Challenges and Future Prospects of Targeting Myostatin/Activin A Signaling to Treat Diseases of Muscle Loss and Metabolic Dysfunction

Abstract

Over the past 25 years, considerable progress has been made in terms of elucidating the regulatory and signaling mechanisms underlying the control of skeletal muscle mass by myostatin and other secreted proteins belonging to the transforming growth factor-β superfamily. Preclinical studies demonstrating the potential benefits of targeting the activities of these ligands have fueled the development of numerous biologics capable of perturbing this signaling pathway and increasing muscle mass and function. These biologics have been tested in numerous clinical trials for a wide range of indications characterized by muscle loss and excess adiposity. Here, we review the results of these trials and discuss some of the challenges and future prospects for targeting this signaling pathway to treat muscle and metabolic diseases. Myostatin inhibitors may improve metabolic outcomes by increasing muscle mass, and metabolic disorders may be attractive potential indications for these molecules.

Myostatin Regulatory System

Myostatin (MSTN, also known as GDF-8)) was originally identified in a screen for new members of the transforming growth factor-β (TGF-β) superfamily (for review, see ref (1)). Mstn was shown to be expressed specifically in the skeletal muscle lineage both during embryogenesis and in adult mice, and the function of MSTN as a negative regulator of skeletal muscle mass became apparent when mice carrying a targeted deletion of Mstn were found to have about a doubling of skeletal muscle mass throughout the body. The increased muscle mass in Mstn null mice was shown to result from a combination of increased muscle fiber numbers and size, reflecting the two distinct roles of MSTN: to regulate the numbers of fibers that form during development and to regulate the growth of those fibers postnatally. Both engineered and naturally occurring MSTN mutations were subsequently shown to cause increased muscling in many other mammalian species, including humans, as well as in avian and piscine species.

MSTN is synthesized in a precursor form consisting of an N-terminal propeptide and C-terminal mature peptide that contains the cystine knot structure characteristic of other TGF-β family members. MSTN and the highly related protein, GDF-11, form their own subgroup within the larger superfamily. Following proteolytic processing of the precursor protein, the propeptide remains non-covalently bound to mature MSTN and maintains it in an inactive, latent complex that can be activated by cleavage of the propeptide by members of the BMP-1/tolloid family of metalloproteases (Figure 1A). MSTN is regulated extracellularly by multiple other inhibitory binding proteins, including follistatin (FST), FSTL-3, GASP-1, and GASP-2. When free of these inhibitory proteins, MSTN signals by binding initially to the type 2 activin receptors, ACVR2 and ACVR2B, followed by engagement of the type 1 receptors, ALK4 and ALK5.

Many of these regulatory and signaling components are shared by other TGF-β family members, including GDF-11 and activin A. Although the role that GDF-11 may play in regulating muscle is still unclear, activin A has a similar activity to MSTN in muscle. Simultaneously targeting both MSTN and activin A has a more substantial effect in inducing muscle growth compared to targeting MSTN alone, demonstrating that these ligands are partially functionally redundant with respect to the control of muscle growth. Signaling by certain other TGF-β family members seems to have the opposite effect, with BMP signaling to myofibers being capable of inducing muscle hypertrophy. Hence, muscle mass is regulated by the overall balance between MSTN/activin A signaling, which has a negative role, and BMP signaling, which has a positive role.

Development of Drugs Targeting the MSTN Signaling Pathway

The elucidation of the biological function of MSTN and the beneficial effects seen upon targeting the MSTN signaling pathway in preclinical studies have led to extensive efforts to develop inhibitors of MSTN, and to a lesser extent other ligands of the pathway, for clinical use (for review, see ref. (1); Figure 1B). Although many therapeutic strategies have been explored and tested in preclinical studies, here we will focus only on those biologics that have been developed by pharmaceutical and biotechnology companies and tested in clinical trials. Most of these biologics have targeted MSTN itself, and these include the monoclonal antibodies MYO-029 (Wyeth/Pfizer, New York, NY), domagrozumab (Pfizer), landogrozumab (Eli Lilly, Indianapolis, IN), and REGN1033 (Regeneron, Tarrytown, NY), as well as the peptibody PINTA-745 (Amgen/Atara, Thousand Oaks, CA) and the adnectin taldefgrobep alfa (Bristol Myers Squibb, New York, NY; Roche, Basel, Switzerland; Biohaven, New Haven, CT). These drugs are all conceptually similar in their ability to bind and neutralize MSTN but vary in their ligand specificity, with REGN1033 being highly specific for MSTN and domagrozumab and landogrozumab being capable of also binding the highly related protein GDF-11; the binding specificities of PINTA-745 and taldefgrobep alfa have not been published. In addition to these biologics that target the mature form of MSTN, 2 monoclonal antibodies, apitegromab (Scholar Rock) and RO7204239 (Roche/Chugai), bind the MSTN propeptide and inhibit its cleavage by BMP-1/tolloid metalloproteases, thereby blocking the activation of latent MSTN. Because the amino acid sequences of the propeptides of MSTN and GDF-11 are significantly different compared to their mature domains, these monoclonal antibodies are highly specific for MSTN. RO7204239 also has a “sweeping” function, leading to clearance of the latent complex from the circulation followed by re-cycling of the antibody.

In addition to these biologics targeting either mature MSTN or its propeptide, several other biologics have been developed to target other components of this signaling pathway. These include 2 ligand traps, ACE-031 (Acceleron), which is a decoy form of the ACVR2B receptor consisting of its extracellular ligand-binding domain fused to an immunoglobulin Fc domain, and ACE-083 (Acceleron), which is a portion of the binding protein FST fused to an Fc domain. Both ACE-031 and ACE-083 are capable of binding multiple members of the TGF-β superfamily, including not only MSTN and GDF-11 but also activins and certain BMPs. Additional preclinical studies targeting activin A with selective monoclonal antibodies have implicated its role in muscle regeneration after acute injury. These findings highlight the integration of immune modulation by targeting a subset of macrophages that facilitate debris removal followed by a myogenic regenerative activity, which offers potential inflammatory muscle wasting conditions as avenues to explore clinical utility (2). A third biologic in this group is bimagrumab (Novartis/Versanis), which is a monoclonal antibody directed against both activin type 2 receptor subtypes. Bimagrumab is capable of blocking signaling by multiple ligands utilizing both ACVR2 and ACVR2B. A key difference in the mechanism of action between a decoy receptor and receptor antagonist is demonstrated by ACE-031 and bimagrumab. ACE-031 is able to target all ligands capable of binding the activin type 2B receptor, while bimagrumab targets both activin type 2 receptors (ACVR2 and ACVR2B) preventing the binding and subsequent activity of the receptor ligands in a given cell type. This distinction may be particularly relevant in the context of BMPs, which are capable of utilizing not only the BMP type 2 receptor (BMPR2), but also activin type 2 receptors for signaling. ACE-031 is known to bind and inhibit BMP-9 and BMP-10 reducing the amount of available ligand to utilize BMPR2 in the key cell types. In contrast, bimagrumab binds to the activin type 2 receptors, not its ligands, resulting in a higher concentration of circulating ligands, such as BMP-9 and BMP-10 to bind to other receptors, such as BMPR2. Other BMPs have low affinity for ACVR2/ACVR2B but can nevertheless utilize these receptors for signaling because they bind first to type 1 receptors (such as ALK3 or ALK6) and then engage the type 2 receptors. This may be particularly relevant in skeletal muscle, where BMP signaling has been shown to induce muscle hypertrophy, which would counteract the negative effects of MSTN/activin A signaling, as discussed above.

Clinical Trials With MSTN Inhibitors

These biologics have been tested in multiple phase 2 clinical trials for a wide range of indications, focused on muscle loss and the associated functional limitations often seen in older adults and people with chronic diseases. Eli Lilly’s LY2495655 was tested in older weak individuals who had fallen during the previous year (3) and in patients following total hip arthroplasty (4). In the former trial, LY2495655 was shown to cause increases in lean body mass, accompanied by a persistent decrease in fat mass, as well as statistical improvements in stair climbing, time to rise from a chair, and gait speed compared to placebo, although clinical relevance was unclear. In the latter trial, LY2495655 failed to achieve statistical significance in endpoints such as the 6-minute walk. However, unlike the former which was conducted in older adults with a lower level of physical function, the trial in people undergoing hip arthroplasty did not show any improvements compared to the standard of care that included post-operative physical therapy. Novartis’s bimagrumab was tested in 2 trials for sarcopenia in older adults and resulted in significant increases in mean lean body mass of 6%–8%. However, improvements in muscle strength and physical function were variable. In the initial trial, an improvement in grip strength was seen among the entire cohort, while clinically meaningful improvements in gait speed and 6-minute walk distance were seen in a subset of individuals with slower walking speeds at baseline (5). The larger second trial, however, failed to duplicate the initial findings and meet the primary endpoint of improved performance on the Short Physical Performance Battery after 6 months of treatment (6). A similar result of an increase in lean body mass with no improvement in physical function was seen following 6 months of treatment in patients undergoing hip fracture repair surgery (7). Although the results of the phase 2 trial with Regeneron’s REGN1033 for age-related sarcopenia have not yet been published, the data that have been reported to date suggest that functional improvements in the older population treated with MSTN inhibitors have been inconsistent and modest.

The results of clinical trials with other indications have sometimes reported increases in lean body mass, but similarly have not shown consistent, clinically meaningful improvements in performance-based measures of physical function. A number of trials have focused on patients with various forms of muscular dystrophy. Wyeth’s MYO-029 monoclonal antibody was tested in a phase 1/2 trial in a mixed population of adults with Becker muscular dystrophy, facioscapulohumeral dystrophy, and limb-girdle muscular dystrophy, which found no improvement in either lean body mass or functional measures (8). Pfizer’s domagrozumab was tested in a phase 2 trial in boys with Duchenne muscular dystrophy with a primary efficacy endpoint of improvement on 4-stair climb time. This trial was terminated early for lack of efficacy, although MRI analysis did show increases in thigh muscle volume in patients who received the drug compared to placebo (9). Roche’s phase 2 trial using their anti-MSTN adnectin taldefgrobep alfa in boys with Duchenne muscular dystrophy was also discontinued based on a preplanned futility analysis that indicated that the drug was unlikely to demonstrate clinical efficacy, although the data from the trial have not yet been published. Finally, Acceleron’s trial with their decoy receptor ACE-031 in boys with Duchenne muscular dystrophy was terminated early due to the development of epistaxis and telangiectasias in some patients, likely due to the drug’s ability to bind BMP-9 (10).

MSTN inhibitors have been tested in patients with other neuromuscular disorders as well. One of the most highly watched was Novartis’s trial of bimagrumab in patients with sporadic inclusion body myositis (sIBM), for which the drug had been given a breakthrough therapy designation by the FDA. Moreover, an initial trial with a small number of sIBM patients that received a single dose of bimagrumab showed promising results in terms of increased thigh muscle volume and lean body mass as well as a 14.6% increase in 6-minute walk distance after 16 weeks (11). A follow-up phase 3 trial with a larger number of patients with sIBM treated with bimagrumab every 4 weeks for 48 weeks, however, was terminated early due to a lack of efficacy. Although the drug resulted in significant increases in lean body mass compared to placebo, treatment with bimagrumab failed to meet the primary endpoint of improved 6-minute walk distance (12,13). Another neuromuscular disorder for which MSTN inhibition is being tested is spinal muscular atrophy (SMA). Scholar Rock tested apitegromab in patients with type 2 and type 3 SMA, and although a summary of the data posted on-line seems to indicate promising results, especially in the more severely affected patients, the results of this trial have not yet been published.

Another indication that has been the focus of considerable investigation is cachexia. Eli Lilly tested LY2495655 in a phase 2 trial in patients with pancreatic cancer, but the drug showed no improvement in survival, which was the primary endpoint (14). Subgroup analysis did show some benefits of drug treatment in terms of lean body mass and some functional measurements in patients with less than 5% weight loss in the preceding 6 months prior to enrollment. Novartis also tested bimagrumab in a phase 2 trial in patients with either lung or pancreatic cancer, although the results of this trial have not yet been published. In patients with cachexia due to chronic obstructive pulmonary disease (15), bimagrumab treatment was associated with significant increases in thigh muscle volume compared to placebo, but no differences were observed in functional assessments, including 6-minute walk distance, stair climbing time, timed up and go, or muscle strength. Finally, Atara tested their biologic, PINTA-745, in patients with end-stage renal disease. Although the results have not yet been published, the trial failed to meet its primary endpoint in terms of change in lean body mass and showed no improvement in physical function as a result of drug treatment.

Perhaps the most promising clinical data with MSTN inhibitors that have been published to date may be the results of Novartis’s phase 2 trial of bimagrumab in obese individuals with type II diabetes (16). Treatment with bimagrumab every 4 weeks for 48 weeks was associated with statistically and clinically significant reductions in total body fat mass by 20% (7.31 kg), waist circumference by 9.5 cm, and HbA1c by 0.76 percentage points and an increase in lean body mass by 4.4% (2.1 kg), compared to placebo, with all p values ≤ .005. Indeed, this effect of MSTN inhibition on reducing fat mass was also observed in many of the clinical trials with bimagrumab, LY2495655, AMG-745, and ACE-031 previously mentioned. This finding of a reduction in adipose tissue is consistent with observations in mice in which this signaling pathway has been targeted. Genetic disruption of myostatin in Ldlr null mice is associated with hypermuscularity, and resistance to fat accumulation, development of hepatic steatosis, insulin resistance, proatherogenic dyslipidemia, and the progression of experimentally induced aortic atherogenesis (17). At least some of this fat reduction associated with MSTN inhibition is likely to be an indirect result of the anabolic effect on muscle, as targeting receptors specifically in myofibers have been shown to be sufficient to reduce total body fat content in mice. Several secretory products of the skeletal muscle, such as IL-8, IL-15, follistatin, follistatin-like (FSL) 1 and 3, myostatin, and irisin serve as myokines that may improve metabolism by diverse mechanisms, including regulation of brown fat generation, lipolysis, fatty acid β oxidation, insulin secretion and action, inflammation, and angiogenesis (18). Additional studies will be required to determine whether inhibition of signaling by key ligands directly to adipocytes may also play a role. Whatever the mechanism may be, these findings suggest that targeting this signaling pathway may have applications in the prevention or treatment of metabolic diseases such as obesity and type II diabetes.

Lessons Learned From Clinical Trials With MSTN Inhibitors

In clinical trials with a wide range of MSTN inhibitors that have been tested in a diverse group of indications, blocking the MSTN signaling pathway has consistently increased lean body mass and/or muscle volume. The increases in lean body mass seen in these trials were in the range of 3%–8%, with biologics specific for MSTN/GDF-11 generally producing effects toward the lower end of that range and biologics also capable of targeting activin A generally producing effects toward the upper end. These increases, however, were substantially lower than those observed in preclinical studies in mice, in which increases in muscle weights in the range of 25%–50% have been reported using the murine form of these same types of inhibitors. This raises the question as to whether humans and mice are fundamentally different with respect to the overall capacity for muscle growth that can be induced by targeting this signaling pathway, and a related question moving forward is whether these drugs have tapped the full muscle anabolic potential of targeting this pathway in humans.

In mice, it appears that the capacity for inducing muscle growth may be far greater than what has been achieved to date using biologics (for review, see ref. (1)). In particular, simultaneously targeting the 2 type 1 receptors, ALK4 and ALK5, genetically in myofibers has been shown to cause massive increases in muscle mass, far greater than those seen upon simultaneously targeting the 2 type 2 receptors, ACVR2 and ACVR2B. Although the basis for this discrepancy is not completely understood, a possible explanation may be that targeting the type 1 receptors leads to a loss of signaling by a different subset of ligands compared to targeting the type 2 receptors. In particular, muscle growth has been shown to be regulated by a balance between MSTN/activin A signaling and BMP signaling. Because BMPs utilize multiple type 2 receptors for signaling, including activin type 2 receptors, the anabolic effect of targeting ACVR2 and ACVR2B with respect to blocking MSTN/activin A signaling may be partially counteracted by partial or complete loss of BMP signaling through the same receptors. In contrast, targeting ALK4 and ALK5 would completely block MSTN/activin A signaling without affecting signaling by BMPs, which utilize different type 1 receptors, namely ALK3 and ALK6. The promiscuity of activin type 2 receptors may also potentially limit the anabolic potential of bimagrumab compared to that of ACE-031, the decoy ACVR2B/Fc receptor, which at high doses in animals has been shown to cause over 50% muscle growth in just 2 weeks. One could speculate that the decoy receptor, with its relatively low affinity for classic BMPs, such as BMP-2 and BMP-4, acts as a ligand trap primarily for MSTN/GDF-11/activin A without affecting BMP signaling to myofibers, whereas the ability of BMPs to engage the activin type 2 receptors following initial binding to ALK3 and/or ALK6 may render BMP signaling to myofibers at least partially susceptible to inhibition by bimagrumab.

It seems clear based on both mouse and human studies that in order to generate the greatest muscle anabolic effects, the range of ligands being targeted has to extend beyond MSTN itself, certainly to include activin A and perhaps other unidentified ligands as well. Of course, the expanded ligand specificity carries the risk of adverse effects on other tissues, and this has already been borne out in the clinical trial results using ACE-031. Specifically, treatment with ACE-031 led to the development of epistaxis and telangiectasias in some patients, which have been attributed to ACE-031’s ability to bind BMP-9 and/or BMP-10, which appear to signal through different type 1 receptors than those predominantly utilized by classical BMPs. Notably, these coagulopathy-related adverse effects have not been reported in any of the multiple trials using bimagrumab. One possible explanation, as discussed above, may be that bimagrumab is a receptor antagonist and therefore may have no direct effect on BMP-9 and BMP-10 signaling, which likely utilizes a different type 2 receptor, namely BMPR2, in the relevant cell types. Comparatively, ACE-031 is a ligand trap capable of binding and inhibiting BMP-9 and/or BMP-10 regardless of which receptors are utilized in vivo.

Despite the consistent increases in lean body mass and muscle volume that were achieved with MSTN and activin antagonists in most of the clinical trials, the increases in lean body mass and muscle volume have not been associated with clinically meaningful improvements in performance-based measures, such as 6-minute walk distance, or self-reported measures of physical function. One reason for the failure of these drugs to produce robust functional benefits could be related to the modest increases in muscle volume typically in the range of 3%–8%. Indeed, the degree to which muscle mass correlates with metrics such as strength, distance walked, stair climbing, time to rise from a chair, etc. is not clear. It is also not clear as to how much muscle mass would have to be increased in order to produce significant improvements in strength and on these types of assessments, although it seems reasonable to expect that clinically meaningful outcomes would be more likely if the anabolic effects could be increased.

It is also possible that the benefits of an increase in muscle mass are limited by other components of the neuromotor system. Many older adults and patient populations with limitations in physical function have compromised neuromotor mechanisms (ie, decreased motor unit number, reduced neural drive to the muscles innervated) that may attenuate the translatability of larger muscle fibers to improve physical function. Countering a decline in the neuromuscular capacity needed to translate increased muscle mass into functional improvements may require exercise training. In this regard, a key question is the extent to which increased signaling through this pathway may contribute to functional decline during aging. Several groups have attempted to address this question by measuring circulating levels of MSTN, GDF-11, and activin A during aging in rodents and humans (19–22). Although not all of the findings were completely consistent with one another, the general consensus seems to be that circulating levels of these key ligands do not change appreciably during aging, with MSTN levels perhaps even decreasing slightly, likely reflecting the decreased amount of overall muscle mass. It is certainly possible, however, that circulating levels may not reflect actual signaling activity in muscle, given that these ligands are regulated by a complex network of inhibitory binding proteins and given that local autocrine or paracrine signaling may play a significant role in regulating muscle homeostasis.

In considering the results of these trials, it is important to keep in mind that these trials not only utilized a wide range of MSTN inhibitors, including ones with different biological properties as well as mechanisms of action but also targeted a wide range of clinical indications. Hence, it may be misleading to consider the results of all of the trials collectively, as any individual failed trial could have reflected the specific drug, indication, or method for assessing clinical benefit. One indication that is currently being pursued by several companies is SMA, with Scholar Rock, Roche, and Biohaven each launching phase 3 trials with MSTN inhibitors in 2022. SMA is somewhat unique in that the functional status of SMA patients is standardly assessed using the Hammersmith scale, which measures performance on dozens of individual tasks, with patients receiving a score of 0, 1, or 2 on each task, and overall improvements of 3 points out of a possible 66 points are considered by clinicians to be meaningful improvements. Hence, it may be that this approach for documenting the functional improvement in individuals with SMA may be more sensitive than many of the tests used in other patient populations in the MSTN trials, such as 6-minute walk distance, time to rise from a chair, etc. In the case of age-related sarcopenia, a key question is whether the inconsistent effects seen in those trials would have been more clearly documented had subjects been assessed using a clinical outcome similar to the Hammersmith scale adapted for a sarcopenic population. Discussions regarding assessments that may be acceptable for regulatory approval of drugs that treat muscle loss and decreased physical function are ongoing.

Other indications that warrant further investigation would be metabolic diseases, including obesity and type 2 diabetes. Indeed, the effects seen in reducing fat content and improving glucose metabolism were consistently seen in many trials, and if these beneficial effects are the direct result of the anabolic stimulus in skeletal muscle caused by MSTN inhibition, the clinical data suggest that even 3%–8% increases in lean body mass may be sufficient to have substantial metabolic benefits. The results of the bimagrumab trial in obese patients with type 2 diabetes were particularly striking in the magnitude of the observed effects. If similar results are observed in Versanis’s ongoing phase 2b trial for bimagrumab in subjects with obesity, bimagrumab, or potentially other MSTN inhibitors as well, could be an attractive alternative to other classes of obesity drugs where weight reduction is the result of a loss of both adipose tissue and muscle mass. Inhibiting negative muscle growth regulators, such as MSTN, may achieve weight loss without affecting gastric emptying or appetite control, while reducing adipose tissue mass and preserving or even increasing lean body mass. Similar findings were observed using the LY 2495655, in sustained fat mass decreases in the population of older adults with a history of falls. Indeed, it is possible that regulatory approval for the use of MSTN inhibitors as an approach to increase muscle mass in the older adult population may result from the focus on metabolic diseases as indications rather than age-related sarcopenia per se.

Acknowledgments

The views expressed in this manuscript are based on discussions that occurred during the NIA virtual workshop on Development of Function Promoting Therapies: Public Health Need, Molecular Targets, and Drug Development held on March 20-22, 2022.

Funding

S.-J.L.’s work was supported by National Institutes of Health grants R01AG052962 and R01AR081659. S.B.’s work was supported by National Institutes of Health grants RO1 DK49296, R56AG052972, and RO1AG072087 and by the infrastructural resources of the Boston Claude D. Pepper Older Americans Independence Center P30AG31679.

This supplement is sponsored by the National Institute on Aging (NIA) at the National Institutes of Health (NIH).

Conflict of Interest

S.-J.L. is a consultant to Biohaven Pharmaceuticals, Inc. S.B. reports receiving research grants from National Institute on Aging, National Institute of Child Health and Human Development-National Center for Medical Rehabilitation Research, National Institute of Nursing Research, AbbVie, Function Promoting Therapies, and Transition therapeutics, and consulting fees from Aditum and OPKO. L.K. is an employee of Versanis Bio, Inc. V.K. is an employee of Eli Lilly, Inc. D.R. is an employee of Novartis Institute for BioMedical Research, Inc.

References