthetized with ketamine (80 to 100 mg/kg of body weight, intraperitoneally [i.p.]) and xylazine (10 mg/kg, i.p.), and then the larynx was well visualized under a fiber optic light source before endotracheal intubation with a 3/400 24-gauge plastic catheter. Compounds were diluted in PBS and administered to mice at 1 mg/kg through an i.p. injection 10 min after mice were given with LPS (i.t. 3 mg/kg). Mice were then euthanized 18 h later; Lavage fluids were collected from mice to measure protein and cytokine concentration. Lungs were isolated from mice for H&E staining.

Male C57LB/6 mice (purchased from Jackson Laboratories) were acclimated at the University of Pittsburgh Animal Care Facility and maintained according to all federal and institutional animal care guidelines and under a University of Pittsburgh Institutional Animal Care and Use Committee (IACUC)-approved protocol (14023127). Male C57LB/6 Mice were deeply anesthetized with ketamine (80 to 100 mg/kg of body weight, intraperitoneally [i.p.]) and xylazine (10 mg/kg, i.p.), and then the larynx was well visualized under a fiber optic light source before endotracheal intubation with a 3/400 24-gauge plastic catheter. 50 ul of H1N1 (A/ PR/8/34, 10 5 pfu/mouse) was instilled i.t. and mice monitored for up to 6 d. For compound treatment, a stock solution (5 mg/ml) was added to the drinking water (containing 2% sucrose) to a final concentration of 30 mg/ml. For survival studies, mice were carefully monitored and weighted over time; mice that lost .20% weight and moribund, preterminal mice were immediately euthanized and recorded as deceased. Lavage fluids were collected from mice to measure protein and cytokine concentration. Lungs were isolated from mice for H&E staining.

In silico docking study 

A PDE activity assay was carried out using the PDE-Glo Phosphodiesterase Assay per manufacturer's instruction (Promega). Briefly, 5 ul of 1X PDE-Glo Reaction buffer containing 10 mU of purified human recombinant PDE4A4 (Millipore) was added to 96 well plate. All test compounds were dissolved in DMSO and a serial dilution (110) of the inhibitors were performed using 1X PDE-Glo Reaction buffer. 7.5 ul of test compounds were added to the PDE enzyme with the final concentrations in each assay of 100, 10, 1, 0.1, 0.01, 0.001, 0.0001 mM, and incubated at room temperature for 10 min. To initiate the PDE reaction, 12.5 ul of 2 mM cAMP was added to to each reaction, mixed well and incubate at room temperature for 10 min. The reactions were terminated using 12.5 ul of PDE-Glo Termination Buffer. 12.5 ul of PDE-Glo Detection Solution containing PDE-Glo Detection Buffer and Protein Kinase A was added to each reaction. The reactions were carried at room temperature for 20 min. 50 ul of Kinase-Glo Reagent was then added to each reaction, mixed well, and incubate for 10 minutes at room temperature. The chemical luminescence signals were measured, quantified and graphed. Compounds 6, 7, 9 and 13 were further tested similarly using 25 mU purified PDE3, 5, 7, 10 (Millipore) enzyme.

LPS induced cytokine release in RAW264.7 and PBMC cells 5610 5 RAW264.7 cells were seeded in 96 wells for 18 h. Cells were primed with compounds at different concentration for 3 h before treated with LPS (1 ng/ml) for additional 8 h. TNF cytokine releases were monitored by ELISA. PBMC (0.2 ml at 1610 5 /ml) were primed with compounds at different concentration for 3 h before treated with LPS (1 ng/ml) for additional 8 h. TNF cytokine releases were monitored by ELISA. % of TNF secretion were calculated and graphed.

Statistical comparisons were performed with the Prism program, version 4.03 (GraphPad Software, Inc., San Diego, CA) using an ANOVA 1 or an unpaired 2 ttest with p,0.05 indicative of significance. 

Phosphodiesterase 4B mediates extracellular signal-regulated kinase-dependent up-regulation of mucin MUC5AC protein by Streptococcus pneumoniae by inhibiting cAMP-protein kinase A-dependent MKP-1 phosphatase pathway

Cyclic nucleotide phosphodiesterases: functional implications of multiple isoforms

Cyclic nucleotide research -still expanding after half a century

Structural determinants for inhibitor specificity and selectivity in PDE2A using the wheat germ in vitro translation system

Phosphodiesterase: overview of protein structures, potential therapeutic applications and recent progress in drug development

Cyclic nucleotide phosphodiesterases

Insights into the regulation of TNFalpha production in human mononuclear cells: the effects of non-specific phosphodiesterase inhibition

Dual expression and differential regulation of phosphodiesterase 3A and phosphodiesterase 3B in human vascular smooth muscle: implications for phosphodiesterase 3 inhibition in human cardiovascular tissues

Antiplatelet therapy using cilostazol, a specific PDE3 inhibitor

Structure of the catalytic domain of human phosphodiesterase 5 with bound drug molecules

The future of the oral pharmacotherapy of male erectile dysfunction: things to come

Comparative relaxing effects of sildenafil, vardenafil, and tadalafil in human corpus cavernosum: contribution of endogenous nitric oxide release

Phosphodiesterase 5 inhibition in essential hypertension

Improvement of exercise capacity in monocrotaline-induced pulmonary hypertension by the phosphodiesterase-5 inhibitor Vardenafil

Discovery of novel PDE10 inhibitors by a robust homogeneous screening assay

The effect of oxidative stress on macrophages and lung epithelial cells: the role of phosphodiesterases 1 and 4

Role of PDE4 in superoxide anion generation through p44/42MAPK regulation: a cAMP and a PKA-independent mechanism

The influence of infection sites on development and mortality of ARDS

Historical insights into cytokines

Preparative isolation and purification of phillyrin from the medicinal plant Forsythia suspensa by high-speed counter-current chromatography

Antioxidant and antibacterial activity of two compounds (forsythiaside and forsythin) isolated from Forsythia suspensa

Lignans from the fruits of Forsythia suspensa

Lignans from the Fruits of Forsythia suspensa (Thunb.) Vahl Protect High-Density Lipoprotein during Oxidative Stress

ZINCPharmer: pharmacophore search of the ZINC database

Targeting F box protein Fbxo3 to control cytokine-driven inflammation

Cyclic AMP: master regulator of innate immune cell function

Elevated cyclic AMP and PDE4 inhibition induce chemokine expression in human monocyte-derived macrophages

Highly selective phosphodiesterase 4 inhibitors for the treatment of allergic skin diseases and psoriasis

Evidence for a role of phosphodiesterase 4 in lipopolysaccharide-stimulated prostaglandin E2 production and matrix metalloproteinase-9 activity in human amniochorionic membranes

Inhibitory effects of flavonoids on phosphodiesterase isozymes from guinea pig and their structure-activity relationships

Absorption and metabolism of cyanidin-3-glucoside and cyanidin-3-rutinoside extracted from wild mulberry (Morus nigra L.) in rats

Absorption and metabolism of cyanidin 3-O-beta-D-glucoside in rats

Preparing for the next pandemic

Lung pathology of fatal severe acute respiratory syndrome

Viral pneumonia

Innate immune response of human alveolar macrophages during influenza A infection

A combinatorial F box protein directed pathway controls TRAF adaptor stability to regulate inflammation

Different subsets of enteric bacteria induce and perpetuate experimental colitis in rats and mice

Phosphodiesterase-4 inhibitor therapy for lung diseases

Roflumilast: a selective phosphodiesterase 4 inhibitor