Rigosertib Reverses Hypertrophic Cardiomyopathy in Noonan Syndrome

Abstract

Background:

RASopathies constitute a group of rare genetic disorders caused by mutations in genes along the canonical RAS/MAPK signaling pathway, affecting cell growth and differentiation. These syndromes, which include Noonan syndrome (NS), are characterized by developmental delays, distinctive facial dysmorphia, and cardiac defects, notably hypertrophic cardiomyopathy (HCM). Despite their prevalence and impact, therapeutic options for RASopathies remain limited. Rigosertib, a novel dual RAS/MAPK and PI3K/AKT pathway inhibitor, is currently in clinical trials for treatment of melanoma and recessive dystrophic epidermolysis bullosa. Here, we identify rigosertib as a candidate therapy for RAF1-associated HCM.

Methods:

We performed a drug screen of clinically relevant compounds in transgenic Drosophila models of RASopathies to identify candidate therapeutics. Cardiac-targeted Drosophila models expressing RASopathy-associated transgenes were used to evaluate the effects of rigosertib on cardiac hypertrophy and compare its efficacy with the MEK inhibitor trametinib. Therapeutic efficacy was further assessed in a mammalian model using Raf1^L613V/+ knock-in mice treated with rigosertib for six weeks. Cardiac structure and function were evaluated by echocardiography, histology, and molecular analyses, including assessment of cardiomyocyte (CM) size, fetal gene expression, and ERK/AKT signaling. Additional Noonan syndrome-associated phenotypes, including skeletal growth and craniofacial abnormalities, were also evaluated.

Results:

Rigosertib was effective across a panel of transgenic Drosophila RASopathy models, suggesting activity against multiple disease variants. In cardiac-targeted fly models, rigosertib reduced cardiac hypertrophy and outperformed trametinib. In Raf1^L613V/+ mice, six weeks of treatment significantly improved left ventricular chamber dimension, posterior wall thickness, heart mass, and CM size, resulting in reversal of cardiac hypertrophy. Rigosertib also normalized fetal gene expression and inhibited ERK and AKT signaling in primary CMs. In addition to reversing cardiac pathology, rigosertib significantly improved other Noonan syndrome-associated features, including increased bone growth and correction of craniofacial abnormalities.

Conclusions:

Together, our findings suggest rigosertib normalizes and reverses RASopathy-associated HCM and other NS-associated syndromic features, supporting its development as a promising treatment for RAF1-associated HCM and, potentially, other RASopathy-dependent pathologies. This study not only highlights the therapeutic potential of rigosertib but also demonstrates the utility of an integrated approach using Drosophila and mammalian models to elucidate drug effects across complex biological systems.

Introduction

RASopathies are a group of rare, autosomal dominant genetic disorders caused by pathogenic variants in the canonical RAS/mitogen-activated protein kinase (MAPK) signaling pathway ¹. RASopathies affect ~1 in 1000–2000 live births, with the most common being Noonan syndrome (NS) ². Typical characteristics of NS include short stature, dysmorphic facial features, intellectual delays, and cardiovascular issues; the latter include congenital heart disease (CHD)—most prevalently pulmonary valve stenosis—and hypertrophic cardiomyopathy (HCM) ². Indeed, HCM is the second most common cardiovascular manifestation in RASopathies, with a prevalence of 20–25%, exhibiting unique features that include early onset, rapid progression with congestive heart failure, and co-occurrence of CHD, leading to significant mortality ³.

NS is caused by gain-of-function (GOF) variants in multiple genes that affect RAS/MAPK signaling, including PTPN11, SOS1, RAF1, and RIT1 genes, although pathogenic variants in other genes have also been described ^4,5. There are several other genetic traits with overlapping clinical characteristics, such as Costello syndrome (CS), Noonan syndrome with multiple lentigines (NSML), and cardiofaciocutaneous syndrome (CFC). Like NS, these traits are caused by pathogenic alleles—typically GOF—in RAS/MAPK-related genes. NSML is the outlier, as it is most commonly associated with loss-of-function or dominant-negative mutations in PTPN11 ^6–9.

NS exhibits several well-established genotype-phenotype associations ⁴. For example, PTPN11-associated NS, accounting for greater than 50% of all cases of NS ¹⁰, has an HCM prevalence of only 6%, significantly lower than the overall frequency of HCM in other NS-causing genes ³. In contrast, the prevalences of HCM in RAF1- and RIT1-associated NS are 65% and 36%, respectively. While these two genetic forms of NS account for smaller percentages of NS pathogenesis (3–17% and 5–9%, respectively), they account for nearly 50% of HCM in NS and an even higher proportion of cases with early onset of congestive heart failure ^3,11. These observations emphasize that different NS genes have distinct pathological effects and highlight the need for selective therapeutic approaches to treating NS-associated HCM.

Knock-in (KI) mice heterozygous for the NS-associated Raf1 L613V allele (Raf1^L613V/+) develop HCM, with upregulated cardiac fetal gene expression and left-ventricular decompensation in response to pressure overload ¹². Agonist-evoked MAPK/extracellular signal-regulated kinase (ERK) activation is increased in Raf1^L613V/+ cardiomyocytes (CMs) and fibroblasts, suggesting it may play a critical role in growth factor-dependent modulation of ERK1/2-independent pathways ^11–13. RAS pathway inhibition, using the ATP-noncompetitive MAPK kinase (MEK) inhibitor mirdametinib, prevents HCM in Raf1-mutant mice ¹², but treatment was only tested with initiation before the onset of hypertrophy. Based on this pre-clinical result, a limited number of infants and children with NS with severe HCM have been treated with the MEK inhibitor trametinib on a compassionate-care basis, with some evidence of regression of cardiac hypertrophy ^14–17. Clinical trial results are still needed to demonstrate formally the efficacy of this approach ¹⁸. Experience to date has shown some adverse effects of trametinib, particularly dermatologic ones, which can be limiting, and that not all affected patients respond to treatment ^14,15,17. Thus, there is a need to identify other therapies that are efficacious and, ideally, less morbid, for NS-associated HCM.

To find additional small-molecule therapies with potential efficacy for the RASopathies, we developed a pre-clinical drug repurposing pipeline relying upon transgenic Drosophila models and identified that rigosertib could rescue pupal lethality in a RAF1-mutant RASopathy fly model ¹⁹. Rigosertib (ON-01910 sodium salt) is a synthetic benzyl styryl sulfone with novel activities as a dual RAS/MAPK and phosphatidylinositol 3-kinase (PI3K)/AKT inhibitor ^20,21. The primary mechanism of action for rigosertib remains unclear, with evidence supporting direct RAS pathway inhibition ^22,23, while other data indicate it acts as an inhibitor of microtubules ^24–26. Based on phase 1/2 clinical oncologic trials, rigosertib is well tolerated, with few side effects, the commonest being hemorrhagic cystitis, particularly in older men ²⁷. Of note, the range of side effects typically associated with anti-cancer agents that inhibit microtubules, such as peripheral neuropathies, has not been observed in trials of rigosertib. A Phase 3 clinical trial of rigosertib for myelodysplastic syndrome was unsuccessful ²⁸. Additional oncologic trials using rigosertib, alone or in combination with other agents, are ongoing or recently completed ^29,30; it showed efficacy in a Phase 2 trial for recessive dystrophic epidermolysis bullosa-associated squamous cell carcinoma ³¹.

In this study, we used a large panel of Drosophila RASopathy models to identify rigosertib as acting more broadly than other RAS pathway inhibitors, notably the MEK inhibitor trametinib. We demonstrated its efficacy in rescuing the cardiac hypertrophic phenotype in RAF1 S257L, RAF1 L613V, and PTPN11 Q510E fly hearts. Finally, we undertook a trial of rigosertib in reversing the HCM phenotype in Raf1^L613V/+ mice, demonstrating that the drug both normalized cardiac status and restored linear growth and craniofacial anomalies. Our data support the view that rigosertib is a strong candidate as a highly effective, well-tolerated, and specific treatment for HCM in RASopathies, particularly for RAF1-associated HCM, as well as a potentially effective therapy for treating other NS-associated phenotypic effects.

Materials and Methods

Data Sharing.

The data, analytic methods, and study materials that support the findings of this study will be made available to researchers upon reasonable request to the co-corresponding authors. The authors declare that all supporting data are available within the article and its online supplementary files.

Testing efficacy of rigosertib using transgenic Drosophila RASopathy models.

Drosophila testing:

We obtained rigosertib for use in these studies from Traws Pharma (formerly Onconova Therapeutics). We assessed rigosertib and trametinib for therapeutic efficacy using a quantitative Drosophila survival assay, as previously reported ^32–34. Briefly, rigosertib was dissolved in water (200 mM stock), and trametinib was dissolved in DMSO (200 mM stock). Drug was diluted in molten (~50–60 °C) enriched fly media and then aliquoted into 5-ml vials to obtain final drug concentrations of 1, 10 or 50 μM. Based on previous analyses, after larvae consume drug-containing media, the circulating concentration of drug is ~250-fold lower than the media concentration.

We used the driver 765-GAL4 line at 25 °C and crossed flies to UAS-RASopathy transgenic fly lines representing 14 different RASopathies variants observed in patients. Each transgenic line was fed rigosertib or trametinib, mixed into the media. Water only and DMSO only were used as untreated controls for rigosertib and trametinib, respectively. The ratio of treated:untreated surviving pupae/adults (expressed as “% rescue”) was used to assess drug efficacy. Biological replicates using 4–5 vials were performed per experimental condition. Each vial contained between 30–80 developing embryos, and absolute numbers of surviving pupae were counted.

Heart-specific testing in Drosophila:

Previously published data indicate that expressing a RASopathy-associated transgene specifically in the Drosophila heart can recapitulate aspects of HCM ^35,36. We studied transgenic flies with cardiac-specific expression of hRAF1^L613V, hRAF1^S257L, and hPTPN11^Q510E because these alleles are strongly associated with HCM in NS (two RAF1 alleles) and NSML (PTPN11 allele) and pupal lethality from them could be rescued by rigosertib. We used the cardiac-specific driver line tin-GAL4 and crossed them to our UAS-hRaf1^L613V flies (tin>hRaf1^L613V), UAS-hRAF1^S257L flies (tin>hRAF1^S257L), and UAS-hPTPN11^Q510E flies (tin>hPTPN11^Q510E). Flies were raised on standard fly media or media containing 100 μM rigosertib or trametinib for one week starting at enclosure. Fly heart phenotypes were assessed with live imaging or with semi-intact hearts fixed in diastole and then stained with mouse anti-collagen IV (pericardia) 1:100, and/or phalloidin-cy3 1:1000 as previously described ^37,38.

Statistics in Drosophila studies:

Effects of rigosertib and trametinib treatment were assessed across 14 Drosophila RASopathy models using linear models with treatment group as a factor. Differences across groups were assessed via F-test and, if p < 0.05, three hypotheses were tested: treatment effect for treatment 1 [i.e., rigosertib vs. control for rigosertib (water only)], treatment 2 [i.e., trametinib vs. control for trametinib (DMSO only)], and the differences in the treatment effects of rigosertib vs. trametinib. Such hypotheses were tested via contrast and adjusted for multiple hypothesis using Bonferroni correction (i.e., nominal p value < 0.05 corresponding to rejection of the null hypothesis).

Testing the efficacy of rigosertib in a KI Raf1^L613V/+ mouse model.

Mice:

Raf1^L613V/+ (B6N;129S6-Raf1tm1.1Bgn/Mmucd, 036521-UCD) were obtained from the Mutant Mouse Resource & Research Centers (MMRRC) at The Jackson Laboratory, an NIH-funded strain repository, and were donated to the MMRRC by Benjamin G. Neel, M.D., Ph.D., New York University. Clippings were taken from the tip of the tails (1 mm) of 18–21-day old mice, DNA was extracted and subjected to PCR to assay for the Raf1^L613V allele using the primers: sense 5’-ATC CCC TGA TCT CAG CAG GCT CTAC-3’; antisense 5’-AGT AGT CTA GGT CCT TAG CAG CAGC-3’, with subsequent restriction enzyme digestion using DraIII (New England Biolabs) to determine genotype. Animals were kept on a 12-hour light-dark cycle, at 22 ^oC and food was distributed ad libitum. Animals were maintained and utilized for experiments herein by crossing Raf1^L613V/+ mice with mixed C57BL/6 × 129 wildtype (WT) mice (Jackson Labs). All animal procedures were approved by the Masonic Medical Research Institute (MMRI) Animal Care and Use Committee. MMRI’s PHS assurance number is D16–00144(A3228–01), and it is an AAALAC-accredited institution (#001865).

Treatment of mice with rigosertib:

Randomization and blinding were performed in accordance with ARRIVE 2.0 guidelines. Briefly, animals were randomly allocated to treatment groups. Cage position and testing order were also randomized to minimize confounding effects. Investigators performing experiments, outcome assessments, and statistical analyses were blinded to treatment allocation until completion of the analysis. Rigosertib was provided by Onconova Therapeutics (now Traws Pharma). Rigosertib was dissolved in sterile filtered water at 0.01 mg/uL and injected intraperitoneally (IP) at 100 mg/kg body weight, twice daily, for either three or six weeks, as indicated. Prior to sacrifice, mice were fasted for 4 h, following the final round of injections, to normalize signaling effects for subsequent molecular assessments.

Echocardiography:

Transthoracic echocardiography and left ventricular functional assessments were conducted on non-anesthetized animals as described previously ^8,9, using a Visual Sonics Vevo 3100^® high-frequency ultrasound rodent imaging system. Briefly, mice were trained for three consecutive days prior to baseline recordings to acclimate to the echocardiography procedure. Baseline measurements were made at 8 weeks of age and were repeated at 11 weeks and 14 weeks of age, to assess effects of three and six weeks of rigosertib treatment, respectively. Hearts were imaged in the two-dimensional parasternal short-axis view, and an M-mode echocardiogram of the mid left ventricle was recorded at the level of papillary muscles. Heart rate, interventricular septum (IVS), left ventricular posterior wall thickness (LVPW), and end-diastolic and end-systolic internal diameters of the left ventricle (LVIDd and LVIDs, respectively) were measured from M-mode images.

Anatomical measurements:

Body weights for each mouse were recorded weekly to determine dosing. At the time of sacrifice, mice were weighed and then measured from nose tip to tail base to determine body length. At harvest, the heart and lungs were removed, washed with PBS, blot dried and weighed. The right tibia of each mouse was also removed, and length measurements were taken.

Craniofacial morphometry:

We utilized μCT scans using a Locus Ultra MicroCT Scanner (GE Healthcare) to measure changes in craniofacial morphometry. Three-dimensional images of the craniofacial skeleton were generated and analyzed with GEHC MicroView Software (GE Healthcare). Skull measurements were made according to the Standard Protocol and Procedures from The Jackson Laboratory (http://craniofacial.jax.org/standard_protocols.html).

Microscopy and histology:

Hearts for morphometry and histochemistry were flushed with PBS and then fixed in 4% paraformaldehyde for 24 h before paraffin embedding. Serial sections (5-μm thick) of heart tissue were taken in the transverse plane from each mouse at the level of the papillary muscle and then stained with either reticulin or hematoxylin/eosin (H&E). Images of these sections were taken using the Keyence BZ-X810 fluorescence microscope at both 40x and 400x magnification, with an exposure of 1/60 s at 50% transmittance, and then labeled with a 50-mm measure bar.

Tissue sections of hearts were used to assess the cross-sectional length, width, and area of C, where centrally located nuclei (to ensure the same plane of sectioning) were measured using ImageJ 1.41 software (developed by Wayne Rasband; http://rsbweb.nih.gov/ij/). Three individual samples were analyzed for each genotype, with five different fields from four mice per group and three sections per heart. The total number of myocytes counted was 1–2 × 10³ cells per mouse.

Adult primary CM isolation:

Mice were injected with intraperitoneal heparin (40 units/mouse), and their hearts were isolated and perfused via the aorta. The perfusion buffer consisted of KCl 14.7 mM, NaCl 120.4 mM, KH₂PO₄ 0.6 mM, Na₂HPO₄ 0.6 mM, MgSO₄7H₂O 1.2 mM, 2,3 butanedione monoxime 10 mM, taurine 30 mM, HEPES 10 mM, and glucose 5.5 mM. Hearts were digested with collagenase II digestion buffer (2 mg/ml) for approximately 8–10 minutes. Each heart was then cut from the cannula and placed in the dish with digestion buffer and stopping buffer [12.5 μM CaCl₂ and 10% exosome-depleted FBS in perfusion buffer]. Isolated CMs were then cultured in 6-well plates (treated with laminin), with myocyte culture medium [ScienCell #6201], 5% exosome-depleted FBS, and 1% penicillin/streptomycin.

For signaling experiments, isolated CMs were starved and either treated with vehicle (water) or rigosertib (10 nM) for 12 h and then stimulated with either vehicle (DMSO), angiotensin II (20 ng/ml) or insulin-like growth factor (IGF) (20 ng/ml), as indicated, for 5 minutes.

Biochemical studies associated with Raf1^L613V/+ KI mice.

Immunoblotting:

Tissue or cell lysates were prepared by homogenizing the tissue or lysing the cells, respectively, in radioimmunoprecipitation (RIPA) buffer (25 mmol/l Tris-HCl [pH 7.4], 150 mmol/l NaCl, 0.1% SDS, 1% NP-40, 0.5% sodium deoxycholate, 5 mmol/l EDTA), 1 mmol/l sodium fluoride, 1 mmol/l sodium orthovanadate, and a protease cocktail at 4 °C, followed by sonication. The homogenate tissue was centrifuged for 10 minutes at 15000 × g at 4 °C, after which the supernatant was collected. Protein concentration was estimated using a “Pierce Rapid Gold BCA Protein Assay Kit” as directed by the manufacturer (Thermofisher).

For immunoblots, 30 μg of total protein lysates was resolved through a 10% polyacrylamide gel by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to a 0.2-μm nitrocellulose membrane. The resulting membranes were immunoblotted using antibodies against anti-AKT (#9272S; 1:1000 dilution in 5% bovine serum albumin (BSA)/Tris-buffered saline with 0.1% Tween 20 (TBST) at 31 μg/ml), anti-phospho-AKT (Ser473) (#4060S; 1:1000 dilution in 5% BSA/TBST at 91 μg/ml), anti-ERK1/2 (#9102L; 1:1000 dilution in 5% BSA/TBST at 50 μg/ml ), and anti-phospho-ERK 1/2 (Thr202/Tyr204) (#9101S; 1:1000 dilution in 5% BSA/TBST at 191 μg/ml)) (Cell Signaling Technology). Visualization of the primary antibody binding was performed using the Li-Cor anti-rabbit 680 IR-Dye system in conjunction with the Li-Cor Odyssey FC. Quantification of secondary antibody fluorescence was performed using ImageStudio Lite Ver 5.2.

RT-qPCR:

The apex of the left ventricle was utilized for extraction of total RNA using the RNeasy fibrous tissue kit (Qiagen). Reverse transcription of 1 μg of RNA was conducted using the iScript cDNA synthesis kit (BioRad). RT-qPCR samples were tested in triplicate. SYBR green (Applied biosystems) was used per manufacturer’s instructions. Primers targeting expression of fetal genes Anp, Bnp, Myh6, and Myh7 were generated, and all levels of resulting expression were normalized to the average of two housekeeping genes β-actin and 18s ribosomal RNA (18S). Data were quantified using the comparative C_T method (ΔΔCT). For primer sequences and PCR conditions, please see the supplemental information (Table S1).

MEF2 luciferase assay:

Luciferase activity to assess downstream activity of ERK5 signaling was measured using the Luciferase Reporter Assay System (Thermo Scientific, #16176) according to the manufacturer’s instructions and as previously described ¹¹. Briefly, adult cardiomyocytes were isolated and plated in 12-well plates for 2 h in M199+5% FBS +1%P/S. Cells were then infected with 100 vp/ml of Ad-MEF2-Luciferase (Seven Hills Bioreagents, #JMAd-33). After 2 h of infection, cells were treated with vehicle or rigosertib (5 μM) and cultured for an additional 18 h. Cells were then subsequently lysed and luciferase activity was measured in a 96-well plate. Luciferase signals were normalized to total protein concentration.

Statistical analyses.

Animal sample sizes were determined based on prior experience with the model, previously published studies using similar experimental paradigms, study feasibility, and ethical considerations to minimize animal use. Sufficient statistical power to detect biologically meaningful differences between groups was maintained. Predefined inclusion/exclusion criteria were applied uniformly across groups, and no mice were excluded from the analyses.All values in graphs are expressed as the mean ± SEM. Normality was tested with the D'Agostino-Pearson normality test. If the data showed a normal distribution, pairwise testing was performed with the student’s t test or multiple group comparisons were performed by 1- or 2-way ANOVA, followed by post-hoc Bonferroni correction (GraphPad Prism 9). Significance level was set to 0.05.

Results

Rigosertib treatment showed efficacy against most RASopathy Drosophila models.

Our goal was to identify a lead therapeutic compound effective against multiple RASopathy isoforms. Given the challenges of recruiting a sufficient number of patients with the individual RASopathy genes, we tested candidate drugs/compounds using 14 previously generated Drosophila models, each of which expressed a different human or Drosophila RASopathy gene variant under the control of an inducible UAS promoter ¹⁹. In our screen of RAS relevant compound and drugs (¹⁹), rigosertib proved unique in its broad efficacy. When treated with rigosertib at three different doses (1, 10, and 50 μM) in the fly media, we observed significant pupal rescue for 8/14 (57%) of the RASopathy fly models (Figure 1A). By comparison, feeding flies the potent MEK inhibitor trametinib led to rescue of pupal lethality in 3/14 (21%) models; all three were also rescued by rigosertib (Figure 1B). For two RASopathy fly models—KRAS^G12D and PTPN11^Y279C—the treatment effect-size rescue was statistically greater for rigosertib than trametinib; no fly model showed a greater effect size rescue by trametinib than rigosertib (Figure 1B). These data together suggest that rigosertib is especially effective in rescuing pupal lethality in Drosophila RASopathy models, outperforming the RASopathy candidate therapeutic trametinib.

Figure 1. Rigosertib treatment shows efficacy against some, but not all, RASopathy Drosophila models.

A. The percent survival of 1, 10, or 50 μM rigosertib treatment in RASopathies Drosophila pupae as compared to untreated pupae of the same genotype. Percent survival is indicated in the color code bar at the right of the table. For pupal viability analysis, mean and standard error of the mean (SEM) were calculated from 4–5 vials (biological replicates) per experimental condition. Each vial contained between 30–80 developing embryos, and absolute numbers of surviving pupae/adults were compared to diluent-only treatment to obtain % rescue. B. Maximal fold increase in pupal survival following treatment with either rigosertib or trametinib as compared to untreated RASopathy Drosophila models. Fold change is indicated in the color code bar at the right of the table. Data was evaluated for Gaussian distribution and all data sets passed the Shapiro-Wilk test for normality. Data represent mean ± SEM; *p < 0.05, where p values were derived from ANOVA with Tukey’s multiple comparisons post-test when ANOVA was significant.

Rigosertib rescues the HCM phenotype in three RASopathy Drosophila models.

RAF1-associated NS and PTPN11-associated NSML are strongly associated with HCM, which is a significant source of morbidity and mortality in these patients. To assess the effects of rigosertib on cardiac pathology in Drosophila, we used the heart-specific tin-Gal4 promoter to drive three UAS-human RASopathy transgenes (tin>hRAF1^L613V, tin>hRAF1^S257L, and tin>hPTPN11^Q510E), resulting in cardiac-specific expression of three variants strongly associated with HCM (Figure 2A). We found that adult Drosophila hearts expressing these three RASopathy alleles showed multiple deficits, including significant increases in heart wall thickness and end-diastolic diameter (Figure 2B–D). The fibrous extracellular collagen IV (pericardin) network associated with the heart tube was also significantly increased (Figure 2E–F). These data suggest that the three tin>RASopathy allele fly models recapitulate key aspects of HCM observed in NS patients with inherited RAF1 pathogenic alleles.

Figure 2. Rigosertib rescues the Drosophila HCM phenotype in three RASopathy models.

(A) Using targeted heart-specific tin-Gal4 promoter to drive the UAS-human RAF1^L613V, hRAF1^S257L, and hPTPN11^Q510E transgenes (resulting in tin>hRaf1^L613V, tin>hRAF1^S257L, and tin>hPTPN11^Q510E flies, respectively), hearts were optically imaged. We observed significant changes in heart wall thickness (B and C), end-diastolic diameter (D), and normalized extracellular collagen IV area (E and F), indicative of pathological hypertrophy. These changes were reversed in the presence of rigosertib more efficaciously than in the presence of trametinib, both at the concentration of 100 μM in the fly media. Data represent mean ± SEM; *p < 0.01, **p < 0.005, ***p < 0.001,****p < 0.0005, and all p values were derived from ANOVA with Bonferroni post-test when ANOVA was significant. Bars in panels B (vertical red in the lower left corners) and F (horizontal white in the lower right corners) indicate sizes of 5 and 25 μm, respectively.

Next, we tested if feeding adult flies with rigosertib or trametinib for one-week post-eclosion significantly improved heart wall thickness, end-diastolic diameter, and extracellular collagen IV. As shown in Figure 2C, rigosertib rescued heart wall thickness for all three RASopathy models, achieving thicknesses similar to wild-type fly heart (TinHE/+). By comparison, trametinib only rescued two of the three RASopathy models, did not reduce thickness to the level of wild-type and was statistically inferior to rigosertib for all three fly lines. In assessing end-diastolic diameter (Figure 2D), we observed that rigosertib and trametinib both significantly rescued this phenotype in all three RASopathy models. The level of rescue normalized to the level of the wild-type fly hearts for the two NS models but failed to completely normalize for the NSML model. The efficacy of rigosertib and trametinib were similar (rigosertib was superior for one line, trametinib was superior for another line, and they were not significantly different for the third line). With respect to the increased extracellular collagen IV, rigosertib significantly rescued this phenotype in two of the three lines, achieving levels comparable to wild-type, and reduced collagen in the third, albeit non-significantly. In contrast, trametinib treatment did not reduce the collagen content in two of the three lines and did not achieve significance in the third, for which reduction to the level of the wild-type fly heart was not observed. In summary, rigosertib was more effective than trametinib in rescuing the increased heart wall thickness and increased end-diastolic diameter, as well as increasing the extracellular collagen IV, in all three RASopathy models tested.

Rigosertib treatment of KI Raf1^L613V/+ mice reversed hypertrophic cardiomyopathy.

To validate the Drosophila HCM findings and to better investigate the impact of rigosertib on mammalian HCM, we next examined the effects of rigosertib on Raf1^L613V/+ KI mice. As previously described and as shown here, both male and female Raf1^L613V/+ mice develop measurable cardiac hypertrophy by 8 weeks of age ¹² (Tables 1–4; Tables S2-S9). To assess the effects of rigosertib in these mice, we injected male and female 8-week-old mice intraperitoneally (IP) with either vehicle (water) or rigosertib at a dose of 100 mg/kg body weight, twice daily; we then assessed cardiac physiological and functional parameters after three and six weeks of treatment. Our experimental dose was selected based on maximum tolerable dose (MTD) studies previously conducted at Traws Pharma (formerly Onconova Therapeutics), where 100 mg/kg/day was found to be both sub-MTD and efficacious in a long-term study ²². No mortality occurred in our cohort with treatment of rigosertib at this dose for this time period. We found that rigosertib reverted and completely normalized the HCM cardiac phenotype in both male and female Raf1^L613V/+ mice (Figure 3A-B).

Table 1.

Rigosertib treatment normalized Raf1^L613V/+ NS-associated anatomical features.

Male and Female	Wildtype Vehicle		Wildtype Rigosertib		L613V/+ Vehicle		L613V/+ Rigosertib
Tx Period	3 wks	6 wks	3 wks	6 wks	3 wks	6 wks	3 wks	6 wks
Body Weight (g)	22.13 ± 1.31	23.26 ± 1.48	21.63 ± 1.27	21.77 ± 1.41	21.27 ± 1.46	21.95 ± 0.61	21.53 ± 1.11	22.88 ± 1.07
Body Length (cm)	9.18 ± 0.13	9.15 ± 0.21	9.25 ± 0.14	9.02 ± 0.19	8.78 ± 0.12*	8.62 ± 0.05*	9.02 ± 0.07	9.02 ± 0.13
Tibia Length (mm)	18.84 ± 0.50	18.67 ± 0.51	19.17 ± 0.44	18.67 ± 0.45	16.50 ± 0.39*	17.33 ± 0.19*	18.50 ± 0.31†	17.75 ± 0.52†
HW:TL (mg/mm)	5.77 ± 0.27	6.97 ± 0.31	5.21 ± 0.14	6.56 ± 0.48	9.56 ± 0.59*	10.39 ± 0.45*	7.27 ± 2.45	8.56 ± 0.54†
HW:BW (mg/mm)	4.94 ± 0.2	5.64 ± 0.20	4.65 ± 0.12	5.61 ± 0.16	7.65 ± 0.81*	8.20 ± 0.24 *	6.29 ± 0.44	6.76 ± 0.29†
LW:TL (mg/mm)	10.30 ± 0.71	9.46 ± 0.64	9.41 ± 0.25	9.29 ± 0.56	11.76 ± 0.44	10.78 ± 0.53	10.73 ± 0.35	11.59 ± 0.57
LW:BW (mg/g)	8.80 ± 0.45	7.69 ± 0.55	8,43 ± 0.35	8.11 ± 0.64	9,39 ± 0.79	8.57 ± 0.51	9.30 ± 0.40	9.17 ± 0.46

Heart weight (HW), body length, tibia length (TL), HW:TL ratio, HW to body weight (HW:BW) ratio, lung weight (LW) to TL, and LW:BW ratios were all assessed from male and female WT or Raf1^L613V/+ mice that were either intraperitoneally (IP) injected with vehicle (water) or rigosertib (100 mg/kg), administered twice a day, for a total period of 3-weeks or 6-weeks, as indicated, starting at 8 weeks of age. N = 8 mice per group, per time period. Data in tables represent mean ± SEM; statistical significance was determined by 2-way ANOVA with Bonferroni correction, where P<0.05 was deemed significant. Significance is indicated with *, comparison between vehicle-treated WT vs. each of the other groups at the same time points, i.e. vehicle-treated WT vs. rigosertib-treated WT; vehicle-treated WT vs. vehicle-treated Raf1^L613V/+; vehicle-treated WT vs. rigosertib-treated Raf1^L613V/+; ^†, comparisons between vehicle-treated Raf1^L613V/+ vs. rigosertib-treated Raf1^L613V/+ at the same time points.

Table 4.

Echocardiographic analysis of WT vs Raf1^L613V/+ mice.

Echo Data	WT					L613V/+
Male and Female	Baseline	Vehicle		Rigosertib		Baseline	Vehicle		Rigosertib
Tx Period	0 wks	3 wks	6 wks	3 wks	6 wks	0 wks	3 wks	6 wks	3 wks	6 wks
Heart Rate (BPM)	705.77 ± 31.00	626.2 ± 18.27	681.32 ± 14.09	623.27 ± 10.41	646.94 ± 22.31	666.97 ± 9.83	620.12 ± 21.13	695.62 ± 9.87	589.04 ± 18.02	656.92 ± 19.28
Diameter;s (mm)	1.94 ± 0.10	2.24 ± 0.15	2.03 ± 0.07	1.93 ± 0.06	2.12 ± 0.08	1.91 ± 0.05	1.68 ± 0.09†	1.71 ± 0.10	2.45 ± 0.12†‡	2.07 ± 0.07
Diameter;d (mm)	3.45 ± 0.09	3.77 ± 0.11	3.60 ± 0.08	3.58 ± 0.04	3.64 ± 0.07	3.58 ± 0.06	3.49 ± 0.15	3.51 ± 0.09	4.13 ± 0.13‡	3.74 ± 0.10
SV (uL)	37.60 ± 1.89	43.37 ± 2.05	41.38 ± 1.69	41.85 ± 0.64	41.10 ± 1.75	42.49 ± 1.87	53.80 ± 3.42	42.81 ± 2.12	43.06 ± 4.37	45.81 ± 2.43
EF (%)	76.06 ± 1.49	71.72 ± 3.01	75.80 ± 1.05	78.19 ± 1.06	73.45 ± 1.65	78.19 ± 1.26	83.01 ± 1.59†	83.10 ± 1.63	71.24 ± 1.74‡	76.53 ± 0.79
FS (%)	44.45 ± 1.36	40.96 ± 2.53	43.92 ± 0.89	4.25 ± 1.00	41.97 ± 1.38	46.76 ± 1.37	51.82 ± 1.81†	51.55 ± 1.71	41.02 ± 1.42‡	44.76 ± 0.65
LV Mass (mg)	73.86 ± 3.75	84.59 ± 5.40	90.47 ± 4.06	86.59 ± 2.86	94.18 ± 4.23	100.72 ± 4.25*	112.55 ± 6.30†	146.57 ± 9.68†	100.75 ± 6.74‡	110.07 ± 7.13†
LVAW;s (mm)	1.45 ± 0.03	1.44 ± 0.03	1.58 ± 0.02	1.57 ± 0.03	1.50 ± 0.03	1.58 ± 0.04	1.73 ± 0.04	1.88 ± 0.06†	1.48 ± 0.05	1.65 ± 0.03
LVAW;d (mm)	0.68 ± 0.02	0.66 ± 0.02	0.76 ± 0.01	0.72 ± 0.01	0.75 ± 0.01	0.73 ± 0.03	0.82 ± 0.03†	1.05 ± 0.04†	0.64 ± 0.02‡	0.84 ± 0.02†
LVPW;s (mm)	0.97 ± 0.03	0.91 ± 0.04	0.97 ± 0.04	0.96 ± 0.05	0.98 ± 0.03	1.23 ± 0.04*	1.38 ± 0.07†	1.59 ± 0.06†	1.07 ± 0.04‡	1.18 ± 0.05†
LVPW;d (mm)	0.67 ± 0.01	0.68 ± 0.02	0.75 ± 0.02	0.73 ± 0.02	0.77 ± 0.01	0.89 ± 0.02*	0.99 ± 0.02†	1.10 ± 0.03†	0.72 ± 0.02‡	0.85 ± 0.02†

WT or Raf1^L613V/+ male and female mice were either treated with intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day, for a total period of 3-weeks or 6-weeks, as indicated, starting at 8 weeks of age. Echocardiographic measurements in M-mode to measure cardiac function were taken at baseline and at both 3- and 6-week time points. Tx=treatment; BPM=beats per minute; diameter;d= diameter in diastole; SV= stroke volume, EF= ejection fraction, FS= fractional shortening, LV= Left Ventricular; LVAW;d= left ventricular anterior wall thickness in diastole; LVPW;d= left ventricular posterior wall thickness in diastole. N=8 mice per group, per time point. Data in tables represent mean ± SEM; statistical significance was determined by 2-way ANOVA with Bonferroni correction, where p < 0.05 was deemed significant. Significance is indicated with *, comparison between baseline WT vs. Raf1^L613V/+; ^†, comparison between vehicle-treated WT vs. each of the other groups at the same time points, i.e. vehicle-treated WT vs. rigosertib-treated WT, vehicle-treated WT vs. vehicle-treated Raf1^L613V/+, vehicle-treated WT vs. rigosertib-treated Raf1^L613V/+; ^‡, comparison between vehicle-treated Raf1^L613V/+ vs. rigosertib-treated Raf1^L613V/+ at the same time points. Note: no differences in WT groups were noted at either time point, either in the presence or absence of rigosertib, suggesting no adverse effects of the inhibitor in normal hearts.

Figure 3. Rigosertib treatment of Raf1^L613V/+ KI mice normalizes heart physiology.

A. Representative photographs of hearts from 14-week-old WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of 6 weeks, scale bar=50 mm. B. Representative H&E whole heart cross-sections from 14-week-old WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of 6 weeks, weeks. Scale bar=1 mm. C. Heart weight to tibiae length ratio generated from male and female 14-week-old WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of either 3- or 6-weeks. N = 8/group. Data in graphs represent mean ± SEM; statistical significance was determined by 2-way ANOVA with Bonferroni correction, where *p < 0.05 and ***p < 0.005. D. Representative H&E whole heart cross-sections from 14-week-old WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of 6 weeks, weeks. Scale bar = 50 μm. E. Reticulin staining from heart cross-sections of 14-week-old WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of 6 weeks, Scale bar = 20 μm. F. Frequency distribution of CM area from cross-sections of 14-week-old male and female WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of 6 weeks, N = 8 mice per group, with at least 1×10³ cell counts per heart. Data in graph represents mean ± SEM; statistical significance was determined by 2-way ANOVA with Bonferroni correction, where ****p < 0.0005.

Raf1^L613V/+ mice have increased heart weight (HW) to tibia length (TL) ratios, indicative of their pathological hypertrophy ¹². To determine whether rigosertib normalized heart size, we measured HW:TL ratios in both 3- and 6-week vehicle or rigosertib-treated male and female WT and Raf1^L613V/+ mice (Figure 3C, Table 1, Table S2-S3). We observed a progressive improvement in HW:TL ratios in response to rigosertib treatment in both male and female RAF1^L613V/+ mice; no changes in HW:TL were observed in WT mice, with or without rigosertib treatment, at either time point (Figure 3C, Table 1, Table S2-S3).

H&E or reticulin staining of whole heart sections from both male and female Raf1^L613V/+ mice showed that the mutant mouse hearts had enlarged individual CMs with increased overall cross-sectional cell surface area. Six-week treatment of Raf1^L613V/+ mice with rigosertib normalized these phenotypes to levels similar to WT (Figure 3D-F; Table 2; Table S4-S5). To confirm these results, we isolated individual CMs from male or female Raf1^L613V/+ mice treated with either vehicle or with three or six weeks of rigosertib. We did not observe significant changes in CMs after three weeks of treatment; however, by six weeks, rigosertib normalized length, width, as well as overall cell surface area in both male and female Raf1^L613V/+ isolated CMs to levels similar to WT CMs (Figure 4; Figure S1, Table 3, Table S6-S7).

Table 2.

Quantification of CM cross-sectional areas in response to rigosertib treatment in KI Raf1^L613V/+ mice.

Male and Female (Tx = 6 wks)	Cross-sectional Area (μm2)
WT Vehicle	457.34 ± 7.38
WT Rigosertib	456.55 ± 9.03
L613V/+ Vehicle	645.07 ± 20.98*
L613V/+ Rigosertib	468.20 ± 6.26†

Quantification of cross-sectional CM area from 14-week-old male and female WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of 3-weeks or 6-weeks, as indicated. N = 6 mice per group, with at least 1×10³ cell counts per heart. Data in tables represent mean ± SEM; significance is determined by 2-way ANOVA, with Bonferroni correction, where p < 0.05 was deemed significant. Significance is indicated with *, comparison between vehicle-treated WT vs. each of the other groups at the same time points, i.e. vehicle-treated WT vs. rigosertib-treated WT; vehicle-treated WT vs. vehicle-treated Raf1^L613V/+; vehicle-treated WT vs. rigosertib-treated Raf1^L613V/+; ^†, comparison between vehicle-treated Raf1^L613V/+ vs. rigosertib-treated Raf1^L613V/+ at the same time points.

Figure 4. Rigosertib treatment of Raf1^L613V/+ KI mice decreases individual CM area, reducing cardiac hypertrophy.

Representative photomicrograph of ventricular myocytes isolated from 14-week-old WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of either 3- (A) or 6-weeks (B). Scale bar = 50 μm. Quantitative assessment of CM length (C), width (D), and total area (E), as assessed from CMs isolated from male and female 14-week-old WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of either 3- or 6-weeks. A mixed-effects model was used to take measurements from each heart, where at least 100 cells per heart were evaluated from N=8 hearts per group, and then analyzed by comparing the average measurements between each group to determine non-independence and statistical significance Data in graphs represent mean ± SEM; statistical significance was determined by 2-way ANOVA with Bonferroni correction, where **p < 0.01, ***p < 0.005, ****p < 0.0005.

Table 3.

Measurements of CM cross-sectional areas in response to rigosertib treatment in KI Raf1^L613V/+ mice.

TX period	Wildtype Vehicle		Wildtype Rigosertib		L613V/+ Vehicle		L613V/+ Rigosertib
	3 wks	6 wks	3 wks	6 wks	3 wks	6 wks	3 wks	6 wks
Length (μm)	120.44 ± 1.54	126.55 ± 1.99	124.64 ± 1.68	122.49 ± 1.85	140.98 ± 2.04*	140.59 ± 2.32*	133.49 ± 1.98†	122.48 ± 2.05†
Width (μm)	23.15 ± 0.40	28.51 ± 0.58	24.81 ± 0.49	27.39 ± 0.64	29.44 ± 0.61*	34.73 ± 0.70*	25.98 ± 0.55†	27.89 ± 0.60†
Area (μm2)	2782.36 ± 57.52	3656.59 ± 109.85	3082.61 ± 71.54	3388.76 ± 108.79	4196.21 ± 117.81*	4942.47 ± 141.08*	3521.91 ± 100.26†	3509.44 ± 121.22†

Measurements of length, width, and total area of individually isolated CMs from 14-week-old male and female WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of 3-weeks or 6-weeks, as indicated. N = 6 mice per group, with at least 1×10³ cell counts per heart. Data in tables represent mean ± SEM; significance is determined by 2-way ANOVA, with Bonferroni correction, where p < 0.05 was deemed significant. Significance is indicated with *, comparison between vehicle-treated WT vs. each of the other groups at the same time points, i.e. vehicle-treated WT vs. rigosertib-treated WT; vehicle-treated WT vs. vehicle-treated Raf1^L613V/+; vehicle-treated WT vs. rigosertib-treated Raf1^L613V/+; ^†, comparison between vehicle-treated Raf1^L613V/+ vs. rigosertib-treated Raf1^L613V/+ at the same time points.

Rigosertib treatment of KI Raf1^L613V/+ mice normalized cardiac function.

To determine the effects of rigosertib on cardiac function, we conducted echocardiographic analysis in these mice and found that the vehicle-treated Raf1^L613V/+ mice showed significant progressive left ventricular hypertrophy at both 11- and 14-weeks of age, as indicated by increased LV mass as well as thickened IVS and LVPW, as compared to WT (Figure 5, Table 4, Table S8-S9). Treatment of Raf1^L613V/+ mice with rigosertib, however, normalized the hypertrophy phenotype by the end of the 6-week treatment period in both males and females (Figure 5, Table 4, Table S8-9). Interestingly, we also observed increased fractional shortening (FS) and ejection fraction (EF) in the Raf1^L613V/+ mutant mice, consistent with HCM, which was reduced in the rigosertib-treated mice (Table 4, Table S8-9), suggesting overall cardiac functional normalization. No functional changes or effects were observed in WT hearts over the 3- or 6-week treatment period, either with or without rigosertib (Table 4, Table S8-9).

Figure 5. Rigosertib treatment of Raf1^L613V/+ KI mice normalizes cardiac function.

A. Representative echocardiography of 14-week-old WT or RAF1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of 6-weeks. Quantification of B. EF%; C. FS%; D. LVPWd and E. LV mass (n=8/group) from 14-week-old male and female WT or RAF1^L613V/+ mice. EF%= ejection fraction percentage; FS%= fractional shortening percentage; LVPW;d=left ventricular posterior wall thickness in diastole; LV mass=left ventricular heart mass. Data in graphs represent mean ± SEM; statistical significance was determined by 2-way ANOVA with Bonferroni correction, where *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Of note, rigosertib significantly increased the LVIDd of the male Raf1^L613V/+ mice at three weeks of treatment; however, this effect appeared transient, as the diameter dimension normalized to that of WT by six weeks of rigosertib treatment (Table S8-9). In contrast, three weeks of rigosertib treatment was sufficient to normalize LV mass in the female Raf1^L613V/+ mice, (Table S8-9), a sexual dimorphism consistent with the fact that HCM is generally milder in females than in males with the Raf1^L613V/+ genotype.

Rigosertib treatment of KI Raf1^L613V/+ mice reduced fetal gene expression, suggesting reversal of pathological hypertrophy.

Given the strong phenotypic and functional improvements following rigosertib treatment of Raf1^L613V/+ mutant mice, we next examined its effects on the molecular signaling of treated hearts. To assess this, we first conducted quantitative PCR (qPCR) analysis to measure changes in fetal genes, including alpha myosin heavy chain (Myh6) and beta myosin heavy chain (Myh7) (Figure S2A). While the mRNA levels for Myh6 decreased in Raf1^L613V/+ mouse hearts, they were normalized to levels similar to WT in Raf1^L613V/+ rigosertib-treated hearts. Consequently, the ratio of Myh7 to Myh6 was also increased, indicating improvement in overall cardiac function in response to rigosertib. In addition, we observed a trend towards normalized fetal gene expression of atrial natriuretic factor (Anp) in response to rigosertib treatment of Raf1^L613V/+ mice, though significance was not reached (Figure S2B).

Rigosertib treatment of KI Raf1^L613V/+ mice reduced hypertrophy-induced activation of Erk and Akt.

Rigosertib is a dual PI3K/AKT and ERK/MAPK pathway inhibitor, and both can be altered in RASopathy patients. To assess the effects of these pathways in rigosertib-treated hearts, we isolated whole hearts from either WT or RAF1^L613V/+ mice treated with vehicle (water) or rigosertib (100 mg/kg). Drug was administered twice a day, beginning at 8 weeks of age, for a period of six weeks. While treatment of rigosertib trended towards reducing Akt activation in lysates from Raf1^L613V/+ hearts vs. vehicle treatment, the results did not reach significance (Figure S2C-D); rigosertib also did not have any effect on Akt activation in WT hearts (Figure S2C-D). Similarly, no changes in Erk activation were detected in whole heart lysates isolated from either WT or Raf1^L613V/+ mice, with or without rigosertib treatment (Figure S2D, E).

One possible explanation for these results could be that treatment with rigosertib over 6 weeks normalizes hyperactivation of Erk and Akt, making acute changes in signaling difficult to observe. Alternatively, other cell types in whole heart could also mask effects of rigosertib-induced inhibition of these pathways in CMs. Therefore, to more fully elucidate the signaling effects of rigosertib directly in CMs, we isolated primary CMs from WT and Raf1^L613V/+ mutant mice, treated them with rigosertib for 12 h in vitro, and then subsequently stimulated them with either vehicle, AngII (20 ng/ml) or IGF (50 ng/ml) for 5 minutes, to determine acute effects of drug treatment on Erk and Akt signaling, respectively. At baseline, we did not observe any significant changes in either Erk or Akt activity, consistent with previous findings ¹². In response to treatment with AngII, an upstream activator of the ERK/MAPK signaling cascade, Erk1/2 phosphorylation was significantly increased in Raf1^L613V/+ CMs as compared to WT CMs (Figure 6A-B). Similarly, in response to IGF stimulation, a mediator of PI3K/AKT signaling, Raf1^L613V/+ CMs exhibited significantly increased Akt phosphorylation as compared to WT controls (Figure 6A, 6C). Rigosertib treatment of mutant cardiomyocytes, however, was able to normalize the Erk and Akt activities of Raf1-mutant CMs to levels similar to those of WT (Figure 6A-C).

Figure 6. Rigosertib treatment pf Raf1^L613V/+ KI mice inhibits hyperactivation of ERK and AKT activities in response to AngII and IGF stimulation, respectively, and negatively regulates Erk5 signaling.

Representative immunoblots of isolated primary cardiomyocyte lysates isolated from WT or RAF1^L613V/+ mice littermates, cultured for 12 h in the absence or presence of rigosertib (5 μM), and then subjected to 5 minutes of stimulation with either vehicle, AngII (20 ng/ml) or IGF (50 ng/ml). The heart lysates were immunoblotted with phospho-ERK1/2 or phospho-AKT, as indicated, followed by total ERK1/2 and total AKT to control for total protein expression. Experimental samples were also immunoblotted with anti-GAPDH to normalize loading within each gel. Quantification of B. p-ERK/total ERK and C. p-AKT/total AKT immunoblotting data from n = 3 experimental replicates are indicated. D. Adult cardiomyocytes isolated from WT or RAF1^L613V/+ mice were treated with vehicle or rigosertib (5 μM) for 18 h. Luciferase signals were measured and normalized to total protein concentration. Data were obtained from three independent experiments and in triplicate (n = 3 for both). Data represent mean ± SEM; statistical significance was determined by 2-way ANOVA with Bonferroni correction, where *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns: non-significant.

Rigosertib treatment of KI Raf1^L613V/+ affects MEF2 activity.

We previously identified ERK5 as a target of RAF1 kinase activity in RAF1^S257L/+ inducible pluripotent stem cell-derived CMs ¹¹. To identify whether rigosertib affects downstream signaling to the ERK5 pathway in primary Raf1^L613V/+ CMs, we measured MEF2 transcriptional activity, one major downstream target of MEK5/ERK5. We found that MEF2 activity significantly increased in Raf1^L613V/+ CMs, as compared to WT, and that treatment of these cells with rigosertib drastically reduced MEF2 activity in both WT and Raf1^L613V/+ CMs (Figure 6D).

Rigosertib reduces craniofacial abnormalities and induces linear growth in Raf1^L613V/+ mice.

In addition to HCM, Raf1^L613V/+ mice display several other NS phenotypes, including craniofacial abnormalities, reduced stature, and shortened tibias ¹². To determine whether rigosertib affects non-cardiac NS-associated phenotypes, we conducted μCT scans of the craniofacial areas and showed that treatment with rigosertib normalized craniofacial inner canthal width in Raf1^L613V/+ mice to levels similar to those in WT (Figure 7A-B, Table S10). Skull lengths, widths, and nose lengths from treated Raf1^L613V/+ mice also trended towards normalized WT measures but did not reach significance (Table S10).

Figure 7. Rigosertib functions to ameliorate multiple NS syndromic features in RAF1^L613V/+ KI mice.

A. Representative μCT scans showing measures of skull lengths, widths, inner canthal width, and nose lengths; Quantification of the inner canthal width (B), body length (C), and tibia length (D) from male 14-week-old WT or Raf1^L613V/+ mice, either in the presence or absence of intraperitoneally (IP) injected vehicle (water) or rigosertib (100 mg/kg), administered twice a day for a total period of either 3- or 6-weeks. N = 4/group. Data in graphs represent mean ± SEM; statistical significance was determined by 2-way ANOVA with Bonferroni correction, where *p < 0.05, **p < 0.01, and ***p < 0.005.

In addition to assessing changes in facial dysmorphia, we also measured total body length and tibia length in both 3- and 6-week treated male and female Raf1^L613V/+ mice (Figure 7C-D, Table 1, Table S2-S3). Rigosertib treatment led to a reversal of the short stature phenotype and normalized tibia lengths in both male and female rigosertib-treated Raf1^L613V/+ mice (Figure 7C-D, Table 1, Table S2-S3). This indicates that rigosertib treatment not only ameliorates the HCM phenotype in RAF1-assoicated NS but also leads to positive benefits on non-cardiac aspects of this disorder as well.

Discussion

Treatments for RASopathies remain limited. Our data indicate that rigosertib, a compound with demonstrated low toxicity ²⁷, is a strong candidate for the treatment of RASopathy-associated HCM. Treatment with rigosertib led to increased survival of multiple transgenic RASopathy Drosophila models. We also observed improved adult heart wall thickness, diastolic diameter, and extracellular type IV collagen in three cardiac-specific RASopathy Drosophila lines, two modeling RAF1-associated NS and one modeling PTPN11-associated NSML, that were fed rigosertib. Of note, overall efficacy of rigosertib was greater than trametinib. These findings were validated in Raf1^L613V/+ mice, where rigosertib treatment reversed HCM, reduced CM size, improved cardiac function, and normalized cardiac pathology. Further, and important for clinical considerations, rigosertib had significant positive effects on other NS-associated pathologies, including craniofacial anomalies, short stature and tibia length. Overall, our data suggest rigosertib is a promising treatment for RAF1-associated HCM and, perhaps, other RASopathies and/or RASopathies-associated pathologies as well.

Rigosertib was initially developed, alongside several other styryl benzyl sulfones, to address elevated cyclin activity that leads to tumorigenesis as a consequence of elevated Polo-like kinase 1 (PLK1) ²². The treatment showed promise early on, as the drug-induced cytotoxicity in tumor cells while leaving healthy cells unharmed ²⁷. Having both a high level of efficacy against a multitude of tumor types and a favorable toxicity profile ³⁹, rigosertib has been a compound of great interest for treatment of disease. Interestingly, the mechanism for its regulation was later found to not involve binding to PLK1 directly; instead, rigosertib functions as a non-ATP competitive inhibitor, acting as a RAS binding domain mimetic ⁴⁰ and inhibiting multiple kinase pathways downstream of RAS, including PI3K/Akt and RAS/MAPK signaling ²². Recently, rigosertib has been shown to modulate the PLK-1 pathway in squamous cell carcinoma in the setting of recessive dystrophic epidermolysis bullosa (RDEB-SCC) ²⁹, triple negative breast cancer ^39,41 and hepatitis C virus proliferation ⁴².

Increased ERK and MEK activities have been demonstrated in RAF1^L613V/+ mouse hearts in response to transverse aortic constriction (TAC), a chronic pathological stressor that leads to cardiac hypertrophy and heart failure ¹². Interestingly, treatment with the MEK inhibitor mirdametinib normalized these cardiac defects, suggesting a role for aberrant ERK signaling in RAF1-associated HCM development ¹². In addition, we previously showed that RAF1-mutant iPSC-derived CMs mediate HCM through hyperactivation of MEK1/2, but not ERK1/2, mediating myofibrillar disarray ¹¹. In this case, we found that the enlarged CM phenotype occurred as a consequence of increased extracellular regulated kinase 5 (ERK5) signaling, a pathway not previously known to be involved in NS ¹¹. When measured here in Raf1^L613V/+ CMs, we similarly found that rigosertib negatively regulates MEF2 activity, suggesting non-canonical MEK5/ERK5 signaling, one downstream target of MEF2, may also be modulating cardiac hypertrophy.

Rigosertib is a dual RAS/MAPK and PI3K/AKT inhibitor and herein displays a positive effect on reversing HCM in RAF1-associated RASopathy. Indeed, several studies similarly suggest positive effects of RAS/MAPK and PI3K/AKT inhibition in the heart. For example, treatment of CMs with an AKT inhibitor prevents hypertrophy evoked by stimulation by various agonists ^43–46. Inhibition of AKT/mTOR also attenuates or reverses pressure overload-associated cardiac hypertrophy ^44,47,48. In RASopathies, we generated an NSML KI mouse model of the PTPN11 mutation Y279C (Ptpn11^Y279C/+) that recapitulated the human disorder, with short stature, craniofacial dysmorphia, and HCM; interestingly, these mice showed aberrant agonist-evoked ERK/MAPK signaling, but increased basal and agonist-induced AKT/mTOR activity ⁸. Moreover, in these mice, the HCM-associated cardiac defects were completely reversed with treatment by rapamycin, an inhibitor of mTOR ⁸. Similarly, NSML iPSC-derived CMs demonstrated higher sarcomeric disorganization and increased cell size, indicative of cardiac hypertrophy ⁴⁹. Use of a specific AKT inhibitor, ARQ 092, in another NSML iPSC-derived CM line blocked pathological HCM ⁴⁹. Together, these data support the notion that PI3K/AKT signaling plays a regulatory role in the development of pathological HCM, and that inhibition of this pathway may positively affect and reverse pathological cardiac outcomes.

As with PI3K/AKT signaling, preclinical studies have also suggested that chronic inhibition of RAS/MAPK signaling may have beneficial and cardioprotective effects against development of cardiac hypertrophy and heart failure ^50,51. Blocking RAS/MAPK signaling also leads to decreased cardiac fibrosis in response to pathological stress, including ischemia/reperfusion (I/R) injury ^52,53. Indeed, MEK inhibitors have recently gained traction in RASopathies. Specifically, the FDA approved selumetinib to treat children with NF1 with symptomatic, inoperable plexiform neurofibromas and low-grade gliomas (LGGs); the phase 2 trial of selumetinib for treatment of LGGs resulted in 40% partial response and 96% of patients with 2 years of progression-free survival ⁵⁴. More recently, mirdametinib, an oral allosteric MEK inhibitor, was approved for treatment of NF1-associated plexiform neurofibromas ⁵⁵. While these studies appear promising, at least one study suggests that long-term inhibition of RAS/MAPK signaling may elicit adverse cardiovascular outcomes, through increased activation of cAMP and PKA ⁵⁶.

In summary, our data identify rigosertib as a potent therapeutic candidate that can reverse RAF1-associated HCM. Although these findings hold promise for the treatment of hypertrophy in this and other RASopathies, additional experiments, including optimization of dose and treatment duration, as well as long-term efficacy and toxicity, will be needed to fully assess the clinical implications for its use in patients with these disorders. This is especially true given RAS/MAPK and PI3K/AKT pathways are critically involved in important biological processes that regulate growth and differentiation. Taken together, we find rigosertib to be a highly promising, well-tolerated, and highly effective therapeutic for treatment of RAF1-associated HCM and, potentially, other RASopathies-dependent pathologies as well.

Supplementary Material

Checklist:

Tables S1-S10

Figures S1-S2

Clinical Perspective.

What is new?

• Rigosertib, a dual RAS/MAPK and PI3K/AKT pathway inhibitor, demonstrated robust efficacy across multiple Drosophila models of NS and related RASopathies, rescuing pupal lethality and reversing an HCM-like phenotype; importantly, rigosertib outperformed trametinib, a MEK inhibitor currently used on a compassionate basis, highlighting the superiority of dual-pathway inhibition over single-node targeting strategies.

• Rigosertib treatment of mice with an NS knock-in allele in Raf1 (Raf1^L613V/+) not only prevented progression but also reversed established HCM, normalizing cardiac morphology, CM size, fetal gene expression, and ERK/AKT pathway hyperactivation; functional recovery was confirmed by improved systolic performance, demonstrating true disease modification rather than partial suppression.

What are the clinical implications?

• HCM remains a leading cause of morbidity and mortality in patients with RAF1-associated NS, with no approved targeted therapies capable of reversing established disease; these data position rigosertib as a compelling candidate for disease-modifying treatment of RAF1- and other RASopathy-driven HCM.

• Unlike MEK inhibitors, which incompletely suppress downstream signaling and require chronic dosing with tolerability limitations, rigosertib’s dual inhibition of RAS/MAPK and PI3K/AKT signaling provides broader pathway control and superior phenotypic rescue in preclinical models.

• Although not yet FDA approved, rigosertib’s clinical trial experience in oncology provides pharmacokinetic and safety data that may accelerate repurposing for rare genetic cardiovascular disease.

Non-standard Abbreviations and Acronyms

ANF

Atrial natriuretic factor

BNP

Brain natriuretic peptide

BSA

Bovine serum albumin

CHD

Congenital heart disease

CFC

Cardiofaciocutaneous syndrome

CM

Cardiomyocyte

CS

Costello syndrome

cTNT

Cardiac troponin T

DMSO

Dimethyl sulfoxide

EBs

Embryoid bodies

EF

Ejection fraction

ERK

Extracellular signal regulated kinase

FACS

Fluorescence-activated cell sorting

FS

Fractional shortening

HCM

Hypertrophic cardiomyopathy

H&E

Hematoxylin/eosin

HW

Heart weight

IP

Intraperitoneally

I/R

ischemia/reperfusion

KI

Knock-in

LVIDd

Left ventricular internal diameter end-diastole

LVIDs

Left ventricular internal diameter end-systole

LVPW

Left ventricular posterior wall thickness

MAPK

Mitogen-activated protein kinase

MEK

Mitogen-activated protein kinase kinase

MMRRC

Mutant Mouse Resource & Research Centers

MYH6

α-myosin heavy chain

MYH7

β-myosin heavy chain

NF1

Neurofibromatosis, type 1

NS

Noonan syndrome

NSML

Noonan syndrome with multiple lentigines

PI3K

Phosphatidylinositol 3-kinase

RIPA

Radioimmunoprecipitation

Raf1 ^L613V/+

RASopathy mice heterozygous for the Raf1 L613V allele

Rigosertib

2-[2-methoxy-5-[[(E)-2-(2,4,6-trimethoxyphenyl) ethenyl] sulfonylmethyl] anilino] acetic acid (ON01910

SEM

Standard error of the mean

SDS-PAGE

Sodium dodecyl-sulfate polyacrylamide gel electrophoresis

SV

Stroke volume

tin>hRaf1L613V

Cardiac-specific tin-GAL4 driving UAS-hRaf1 L613V transgene

TBST

Tris-buffered saline with 0.1% Tween 20

TL

Tibia length

WT

Wild type