Inflammatory Glycoprotein 130 Signaling Links Changes in Microtubules and Junctophilin-2 to Altered Mitochondrial Metabolism and Right Ventricular Contractility

Abstract

Background:

Right ventricular dysfunction (RVD) is the leading cause of death in pulmonary arterial hypertension (PAH), but no RV-specific therapy exists. We showed microtubule-mediated junctophilin-2 dysregulation (MT-JPH2 pathway) causes t-tubule disruption and RVD in rodent PAH, but the druggable regulators of this critical pathway are unknown. Glycoprotein 130 (GP130) activation induces cardiomyocyte microtubule remodeling in vitro, however the effects of GP130 signaling on the MT-JPH2 pathway and RVD resulting from PAH are undefined.

Methods:

Immunoblots quantified protein abundance, quantitative proteomics defined RV microtubule-interacting proteins (MT-interactome), metabolomics evaluated the RV metabolic signature, and transmission electron microscopy (TEM) assessed RV cardiomyocyte mitochondrial morphology in control, monocrotaline (MCT), and MCT-SC-144 (GP130 antagonist) rats. Echocardiography and pressure-volume loops defined the effects of SC-144 on RV-pulmonary artery coupling in MCT rats (8–16 rats per group). In 73 PAH patients, the relationship between interleukin-6, a GP130 ligand, and RVD was evaluated.

Results:

SC-144 decreased GP130 activation, which normalized MT-JPH2 protein expression and t-tubule structure in the MCT RV. Proteomics analysis revealed SC-144 restored RV MT-interactome regulation. Ingenuity pathway analysis of dysregulated MT-interacting proteins identified a link between microtubules and mitochondrial function. Specifically, SC-144 prevented dysregulation of electron transport chain, Krebs cycle, and the fatty acid oxidation pathway proteins. Metabolomics profiling suggested SC-144 reduced glycolytic dependence, glutaminolysis induction, and enhanced fatty acid metabolism. TEM and immunoblots indicated increased mitochondrial fission in the MCT RV, which SC-144 mitigated. GP130 antagonism reduced RV hypertrophy and fibrosis and augmented RV-pulmonary artery coupling without altering PAH severity. In PAH patients, higher interleukin-6 levels were associated with more severe RVD (RV fractional area change 23±12% vs. 30±10%, p=0.002).

Conclusions:

GP130 antagonism reduces MT-JPH2 dysregulation, corrects metabolic derangements in the RV, and improves RVD in MCT rats.

INTRODUCTION

Pulmonary arterial hypertension (PAH) is a progressive vasculopathy that increases pulmonary arterial pressures and reduces pulmonary arterial compliance (PAC)¹. The pathological alterations in the pulmonary circuit ultimately manifest as right ventricular dysfunction (RVD). Although RVD is the strongest predictor of mortality in PAH², the molecular mediators of RVD are understudied³. This knowledge gap may explain the absence of pharmaceuticals that directly combat RVD pathophysiology. Unfortunately, medications used for left ventricular failure have not yielded similar success when applied to RVD³. Thus, there is an urgent need to develop RV-directed therapies.

We previously demonstrated the importance of microtubules in contributing to the genesis of RVD by a mechanism involving dysregulation of junctophilin-2 in preclinical models of PAH with RVD⁴ (MT-JPH2 pathway). In monocrotaline (MCT) PAH, there is a chamber-specific microtubule stabilization in RV cardiomyocytes, which lowers JPH2 levels and leads to t-tubule structural derangements that impair RV contractility⁴. Importantly, colchicine-mediated microtubule depolymerization increases JPH2 levels, combats pathological t-tubule remodeling, and augments RV function in MCT rats⁴, a validated model of RV failure³. Furthermore, Xie showed the degree of JPH2 reduction and t-tubule derangements correlates with the severity of RV failure in MCT rats⁵. Thus, the MT-JPH2 pathway is critical for RV cardiomyocyte function via its impact on modulation of t-tubule structure, calcium handling, and cardiomyocyte contractility⁶. Unfortunately, our colchicine results may not readily translate to human RVD because the equivalent human dose may have toxicities⁷. To capitalize on the therapeutic potential of restoring the MT-JPH2 pathway, we chose to focus on upstream regulators to identify a druggable target for RVD.

Glycoprotein 130 (GP130) is the master membrane receptor of the interleukin-6 (IL-6) cytokine superfamily. GP130 downstream signaling molecules include the Janus kinase/signal transducer and activator of transcription (JAK/STAT), phosphatidylinositol 3-kinase (PI3K), and mitogen activated protein kinase (MAPK) pathways⁸. However, STAT3 is believed to be the predominant intracellular effector protein⁹. GP130 has direct relevance to the MT-JPH2 pathway as GP130 stimulation stabilizes microtubules in neonatal cardiomyocytes¹⁰. Importantly, STAT3 phosphorylation is essential for GP130-mediated microtubule remodeling¹⁰. At present, the role of GP130 signaling in RVD due to PAH is undefined. However, clinical studies suggest GP130 might modulate RV function as IL-6 levels are independently associated with RVD in PAH patients¹¹. Moreover, multiple PAH cohort studies show elevated IL-6 levels predict worse survival rates despite minimal differences in pulmonary vascular disease severity^11–14.

Remodeling of cardiomyocyte microtubules alters t-tubule integrity and contractility¹⁵, but the detrimental effects of microtubule perturbations on other areas of cardiomyocyte cell biology are unexplored. Chemical modulation of microtubules disrupts the balance of mitochondrial fission and fusion in budding yeast¹⁶, which ultimately results in mitochondrial dysfunction. This finding may be applicable to the RV because excess mitochondrial fission promotes RVD in rodent PAH¹⁷. A recent transcriptomic study of the RV in MCT rats and human PAH identifies mitochondrial metabolic dysfunction and inflammation as the two most dysregulated pathways in both species¹⁸, but the molecules that promote crosstalk between these two pathological entities are unknown. We speculate GP130-induced microtubule dysregulation mediates the intersection of inflammation and the acquired imbalance of mitochondrial fission/fusion and subsequent metabolic dysfunction of the RV in PAH.

In this study, we hypothesized increased GP130 signaling induces pathological microtubule remodeling, which causes RVD by induction of JPH2 downregulation and mitochondrial metabolic dysfunction. We used a small molecule GP130 antagonist, starting two weeks after MCT injection, to reverse microtubule-mediated JPH2 downregulation and t-tubule disruption in MCT rats. In addition, we employed quantitative proteomics to define the effects of GP130 antagonism on the microtubule-interacting protein (MT-interactome) fraction of the RV. We also used transmission electron microscopy (TEM) and global metabolomics profiling to determine how GP130-mediated microtubule remodeling altered mitochondrial structure and metabolic function. We subsequently assessed how inhibition of GP130 signaling affected RV-pulmonary artery coupling using both echocardiography and high-fidelity cardiac catheterization generated pressure-volume loops. Finally, we examined the relationship between serum IL-6 levels and RV function in human PAH. Our findings establish a role for GP130 signaling in RVD and show that this inflammatory pathway dysregulates mitochondrial form and function in a manner which contributes to RVD.

METHODS

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Detailed methods are described in the Supplemental Appendix. Animal studies were approved by the University of Minnesota Institutional Animal Care and Use Committee and procedures were followed in accordance with institutional guidelines. Briefly, male Sprague Dawley rats received a single subcutaneous injection of MCT or phosphate buffered saline. Two weeks after MCT injection (60 mg/kg), rats received either daily intraperitoneal injections of the GP130 antagonist, SC-144 (10 mg/kg), or a dimethyl sulfoxide and propylene glycol vehicle for 10 days. Immunoblots quantified levels of cardiac protein abundance¹⁹. Antibodies used are described in Supplemental Table I.

Cosedimentation of the MT-interactome in RV extracts was completed as described²⁰. Metabolomic profiling of frozen RV free wall specimens was performed by Metabolon Inc. (Durham, NC)¹⁹ and detailed in the supplement. Echocardiography, pressure-volume loops, and pulmonary vascular histology assessed the effects of SC-144 on RV function and pulmonary vascular disease severity. Finally, we examined the relationship between the GP130 agonist, IL-6 and RVD in 73 PAH patients (Supplemental Table II). This study was approved by the University of Minnesota Institutional Review Board and the participants gave written informed consent.

Complete description of the statistical methods is in the supplement. Statistical analysis and graphing were performed on GraphPad Prism version 9 except for principal component analysis, hierarchical cluster analysis, and random forest analysis. Data are presented as mean±standard deviation.

RESULTS

SC-144 mitigated GP130 activation in the MCT RV

First, we investigated how MCT-PAH affected GP130 pathway activation in the RV. MCT-Vehicle (MCT-V) rats had elevated expression of GP130, STAT3, phosphorylated STAT3 (pSTAT3), and the pSTAT3/STAT3 ratio as compared to controls. However, SC-144 reduced RV expression of GP130, STAT3, pSTAT3, and the pSTAT3/STAT3 ratio to near control levels (Figure 1A–B). Immunofluorescence analysis of RV cardiomyocytes showed increased GP130 membrane localization (Figure 1C–D) and pSTAT3 positive nuclei (Figure 1E–F) in MCT-V as compared to control. SC-144 significantly decreased the immunoreactivity of both GP130 and pSTAT3 (Figure 1C–F) when compared to MCT-V.

Figure 1: GP130 antagonism blunted RV STAT3 activation, normalized the MT-JPH2 pathway, and restored t-tubule architecture.

(A) Representative Western blots and (B) quantification of protein abundance in RV extracts from control, MCT-V, and SC-144 rats demonstrated GP130 inhibition normalized expression of GP130, STAT3, pSTAT3, and the ratio of pSTAT3/STAT3. Data shown as expression relative to control (n=4 per group). Representative confocal micrographs of RV free wall sections showed SC-144 reduced the amount of (C) GP130 receptors (white arrows) (green: GP130, blue: DAPI, red: wheat germ agglutinin) at the cell membrane as quantified in (D) as amount of GP130 expression per area (%) and amount of (E) pSTAT3 (white arrows) (green: pSTAT3, blue: DAPI) in cardiomyocyte nuclei as quantified in (F) as percent pSTAT3 positive nuclei per area. n=3–4 animals per group. (G) Representative Western blots and (H) protein quantification showed GP130 antagonism reduced expression of α- and β-tubulin, detyrosinated α-tubulin, and MAP4 and increased JPH2. Data shown as expression relative to control (n=4 per group). (I) Representative confocal images of RV free wall sections stained with Alexa Fluor-633 conjugated wheat germ agglutinin. SC-144 restored RV t-tubule architecture (red arrows). (J) Quantification of t-tubule architecture and organization by TTorg (arbitrary units are TTpower). n=3 animals per group, ≥7 cardiomyocytes were quantified per animal. Unpaired t-test was used to compare MCT-V vs. SC-144 in (B) and (H). ***p<0.001, ****p<0.0001, and (ns) not significantly different as assessed by Kruskal-Wallis ANOVA with Dunn post-hoc test in (D) and (J) and Brown-Forsythe and Welch ANOVA with Dunnett post-hoc analysis in (F).

SC-144 prevented dysregulation of the MT-JPH2 pathway and adverse t-tubule remodeling in the RV

Next, we performed an in silico analysis of potential STAT3-regulated microtubule proteins by querying the human STAT3 ChIP-seq database²¹. STAT3 was predicted to induce transcription of multiple tubulin isoforms and microtubule associated protein 4 (MAP4), a microtubule stabilizing protein²² (Supplemental Table III). Quantitative RT-PCR showed most tubulin isoforms and Map4 transcripts were elevated in MCT-V compared to control, but SC-144 only slightly and non-statistically reduced expression of Tuba1a and Map4 (Supplemental Table IV). Therefore, we investigated whether pSTAT3 may be physically associated with microtubules. Western blot analysis of the MT-interacting proteins identified higher levels of pSTAT3 in MCT-V compared to control and SC-144 (Supplemental Figure I).

In RV extracts, α-tubulin, β-tubulin, detyrosinated α-tubulin (tubulin found in stabilized microtubules²³), and MAP4 protein levels were elevated in MCT-V when compared to control (Figure 1G–H). Microtubule remodeling resulted in downregulation of JPH2 in the MCT-V RV, consistent with previous findings^{4, 24}. SC-144 prevented upregulation of all tubulin isoforms and MAP4 and normalized JPH2 expression levels (Figure 1G–H). We subsequently evaluated RV t-tubule architecture as a structural readout of altered JPH2 regulation. Control RV cardiomyocytes displayed a highly regular, striated t-tubule staining pattern, but MCT-V RV cardiomyocytes exhibited near complete loss of organized t-tubule structure (Figure 1I–J). However, RV t-tubule morphology was restored with SC-144 (Figure 1I–J).

In contrast to the RV, the GP130 and MT-JPH2 pathways were minimally altered in the MCT-V left ventricle (LV) (Supplemental Figure II). Expression of GP130, STAT3, tubulins, and JPH2 were not different between MCT-V and control LV. SC-144 did not change the expression of GP130, STAT3, pSTAT3, or pSTAT3/STAT3 in the LV. These results suggested GP130-mediated MT-JPH2 dysregulation is confined to the RV, and it is not a systemic effect of MCT.

Quantitative proteomics identified a link between microtubule remodeling and mitochondrial metabolic dysregulation

To delineate the effects of microtubule remodeling on other aspects of RV cardiomyocyte biology, we used quantitative mass spectrometry to define the RV MT-interactome using a tissue-based microtubule cosedimentation assay. We identified 2854 MT-interacting proteins in RV extracts and 1032 displayed significant variation in expression when comparing the three groups. In MCT rats, GP130 antagonism shifted the RV MT-interactome signature towards control as depicted by principal component (Figure 2A) and hierarchical cluster analyses (Figure 2B). Ingenuity pathway analysis of the dysregulated MT-interacting proteins revealed mitochondrial function and oxidative phosphorylation were the two most significantly enriched pathways (Figure 2C). Hierarchical cluster analysis demonstrated altered regulation of proteins in complexes I-V of the electron transport chain, the tricarboxylic acid (TCA) cycle, and the fatty acid oxidation pathway in MCT-V RVs. SC-144 normalized expression levels of nearly all these key mitochondrial metabolic proteins (Figure 2D–J). Collectively, these data suggested microtubule remodeling in the MCT-V RV modulated the expression of mitochondrial proteins that bound microtubules, and SC-144 corrected these changes.

Figure 2: Quantitative proteomic analysis of the MT-interactome in the RV revealed a link between microtubule remodeling and mitochondrial protein expression.

(A) Principal component analysis showed GP130 antagonism partially restored the global expression signature of the MT-interactome. (B) Hierarchical cluster analysis of MT-interacting proteins demonstrated the expression pattern of the MT-interactome in the RV of SC-144 rats more closely resembled control than MCT-V. (C) Ten most significantly enriched pathways identified using Ingenuity pathway analysis of dysregulated MT-interacting proteins in MCT-V RV when compared to control. The two most enriched pathways were mitochondrial dysfunction and oxidative phosphorylation. Hierarchical cluster analysis of proteins in complex I (D), complex II (E), complex III (F), complex IV (G), and complex V (H), the TCA cycle (I), and fatty acid oxidation (J). SC-144 corrected dysregulation of nearly all mitochondrial metabolic proteins.

GP130 antagonism corrected RV metabolism

Global metabolomic profiling of 767 metabolites in RV free wall specimens defined the impact of SC-144 on RV metabolism. Hierarchical cluster analysis showed a distinct metabolic profile between control and MCT-V RVs, but most SC-144 rats clustered with controls (Figure 3A). Random forest analysis identified 15 metabolites with the highest potential to differentiate control, MCT-V, and SC-144. There were dysregulated metabolites in multiple metabolic pathways in MCT-V RV, which were corrected with SC-144 (Figure 3B). Specific pathway profiling suggested increased glycolytic metabolism and glutaminolysis induction in the MCT-V RV, which SC-144 blunted (Figure 3C–D). Furthermore, there was evidence of impaired fatty acid oxidation in the MCT-V RV as nearly all acylcarnitines were reduced. SC-144 actually increased acylcarnitine abundance, suggestive of augmented fatty acid metabolism (Figure 3E). Consistent with our targeted analysis, computational integration of our proteomics and metabolomics analyses identified the TCA cycle, purine metabolism, pyruvate metabolism, glycolysis or gluconeogenesis, and fatty acid degradation as the most altered metabolic pathways in the MCT-V RV (Supplemental Figure III). In summary, these data showed GP130 antagonism corrected deficits in multiple metabolic pathways in the RV.

Figure 3: SC-144 improved the RV metabolic signature, restored mitochondrial morphology, and corrected mitochondrial fission/fusion imbalance.

(A) Hierarchical cluster analysis demonstrated SC-144 normalized the global RV metabolic signature. (B) Random forest analysis highlighting the 15 metabolites that differentiate control, MCT-V, and SC-144. Hierarchical cluster analysis of (C) glycolysis, and (D) glutaminolysis identified a normalization of multiple metabolic pathways with SC-144 while most of the (E) acylcarnitine metabolites were increased by SC-144. (F) Representative electron micrographs of mitochondria. GP130 antagonism partially corrected the RV mitochondrial morphology as assessed by (G) mitochondrial area and (H) eccentricity index. GP130 antagonism reduced the amount of large, swollen mitochondria as quantified in (G) and restored the normal elongated mitochondrial shape as assessed in (H). n=3 RV per group, with >70 mitochondria measured per animal. (I) Representative Western blots and (J) quantification of protein abundance in RV extracts from control, MCT-V, and SC-144 demonstrated GP130 inhibition did not affect MFN1 expression and minimally changed MFN2 but normalized OPA1 (non-statistical change), FIS1, and DRP1 expression. *p<0.05, ****p<0.0001, and (ns) no statistical difference as determined by Brown-Forsythe ANOVA with Dunnett post-hoc test after transformation in (G) and (H) and unpaired t-test in (J).

SC-144 restored mitochondrial morphology via normalization of fission/fusion balance

Next, we used TEM to examine RV cardiomyocyte mitochondrial morphology to determine if disruption of the mitochondrial fission/fusion balance contributed to the observed metabolic defects. As compared to control mitochondria, MCT-V mitochondria were significantly larger and more spherical in shape (Figure 3F–H), consistent with excess fission. Importantly, SC-144 corrected mitochondrial morphology (Figure 3F–H). To supplement our TEM analysis, we evaluated the effects of SC-144 on RV mitochondrial fission/fusion protein regulation. When compared to controls, MCT-V RV protein expression profile favored mitochondrial fission as there was reduced abundance of the pro-fusion protein, optic atrophy protein 1 (OPA1) and increased expression of the pro-fission proteins, mitochondrial fission 1 (FIS1) and dynamin-related protein 1 (DRP1) (Figure 3I–J). There was no change of expression of the fusion mediators, mitofusin (MFN)-1 and MFN-2, in MCT-V RV. GP130 antagonism led to a small but significant decrease in MFN2, prevented downregulation of OPA1 (non-statistically significant difference) and significantly reduced upregulation of both FIS1 and DRP1 (Figure 3I–J). Thus, the summation of our TEM findings and protein expression changes suggested SC-144 prevented excess mitochondrial fission.

Microtubule stabilization impaired mitochondrial metabolism in vitro

To examine the effects of microtubule remodeling on mitochondrial metabolic capacity, we quantified the effects of paclitaxel, a microtubule-stabilizing compound, on Seahorse micropolarimetry-defined mitochondrial function. Paclitaxel depressed H9c2 cardiomyocyte mitochondrial metabolic activity, evident in that maximal respiration, spare capacity, ATP production, and coupling efficiency were reduced and proton leak was enhanced (Figure 4A–B). These results linked microtubule stability and mitochondrial metabolic dysfunction.

Figure 4: Chemical stabilization of microtubules with paclitaxel in H9c2 cardiomyocytes altered mitochondrial metabolic function.

(A) Seahorse analysis of the oxygen consumption rate profile and (B) individual parameters of mitochondrial respiration. Con: control; OCR: oxygen consumption rate; Pac: paclitaxel. Data presented in (A) as mean ± standard deviation. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, and (ns) no statistical difference as determined by t-test or Mann-Whitney U-test.

Inhibition of GP130 signaling did not affect pulmonary vascular disease

Because studies showed IL-6 promotes adverse pulmonary vascular remodeling²⁵, we quantified the effects of SC-144 on PAH severity. MCT-V and SC-144 rats displayed a nearly identical PAH phenotype, evident by similar pulmonary artery acceleration time (Control: 34.0±5.8, MCT-V: 14.1±3.3, SC-144: 15.0±3.9 ms, p=0.87 between MCT-V and SC-144), right ventricular systolic pressure (RVSP) (Control: 29.2±7.4, MCT-V: 71.0±17.4, SC-144: 64.5±32.5 mmHg, p=0.86 between MCT-V and SC-144), and effective arterial elastance (Ea) (Control: 0.20±0.07, MCT-V: 0.61±0.36, SC-144: 0.58±0.39 mmHg/μl, p>0.999 between MCT-V and SC-144) (Supplemental Figure IV A–C). Furthermore, the percent medial thickness of small pulmonary arterioles was similarly increased in MCT-V and SC-144 rats as compared to controls (Control: 27.1±7.0, MCT-V: 51.6±16.7, SC-144: 52.0±15.4%, p>0.999 between MCT-V and SC-144) (Supplemental Figure IV D–E). Thus, SC-144 did not alter PAH severity in established PAH.

GP130 antagonism decreased RV hypertrophy and fibrosis

Next, we evaluated the effects of SC-144 on RV hypertrophy. Fulton index (Control: 0.16±0.05, MCT-V: 0.42±0.10, SC-144: 0.28±0.11) and RV mass normalized to body mass (RV/BW) (Control: 0.33±0.11, MCT-V: 0.89±0.21, SC-144: 0.57±0.26 mg/g) were significantly reduced in the SC-144 rats compared to MCT-V (Figure 5A–B). Furthermore, SC-144 significantly decreased cardiomyocyte cross-sectional area compared to MCT-V (Control: 295±116, MCT-V: 493±172, SC-144: 323±90 μm²) (Figure 5C–D). However, SC-144 did not completely prevent RV hypertrophy as Fulton index was higher than control (Figure 5A).

Figure 5: GP130 antagonism reduced RV hypertrophy and fibrosis.

SC-144 reduced RV hypertrophy as assessed by the Fulton index (RV/LV+S) (A) and RV weight normalized to body weight (B). n=10–26 animals per group. (C) SC-144 decreased cardiomyocyte area. (D) Representative images of cardiomyocytes in H&E stained RV free wall sections. n=2–3 RV per group, at least 21 cardiomyocytes measured per animal. (E) Representative Western blots and (F) quantification of collagen I and III protein abundance in RV extracts. n=4 per group. SC-144 decreased expression of collagen I and III protein expression compared to MCT-V, which corresponded with less RV fibrosis observed in trichrome staining of RV free wall as quantified in (G). n=2–3 RV per group, 3–4 areas sampled per animal. (H) Representative trichrome stained RV sections. White arrows highlight RV fibrosis. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, and (ns) no statistical difference as determined by one-way ANOVA with Tukey post-hoc analysis in (A) and Kruskal-Wallis ANOVA with Dunn post-hoc test in (B), (C), and (G). Unpaired t-test between MCT-V and SC-144 was completed in (F).

Then, we examined how SC-144 regulated RV fibrosis. MCT-V rats had higher RV collagen I and III protein levels than controls, but SC-144 prevented collagen I/III protein accumulation (Figure 5E–F). Likewise, GP130 antagonism reduced RV fibrosis (Control: 1.5±0.9, MCT-V: 5.5±3.2, SC-144: 2.3±1.4%) (Figure 5G–H).

Inhibition of GP130 signaling enhanced RV systolic function and survival

Both echocardiography and pressure-volume loop analysis demonstrated impaired RV systolic function in MCT-V rats as compared to controls (Figure 6). However, SC-144 augmented RV function as all the following parameters were higher in SC-144 than MCT-V rats: tricuspid annular plane systolic excursion (TAPSE) (Control: 2.8±0.4, MCT-V: 1.9±0.4, SC-144: 2.4±0.5 mm), percent change in RV free wall thickness (Control: 96±34, MCT-V: 25±17, SC-144: 75±33%), stroke volume (Control: 0.38±0.10, MCT-V: 0.15±0.07, SC-144: 0.28±0.11 mL), cardiac output (Control: 131±30, MCT-V: 49±27, SC-144: 101±45 mL/min), and cardiac output normalized to body weight (Control: 0.28±0.06, MCT-V: 0.15±0.08, SC-144: 0.31±0.12 mL/min/g) (Figure 6A–E). Moreover, invasive hemodynamic studies showed GP130 antagonism improved RV ejection fraction (Control: 91±9, MCT-V: 63±12, SC-144: 79±16%), RV end-systolic elastance (Ees) (Control: 0.3±0.1, MCT-V: 0.2±0.1, SC-144: 0.5±0.4 mmHg/μl), and RV-pulmonary artery coupling (Ees/Ea) (Control: 1.8±0.6, MCT-V: 0.4±0.3, SC-144: 0.9±0.3) (Figure 6F–H). SC-144 prevented premature mortality (Supplemental Figure V).

Figure 6: SC-144 improved RV function.

(A) TAPSE, (B) percent RV free wall thickness change, (C) stroke volume, (D) cardiac output, and (E) cardiac output normalized to body weight were measured by echocardiography. n=9–14 rats per group. SC-144 improved RV function in all echocardiographic measures. Invasive hemodynamics and pressure-volume (PV) loops demonstrated that SC-144 augmented (F) RV ejection fraction (RVEF), (G) end-systolic elastance (Ees), and (H) RV-PA coupling (Ees/Ea). n=8–16 rats per group. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, and (ns) no statistical difference as determined by one-way ANOVA with Tukey post-hoc analysis for (A), (B), (C), and (F), Kruskal-Wallis ANOVA with Dunn post-hoc test in (D), (E), and (H), and Brown-Forsythe ANOVA with Dunnett post-hoc analysis in (G).

Elevated IL-6 levels in PAH patients were associated with worse RV function

Finally, we addressed the translatability of our preclinical studies by examining how serum levels of the GP130 ligand, IL-6, were related to RV function in 73 PAH patients (Supplemental Table II). After dichotomizing our cohort by median IL-6 level, patients with elevated IL-6 levels had higher N-terminal pro B-type natriuretic peptide levels (NT-proBNP) (median level 2576 vs. 599 pg/ml, p=0.001) and lower RV fractional area change (RVFAC) (23±12 vs. 30±10%, p=0.005) (Figure 7A–B). When we plotted the relationship between RVFAC and mean pulmonary arterial pressure (mPAP), patients with higher IL-6 levels had lower RVFAC at all corresponding mPAPs (p=0.02) (Figure 7C). Likewise, patients with high IL-6 levels had decreased RVFAC at all pulmonary vascular resistance (PVR) values (p=0.02) (Figure 7D). These results suggested IL-6 had negative RV inotropic properties. Consistent with an RV-predominant effect of IL-6, there were no significant differences in mPAP (48±16 vs. 45±14 mmHg, p=0.35), PVR (9.2±5.6 vs.7.5±4.4 Wood units, p=0.28), and PAC (1.6±1.1 vs.1.9±1.1 mL/mmHg, p=0.14) between the two groups (Figure 7E–G). Lastly, when we divided our cohort into tertiles, patients with the highest IL-6 levels had elevated NT-proBNP levels and reduced RVFAC compared to the middle and lowest IL-6 tertiles (Supplemental Figure VI A–B). When we defined the relationships between RVFAC and measures of RV afterload (mPAP and PVR), patients with the highest IL-6 levels had the lowest predicted RVFAC of the three tertiles (Supplemental Figure VI C–D).

Figure 7: Higher IL-6 levels in PAH patients were associated with worse RV function independent of changes in PAH severity.

Higher IL-6 levels are associated with higher NT-proBNP levels (A) and lower RVFAC (B). (C) Relationship between RVFAC and mPAP. Patients with higher IL-6 levels had lower RVFAC at each mPAP compared to patients with lower IL-6 levels (statistical difference in y-intercept, p=0.02; no difference in slope, p=0.84). (D) Relationship between RVFAC and PVR. PAH patients with higher IL-6 had reduced RVFAC at each PVR compared to patients with lower IL-6 (statistical difference in y-intercept, p=0.02; no difference in slope, p=0.56). n=73 total patients with the two groups stratified by median IL-6 level. There were no differences in (E) mPAP (p=0.35), (F) PVR (p=0.28), or (G) PAC (p=0.14) between patients with higher IL-6 levels compared to those with lower IL-6 levels. **p<0.01 and (ns) not significant as determined by Mann-Whitney U-test in (A), (B), (F), and (G) and unpaired t-test in (E). Linear regression evaluated differences between the lower and higher IL-6 curves in (C) and (D).

DISCUSSION

In this study, we demonstrate GP130 signaling causes many of the observed pathological subcellular changes in the PAH RV (microtubule remodeling, JPH2 downregulation, t-tubule derangements, and mitochondrial dysfunction) that cause RVD (Figure 8). Small molecule antagonism of GP130 mitigates STAT3 activation and restores expression of α- and β-tubulin, detyrosinated α-tubulin, and MAP4. The normalization of the microtubule cytoskeleton reverses pathological t-tubule remodeling, restores mitochondrial morphology and metabolic function, decreases RV hypertrophy, enhances RV function, and improves survival. This study also establishes the effects of the GP130 pathway on mitochondrial metabolism. Dysregulation of proteins in complexes I-V of the electron transport chain, the TCA cycle, and the fatty acid oxidation pathway in MCT-V RVs occurred, consistent with increased glycolytic metabolism and glutaminolysis and impaired fatty acid oxidation. The metabolic changes were all improved by GP130 antagonism. Importantly, all these molecular and physiological changes occur independently of pulmonary vascular disease severity, demonstrating a RV-specific effect of GP130 antagonism. Finally, in PAH patients, elevated serum levels of the GP130 agonist, IL-6 are associated with more severe RVD, suggesting targeting GP130 signaling may have translational utility.

Figure 8: GP130-mediated RV dysfunction in PAH.

Heightened GP130 activation may induce STAT3-mediated gene transcription of microtubule proteins and pSTAT3 may interact with microtubules. GP130 activation causes microtubule remodeling, which leads to JPH2 dysregulation, t-tubule derangements, and mitochondrial dysfunction via imbalance of mitochondrial fission/fusion. These molecular changes ultimately manifest as RV dysfunction.

The importance of the MT-JPH2 pathway and t-tubule remodeling in cardiac dysfunction are well-documented⁶. Microtubule-mediated JPH2 disorganization causes t-tubule disruption and LV failure in rodent pressure-overload²⁴. Furthermore, overexpression of JPH2 prevents loss of t-tubules and improves LV function in pressure-overloaded mice²⁶. JPH2 is not as extensively studied in RVD, but JPH2 expression is reduced by 35% in MCT RV cardiomyocytes⁵. Additionally, sildenafil, which restores RV JPH2 expression and t-tubule architecture, improves RV contractility in MCT rats⁵. Moreover, we previously showed colchicine reduces microtubule density and increases JPH2 expression, which augments RV function⁴. Collectively, these findings highlight the crucial role of JPH2 for proper LV and RV function. However, in PAH, the LV is relatively unaffected and thus the GP130-STAT3-JPH2 microtubular pathway is primarily dysregulated in the RV.

The direct link between GP130 signaling and microtubule remodeling is likely pleiotropic as there are multiple mechanisms by which the GP130-STAT3 axis could modulate MT dynamics. As discussed above, STAT3 is predicted to regulate expression of isoforms of α- and β-tubulin and Map4²¹ (Supplemental Table III). However, SC-144 only slightly decreased transcript levels of several tubulin isoforms and Map4 in the RV (Supplemental Table IV), suggesting transcriptional modulation may not be the most important mechanism of microtubule regulation. Interestingly, there are higher levels of pSTAT3 associated with microtubules in MCT-V as compared to control, which SC-144 mitigates (Supplemental Figure I). These results are in agreement with a previous study showing STAT3 directly binds and stabilizes microtubules in vitro²⁷. Thus, a pSTAT3-microtubule interaction may contribute to RV microtubule remodeling (Figure 8). Additionally, PI3K and MAPK signaling may also alter the microtubule cytoskeleton and these pathways merit future investigation. Finally, posttranslational modifications of tubulins enhance microtubule stability and promote cardiac dysfunction^{15, 23}, and these modifications may also be regulated by GP130 signaling.

Our finding that GP130 modulates RV fibrosis is congruent with multiple previous publications. First, genetic deletion of IL-6 reduces fibrosis in diabetic cardiomyopathy²⁸. Moreover, inhibition of STAT3 activity with parthenolide decreases LV fibrosis via modulation of fibroblast activation²⁹. Additionally, β_IV-spectrin knockout in fibroblasts activates STAT3 signaling and causes cardiac fibrosis and dysfunction. Importantly, these pathological changes are reversed by pharmacological inhibition of STAT3³⁰. Thus, there are multiple lines of evidence linking the GP130-STAT3 axis to pathological cardiac fibrosis, which likely explains our finding that SC-144 decreases RV fibrosis (Figure 5).

The proteomics/metabolomics analyses (Figures 2 and 3) and Seahorse experiments (Figure 4) demonstrate that microtubules modify mitochondrial metabolism. This hypothesis is supported by other publications^{31, 32}. In particular, microtubules are proposed to be critical for mitochondrial movement and activity as metabolites and mitochondria are enriched in discrete subcellular locations via microtubule-mediated trafficking^{33, 34}. Additionally, microtubules may facilitate inter-mitochondrial interactions in cardiomyocytes, which are important for maintenance of proper cardiac function³⁵. Finally, in yeast, the association of mitochondria with microtubules prevents mitochondrial fission by inhibition of binding of Dnm1³², the yeast analogue of DRP1. This alters mitochondrial fission/fusion balance and ultimately mitochondrial function. In summary, the microtubule cytoskeleton is important for proper mitochondrial metabolic function.

The correction of multiple metabolic pathways with SC-144 further validates the pathogenic effects of impaired metabolism on RV function². Consistent with a previous study³⁶, we show glutaminolysis is activated in the MCT-V RV (Figure 3). The attenuation of glutaminolysis with SC-144, and the fact that pharmacological inhibition of glutaminolysis enhances RV function in MCT rats³⁶ demonstrates this metabolic pathway has a maladaptive effect on RV function. Importantly, glutaminolysis is also induced in human RVD³⁶, so this is relevant to human disease. In addition, we show alteration of fatty acid metabolism is associated with RVD. MCT-V rats have reduced RV acylcarnitine and acetyl Co-A levels (surrogate measures of fatty acid metabolism), but those metabolites are elevated with SC-144 (Figure 3). The finding of impaired fatty acid oxidation in the PAH RV is consistent with prior work in fawn hooded rats, which develop spontaneous pulmonary hypertension and RV failure³⁷. Interestingly, human PAH RV samples also have decreased acylcarnitine levels³⁸. Thus, our data support an important role for fatty acid metabolism in proper RV function. In conclusion, the ability of SC-144 to combat glutaminolysis induction and enhance fatty acid metabolism, two pathways also altered in human RVD, highlights the importance of metabolism for proper RV function.

The fact that GP130 antagonism did not alter the pulmonary vasculature might seem to contradict the findings of Tamura et al.³⁹ This group used an IL-6 receptor/soluble IL-6 receptor antagonist, 20S,21-epoxy-resibufogenin-3-formate, and showed it lowered mPAP and attenuated adverse pulmonary vascular remodeling in MCT rats. However, their molecule more exclusively targets IL-6, while SC-144 has broader targets for antagonism. The differences in timing of treatment may also explain why our results are dissimilar. Tamura et al.³⁹ started therapy one week after MCT, whereas we initiated therapy two weeks after MCT, a timepoint when PAH is more established⁴⁰. Consistent with our findings, TRANSFORM-UK (NCT02676947) showed IL-6-specific inhibition, achieved by tocilizumab, did not alter PVR²⁵, suggesting IL-6 inhibition does not reverse established pulmonary hypertension. Moreover, our study along with Soon et al.¹² and Simpson et al.’s¹³ do not show any relationship between serum IL-6 levels and pulmonary vascular hemodynamics in PAH. Perhaps the association between elevated IL-6 levels and mortality in PAH may relate to the severity of RV dysfunction.

Finally, the pathogenic effects of GP130 signaling may be more relevant to RVD than LV dysfunction as analysis of the publicly available human heart single nucleus RNA sequencing database⁴¹ shows expression of GP130, STAT3, and many tubulin isoforms are higher in RV cardiomyocytes than LV cardiomyocytes (Supplemental Table V). We speculate that this pathway may be more active in the RV, and thus blocking the GP130-microtubule axis may have accentuated beneficial effects in RVD, as compared to LV failure.

Limitations and Future Directions

We acknowledge several important limitations. First, we did not study the effects of GP130 antagonism in Sugen-hypoxia rats. We chose the MCT rat model because it develops significant RV hypertrophy and failure, paralleling the phenotype observed in PAH patients with RV decompensation⁴². However, a recent RV transcriptomic analysis of MCT and Sugen-hypoxia rats demonstrates involvement of nearly identical biological pathways, including IL-6 and STAT3 signaling⁴³. Additionally, we only studied male rats as potential sex differences in GP130 signaling were not the focus of this study. However, our human cohort shows elevated IL-6 is associated with RVD in female PAH patients.

Many inflammatory mediators are associated with adverse RV remodeling and dysfunction⁴⁴. While serum IL-6 levels are associated with severity of RV dysfunction in PAH^{11, 12}, the robust beneficial effects of GP130 antagonism may also be due to inhibition of other GP130 ligands (e.g. IL-11, leukemia inhibitory factor, oncostatin M, ciliary neurotrophic factor, and cardiotrophin-1)⁴⁵. This is supported by the observation that blockade of oncostatin M prevents cardiac dysfunction in inflammatory dilated cardiomyopathy⁴⁶. Moreover, small molecule inhibition of GP130 signaling likely affects multiple cell types beyond the RV cardiomyocytes, which may also contribute to the beneficial effects we observed.

Finally, the identification of mitochondrial inner membrane and matrix proteins bound to microtubules will require further exploration in the future. Perhaps microtubules are important for mitochondrially targeted protein translation or the trafficking of proteins translated in the cytoplasm to mitochondria. Another possibility is that we detected partially intact mitochondria in our tissue microtubule co-sedimentation assay, which is plausible because mitochondria are known to be trafficked along microtubules^{33, 34}.

Conclusions

Inhibition of GP130 signaling enhances RV function independent of changes in the pulmonary vasculature. Modulation of pathological microtubule remodeling with GP130 antagonism leading to enhanced t-tubule structure and normalization of mitochondrial metabolism likely underlies the improved RV function.

Potential Clinical Impact:

What is new?

• We show small molecule inhibition of glycoprotein 130 (GP130) signaling mitigates pathological microtubule remodeling in the right ventricle, which subsequently corrects t-tubule architecture and enhances mitochondrial metabolic activity. At the organ level, GP130 antagonism augments right ventricular function independent of pulmonary vascular disease severity in rodent pulmonary arterial hypertension.

What are the clinical implications?

• Compromised right ventricular function is the greatest predictor of death in pulmonary arterial hypertension. However, no currently approved therapy directly targets the failing right ventricle. Our findings suggest GP130 antagonism could be a novel therapeutic approach to enhance right ventricular function and hopefully improve survival in this rare but deadly disease.

NONSTANDARD ABBREVIATIONS AND ACRONYMS

ANOVA

Analysis of variance

CBB

Coomassie brilliant blue

DRP1

Dynamin-related protein 1

Ea

Effective arterial elastance

Ees

End-systolic elastance

FIS1

Mitochondrial fission 1

GP130

Glycoprotein 130

IL-6

Interleukin-6

JAK/STAT

Janus kinase/signal transducer and activator of transcription

JPH2

Junctophilin-2

LV

Left ventricle/ventricular

MAP4

Microtubule associated protein 4

MAPK

Mitogen activated protein kinase

MCT

Monocrotaline

MCT-V

Monocrotaline-vehicle

MFN1

Mitofusin-1

MFN2

Mitofusin-2

mPAP

Mean pulmonary arterial pressure

MT

Microtubule

NT-proBNP

N-terminal pro B-type natriuretic peptide

OCR

Oxygen consumption rate

OPA1

Optic atrophy protein 1

PAC

Pulmonary arterial compliance

PAH

Pulmonary arterial hypertension

PI3K

Phosphatidylinositol 3-kinase

pSTAT3

Phosphorylated signal transducer and activator of transcription 3

PVR

Pulmonary vascular resistance

RV

Right ventricle/ventricular

RV/BW

Right ventricular weight normalized to body weight

RVD

Right ventricular dysfunction

RVEF

Right ventricular ejection fraction

RVFAC

Right ventricular fractional area change

RVSP

Right ventricular systolic pressure

STAT3

Signal transducer and activator of transcription 3

TAPSE

Tricuspid annular plane systolic excursion

TCA

Tricarboxylic acid

TEM

Transmission electron microscopy