Elevated Cyclic AMP and PDE4 Inhibition Induce Chemokine Expression in Human Monocyte-Derived Macrophages

Home , Cilostamide, Piclamilast, Roflumilast

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

Angie L. Hertza,1, Andrew T. Bendera,1, Kimberly C. Smitha, Mark Gilchristb, Paul S. Amieuxa, Alan Aderemb, and Joseph A. Beavoa,2

aDepartment of Pharmacology, University of Washington Medical School, Seattle, WA 98195; and bInstitute for Systems Biology, Seattle, WA 98103

Contributed by Joseph A. Beavo, October 14, 2009 (sent for review April 14, 2009) Macrophages are central mediators of the innate immune system where they are exposed to a distinct milieu of cytokines and that can be differentiated from monocytes upon exposure to growth factors that affect their differentiation and activation cytokines. While increased cyclic adenosine monophosphate states (15–17). For example, monocytes differentiated in the (cAMP) levels are known to inhibit many lipopolysaccharide-elic- presence of macrophage-colony stimulating factor (CSF) carry a ited macrophage inﬂammatory responses, the effects of elevated much different PDE profile than those differentiated in granu- cAMP on monocyte/macrophage differentiation are not as well locyte macrophage-colony stimulating factor (GM-CSF) (18). In understood. We show here that during differentiation, cAMP the lung, infiltrating monocytes are exposed to high levels of agonists can cause a large increase in the mRNA and protein levels GM-CSF, causing them to differentiate into an alveolar-type of several of the pro-inﬂammatory CXCL and CCL chemokines. The macrophage, characterized by low levels of CD11b and high cAMP mediator-exchange protein activated by cAMP (Epac) con- levels of IL-10 and nitric oxide (19, 20). In the inflamed lung, as tributes substantially to the increase in these chemokines. These seen with disorders such as asthma or chronic obstructive chemokines are known to play an important role in the regulation pulmonary disease (COPD), the levels of cAMP-increasing of immune responses, particularly regarding the pathogenesis of agonists like PGE2, histamine, and adenosine are high (21–23).

asthma and chronic obstructive pulmonary disorder. We also found Administration of a PDE4 inhibitor, such as those now in clinical PHARMACOLOGY that a selective cAMP-degrading phosphodiesterase (PDE) 4 inhib- trials for asthma and COPD treatment, can drive intracellular itor can potentiate the chemokine expression elicited by low-dose cAMP levels even higher, potentially altering both the activation forskolin or Prostaglandin E2 (PGE2). These data suggest that and differentiation states of monocytes and macrophages. How- chemokine receptor antagonists administered in conjunction with ever, the efficacy of PDE4 inhibition in these trials has been a PDE4 inhibitor may improve both the efficacy and safety of limited by concerns about the doses that can be safely admin- PDE4-inhibitor therapy for chronic inflammatory disorders. istered (24). In the present study we determined the effects of cAMP on inflammation ͉ phosphodiesterase monocyte to macrophage differentiation, in the absence of macrophage-activating factors like LPS, and found that several yclic nucleotides are ubiquitous second messengers that can proinflammatory chemokine mRNAs and protein levels were Cdifferentially regulate multiple cell processes via their var- increased in a PDE4-dependent manner. Further analysis ious downstream effectors. cAMP can exert its effects through showed that this increase in chemokine expression depended on any of several distinct pathways, including cyclic nucleotide- both Epac-mediated cAMP-signaling microdomains that re- gated ion channels, cAMP-activated protein kinases (PKA), or spond to PDE4, as well as NF-␬B and ATF3-dependent tran- exchange proteins directly activated by cAMP (Epac) (1, 2). scription factor mechanisms. Phosphodiesterases (PDEs) control the amplitude and duration of the cAMP signal through hydrolysis to 5Ј-AMP, and so can Results regulate all of these cellular processes through control of the DNA Microarray Analysis of Forskolin-Treated Monocyte/Macro- availability of this second messenger (3). Current studies in the phages. We wished to evaluate the role of elevated cAMP in the field indicate that there are distinct cAMP-signaling microdo- GM-CSF-induced macrophage to determine its potential rele- mains within many cells that are controlled by specific PDEs, vance to current PDE4-inhibitor treatments for inflammatory which can differentially regulate independent processes con- disorders. To look at changes in gene expression increased by trolled by this common second messenger (4–7). cAMP, we differentiated monocytes for 6 days with GM-CSF (50 cAMP plays a key role in regulation of activated macrophage ng/ml) in the presence or absence of 50 ␮M forskolin and inflammatory responses. Increases of cAMP through either PDE4 performed a DNA microarray analysis. The two conditions were inhibition, PDE4 gene knockout, or cAMP-agonist administration compared using a large-oligo human gene array as described in have been widely shown to dramatically decrease much of the Materials and Methods (see also SI Materials and Methods). Of inflammatory response of macrophages to activating agents such as lipopolysaccharide (LPS), including cytokine expression, adhesion, and Fc␥R-mediated phagocytosis (8, 9). This recognition has led to Author contributions: A.L.H., A.T.B., M.G., P.S.A., A.A., and J.A.B. designed research; A.L.H., A.T.B., K.C.S., and M.G. performed research; A.A. contributed new reagents/analytic tools; a great interest in PDE4 inhibitors as therapeutic agents (10). A.L.H., K.C.S., M.G., P.S.A., and J.A.B. analyzed data; and A.L.H. and J.A.B. wrote the paper. Previous work in this laboratory has shown that both cAMP and The authors declare no conflict of interest. cGMP can arrest differentiation from a monocyte to dendritic cell Data deposition: The data reported in this paper have been deposited in the Gene (11). While some work has been carried out on the effects of cAMP Expression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo (accession no. GSE and PDEs during monocytic differentiation to macrophages (12– 18654). 14), the roles of these regulators in the long-term macrophage 1A.L.H. and A.T.B. contributed equally to this work. responses of the body remain insufficiently answered. 2To whom correspondence should be addressed. E-mail: [email protected]. Monocytes provide a first line of defense against infection. As This article contains supporting information online at www.pnas.org/cgi/content/full/ such, they constantly circulate in the blood and enter tissues, 0911684106/DCSupplemental.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0911684106 PNAS Early Edition ͉ 1of6 Downloaded by guest on September 28, 2021 Table 1. Genes with chemokine function showing 10-fold or greater change in mRNA expression Gene name Fold increase P value Common name and receptor

CXCL5 106.0 0.00537 ENA-78, binds CXCR2 CXCL7 52.7 0.00030 PPBP, binds CXCR2 CXCL2 18.0 0.00011 GRO-␤, binds CXCR2 CCL13 16.9 0.00021 MCP-4, binds CCR2 CXCL3 16.9 0.00021 GRO-␥, binds CXCR2 CCL2 15.1 0.00015 MCP-1, binds CCR2 CXCL1 12.1 0.00038 GRO-␣, binds CXCR2 CCL18 11.9 0.0197 PARC, unknown receptor

DNA microarray analysis was performed on ﬁve separate donors in duplicate. Fold-change is reported as change with 50 ␮M forskolin treatment over control cells. P value indicates how over-represented each gene is in relation to the whole genome.

the 8,530 genes that were reliably detected (seen in 8 of 10 -4 as the major PDEs controlling cAMP degradation in these samples), a total of 334 genes had expression increased greater macrophages (see Fig. S1). Therefore, we treated differentiating than 2-fold, while 356 genes had expression decreased greater monocytes with combinations of PDE3- or PDE4-specific in- than 2-fold upon treatment with forskolin (see Table S1). hibitors or the nonselective PDE inhibitor, 3-isobutyl-1- One class of genes in particular stood out because of their very methylxanthine (IBMX) and forskolin to determine which spe- large and significant fold increases. This group, listed in Table 1, cific functional compartments of cAMP and their associated contained a number of chemokine gene products that primarily PDEs were important for controlling expression of these genes. bind to either of two distinct chemokine receptors, CXCR2 or The selective PDE inhibitors used were cilostamide (PDE3) and CCR2. These chemokine receptors largely regulate recruitment rolipram (PDE4) (3). In the presence of a low dose of forskolin, of neutrophils and monocytes, respectively, and their production a high, but selective, dose of PDE inhibitor (a dose 10-times would be expected to increase the number of these leukocytes larger than the EC50) should have the effect of shifting the trafficking to an area of inflammation. Additionally, several of dose-response curve to the left. One expects that when a specific these chemokines, including CCL18 and CCL13, also attract PDE is inhibited, there should be an increase in cAMP levels in eosinophils and T cells. These large increases in chemokine the compartments to which it is localized, resulting in a larger production induced with forskolin treatment were particularly change in gene expression to the same low dose of agonist. unexpected, because cAMP is widely known to have anti- We first looked at expression of a number of surface markers to inflammatory effects on cytokine and chemokine production by determine the macrophage phenotype (Fig. S2). We found two activated macrophages. As such, we decided to focus on the surface markers that were up-regulated with forskolin treatment, CXCR2 and CCR2 classes of chemokines for subsequent exper- CD14 and CD163. By carrying out dose-response curves on these iments. cells, we found that low doses of forskolin were 5 to 10 ␮M(Fig. S3). Treatment with 5 ␮M forskolin showed a slight increase in expres- PDE4 Controls Surface Marker and Chemokine Expression. Using sion of these surface markers after 6 days (Fig. 1A). When the isozyme-selective PDE activity analysis, we identified PDE3 and monocytes were treated with forskolin plus PDE-specific inhibitors,

Fig. 1. PDE control of gene expression. (A) Cells were treated for 6 days as indicated and analyzed by ﬂow cytometry for surface marker expression. Cil, Cilostamide (1 ␮M); FSK, forskolin (5 ␮M); IBMX, isobutylmethylxanthine (100 ␮M); Rol, Rolipram (10 ␮M) . n ϭ 4–5. (B and D) Cells were treated for 6 days as indicated and the mRNA was analyzed by RT-PCR for changes in chemokine expression and expressed as fold-change induction over control cells. n ϭ 4–12. (C and E) ELISA measurements of secreted protein expression in the supernatant of cells treated for 6 days as indicated. (n ϭ 4–12). PGE2, prostaglandin E2 (10 nM); ␮ Ͻ Ͻ Ͻ FSK (10 M); *, P 0.05; **, P 0.01 vs. control cells; #, P 0.05 vs. FSK (C) or vs. PGE2 (E).

2of6 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0911684106 Hertz et al. Downloaded by guest on September 28, 2021 we observed a marked potentiation of surface marker expression with the PDE4 inhibitor, rolipram, but minimally increased expression with PDE3 inhibitors. We also treated the cells with IBMX alone to inhibit multiple PDEs, and saw a modest increase in expression. However, when used with forskolin, IBMX causes massive cell death, implying some additional PDE may affect cell viability. From these data we can conclude that PDE4, but not PDE3, coordinates a signaling microdomain involved in regulation of CD14 and CD163 expression. We also evaluated the regulation by PDEs of the expression of the C-X-C and C-C classes of chemokines using specific PDE inhibitors. The PDE4 inhibitors were of particular interest not only because this family of PDEs appears to be important for both differentiation and function of the macrophages, but also because they are a class of drugs currently being evaluated clinically as treatment for COPD and several other inflammatory disorders. Because several of the chemokines up-regulated by forskolin are found in high levels in the lungs of patients with COPD (23, 25), we wanted to determine if a PDE4 could be regulating one or more signaling microdomains that control the expression of these chemokines. We saw an increase in both the mRNA (Fig. 1 B and D) and protein expression (Fig. 1 C and E) of several of these chemokines over basal upon treatment with either forskolin or the endogenous cAMP agonist PGE2 when coadministered with rolipram. In contrast, the PDE3-selective inhibitor cilostamide showed little or no potentiation of expression for either chemokine when

given in conjunction with a low dose of forskolin or PGE2, PHARMACOLOGY demonstrating a minimal involvement of PDE3 in controlling chemokine expression (Fig. S4). Therefore, it would appear that PDE4 is a major regulator of the cAMP-dependent pathways that can increase chemokine expression.

Epac Mediates Much of the cAMP-Stimulated Increase in Chemokine Expression. It seemed likely that cAMP is acting through either the Epac or PKA signaling pathways to exert its effects on chemokine expression. To determine which isoform of Epac was Fig. 2. Role of Epac and PKA in chemokine signaling. Chemokine levels in present in these cells, we examined the microarray data and cells treated for 6 days with indicated agonist. Sp-8-pCPT ϭ Sp-8-pCPT-2Ј-O- found that mRNA for Epac1, but not Epac2, was present in our Me-cAMPS, 6Bz ϭ 6-Bnz-cAMP (1 ␮M), FSK ϭ forskolin (25 ␮M). mRNA levels differentiated macrophages. To address how chemokine expres- analyzed by RT-PCR for CXCL7 (A), CXCL5 (C) and CCL2 (E) Secreted protein sion is regulated, we treated monocytes for 6 days during levels after 6 days measured from cell supernatant by ELISA for CXCL7 (B), differentiation with cAMP analogues that can specifically acti- CXCL5 (D) and CCL2 (F). *, P Ͻ 0.05; **, P Ͻ 0.01 vs. control cells. n ϭ 4–12. vate either PKA or Epac. Using the nonhydrolyzable Epac- selective activator, Sp-8-pCPT-2Ј-O-Me-cAMPS (Sp-8-pCPT), binding sites. The sequence 2,000 bp upstream of the start codon we observed a dose-dependent increase in CXCL7, CXCL5, and for 64 of the immune relevant genes, including all of the CCL2 expression at both the mRNA (Fig. 2 A, C, and E) and chemokines, identified in the microarray analysis, was searched protein levels (Fig. 2 B, D, and F) in monocyte-differentiated for transcription-factor binding sites using the program Clover macrophages. The magnitude of the increase was usually as great as or greater than the maximal seen with either forskolin or (26). Through comparison to a background list of unchanged 6 genes, six transcription-factor binding sites were found to be PGE2. When we used the PKA-specific activator, N - Benzoyladenosine-3Ј,5Ј-cyclic monophosphate (6-Bnz-cAMP), over-represented (Fig. 3A), with a P value indicating the statis- at a concentration of 1 ␮M, we also saw small increases in tical significance of over-representation and a score indicating chemokine mRNA levels, but no significant increases in secreted the strength of the factor’s presence in the whole sequence set. ␬ protein levels. Unfortunately, higher levels of 6-Bnz-cAMP The NF- B family represents the top three sites, suggesting that caused cell death when present during differentiation, and this transcriptional regulator might be common to many of the therefore could not be tested. This suggests that chemokine genes regulated by forskolin. We also graphed the motif against expression is largely controlled by Epac, with a possible smaller its corresponding score for each of three up-regulated groups: involvement of PKA. However, we have noted that the magni- those increased more than 5-fold, those increased 3- to 5-fold, tude of induction of these chemokines in response to treatment and those increased 2- to 3-fold (Fig. 3B). Again, we found that with Sp-8-pCPT was highly variable among patients, perhaps the most highly up-regulated genes contained NF-␬B sites in suggesting an additional level of regulation or alteration of basal their promoters. activation state in these cells. In a separate set of experiments, we were able to show that the forskolin-induced increases in CD14 and CD163 were dependent cAMP Controls Surface Marker Expression in an NF-␬B-Dependent on NF-␬B by using the cell-permeable peptide inhibitor SN50 to Manner. To determine mechanistically how cAMP/Epac might be block the nuclear translocation of p50 NF-␬B. The inhibitory affecting the expression of these genes, the putative promoter effect of SN50 was dose-dependent for increasing concentrations regions of many of the genes identified by clustering analysis of of SN50 on both CD14 and CD163 surface marker expression the array data were searched for common transcription-factor (Fig. 3C). Surprisingly, SN50 had little or no effect on forskolin-

Hertz et al. PNAS Early Edition ͉ 3of6 Downloaded by guest on September 28, 2021 Fig. 3. NF-␬B control of gene expression. (A) Promoter analysis of transcription factor binding sites for all genes up-regulated greater than 5-fold on the microarray. (B) Transcription factor binding-site prediction grouped by fold induction. (C) The forskolin-induced increase in CD163 and CD14 expression was dose-dependently inhibited by the NF-␬B inhibitor, SN50, after 6 days of treatment. n ϭ 5. (D) The increase in chemokine expression seen with 25 ␮M forskolin treatment was not inhibited by treatment with SN50 (50 ␮g/ml) after 6 days of treatment. n ϭ 4. (E) Cells were differentiated for 6 days in the presence of GM-CSF, the media changed and either FSK (25 ␮M) or FSK ϩ BMS-345541 (10 ␮M) was added for 6 h and the mRNA harvested. CXCL7 and CXCL5 mRNA levels were calculated relative to GAPDH. n ϭ 4. *, P Ͻ 0.05; **, P Ͻ 0.01.

induced chemokine expression (see Fig. 3D). One possibility to ChIP analysis using an antibody to ATF3 and gene primers explain this lack of effect is that SN50 does not block the specific for the predicted ATF3 binding region in the CXCL7 alternative p52/RelB pathway, which could also be regulated by promoter to measure the DNA binding activity of ATF3 in these cAMP. To address this possibility, we used BMS-345541, a cells. After 6 days of forskolin treatment, we observed a marked selective allosteric inhibitor of the I␬B kinases, IKK1 and IKK2 decrease in the amount of ATF3 bound to the promoter of (27). When we treated fully differentiated macrophages with CXCL7 (Fig. 4B). Our results are consistent with a relief of forskolin and a moderate dose (10 ␮M) of BMS-345541 for 6 h, inhibition by ATF3, but more work will need to be done to we saw a marked decrease in CXCL5 and CXCL7 expression determine the exact mechanism by which cAMP and ATF3 may with inhibitor treatment, demonstrating that the NF-␬B pathway be contributing to the increase in chemokine levels. also can control forskolin-induced expression of these chemo- Discussion kine genes. GM-CSF is the major cytokine driving differentiation of mono- Role of ATF3 in Transcriptional Regulation of CXCL7 Expression. A cytes into macrophages in the lung and several other tissues. further inspection of the down-regulated genes in the microarray GM-CSF promotes the conversion of monocytes into macro- suggested a possible additional mechanism for regulation of phages that can then be strongly activated to promote inflam- chemokine genes by forskolin: a 7-fold decrease in mRNA levels matory responses when they come in contact with external for the CREB-family transcription factor, ATF3. ATF3 has stimuli, such as LPS or cigarette smoke. To date, most studies on recently been shown to be a negative regulator of NF-␬B acting the effects of cAMP on macrophage function have focused on its through an increase in histone deacetylase activity (28, 29). anti-inflammatory effects on fully activated macrophages. For Therefore, decreases in ATF3 levels should have the effect of example, in LPS-activated macrophages, agents that increase cAMP levels decrease inflammatory cytokine release, inhibit increasing transcription of NF-␬B-regulated genes. We con- phagocytosis, and reduce reactive oxygen and bactericidal ac- firmed the forskolin-dependent decrease in ATF3 mRNA ex- tivity (8). It is important to note that cAMP is itself not the pression using RT-PCR (Fig. 4A). Additionally, we performed primary mediator of these processes but is instead a modifying stimulus. From the present results, it appears that cAMP can also act as a positive regulator of a number of immune function genes, including some normally considered to be proinflammatory. It has been shown in several systems that LPS can increase cAMP in macrophages (9, 30), and in this regard a subset of cAMP- induced gene transcription could be viewed as part of a normal inflammatory response. If so, it points once again to the com- partmentalized mechanisms by which cAMP functions in many cells, given that the same signaling molecule can have opposing effects within the same cell. From the results of this study, it is clear that PDE4 inhibition can modulate monocyte to macrophage differentiation by potentiating Fig. 4. ATF3 mRNA expression and DNA binding activity. (A) Real-time PCR analysis of ATF3 mRNA levels, relative to 18 s mRNA in macrophages treated the effects of forskolin and PGE2. Although PDE3 and PDE4 for 6 days with FSK (25 ␮M). n ϭ 6–8.*,P Ͻ 0.05. (B) ATF3 binding on the CXCL7 constitute nearly equal amounts of the cAMP PDE activity in these promoter in macrophages was measured by ChIP assay using an anti-ATF3 macrophages, only the PDE4 selective inhibitors had an appreciable antibody. Data shown is representative of three independent experiments. effect on the macrophage phenotype observed.

4of6 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0911684106 Hertz et al. Downloaded by guest on September 28, 2021 PDE4 inhibitors have a long history as potential clinical agents Supporting this hypothesis is the fact that the transcription and are currently under investigation for treatment of depres- factor ATF3, which is down-regulated by forskolin treatment, sion, memory disorders, and several inflammatory conditions (3, has recently been reported as a negative regulator of NF-␬B- 24). These drugs can often act as anti-inflammatory agents dependent transcription (28). ATF3 has also been implicated in against several types of activated immune cells both in vitro and control of chemokine expression in a mouse asthma model (29). in vivo. However, as it is very difficult to obtain material to In these studies, Gilchrist et al (29) showed a marked increase in measure monocyte- and macrophage-specific activity in humans many of the same chemokines we see up-regulated with cAMP: treated in vivo, ex vivo stimulation of TNF-␣ release by LPS has CXCL1, -2, and -5, and CCL2 in the ATF3-null mice compared been the predominant surrogate measure of their function. to wild-type controls. Despite their anti-inflammatory effects against monocytes Finally, it seems that the Epac signaling pathway plays a and macrophages in vitro and in mouse models, to date the prominent role in cAMP regulation of chemokine expression in clinical trials for PDE4 inhibitors in the treatment of COPD and the GM-CSF macrophage differentiation model. In previous asthma have met with mixed success. One possible reason is that studies, it has been shown that many of cAMP’s anti- PDE4-inhibitor treatment in humans is dose-limited by side inflammatory effects are mediated through the PKA pathway, effects, such as emesis (24). However, a few earlier in vivo although a role for Epac has been established in modulating inflammation studies have suggested that under some conditions microbicidal activity and Fc␥R-mediated phagocytosis (8, 29). treatment with PDE4 inhibitors can actually promote macro- Specific isoforms of PDE4, working through PKA, have been phage accumulation, which could be counterproductive for implicated in regulation of both cytokine and adhesion molecule resolving inflammation. For example, roflumilast treatment expression using PDE4A, PDE4B, and PDE4D knockout mice slightly increased the number of macrophages in lung bronchoal- (9, 40). Additionally, using a monocytic differentiation protocol veolar lavage fluid using a mouse model of acute cigarette smoke similar to that used in this article, Bryn et al showed that TNF-␣, exposure, although neutrophil numbers were decreased (31). IL-12, MIP1-␤, integrin, and surface marker expression were Another study reported that roflumilast and piclamilast reduced controlled by PKA and not Epac (41). In these studies, Epac KC (mouse CXCL1) and TNF-␣ release and cell infiltration into expression was increased upon differentiation; however, there lungs of mice challenged with LPS. However, in the absence of were no major anti-inflammatory changes attributed to Epac LPS challenge, the inhibitors actually had the opposite effect by alone. These data fit nicely with the present observations that inducing release of KC and greatly increasing neutrophil infil- activation of Epac appears to have proinflammatory effects, not tration into lung tissue (32). These investigators postulated that anti-inflammatory effects, on monocyte/macrophages. It has PHARMACOLOGY chemokine release was likely from endothelial cells, but the data also been reported that Epac can shuttle in and out of the shown here suggest that macrophages may also play an equally nucleus, which may be relevant for how the changes in chemo- important role in PDE4-mediated chemokine release. kine genes are regulated by Epac (42). Taken together, these From the results presented, it seems quite possible that data suggest an additional role for Epac in the inflammatory increased chemokine production by macrophages in the pres- process and a possible additional site of therapeutic intervention, ence of a cAMP agonist and a PDE4 inhibitor may exacerbate if Epac antagonists can be developed. the underlying chronic inflammation in lung diseases, such as In summary, we suggest that the well-established effectiveness asthma or COPD (33, 34). Given the large induction of chemo- of PDE4 inhibitors to decrease acute inflammatory responses in kines generated by cAMP, it seems likely that many neutrophils, activated macrophages may be limited by an opposing effect on eosinophils, T cells, and monocytes, responding to CXCR2 and chemokine induction. Addition of a chemokine receptor antag- CCR2 receptors, would be recruited to sites of lung inflamma- onist, glucocorticoid, or Epac inhibitor to the PDE4 inhibitor tion and contribute to underlying chronic inflammation through treatment may not only make the PDE4 inhibitor more effica- immune cell recruitment. These chemokines can also directly cious, but also have the added benefit of reducing concerns about modulate immune cell functions, further contributing to the inflammatory side effects, such as the mesenteric vasculitis pathogenesis of these diseases. It will be of great interest to sometimes seen with chronic PDE4-inhibitor treatment. determine if combinations of other PDE inhibitors, perhaps working on PDEs in different functional compartments, can be Materials and Methods identified that do not promote recruitment of immune cells but, Monocyte Isolation and Culture. Monocytes were puriﬁed by positive selection rather, only target the anti-inflammatory functions of the acti- with magnetic CD14 microbeads (Miltenyi Biotec) from buffy coats obtained vated macrophages (35). Finally, these results also suggest the from human donors with consent by the Portland Oregon Red Cross or the possibility that PDE4 inhibitors may be more efficacious if used Stanford Blood Center, as previously described (13). Monocyte differentiation ϫ 5 in conjunction with a glucocorticoid, one of the newly developed was carried out at a concentration of 5 10 cell/ml with 50 ng/ml recombi- chemokine receptor antagonists (25, 36, 37) or, if they can be nant human GM-CSF (leukine, Berlex) for 6 days; 10% volume of additional media was added on Day 3 if necessary. Reagents were added as indicated on developed, Epac antagonists. One would expect that not only Day 0 with GM-CSF, and allowed to incubate for 6 days without supplemen- would the efficacy of each drug be increased by coadministra- tation or added at Day 6 and incubated for 6 h. tion, but also that the safety margins and side-effect profiles of these PDE4 inhibitors may be improved by coadministration of FACS Staining and Analysis. Fluorescently labeled antibodies were from either such agents. Beckman Coulter (mannose receptor, CD64, CD80, CD86,) or BD Biosciences Mechanistically, these proinflammatory cAMP effects may be (CD14, CD71, CD163). Cells were prepared for FACS and analyzed as previously mediated in part through differential regulation of NF-␬B described using a Becton Dickinson FACScanto ﬂow cytometer and FACSdiva activity. That increased cAMP can inhibit induction of many software (43). genes triggered by an inflammatory stimulus, such as LPS, is well established. We have shown here that several of the genes RT-PCR Analysis. Total RNA was isolated using QIAshredder columns and the induced by forskolin have an NF-␬B binding motif in their RNeasy kit from Qiagen. Total RNA was transcribed to cDNA using the Super- ␬ Script III RT kit from Invitrogen according to the manufacturer’s protocol. promoter regions and that NF- B pathway inhibitors block Real-time analysis of PCR was performed using SYBR Green (CXCL7, CCL2, chemokine and surface marker expression. Given the disparate CXCL5) from Applied Biosystems, or Taqman technology (ATF3), on a Strat- data with BMS 345541 and SN50, the NF-␬B pathway regulating agene MX3000P and analyzed using MxPro software. Data are expressed as the chemokine induction is likely to be the alternative, p100/p52 fold-change using the delta-delta Ct method, or as an amount relative to and RelB-mediated, pathway (38, 39). either GAPDH or 18s mRNA.

Hertz et al. PNAS Early Edition ͉ 5of6 Downloaded by guest on September 28, 2021 Cytokine ELISA Measurements. Duoset ELISA kits for CCL2, CXCL7, and CXCL5 bp2520). SN50 was obtained from BioMol (catalog # P-600). BMS-345541 was (R&D Systems catalog #DY279, #DY393, and DY254, respectively) were used to obtained from Sigma (catalog #B-9935). assess secreted protein levels in cell-culture supernatants according to the protocol recommended by the manufacturer. Statistical Analysis. Data were analyzed using the two-tailed Student’s t-test. P Ͻ 0.05 was considered signiﬁcant. Chromatin Immunoprecipitation. Monocytes were incubated for 6 days in the presence of GM-CSF ϩ/Ϫ 25 ␮M forskolin. Cells were isolated as previously ACKNOWLEDGMENTS. We thank Biolog, Inc., for generously providing the described, and the DNA was extracted and subjected to PCR analysis (28). cyclic nucleotide analogs. This study was supported in part by National Insti- tutes of Health Trainee under Pharmaceutical Sciences Training Grant GM007750 (to A.L.H.), National Institutes of Health Trainee under University Reagents. Cyclic nucleotide analogs were obtained from Biolog, Inc. and used of Washington Pathology of Cardiovascular Disease Training Grant HL0732 (to at concentrations that promoted their relative selectivity for the cAMP effec- A.T.B.), and National Institutes of Health Grants DK21723 and GM083296 (to tor molecules (44). Forskolin was obtained from Fisher Bioreagents (catalog # J.A.B.), AI025032 (to A.A.), and HL092547 (to M.G.).

1. de Rooij J, et al. (1998) Epac is a Rap1 guanine-nucleotide-exchange factor directly 23. Barnes PJ (2008) Immunology of asthma and chronic obstructive pulmonary disease. activated by cyclic AMP. Nature 396:474–477. Nat Rev Immunol 8:183–192. 2. Kawasaki H, et al. (1998) A family of cAMP-binding proteins that directly activate Rap1. 24. Spina D (2008) PDE4 inhibitors: current status. Br J Pharmacol 155:308–315. Science 282:2275–2279. 25. Donnelly LE, Barnes PJ (2006) Chemokine receptors as therapeutic targets in chronic 3. Bender AT, Beavo JA (2006) Cyclic nucleotide phosphodiesterases: Molecular regula- obstructive pulmonary disease. Trends Pharmacol Sci 27:546–553. tion to clinical use. Pharmacol Rev 58:488–520. 26. Fu Y, Frith MC, Haverty PM, Weng Z (2004) MotifViz: an analysis and visualization tool 4. McConnachie G, Langeberg LK, Scott JD (2006) AKAP signaling complexes: getting to for motif discovery. Nucleic Acids Res 32:W420–W423. the heart of the matter. Trends Mol Med 12:317–323. 27. Burke JR, et al. (2003) BMS-345541 is a highly selective inhibitor of Ikappa B kinase that 5. Bornfeldt KE (2006) A single second messenger: Several possible cellular responses binds at an allosteric site of the enzyme and blocks NF-kappa B-dependent transcrip- depending on distinct subcellular pools. Circ Res 99:790–792. tion in mice. J Biol Chem 278:1450–1456. 6. Zaccolo M, Movsesian MA (2007) cAMP and cGMP signaling cross-talk: Role of phos- 28. Gilchrist M, et al. (2006) Systems biology approaches identify ATF3 as a negative phodiesterases and implications for cardiac pathophysiology. Circ Res 100:1569–1578. regulator of Toll-like receptor 4. Nature 441(7090):173–178. 7. Fischmeister R, et al. (2006) Compartmentation of cyclic nucleotide signaling in the 29. Gilchrist M, et al. (2008) Activating transcription factor 3 is a negative regulator of heart: The role of cyclic nucleotide phosphodiesterases. Circ Res 99:816–828. allergic pulmonary inflammation. J Exp Med 205:2349–2357. 8. Serezani CH, Ballinger MN, Aronoff DM, Peters-Golden M (2008) Cyclic AMP: Master 30. Moon, E-Y, Pyo S (2007) Lipopolysaccharide stimulates Epac1-mediated Rap1/NF-kB regulator of innate immune cell function. Am J Respir Cell Mol Biol 39(2):127–132. 9. Jin SLC, Lan L, Zoudilova M, Conti M (2005) Specific role of phosphodiesterase 4B in pathway in Raw 264.7 murine macrophages. Immunol Lett 110:121–125. lipopolysaccharide-induced signaling in mouse macrophages. J Immunol 175:1523– 31. Martorana PA, Beume R, Lucattelli M, Wollin L, Lungarella G (2005) Roflumilast fully 1531. prevents emphysema in mice chronically exposed to cigarette smoke. Am J Respir Crit 10. Houslay MD, Schafer P, Zhang KYJ (2005) Keynote review: Phosphodiesterase-4 as a Care Med 172:848–853. therapeutic target. Drug Discov Today 10:1503–1519. 32. McCluskie K, et al. (2006) Phosphodiesterase type 4 inhibitors cause proinflammatory 11. Giordano D, Magaletti DM, Clark EA, Beavo JA (2003) Cyclic nucleotides promote effects in vivo. J Pharmacol Exp Ther 319(1):468–476. monocyte differentiation toward a DC-SIGNϩ (CD209) intermediate cell and impair 33. Calverley PMA, et al. (2009) Roflumilast in symptomatic chronic obstructive pulmonary differentiation into dendritic cells. J Immunol 171:6421–6430. disease: two randomised clinical trials. Lancet 374:685–694. 12. Shepherd MC, Baillie GS, Stirling DI, Houslay MD (2004) Remodelling of the PDE4 cAMP 34. Fabbri LM, et al. (2009) Roflumilast in moderate-to-severe chronic obstructive pulmo- phosphodiesterase isoform profile upon monocyte-macrophage differentiation of nary disease treated with long-acting bronchodilators: two randomised clinical trials. human U937 cells. Br J Pharmacol 142:339–351. Lancet 374:695–703. 13. Bender AT, Ostenson CL, Wang EH, Beavo JA (2005) Selective up-regulation of PDE1B2 35. Giembycz MA (2005) Phosphodiesterase-4: Selective and dual-specificity inhibitors for upon monocyte-to-macrophage differentiation. Proc Natl Acad Sci USA 102:497–502. the therapy of chronic obstructive pulmonary disease. Proc Am Thorac Soc 2:326–333. 14. Gantner F, Kupferschmidt R, Schudt C, Wendel A, Hatzelmann A (1997) In vitro 36. Palmqvist C, Wardlaw AJ, Bradding P (2007) Chemokines and their receptors as differentiation of human monocytes to macrophages: change of PDE profile and its potential targets for the treatment of asthma. Br J Pharmacol 151:725–736. relationship to suppression of tumour necrosis factor-alpha release by PDE inhibitors. 37. Horuk R (2009) Chemokine receptor antagonists: overcoming developmental hurdles. Br J Pharmacol 121:221–231. Nat Rev Drug Discov 8(1):23–33. 15. Gordon S, Taylor PR (2005) Monocyte and macrophage heterogeneity. Nat Rev Immu- 38. Perkins ND (2007) Integrating cell-signalling pathways with NF-␬B and IKK function. nol 5:953–964. Nat Rev Mol Cell Biol 8(1):49–62. 16. Mosser DM, Edwards JP (2008) Exploring the full spectrum of macrophage activation. 39. Hayden MS, Ghosh S (2008) Shared principles in NF-␬B signaling. Cell 132:344–362. Nat Rev Immunol 8:958–969. 40. Ariga M, et al. (2004) Nonredundant function of phosphodiesterases 4D and 4B in 17. Mantovani A, et al. (2004) The chemokine system in diverse forms of macrophage neutrophil recruitment to the site of inflammation. J Immunol 173:7531–7538. activation and polarization. Trends Immunol 25:677–686. 41. Bryn T, et al. (2006) The cyclic AMP-Epac1-Rap1 pathway is dissociated from regulation 18. Bender AT, Ostenson CL, Giordano D, Beavo JA (2004) Differentiation of human of effector functions in monocytes but acquires immunoregulatory function in mature monocytes in vitro with granulocyte-macrophage colony-stimulating factor and macrophage colony-stimulating factor produces distinct changes in cGMP phosphodies- macrophages. J Immunol 176:7361–7370. terase expression. Cell Signal 16:365–374. 42. Huston E, et al. (2008) EPAC and PKA allow cAMP dual control over DNA-PK nuclear 19. Metcalf D (2008) Hematopoietic cytokines. Blood 111:485–491. translocation. Proc Natl Acad Sci USA 105:12791–12796. 20. Akagawa K (2002) Functional heterogeneity of colony-stimulating factor-induced 43. Bender AT, Ostenson CL, Giordano D, Beavo JA (2003) Differentiation of human human monocyte-derived macrophages. Int J Hematol 76(1):27–34. monocytes in vitro with granulocyte-macrophage colony-stimulating factor and mac- 21. Brown RA, Spina D, Page CP (2008) Adenosine receptors and asthma. Br J Pharmacol rophage colony-stimulating factor produces distinct changes in cGMP phophodiester- 153:S446–S456. ase expression. Cell Signal 16:365–374. 22. Montuschi P, Kharitonov SA, Ciabattoni G, Barnes PJ (2003) Exhaled leukotrienes and 44. Poppe H, et al. (2008) Cyclic nucleotide analogs as probes of signaling pathways. Nat prostaglandins in COPD. Thorax 58:585–588. Meth 5:277–278.

6of6 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0911684106 Hertz et al. Downloaded by guest on September 28, 2021