A Small Molecular Activator of Cardiac Hypertrophy Uncovered in a Chemical Screen for Modifiers of the Calcineurin Signaling Pathway

A small molecular activator of cardiac hypertrophy uncovered in a chemical screen for modifiers of the calcineurin signaling pathway

Erik Bush*†, Jens Fielitz‡, Lawrence Melvin*, Michael Martinez-Arnold‡, Timothy A. McKinsey*, Ryan Plichta*, and Eric N. Olson†‡

*Myogen, Incorporated, Westminster, CO 80021; and ‡Department of Molecular Biology, University of Texas Southwestern Medical Center, Dallas, TX 75390-9148

Contributed by Eric N. Olson, January 5, 2004 The calcium, calmodulin-dependent phosphatase calcineurin, regu- ative roles for these proteins in the control of calcineurin activity. lates growth and gene expression of striated muscles. The activity of Overexpression of MCIP1 (also called Down syndrome critical calcineurin is modulated by a family of cofactors, referred to as region 1), for example, suppresses calcineurin signaling (12). In modulatory calcineurin-interacting proteins (MCIPs). In the heart, the contrast, MCIP1 also seems to potentiate calcineurin signaling, MCIP1 gene is activated by calcineurin and has been proposed to as demonstrated by the diminution of calcineurin activity in the fulfill a negative feedback loop that restrains potentially pathological hearts of MCIP1 knockout mice (13). Intriguingly, the MCIP1 calcineurin signaling, which would otherwise lead to abnormal car- gene is a target of NFAT and is up-regulated in response to diac growth. In a high-throughput screen for small molecules capable calcineurin signaling (15), which has been proposed to create a of regulating MCIP1 expression in muscle cells, we identified a unique negative feedback loop that dampens calcineurin activity, which 4-aminopyridine derivative exhibiting an embedded partial structural would otherwise lead to abnormal cardiac growth. motif of serotonin (5-hydroxytryptamine, 5-HT). This molecule, re- In an effort to identify novel small molecules that might prevent ferred to as pyridine activator of myocyte hypertrophy, acts as a pathological cardiac hypertrophy by stimulating MCIP1 expression, selective agonist for 5-HT2A/2B receptors and induces hypertrophy of we performed a high-throughput screen (HTS) of a chemical cardiac muscle cells through a signaling pathway involving calcineurin library for compounds capable of activating the calcineurin͞NFAT- and a kinase-dependent mechanism that inactivates class II histone responsive promoter of the MCIP1 gene in muscle cells. We deacetylases, which act as repressors of cardiac growth. These find- describe a previously uncharacterized 4-aminopyridine that we ings identify MCIP1 as a downstream target of 5-HT2A/2B receptor refer to as pyridine activator of myocyte hypertrophy (PAMH), signaling in cardiac muscle cells and suggest possible uses for which induces MCIP1 expression and, unexpectedly, drives cardio- 5-HT2A/2B agonists and antagonists as modulators of cardiac growth myocyte hypertrophy. PAMH acts as a 5-hydroxytryptamine and gene expression. (5-HT)2A/2B receptor agonist and induces hypertrophy, at least in part, by stimulating nuclear import of NFAT and nuclear export of umerous agonists that act through G protein-coupled re- class II HDACs. These findings shed light on a powerful hypertro- Nceptors trigger calcium-dependent signal transduction path- phic signaling pathway downstream of 5-HT2A/2B receptor signaling ways that stimulate cardiac growth and gene expression (re- and suggest that chemical modulators of this pathway may be viewed in ref. 1). Postnatal cardiac myocytes respond to such efficacious in the control of cardiac growth and gene expression. signals by hypertrophic growth, characterized by an increase in Materials and Methods myocyte size and protein synthesis, assembly of sarcomeres, and activation of a fetal gene program (reviewed in ref. 2). Activation Cardiomyocyte Cultures. Neonatal rat ventricular myocytes of the calcium, calmodulin-dependent phosphatase calcineurin, (NRVMs) were cultured as described (16). For detailed proce- is sufficient and, in many cases, necessary for pathological dures, see Supporting Text, which is published as supporting cardiac hypertrophy (3), a major predictor of human morbidity information on the PNAS web site. and mortality (4). Thus, there has been intense interest in identifying novel small molecules capable of therapeutically Primary HTS. H9c2 cells (American Type Culture Collection no. ͞ modulating cardiac calcineurin signaling. CRL-1446; ref. 17) were cultured in DMEM with 10% (vol vol) ͞ ͞ Many calcineurin-sensitive genes are controlled by members FBS 4mML-glutamine 1% penicillin/streptomycin. Cells at a of the nuclear factor of activated T-cell (NFAT) family of concentration of 50,000 cells per ml were transiently transfected in transcription factors, which translocate to the nucleus when batch with a reporter construct (20 pg per cell) encoding firefly dephosphorylated by calcineurin (reviewed in ref. 5). The cal- luciferase under control of the exon 4 promoter from the human Ϫ ϩ cineurin pathway also stimulates the myocyte enhancer factor-2 MCIP1 gene (base pairs 874 to 30 relative to the beginning of ϫ Ϫ5 ␮ (MEF2) transcription factor by multiple mechanisms (6). We exon 4) and FuGENE transfection reagent (6 10 l per cell). have shown that calcineurin activates a kinase that phosphory- Transfected cells were plated on 96-well plates (Packard) at a lates class II histone deacetylases (HDACs), which act as MEF2 density of 5,000 cells per well. A 20,000-member library of small corepressors (7). Signal-dependent phosphorylation of class II molecule test compounds from the Myogen Library (selected for HDACs triggers their export from the nucleus and activation of molecular diversity and purchased from ChemBridge, San Diego) MEF2 target genes (8, 9). HDAC mutants lacking the signal- were then added by using a BioMEK FX robotic liquid handling responsive phosphorylation sites are refractory to calcium signaling and prevent cardiomyocyte hypertrophy. Conversely, Abbreviations: ANF, atrial natriuretic factor; PAMH, pyridine activator of myocyte hyper- mice lacking class II HDACs are hypersensitive to the growth- trophy; A-PAMH, antagonist of PAMH; CsA, cyclosporine A; HDAC, histone deacetylase; promoting activity of calcineurin (7). 5-HT, 5-hydroxytryptamine; HTS, high-throughput screen; NFAT, nuclear factor of acti- The activity of calcineurin is influenced by cofactors known as vated T cells; NRVM, neonatal rat ventricular myocyte; PE, phenylephrine. modulatory calcineurin-interacting proteins (MCIPs) or calci- †To whom correspondence may be addressed. E-mail: [email protected] or pressins (reviewed in ref. 10). Recent studies in yeast (11) and [email protected]. mammalian cells (12–14) have revealed both positive and neg- © 2004 by The National Academy of Sciences of the USA

2870–2875 ͉ PNAS ͉ March 2, 2004 ͉ vol. 101 ͉ no. 9 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0308723101 Downloaded by guest on September 28, 2021 Fig. 1. Identiﬁcation of PAMH from a screen for small molecules that enhance MCIP1 expression. (A) Schematic diagram of the screen. A luciferase reporter controlled by the calcineurin-responsive exon 4 promoter of the human MCIP1 gene was transiently transfected into the H9c2 muscle cell line, which was screened in 96-well plates for luciferase expression in the presence of 20,000 individual small molecules. Our hypothesis was that activators of MCIP1 would inhibit cardiac hypertrophy by suppressing calcineurin activity. (B) Structure of PAMH and A-PAMH and their similarity to serotonin. An embedded possible structural motif shared by the three molecules is indicated in red. (C) MCIP1 protein was detected by Western blot after exposure of NRVMs to PAMH (1 ␮M) for 24 h. Calnexin was detected as a loading control. (D) MCIP1 protein was detected by Western blot in extracts derived from NRVMs infected with adenoviruses encoding lacZ (as a control), activated calcineurin (Cn), or the 28-kDa form of MCIP1 initiated from exon 4 in the presence or absence CsA for 24 h. (E) MCIP1 protein was detected by Western blot in extracts derived from NRVMs after exposure to PAMH (1 ␮M) in the presence or absence of CsA. CELL BIOLOGY system (Beckman Coulter) (one compound per well, 10 ␮M con- cardiac growth. Toward that end, we performed an HTS for centration). Sixteen control wells per plate received vehicle alone compounds able to stimulate expression of a luciferase reporter (0.1% DMSO, final). Plates were incubated for 48 h and processed gene controlled by the alternative promoter upstream of exon 4 for quantification of luciferase activity on a multiwell luminometer of the human MCIP1 gene in the H9c2 muscle cell line (Fig. 1A). (Packard Fusion). Primary HTS hits were defined as compounds This genomic region contains 15 NFAT-binding sites and confers that produced an increase in luciferase activity Ͼ3 SD from calcineurin responsiveness to MCIP1 (15). Transcripts initiated vehicle-only controls and were verified in secondary screens. from the exon 4 promoter encode a 197-aa MCIP1 protein with Ϸ Ϸ a Mr 28 kDa compared with the 38-kDa protein encoded by Western Blot Analysis and Immunostaining. A peptide correspond- transcripts containing exon 1 (20). ing to the C terminus of the murine MCIP1 protein (GenBank From a screen of 20,000 compounds, we identified 21 that accession no. AAF63486; CRPEYTPIHLS) was synthesized stimulated MCIP1-luciferase expression by 2-fold or greater. The (Sigma Genosys), incorporating an N-terminal cysteine residue strongest hit, named PAMH, is a previously uncharacterized 4- to facilitate conjugation to keyhole limpet hemocyanin (KLH) aminopyridine (5,6,7,8-tetrahydro-3-methyl-2-phenyl[1]benzo- carrier. Rabbits were immunized with KLH-conjugated peptide thieno[2,3-b]pyridin-4-amine), which exhibits an embedded partial according to standard procedures (Lampire Biological Labora- structure motif of serotonin (5-HT) (Fig. 1B). Consistent with its tories, Pipersville, PA). Western blots and methods for immu- ability to stimulate the MCIP1 exon 4 promoter, PAMH selectively nodetection of ␤-myosin heavy chain and atrial natriuretic factor and reproducibly induced a dramatic increase in expression of the (ANF) were performed as described in Supporting Text. 28-kDa form of MCIP1 in NRVMs (Fig. 1C). RNA Isolation and Analysis. Analysis of RNA expression was The sensitivity of the 28-kDa form of MCIP1 to calcineurin performed by microarray and real-time RT-PCR as described in signaling is shown in Fig. 1D. Infection of NRVMs with an Supporting Text. adenovirus encoding activated calcineurin preferentially up- regulates this form of the protein, which comigrates with the MCIP1 protein encoded by a cDNA that initiates from the ATG Receptor-Binding Assays. Binding of PAMH to the 5-HT2B receptor was measured as described in Supporting Text. in exon 4. The 38-kDa form of MCIP that initiates in exon 1 was less responsive but often showed an increase in expression in Results response to PAMH when cells were maintained at high densities. HTS for Molecules That Enhance MCIP1 Expression in Muscle Cells. The variability in responsiveness of the 38-kDa MCIP1 protein Based on the resistance to hypertrophy of transgenic mice that may reflect a posttranslational effect of PAMH independent of overexpress MCIP1 in the heart (18, 19), we sought to identify its effect on the calcineurin-responsive exon 4 promoter. Inhi- small molecules capable of increasing MCIP1 expression in bition of calcineurin activity with cyclosporine A (CsA) dimin- muscle cells as a potential means of suppressing pathological ished the increase in endogenous MCIP1 expression evoked by

Bush et al. PNAS ͉ March 2, 2004 ͉ vol. 101 ͉ no. 9 ͉ 2871 Downloaded by guest on September 28, 2021 Fig. 2. Stimulation of cardiac hypertrophy by PAMH. (A) ␣-Actinin and perinuclear ANF staining of NRVMs in the absence or presence of PAMH (1 ␮M) for 24 h. PAMH induces hypertrophy, ANF, and sarcomere assembly. (B–E) Cell size, protein content, ANF secretion, and ␤-myosin heavy chain expression were detected in NRVMs in the absence or presence of PAMH (1 ␮M) or PE (10 ␮M) for 24 h. (F) Venn diagram representing array elements regulated by PAMH and PE. Array elements that were regulated at least 2-fold by either PAMH or PE were scored for statistically signiﬁcant regulation by either compound by using the Affymetrix statistical parameters. The areas of the Venn diagram represent the number of array elements that were signiﬁcantly regulated by only one drug or by both drugs.

calcineurin and PAMH (Fig. 1 D and E), further indicating that signaling has been implicated in cardiac growth (22–24), we PAMH activates the calcineurin-responsive MCIP1 promoter. compared the effects of PAMH and serotonin on NRVMs. Serotonin (1 ␮M) had no observable effects on myocyte growth Stimulation of Cardiomyocyte Hypertrophy by PAMH. Because or MCIP1 expression. MCIP1 expression is induced in cardiac myocytes by diverse Vigorous contractions were observed after addition of PAMH to prohypertrophic agonists (21), we tested positives from the NRVMs, suggesting that it acted at the cell surface. To identify cell primary screen in a secondary screen for their potential to induce surface receptors that might mediate its actions, we tested the hypertrophy of NRVMs. PAMH potently induced hypertrophy effects of a series of receptor agonists and antagonists on the ability of NRVMs, as measured by assembly of sarcomeres, cell size, of PAMH to induce ANF mRNA in NRVMs (Fig. 3A). Despite the and protein synthesis (Fig. 2 A–C). PAMH also activated the similarity between the biological activities of PE and PAMH, the expression of ANF and ␤-myosin heavy chain, markers of the activity of PAMH was unaffected by the ␣-adrenergic receptor fetal gene program, and promoted hypertrophy as effectively as antagonist prazosin or the ␤-adrenergic receptor antagonist pro- phenylephrine (PE), a potent hypertrophic agonist that acts pranolol. In contrast, the general 5-HT receptor antagonist cypro- ␣ through the -adrenergic signaling pathway (Fig. 2 B–E). heptadine and the 5-HT2A/2C-selective antagonist, ketanserin, The gene expression patterns in the presence of PAMH and PE, blocked induction of ANF and the 28-kDa form of MCIP1 in as determined by microarray analysis, were also remarkably similar. response to PAMH (Fig. 3 A and C). AMI-193 and SB204741, Of 15,866 cDNAs analyzed on microarray chips, 175 were up- which act as 5-HT2A and 5-HT2B selective antagonists, respectively, regulated and 226 were down-regulated by at least 2-fold by both each had partial inhibitory effects on PAMH action, whereas the compounds (Fig. 2F). The magnitude of changes in gene expression 5-HT2B/2C selective antagonist, SB206553, had no effect on PAMH and the rank order of gene responsiveness to the two agonists were activity (Fig. 3A). also similar with both agonists (Table 1, which is published as We further tested the receptor-binding properties of PAMH supporting information on the PNAS web site, and Supporting by measuring its ability to compete with radiolabeled lysergic Text). MCIP1 mRNA was up-regulated Ϸ3-fold by PAMH and PE. acid diethylamide, a standard ligand for the 5-HT2B receptor, in a receptor-binding assay. PAMH bound to the 5-HT2B receptor Signaling by PAMH Through 5-HT2A/2B Receptors. Following the with high affinity, showing a Ki for the receptor of 64 nM (Fig. observation of a partial serotonin pharmacophore embedded in 3B). In contrast, PAMH showed no appreciable binding to the the structure of PAMH (Fig. 1B), and because 5-HT receptor ␣-adrenergic receptor (data not shown). Although PAMH binds

2872 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0308723101 Bush et al. Downloaded by guest on September 28, 2021 Fig. 3. Effects of receptor antagonists on the activity of PAMH. (A) NRVMs were pretreated for 1 h with the indicated compound followed by PAMH (1 ␮M) for 24 h. CELL BIOLOGY The ratio of ANF͞GAPDH mRNA expression was determined. Values are expressed relative to the ANF͞GAPDH ratio in cells treated with PAMH alone, which was assigned a value of 100%. Compounds were used at the concentrations described in Supporting Text. (B) Competition curve showing the percent inhibition of binding of radiolabeled lysergic acid diethylamide to the 5-HT2B receptor by the indicated concentrations of unlabeled PAMH. (C) NRVMs were treated with PAMH (1 ␮M) and the indicated concentrations of cyproheptadine and ketanserin for 24 h and MCIP1 was detected by Western blot. (D) NRVMs were treated with PAMH (1 ␮M) in the presence or absence of A-PAMH at the indicated concentrations and secreted ANF was detected. Values are expressed relative to the maximal level of ANF expression in the presence of PAMH. (E) NRVMs were treated with PAMH (1 ␮M) and A-PAMH at the indicated concentrations, and MCIP1 was detected by Western blot.

the 5-HT2B receptor, the failure of SB206553 to inhibit PAMH (PD98059) had modest or no effects on PAMH activity. Inhi- activity suggests that the 2A and 2B receptors may play redun- bition of calcineurin by CsA had only a partial effect on the dant roles in transmitting PAMH signals. activity of PAMH. The effects of each of the above antagonists on expression of ANF and MCIP1 and on myocyte hypertrophy and sarcomere assembly in the presence of PAMH were comparable (data not shown). These findings suggest that PAMH acts through 5-HT2A ͞ and or 5-HT2B receptors, and that MCIP1 and hypertrophy are regulated in parallel by PAMH. In an effort to identify novel PAMH mimetics or antagonists, we compared the activities of a series of structurally related molecules. One such molecule, 2-phenyl-4-quinolinamine, which we refer to as antagonist of PAMH (A-PAMH, Fig. 1B), blocked hypertrophy and induction of MCIP1 in response to PAMH (Fig. 3 D and E).

Analysis of the PAMH Signaling Pathway. To explore the PAMH signaling pathway, we tested the effects of protein kinase and phosphatase inhibitors on PAMH activity in NRVMs. The hypertrophic activity of PAMH was blocked by staurosporin, a general inhibitor of serine, threonine kinases, and by the tyrosine kinase inhibitors AG556 and AG490 (Fig. 4), whereas inhibitors Fig. 4. Effects of inhibitors on the activity of PAMH. NRVMs were pretreated of PKC (Go¨6983, Bisindolylmaleimide), PKA (H89, HA1004), with the indicated inhibitors for 1 h followed by PAMH (1 ␮M) for 24 h, and PKG (HA1004 and PKG inhibitor), phosphatidylinositol-3 ki- ANF transcripts were detected, as in Fig. 3A. Compounds were used at the nase (wortmannin, LY294002), Raf (ZM336372), or MEK1͞2 concentrations described in Supporting Text.

Bush et al. PNAS ͉ March 2, 2004 ͉ vol. 101 ͉ no. 9 ͉ 2873 Downloaded by guest on September 28, 2021 export of GFP-HDAC5. Export was blocked by A-PAMH (data not shown). To independently test whether HDAC phosphorylation was required for PAMH-mediated hypertrophy, we examined whether a signal-resistant HDAC5 mutant containing alanines in place of serines 259 and 498, which are required for nuclear export, could block the prohypertrophic activity of PAMH. Indeed, NRVMs infected with an adenovirus encoding this HDAC5 mutant showed an increase neither in cell size nor in sarcomere assembly in response to PAMH (Fig. 5C). These findings suggest that hypertrophy in response to PAMH requires transcriptional activation of genes that are repressed by HDAC5, and that PAMH activates a kinase that phosphorylates the regulatory serines that inactivate class II HDACs. Discussion In a chemical screen for regulators of MCIP1 expression, we identified a 5-HT2A/2B receptor agonist with potent hypertrophic activity for cardiac muscle cells. This agonist, called PAMH, stimulates the calcineurin-responsive promoter of the MCIP1 gene and promotes myocyte hypertrophy, at least in part, through the calcineurin-dependent nuclear import of NFAT and the kinase-dependent nuclear export of class II HDACs.

Signaling by PAMH Through 5-HT2A/2B Receptors. We speculate that PAMH and serotonin share a partial pharmacophore (Fig. 1B), which directs the activity of PAMH through 5-HT2 receptors. By comparison of a series of PAMH analogs, it appears that the fused cycloalkyl ring is critical for biological activity, and that substitution on the pyridine ring contributes to the prohypertrophic activity. The PAMH antagonist, A-PAMH, which lacks

Fig. 5. Signaling to NFAT and class II HDACs by PAMH. (A) Stimulation of NFAT nuclear import by PAMH. NRVMs transfected with a GFP-NFATc expression plas- mid were exposed to PAMH (1 ␮M) or PE (10 ␮M) for 24 h, as indicated, and GFP was detected. In unstimulated cells, GFP-NFATc is localized to the cytoplasm. Both agonists drive GFP-NFATc to the nucleus. (B) Stimulation of GFP-HDAC5 nuclear export by PAMH. NRVMs infected with an adenovirus encoding GFP-HDAC5 were exposed to PAMH (1 ␮M) or PE (10 ␮M), as indicated, for 24 h, and GFP was detected. In unstimulated cells, GFP-HDAC5 is localized to the nucleus. Both agonists promote translocation to the cytoplasm, and A-PAMH blocks this effect. (C) Blockade to PAMH activity by signal-resistant HDAC. NRVMs infected with an adenovirus encoding a signal-resistant HDAC5 mutant were exposed to PAMH, as indicated, and stained for ␣-actinin. The HDAC5 mutant prevents hypertrophy and sarcomere assembly in response to PAMH.

Transcriptional Responses to PAMH. In light of the sensitivity of the exon 4 promoter of MCIP1 to calcineurin͞NFAT signaling (15), we tested whether PAMH promoted the nuclear import of NFAT in NRVMs. PAMH and PE both induced the translocation of a GFP-NFATc fusion protein from the cytoplasm to the nucleus, which suggests that they activate the calcineurin pathway (Fig. 5A). We also examined the effect of PAMH on nuclear localization of HDAC5, which, like other class II HDACs, acts as a suppres- Fig. 6. Possible signaling pathways controlled by PAMH. The activity of sor of cardiac hypertrophy (7). Hypertrophic signals lead to the PAMH is blocked by staurosporin and the tyrosine kinase inhibitor AG490, phosphorylation of two critical serine residues in the N-terminal implicating serine, threonine kinases, and tyrosine kinases, respectively, in the regulatory regions of class II HDACs, which results in their signaling pathway controlled by PAMH. PAMH drives nuclear export of export from the nucleus and derepression of the hypertrophic HDAC5 and nuclear import of NFATc and stimulates the expression of MCIP1 gene program (8). As shown in Fig. 5B, a GFP-HDAC5 fusion and other genes involved in cardiac growth and remodeling. Signaling events that appear to play dominant roles in PAMH signaling are shown in bold. protein was localized to the nucleus in unstimulated NRVMs and Those that are documented, but for which the relative importance remains to became distributed diffusely throughout the cells in the presence be established, are shown by thin lines, and those that remain hypothetical are of PAMH, suggesting that PAMH activates the kinase- shown by dashed lines. Prohypertrophic effectors are in green, and antihy- dependent nuclear export of HDAC5. PE also drives nuclear pertrophic effectors are in red.

2874 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0308723101 Bush et al. Downloaded by guest on September 28, 2021 a fused cyclohexyl ring and methyl side chain modification of PAMH activates the calcineurin pathway. Consistent with the PAMH, cannot induce hypertrophy, although it can bind 5-HT notion that PAMH activates calcineurin signaling, it stimulates receptors (data not shown). PAMH is structurally related to a nuclear import of NFAT and activation of the NFAT-responsive family of molecules that possess anxiolytic activity (25); however, MCIP1 promoter. However, we also note that MCIP1 protein is their potential effects on muscle cells have not been reported. up-regulated more dramatically than MCIP1 mRNA by PAMH. The antihypertrophic activity of A-PAMH has potentially Thus, although PAMH was identified on the basis of its ability significant implications for the blockade of pathological cardiac to stimulate the calcineurin-responsive exon 4 promoter of hypertrophy. In this regard, it is interesting to note that ketan- MCIP1, posttranslational mechanisms may also contribute to the serin has been shown to reduce left ventricular hypertrophy while dramatic increase in MCIP1 protein in the presence of PAMH. preserving cardiac function in hypertensive patients (26), and the 5-HT2 receptors couple to calcineurin activation in neurons 5-HT antagonist sarpogrelate attenuates the ability of the hy- (29). Similarly, serotonin signaling controls egg laying in Cae- pertrophic agonists angiotensin II and endothelin-1 to stimulate norhabditis elegans by activating the calcineurin pathway (30), protein synthesis in cardiomyocytes (27). and transgenic overexpression of MCIP1 mimics the calcineurin Based on the effects of 5-HT2 antagonists on the prohypertrophic loss-of-function phenotype (31), consistent with MCIP function- activity of PAMH and on ligand-receptor-binding data, we con- ing as a repressor of calcineurin signaling. Although MCIP1 and clude that PAMH is selective for 5-HT2A/2B receptors. Signaling CsA can suppress hypertrophy due to calcineurin signaling (18), from the 5-HT2B receptor influences cardiac growth and develop- they may fail to do so in the presence of PAMH, because parallel ment (reviewed in ref. 24). Knockout mice lacking the receptor die pathways controlled by PAMH-dependent kinases bypass the from a thin myocardium (22), and overexpression of the receptor requirement of calcineurin. in the adult heart results in hypertrophy (23). Remarkably, sero- In addition to its effect on NFAT nuclear import, the ability tonin, which also binds 5-HT2 receptors, did not promote hyper- of PAMH to promote nuclear export of HDAC5 and of a trophy of cardiac myocytes. This disparity in actions of serotonin signal-resistant HDAC5 mutant to block hypertrophy in re- and PAMH could be explained if serotonin is unstable and rapidly sponse to PAMH suggests that PAMH stimulates the activity of degraded. Alternatively, it might activate opposing pro- and anti- an HDAC kinase that inactivates HDAC5, a conclusion sup- hypertrophic signaling pathways, whereas PAMH activates only a ported by the complete blockade to PAMH activity by stauro- prohypertrophic pathway. 5-HT2A/2B receptors might also exist in sporin. How mutant HDACs suppress the activity of PAMH multiple states (for a review, see ref. 28), providing for the possi- remains to be determined, but we favor the possibility that they bility that PAMH has a receptor-state specific effect not accessible interact with myocyte enhancer factor-2 or other prohypertro- by serotonin. phic transcription factors such as NFAT or GATA4 (32).

Downstream Signaling by PAMH. The 5-HT2 receptor subtypes (2A, ␣ ͞ Implications. The biological activity of PAMH suggests interest- 2B, and 2C) couple to G q G11 and phospholipase C, which ing possibilities for pharmacological modulation of muscle trigger intracellular calcium release via inositol triphosphate and growth and function. In addition to influencing cardiac growth consequent activation of PKC (Fig. 6). The strong inhibition of and development, 5-HT receptor signaling promotes myocyte PAMH activity by staurosporin, but not by inhibitors of PKC or 2B

survival and mitochondrial biogenesis (33). Recent studies have CELL BIOLOGY other kinases implicated in hypertrophic signaling, suggests that also revealed a profound stimulatory effect of PAMH on skeletal a calcium-dependent kinase yet to be identified may mediate the muscle development (unpublished results). Thus, the PAMH effects of PAMH. Tyrosine kinase inhibitors also suppressed signaling pathway provides therapeutic possibilities for manip- PAMH actions. Several tyrosine kinases, including src, JAK ulating striated muscle function in the settings of pathological kinase, and Erb-B, have been implicated in 5-HT receptor signaling (24). The tyrosine kinases and their potential targets in hypertrophy, heart failure, or skeletal muscle wasting disorders. the PAMH signaling pathway remain to be determined. We are grateful to R. Gorcyznski, B. Rothermel, R. Vega, L. Avery, and In contrast to the complete inhibition of PAMH activity by R. Bassel-Duby for advice; A. Tizenor for assistance with graphics; and staurosporin and tyrosine kinase inhibitors, CsA inhibited J. Page for editorial assistance. J.F. is a postdoctoral fellow of the PAMH activity only partially. Thus, we speculate that PAMH Charite, Campus Virchow Klinikum. E.N.O. is supported by grants from acts primarily through kinase-dependent pathways augmented the National Institutes of Health, the D. W. Reynolds Center for Clinical by calcineurin (Fig. 6). Alternatively, the incomplete inhibition Cardiovascular Research, the Robert A. Welch Foundation, and the of PAMH activity by CsA could reflect the potency in which Texas Advanced Technology Program.

1. Frey, N. & Olson, E. N. (2003) Annu. Rev. Physiol. 65, 45–79. 19. Hill, J. A., Rothermel, B., Yoo, K. D., Cabuay, B., Demetroulis, E., Weiss, R. M., Kutschke, 2. MacLellan, W. R. & Schneider, M. D. (2000) Annu. Rev. Physiol. 62, 289–319. W., Bassel-Duby, R. & Williams, R. S. (2002) J. Biol. Chem. 277, 10251–10255. 3. Olson, E. N. & Williams, R. S. (2000) Cell 101, 689–692. 20. Genesca, L., Aubareda, A., Fuentes, J. J., Estivill, X., De La Luna, S. & Perez-Riba, M. 4. Kannel, W. B. & Cobb, J. (1992) Cardiology 81, 291–298. (2003) Biochem. J. 374, 567–575. 5. Hogan, P. G., Chen, L., Nardone, J. & Rao, A. (2003) Genes Dev. 17, 2205–2232. 21. Wang, Y., De Keulenaer, G. W., Weinberg, E. O., Muangman, S., Gualberto, A., Landschulz, 6. McKinsey, T. A., Zhang, C. L. & Olson, E. N. (2002) Trends Biochem. Sci. 27, 40–47. K. T., Turi, T. G., Thompson, J. F. & Lee, R. T. (2002) Am. J. Physiol. 283, H533–H539. 7. Zhang, C. L., McKinsey, T. A., Chang, S., Antos, C. L., Hill, J. A. & Olson, E. N. (2002) 22. Nebigil, C. G., Choi, D. S., Dierich, A., Hickel, P., Le Meur, M., Messaddeq, N., Launay, Cell 110, 479–488. J. M. & Maroteaux, L. (2000) Proc. Natl. Acad. Sci. USA 97, 9508–9513. 8. McKinsey, T. A., Zhang, C. L., Lu, J. & Olson, E. N. (2000) Nature 408, 106–111. 23. Nebigil, C. G., Jaffre, F., Messaddeq, N., Hickel, P., Monassier, L., Launay, J. M. & 9. Grozinger, C. M. & Schreiber, S. L. (2000) Proc. Natl. Acad. Sci. USA 97, 7835–7840. Maroteaux, L. (2003) Circulation 107, 3223–3229. 10. Rothermel, B. A., Vega, R. B. & Williams, R. S. (2003) Trends Cardiovasc. Med. 13, 15–21. 24. Nebigil, C. G. & Maroteaux, L. (2003) Circulation 108, 902–908. 11. Kingsbury, T. J. & Cunningham, K. W. (2000) Genes Dev. 14, 1595–1604. 25. Thompson, M., Forbes, I. T. & Johnson, C. N. (1992) U. S. Patent 5,093,493. 12. Rothermel, B., Vega, R. B., Yang, J., Wu, H., Bassel-Duby, R. & Williams, R. S. (2000) 26. Cobo, C., Alcocer, L. & Chavez, A. (1990) Cardiovasc. Drugs Ther. 4, 73–36. J. Biol. Chem. 275, 8719–8725. 27. Ikeda, K., Tojo, K., Tokudome, G., Hosoya, T., Harada, M. & Nakao, K. (2000) Life Sci. 13. Vega, R. B., Rothermel, B. A., Weinheimer, C. J., Kovacs, A., Naseem, R. H., Bassel-Duby, 67, 2991–2996. R., Williams, R. S. & Olson, E. N. (2003) Proc. Natl. Acad. Sci. USA 100, 669–674. 28. Strange, P. G. (1999) Biochem. Pharmacol. 58, 1081–1088. 14. Ryeom, S., Greenwald, R. J., Sharpe, A. H. & McKeon, F. (2003) Nat. Immunol. 4, 29. Day, M., Olson, P. A., Platzer, J., Striessnig, J. & Surmeier, D. J. (2002) J. Neurophysiol. 87, 874–881. 2490–2504. 15. Yang, J., Rothermel, B., Vega, R. B., Frey, N., McKinsey, T. A., Olson, E. N., Bassel-Duby, 30. Lee, M. H., Han, S. M., Han, J. W., Kim, Y. M., Ahnn, J. & Koo, H. S. (2003) FEBS Lett. R. & Williams, R. S. (2000) Circ. Res. 87, E61–E68. 555, 250–256. 16. Molkentin, J. D., Lu, J. R., Antos, C. L., Markham, B., Richardson, J., Robbins, J., Grant, 31. Bandyopadhyay, J., Lee, J., Lee, J. I., Yu, J. R., Jee, C., Cho, J. H., Jung, S., Lee, M. H., S. R. & Olson, E. N. (1998) Cell 93, 215–228. Zannoni, S., Singson, A., et al. (2002) Mol. Biol. Cell 13, 3281–3293. 17. Kimes, B. W. & Brandt, B. L. (1976) Exp. Cell Res. 98, 367–381. 32. Suzuki, Y. J., Day, R. M., Tan, C. C., Sandven, T. H., Liang, Q., Molkentin, J. D. & Fanburg, 18. Rothermel, B. A., McKinsey, T. A., Vega, R. B., Nicol, R. L., Mammen, P., Yang, J., Antos, B. L. (2003) J. Biol. Chem. 278, 17525–17531. C. L., Shelton, J. M., Bassel-Duby, R., Olson, E. N., et al. (2001) Proc. Natl. Acad. Sci. USA 33. Nebigil, C. G., Etienne, N., Messaddeq, N. & Maroteaux, L. (2003) FASEB J. 17, 98, 3328–3333. 1373–1375.

Bush et al. PNAS ͉ March 2, 2004 ͉ vol. 101 ͉ no. 9 ͉ 2875 Downloaded by guest on September 28, 2021