C-type natriuretic peptide/cGMP/FoxO3 signaling attenuates hyperproliferation of pericytes from patients with … – Nature.com

Unaltered CNP/GC-B/cGMP signaling in pericytes from patients with PAH

To compare CNP/cGMP signaling between pulmonary vascular cell types, we performed experiments with cultured human PASMCs as well as with human lung microvascular ECs and pericytes. CNP did not modulate intracellular cGMP levels of ECs, while ANP provoked a small, 2-fold increase (Fig.1a). In contrast, CNP markedly and concentration-dependently increased cGMP levels in PASMCs and, even more, in pericytes (Fig.1a). As shown, in both types of cells the effects of the highest CNP concentration (100nM) were much greater than the effect of ANP, used at the same concentration.

a Effects of CNP and ANP on cGMP levels of human lung microvascular endothelial cells (n=3 wells from one biological sample), vascular smooth muscle cells (n=6 wells from three biological replicates), and pericytes (n=8 wells from four biological replicates) (1-way ANOVA). b and c Lung pericytes isolated from control individuals and patients with PAH exhibit similar cGMP responses to CNP and similar expression levels of the GC-B receptor and cGKI (b: n=8 wells from four biological replicates (1-way ANOVA). c n=12 from 6 controls and n=10 from 5 PAH patients (unpaired 2-tailed Students t-test). d CNP similarly stimulates the phosphorylation of VASP at Ser239, the cGKI-specific site, in control and PAH pericytes (n=4 biological replicates per group (2-way ANOVA)). *P<0.05 vs. PBS ().

Notably, pericytes isolated from the lungs of PAH patients exhibited similar cGMP responses to CNP as the pericytes from control lungs (Fig.1b). Concordantly, the expression levels of the GC-B receptor and of its downstream target cGKI were similar in control and PAH pericytes (immunoblots in Fig.1c). In both groups, CNP led to cGKI-mediated phosphorylation of the cytoskeleton-associated vasodilator-stimulated phosphoprotein (VASP) at Ser239 (Fig.1d). Together these results demonstrate that among lung vascular cell types, pericytes exhibit the highest responsiveness to CNP. Moreover, CNP/GC-B/cGMP/cGKI signaling is fully preserved in pericytes from patients with PAH.

In agreement with published studies7,8, lung pericytes isolated from PAH patients had slightly enhanced basal proliferation rates, as analyzed by bromodeoxyuridine (BrdU) incorporation assays (Fig.2a). Moreover, the proliferative effect of PDGF-BB (30ng/ml, 24h) was markedly greater in PAH than in control pericytes (Fig.2a). Scratch assays demonstrated that this was associated with enhanced baseline and PDGF-BB-stimulated migration (Fig.2b shows the progressive closing of the scratch-induced wound area during 24h, as ratio of the starting wound area). Moreover, the expression of the PDGF receptor- (PDGFR-) was significantly upregulated in PAH compared to control pericytes (immunoblot in Supplemental Fig.1), which most likely contributes to their augmented responses.

a and b In comparison to control pericytes, PAH pericytes had higher proliferation (a) and migration rates (b) at baseline and in response to PDGF-BB (30ng/ml, 24h) (a BrdU incorporation, n=10 wells from three biological replicates for control and n=16 wells from four biological replicates for PAH; non-parametric KruskalWallis analyses. b Scratch assays with 814 wells from three biological replicates from each group; 2-way ANOVA). c PDGF-BB (30ng/ml, 15min pretreatment) did not significantly alter the cGMP responses of control pericytes to CNP (n=12 wells from four biological replicates per condition; non-parametric KruskalWallis analyses). df CNP (pretreatment for 30min) attenuated PDGF-BB (30ng/ml) induced proliferation (d), cyclin D1 (e) and PCNA protein expression (f) of control and PAH pericytes (d: n=1018 wells from three biological replicates per group; non-parametric KruskalWallis analyses; e and f: n=4 biological replicates from control and PAH pericytes; 1-way ANOVA). g CNP (100nM, 24h) attenuated PDGF-BB-induced migration of control and PAH pericytes. Top panels: representative pictographs of control (left) and PAH pericytes (right) at 0, 16, and 24h of the scratch assay (scale bar: 500mm); Bottom panels, evaluation of the wounding areas in comparison to the initial wound (n=813 wells from three biological replicates per group; 2-way ANOVA). For a: *p<0.05 vs. Controls, #p<0.05 vs. Basal. For b: *p<0.05 vs. PBS-Control, #p<0.05 vs. PDGF-BB-Control. For c: *p<0.05 vs. PBS (). For dg: *p<0.05 vs. PBS (), #p<0.05 vs. PDGF-BB.

Next, we studied whether CNP attenuates the proliferative and promigratory effects of PDGF-BB. Because it was reported that growth factors desensitize the GC-B receptor in fibroblasts16, firstly, we studied in control pericytes whether PDGF-BB influences the cGMP responses to CNP. The cells were pretreated with 30ng/ml PDGF-BB for 15min, followed by CNP stimulation for an additional 15min. As shown in Fig.2c, PDGF-BB slightly reduced the effects of 10 and 100nM CNP on pericyte cGMP levels, but this interaction was statistically insignificant.

To dissect the effects of CNP on PDGF-BBstimulated pericyte proliferation and migration, the cells were pretreated with 10 or 100nM CNP for 30min, followed by incubation with 30ng/ml PDGF-BB for 24h. CNP significantly attenuated PDGF-BB-induced proliferation of control (Fig.2d, upper panel) and PAH pericytes (lower panel). As shown, this inhibitory effect was even stronger in PAH pericytes. Intriguingly, ANP (100nM) tended to have the opposite effect, slightly enhancing the proliferative actions of PDGF-BB in control and especially in PAH pericytes (Fig.2d). Corroborating these results of the BrdU assays, PDGF-BB markedly enhanced pericytes expression of the cell cycle protein cyclin D1 and of the proliferation marker, proliferating cell nuclear antigen (PCNA) and CNP pretreatment attenuated this effect in control and PAH pericytes (Fig.2e and f). Moreover, CNP significantly inhibited the promigratory effects of PDGF-BB in both groups of pericytes (Fig.2g).

Whereas PDGF-BB has been implicated in the enhanced proliferation and migration of pericytes in lungs from patients with PAH, enhanced levels of TGF- and augmented expression of TGF- receptors contribute to their hypercontractile phenotype7,8. As shown in Fig.3a, TGF- pretreatment (10ng/ml, 15min) did not affect pericytes cGMP responses to CNP. Therefore, we studied whether CNP/cGMP signaling attenuates the induction of -SMA by TGF-. In control pericytes, TGF- (10ng/ml, 24h) had very variable effects on the expression of -SMA, which on average were not significant (Fig.3b). In PAH pericytes, the expression of the TGF- receptor II was slightly enhanced (Supplemental Fig.1) and, concordantly, TGF- consistently induced -SMA expression (Fig.3b). Pretreatment with 10 and 100nM CNP for 30min significantly attenuated this TGF- effect in PAH pericytes (Fig.3c).

a TGF- (10ng/ml, 15min) did not significantly alter the cGMP responses of control pericytes to CNP (n=12 from four biological replicates; non-parametric KruskalWallis analyses). b Immunoblotting: TGF- (10ng/ml, 24h) enhanced -smooth muscle actin (-SMA) expression more in PAH pericytes than in controls (n=10 from five biological replicates in each group; non-parametric KruskalWallis analyses). c Immunoblotting: Pretreatment with CNP (10 and 100nM, 15min) significantly attenuated TGF- (10ng/ml, 24h)-induced -SMA expression in PAH pericytes (n=3 biological replicates; 1-way ANOVA). For a: *p<0.05 vs. PBS (). For b: *p<0.05 vs. PBS (), #p<0.05 vs. TGF- in controls. For c: *p<0.05 vs. PBS (), #p<0.05 vs. TGF-.

Our results confirm that pericytes from PAH patients exhibit exacerbated responses to growth factors such as PDGF-BB and TGF-. They add a novel and important piece of information, namely that CNP/cGMP signaling markedly attenuates the effects of these growth factors on pericyte proliferation, migration, and dedifferentiation.

Binding of PDGF-BB to the PDGFR- induces autophosphorylation of the receptor and activates a cascade of intracellular kinases, including AKT and extracellular signal-regulated kinases (ERK), ultimately increasing cell proliferation17. Therefore, next, we studied whether CNP prevents the activation of AKT and/or ERK by PDGF-BB. As before, cells were pretreated with CNP (10 and 100nM, 30min) followed by PDGF-BB (30ng/ml, for an additional 30min). Figure4a shows that the CNP-induced phosphorylation of VASP at Ser239 was preserved in the presence of PDGF-BB and unchanged in PAH pericytes, indicating unaltered activation of cGKI. As expected, PDGF-BB (30ng/ml, 30min) led to strong increases in the phosphorylation of AKT (at Ser473) and ERK 1/2 (at Thr202/Tyr204) (Fig.4b). Notably, CNP significantly prevented these effects in both control and PAH pericytes (Fig.4b).

a CNP (10 and 100nM, 15min) increased VASP (Ser239) phosphorylation in PDGF-BB-pretreated control ad PAH pericytes. b Pretreatment with CNP prevented the PDGF-BB (30ng/ml, 30min)-induced phosphorylation of AKT (Ser473) and ERK1/2 (Thr202/Tyr204) in control and PAH pericytes (n=4 from control and PAH pericytes; 1-way ANOVA). *p<0.05 vs. PBS (), #p<0.05 vs. PDGF-BB.

Among the many downstream targets of AKT and ERK signaling, the Forkhead box O (FoxO) transcription factors FoxO1 and FoxO3 have an important role in the pathogenesis of PAH18,19. Activation of AKT and/or ERK by PDGF-BB leads to FoxO phosphorylation at Thr32 and thereby to its nuclear exclusion and downregulation20,21,22. This pathway contributes to the enhanced proliferation of PASMCs in PAH as well as of fibroblasts in lung fibrosis18,21. Since FoxO3 is the predominant FoxO isoform expressed in pericytes23, we investigated whether CNP prevents PDGF-BB-mediated FoxO3 phosphorylation. PDGF-BB stimulation (30min) led to a significant increase in the phosphorylation of FoxO3 at Thr32 in both control and PAH pericytes, and CNP significantly reduced this effect (Fig.5a). Remarkably, 100nM CNP completely prevented the PDGF-BB-induced FoxO3 phosphorylation in PAH pericytes. In line with these results, immunocytochemistry showed that PDGF-BB (30ng/ml during 6h) significantly reduced nuclear FoxO3 localization in pericytes (Fig.5b depicts quantification of nuclear fluorescence). Pretreatment with CNP (100nM, 30min) prevented this effect, increasing nuclear FoxO3 (Fig.5b).

a CNP (10 and 100nM, 15min) prevented the PDGF-BB (30ng/ml, 30min)-induced phosphorylation of FoxO3 (Thr32) in control and PAH pericytes (n=4 control pericyte samples, n=5 from four PAH pericytes; 1-way ANOVA). b CNP (100nM) pretreatment prevented the PDGF-BB-induced FoxO3 nuclear exclusion as assessed by immunocytochemical staining of FoxO3, followed by nuclear fluorescence intensity measurement by Image J (n=5 from three biological replicates; each value is the mean of three images; 1-way ANOVA). Scale bar: 100mm. c and d Transfection of control pericytes with siFoxO3 reduced FoxO3 protein expression (c: n=6 independent experiments from three biological replicates; 1-way ANOVA), and this prevented the inhibitory effects of CNP on PDGF-BB induced pericyte proliferation (d: n=9 wells from three biological replicates; 2-way ANOVA). For a and b: *p<0.05 vs. PBS (), #p<0.05 vs. PDGF-BB. For c: *P<0.05 vs. untransfected control (), #p<0.05 vs. si-Control. For d: *p<0.05 vs. PBS (), #p<0.05 vs. PDGF-BB, $p<0.05 vs. corresponding vehicle-treated group ().

To determine whether FoxO3 participates in the antagonistic effects of PDGF-BB and CNP on pericyte proliferation, we tested the effect of siRNA-mediated FoxO3 knockdown. Immunoblotting revealed that FoxO3 siRNA (si-FoxO3) transfection for 48h led to a significant, marked reduction of FoxO3 protein in comparison to non-transfected and control siRNA (si-Control) transfected pericytes (Fig.5c). In the si-Control-transfected pericytes, PDGF-BB significantly increased proliferation and CNP prevented this effect (Fig.5d). In contrast, in the FoxO3-deficient cells the baseline proliferation was slightly enhanced, and PDGF-BB only exerted a minor additional effect. We assume that the observed overall duplication of the proliferation rate is the maximal achievable effect in control pericytes (see also Fig.2a). Notably, in such FoxO3 knock-down pericytes, the antiproliferative effect of CNP was fully abolished (Fig.5d). Together these results suggest that CNP antagonizes proliferative PDGF-BB signaling in lung pericytes by inhibiting AKT and ERK activation and subsequent FoxO3 phosphorylation. This stabilizes nuclear FoxO3 expression and fosters the antiproliferative effects of this transcription factor.

To elucidate whether the antiproliferative effect of the CNP/GC-B/cGMP pathway is mediated by activation of cGKI, we applied the specific cGKI inhibitor Rp-8-Br-PET-cGMPS. As shown in Supplemental Fig.2, pretreatment of control pericytes with Rp-8-Br-PET-cGMPS (100M, 20min) fully prevented the effect of CNP on VASP phosphorylation at Ser239 (the cGKI-specific site), demonstrating efficient cGKI inhibition. Notably, such inhibition of cGKI did not prevent the suppressing effect of CNP on PDGF-BB-induced ERKThr202/Tyr204 phosphorylation (Supplemental Fig.2). However, it fully abolished the inhibitory effect of CNP on the PDGF-BB-induced phosphorylation of AKT and FoxO3 (Fig.6a, middle and right panels). In line with these results, Rp-8-Br-PET-cGMPS (10M, 20min) did not alter the baseline or PDGF-BB-stimulated proliferation of pericytes but fully prevented the counter-regulation of such PDGF-BB effects by CNP (Fig.6b and c depict BrdU incorporation assays and PCNA immunoblotting, respectively).

a The cGKI inhibitor Rp-8-Br-PET-cGMPS (100M) prevented the effect of CNP on PDGF-BB (30ng/ml, 30min)-induced phosphorylation of AKT (Ser473) and FoxO3 (Thr32) (n=56 from four biological replicates; 2-way ANOVA). b and c Rp-8-Br-PET-cGMPS (10M) attenuated the inhibitory effects of CNP on PDGF-BB-induced proliferation of control pericytes as analysed by b BrdU incorporation assay (n=910 wells from three biological replicates; non-parametric KruskalWallis analyses) and c immunoblotting for PCNA (n=4 biological replicates; 2-way ANOVA). *p<0.05 vs PBS (), #p<0.05 vs. PDGF-BB, $p<0.05 vs. corresponding vehicle-treated group. Please note that in the samples derived from cells treated with Rp-8-Br-PET-cGMPS the immunoreactive signal obtained for total AKT was markedly increased (second lane in the original westerns depicted in (a)). Presently, we do not have an explanation for this reproducible observation. Due to the brief incubation time (<1h), we believe that this enhanced immunoreactivity is not derived from increased AKT protein expression but related to changes in AKT conformation and enhanced binding of the anti-AKT antibody to its epitope. Due to these changes, the signal of pAKTSer473 was normalized to GAPDH and not to total AKT (middle panel in Fig.6a).

To support these studies with the cGKI inhibitor, we also studied whether 8-Bromo-cGMP, a specific cGKI activator, mimics the effects of CNP. In the presence of PDGF-BB, 8-Bromo-cGMP (0.110M, 30min) increased VASPSer239 phosphorylation in concentration-dependent manner, demonstrating cGKI activation (Supplemental Fig.3a). Concordantly, 8-Bromo-cGMP significantly and concentration-dependently reduced the stimulatory effects of PDGF-BB on AKT and FoxO3 phosphorylation as well as on pericyte proliferation (Supplemental Fig.3b and c). These results, taken together, clearly demonstrate that the inhibitory effects of CNP/GC-B/cGMP signaling on the proliferative PDGF-BB/AKT/FoxO3 pathway are mediated by cGKI. But how can the activation of a kinase prevent the phosphorylation of AKT?

Phosphatase and tensin homolog (PTEN) is a phosphatase that inhibits the PI3K/AKT pathway. Phosphorylation of PTEN at Ser380/Thr382/383 by Rho-associated kinase 1 (ROCK1) or Glycogen synthase kinase-3 beta (GSK-3) reduces its phosphatase activity22,24. On the other hand, cGKI phosphorylates RhoA at Ser188 (essential for activation of ROCK1) and GSK-3 at Ser9, resulting in their inactivation25,26. Therefrom, we hypothesized that CNP, via a cGKI-mediated inhibition of RhoA and/or GSK-3, leads to PTEN activation and thereby attenuates the effects of PDGF-BB on the PI3K-AKT axis (a scheme illustrating this pathway is depicted in Fig.7a). To follow up this hypothesis we studied the phosphorylation of PTEN at Ser380/Thr382/383. As shown in Fig.7b, incubation of control pericytes with PDGF-BB (30ng/ml, 30min) led to an increase in PTEN phosphorylation at Ser380/Thr382/383, which was prevented by CNP (100nM, 30min pretreatment). Inhibition of cGKI with Rp-8-Br-PET-cGMPS evoked a mild increase of PTEN phosphorylation at baseline, which was further increased by PDGF-BB (Fig.7b). As also shown, the inhibitory effect of CNP on PDGF-BB-driven PTEN phosphorylation was fully prevented. Conversely, activation of cGKI with 8-Bromo-cGMP (0.110M, 20min pretreatment) prevented PDGF-BB-induced PTEN phosphorylation, thereby mimicking the effects of CNP (Fig.7c). Our observations indicate that the CNP/cGMP/cGKI pathway indirectly leads to the activation of the phosphatase PTEN, thereby attenuating the PDGF-BBinduced AKT phosphorylation and AKT-mediated inactivation of FoxO3 (scheme in Fig.7a).

a Scheme of the postulated and investigated signaling pathway. PDGF-BB, via its PDGFR- , triggers the activation of the PI3K, which phosphorylates PIP2 to PIP3. PIP3 then activates PDK1-AKT signaling. Subsequently, AKT phosphorylates and inactivates FoxO3, which enhances lung pericyte proliferation. CNP, via GC-B/cGMP signaling, activates cGKI. It has been shown that cGKI elicits inactivating phosphorylations of RhoA at Ser188 and of GSK3b at Ser9, thereby preventing their inhibitory phosphorylations of PTEN27,28. Activated PTEN dephosphorylates PIP3 and prevents AKT activation, resulting in an increase of nuclear FoxO3 and a concomitant reduction in pericyte proliferation. b PDGF-BB stimulated the phosphorylation of PTEN at Ser380/Thr382/383 and CNP prevented this effect in the absence (vehicle) but not in the presence of the cGKI inhibitor Rp-8-Br-PET-cGMPS (100M). c The cGKI activator, 8-Bromo-cGMP (0.0110M), prevented PDGF-BB (30ng/ml)-induced phosphorylation of PTEN in control pericytes (b and c: n=4 biological replicates; b: 2-way ANOVA; c: 1-way ANOVA). *p<0.05 vs. PBS (), #p<0.05 vs. PDGF-BB, $p<0.05 vs. corresponding vehicle-treated group. PDGF-BB platelet-derived growth factor-BB, PDGFR- platelet-derived growth factor beta, PI3K phosphoinositide 3-kinase, PIP2 phosphatidylinositol 4,5-bisphosphate, PIP3 phosphatidylinositol 3,4,5-trisphosphate, PDK1 phosphoinositide-dependent kinase 1, AKT protein kinase B, FoxO3 forkhead box O3, CNP C-type natriuretic peptide, GC-B guanylyl cyclase-B, cGMP cyclic guanosine monophosphate, cGKI cGMP-dependent kinase I, RhoA Ras homolog family member A, PTEN phosphatase and tensin homolog.

Taken together, our results demonstrate that exogenous, synthetic CNP, via GC-B/cGMP signaling, counteracts the growth factor-induced proliferation, migration, and transdifferentiation of pericytes from patients with PAH. This raises the question of whether the endogenous endothelial hormone exerts local protective effects in pulmonary microcirculation and whether such effects are preserved in PAH. To approach this question, firstly, we studied lung CNP and GC-B expression levels in two experimental models of PH: Monocrotaline (MCT)-induced PH in rats27 and milder, chronic hypoxia (HOX)-induced PH in mice28. Quantitative real-time RT-PCR (qRT PCR) revealed that the CNP expression levels were significantly reduced in lung samples from rats and mice with PH in comparison to respective controls (Fig.8a and b, left panels). As also shown, GC-B expression was unchanged in MCT rats, while it was significantly reduced in HOX-exposed mice in comparison to control lungs (Fig.8a and b, right panels). To follow up on the possible clinical relevance, we also studied lung samples from PAH patients. In line with the experimental models, CNP levels were significantly reduced in human PAH as compared to control lungs, whereas GC-B levels were unaltered (Fig.8c).

a and b Lung CNP and GC-B mRNA expression in a monocrotaline (MCT) vs. vehicle-treated rats, b normoxia (NOX) vs. Hypoxia (HOX: 21 days) exposed mice and c PAH patients vs. controls. Values are the ratios of CNP or GC-B mRNA level relative to GAPDH (b) or b2 microglobulin (a and c), determined by qRT-PCR and expressed as x-fold versus vehicle-treated rats (a), NOX mice (b), or human control lungs (c) (a: n=9 samples from the vehicle and n=8 from MCT-treated rats; b: n=10 samples each from NOX vs. HOX exposed mice; a and b: MannWhitney test for CNP and unpaired 2-tailed Students t-test for GC-B; c: n=9 samples from controls and n=12 samples from PAH patients, unpaired 2-tailed Students t-test). *p<0.05.

More here:
C-type natriuretic peptide/cGMP/FoxO3 signaling attenuates hyperproliferation of pericytes from patients with ... - Nature.com