Blocking extracellular activation of myostatin as a strategy for treating muscle wasting

Figures

• 

  Figure 1. Scholar Rock antibodies specifically block proteolytic activation of myostatin and GDF11. (a) Activation of myostatin requires two distinct proteolysis events, generating three major myostatin species. The first cleavage step of proGDF8/proMyostatin (and proGDF11) is carried out by a proprotein convertase to produce the latent complex. Activation and release of the mature growth factor is accomplished after cleavage by an additional protease from the BMP/tolloid family. (b,c) Precursor-specific antibodies raised towards myostatin (GDF8_021, GDF8_086, and SRK-015) or GDF11 (GDF11_010, GDF11_011, and GDF11_015) were evaluated in myostatin (b) and GDF11 (c) activation assays. Results were compared to control reactions to calculate the fraction of proMyostatin or proGDF11 activated in the assay. Antibody function was then expressed as the percent inhibition of signaling.

• 

  Figure 2. Binding to myostatin precursors and blocking activation prevents muscle atrophy in vivo. (a) A panel of ten diverse myostatin-precursor specific antibodies were evaluated in the myostatin activation assay. 200 nM murine proMyostatin (proGDF8) was preincubated with increasing concentrations of the test antibodies and signaling inhibition is expressed as a percent inhibition. EC50 values are reported in Table 2. (b) Evaluation of the efficacy of myostatin precursor-recognizing antibodies in a skeletal muscle atrophy model. Animals were administered dexamethasone in their drinking water, and dosed with 20 mg/kg of test antibodies once per week for two weeks. The percent difference of the gastrocnemius muscle (versus the mean of the IgG control treated animals without dexamethasone, No Dex) is plotted. Individual data points (n = 8–10 animals) along with means +/− standard deviations are shown. Group means were compared by one-way ANOVA followed by a Holm-Sidak test. Calculated P values for test article compared to the IgG control dosed in combination with dexamethasone are shown.

• 

  Figure 3. SRK-015/SRK-015P blocks tolloid protease cleavage of the prodomain and binds endogenous proMyostatin in the muscle. (a) 500 nM human latent myostatin preincubated with increasing amounts of SRK-015P, were analyzed in a myostatin activation assay. Concentrations of antibodies in each sample were 0.12 μM (1), 0.37 μM (2), 1.1 μM (3), 3.3 μM (4), and 10 μM (5). The EC50 was 581 nM (95% CL 344–982 nM). (b) Samples in (a) were run on reducing SDS-PAGE and probed by western blot with an antibody recognizing the prodomain of myostatin. An ~18 kDa band (white box), consisting of the C-terminal cleavage fragment of the prodomain generated after tolloid cleavage, decreased proportionally with increasing doses of SRK-015P. (c) Non-reducing western blot for prodomain (red) and the mature growth factor (green) following immunoprecipitation with indicated antibodies. Separate red and green images for the entire gel are shown in Supplementary Fig. S5a. Lanes loaded with recombinant protein controls show the migration of mature myostatin, proMyostatin and the myostatin prodomain (latent myostatin) on the gel, highlighted by black arrowheads. In serum, both SRK-015 and GDF8-C1 bind to latent…

• 

  Figure 4. Myostatin precursors redistribute during atrophy. (a) Western blot using a polyclonal antibody raised to the prodomain of myostatin. Recombinant protein controls (proMyostatin and latent myostatin) are used to visualize the migration of proMyostatin and myostatin prodomain bands. In muscle from a pilot experiment in which mice were administered dexamethasone for 15 days, proMyostatin levels increase in muscle, while the levels of latent myostatin in plasma (inferred from the prodomain band) decrease. (b) In a follow-on experiment, animals were administered either vehicle (No Dex) or dexamethasone in their drinking water for two weeks, and given a single 20 mg/kg dose of test antibodies (SRK-015 or IgG control) at day 1. Differences in gastrocnemius muscle weights (expressed as percent difference from the mean of the IgG (no Dex) control group) are shown for days 4, 6, 8 and 15. Individual data points (n = 8–10 animals) along with means +/− standard deviations are shown. Group means were compared by one-way ANOVA followed by a Holm-Sidak test. (c–e) Quantitation of proMyostatin and latent myostatin levels in murine muscle and serum at 4, 6, 8, and 15 days following either…

• 

  Figure 5. Administration of muSRK-015P in healthy animals enhances muscle function. (a–c) In vivo Plantarflexor functional performance after four weeks of treatment with muSRK-015P. (a) In vivo maximal force measurements produced from electrical stimulus of the plantarflexor muscle group via the sciatic nerve normalized to limb length. Following 4 weeks of treatment with muSRK-015P, there was an increase of 18% in the maximal force of the plantarflexor group at 100 Hz (Main Effect P = 0.003 by 2-way ANOVA). (b) A significant increase in gastrocnemius weight was measured at the end of the study, and when maximal force was normalized to the gastrocnemius weight (c) no significant changes were observed. (d–f) In vitro extensor digitorum longus (EDL) muscle function after four weeks of treatment with muSRK-015P. (d) Force measurements of the EDL over increasing stimulus frequencies normalized to EDL length. muSRK-015P treatment resulted in a 29% increase in force at 100 Hz (Main effect P = 0.0003 by 2-way ANOVA). (e) A significant increase in EDL weight was measured at the end of the study, and when maximal force was normalized to the EDL weight (f) no significant changes were observed. (g)…