Overcoming resistance to anabolic SARM therapy in experimental cancer cachexia with an HDAC inhibitor

Abstract No approved therapy exists for cancer‐associated cachexia. The colon‐26 mouse model of cancer cachexia mimics recent late‐stage clinical failures of anabolic anti‐cachexia therapy and was unresponsive to anabolic doses of diverse androgens, including the selective androgen receptor modulator (SARM) GTx‐024. The histone deacetylase inhibitor (HDACi) AR‐42 exhibited anti‐cachectic activity in this model. We explored combined SARM/AR‐42 therapy as an improved anti‐cachectic treatment paradigm. A reduced dose of AR‐42 provided limited anti‐cachectic benefits, but, in combination with GTx‐024, significantly improved body weight, hindlimb muscle mass, and grip strength versus controls. AR‐42 suppressed the IL‐6/GP130/STAT3 signaling axis in muscle without impacting circulating cytokines. GTx‐024‐mediated β‐catenin target gene regulation was apparent in cachectic mice only when combined with AR‐42. Our data suggest cachectic signaling in this model involves catabolic signaling insensitive to anabolic GTx‐024 therapy and a blockade of GTx‐024‐mediated anabolic signaling. AR‐42 mitigates catabolic gene activation and restores anabolic responsiveness to GTx‐024. Combining GTx‐024, a clinically established anabolic therapy, with AR‐42, a clinically evaluated HDACi, represents a promising approach to improve anabolic response in cachectic patients.


Appendix Figures
Appendix Figure S1. A) Non-compartmental analysis of single dose AR-42 pharmacokinetics in mouse. Parameters were derived as outlined in the Materials and Methods section. B) AR-42 dose-response. Starting six days after C-26 cell injection, animals received vehicle or AR-42 orally at 1 (n=8), 3 (n=8), 10 (n=13) or 20 mg/kg (n=6) daily or 50 mg/kg (n=5) every other day for 13 days. Individual animal gastrocnemius weights are presented as a percentage of tumor-free control muscle weight. Dashed reference line (100%) represents the mean tumor-free gastrocnemius mass, solid line represents non-linear fit of dose-response data. C) AR-42's in vitro human HDAC isoform inhibition profile. HDAC activity of recombinant human HDAC enzymes was determined in the presence of 1 μM AR-42 as outlined in the Materials and Methods.
Appendix Figure S2. Study 1 (male; n=20) and Study 3 (female; n=43) C-26 tumor volumes. Terminal tumor volume comparisons between Study 1 and Study 3. Statistics: ns, no significant between-study differences in tumor volumes. Student's t-test of combined tumor volumes (Study 1 versus Study 3). ns, no significant differences among treatment groups within each study, one-way ANOVA followed by Tukey's multiple comparison test. Individual p-values displayed in tables.
Appendix Figure S3. Study 4, tumor-bearing male mice receiving AR-42 (10 mg/kg, oral gavage; n=9), TFM-4AS-1 (10 mg/kg, subcutaneous; n=10), Combination of AR-42 and DHT (10 mg/kg oral gavage and 3 mg/kg subcutaneous, respectively; n=10), Combination of AR-42 and TFM-4AS-1 (10 mg/kg, both, n=9) or Vehicle (n=6) and tumor-free mice receiving vehicle (n=6) were treated daily for 12 days starting 6 days post-injection of C-26 cells. A) Tumor volume comparisons between Day 8 and Day 16 post-C-26 cell injection. Statistics: Individual p-values displayed in tables; ns, no significant differences; one-way ANOVA followed by Tukey's multiple comparison test. B) Androgen receptor western blot of C-26 and LLC tumor tissue and associated source cells. Mouse prostate tissue was analyzed as a positive control. C) C-26 cell viability was determined after 48 hours of treatment with DHT or GTx-024 at concentrations of up to 10 µM. Data in panels A and C are presented as means ± SD. Appendix Figure S4. Study 4, Terminal epididymal fat pad mass from mice in Study 4 treated as described in Appendix Figure S7. Statistics: V, T, A indicate statistically significant differences versus tumorbearing vehicle-treated, tumor-bearing TFM-4AS-1-treated and tumor-bearing AR-42-treated groups, respectively. Individual p-values displayed in tables, one-way ANOVA followed by Tukey's multiple comparison test. Data are presented as means ± SD. 5 Appendix Figure S5. Western blot analysis of AR in gastrocnemius muscles from representative mice in Study 1. Black circle -tumor-free, blue square -tumor-free/GTx-024, black triangle -tumor-bearing, blue diamond -tumor/GTx-024, red triangle -tumor/AR-42, green circle -tumor/combo. Figure S6. A) Genes differentially regulated in gastrocnemius muscle by 50 mg/kg AR-42 treatment relative to C-26 tumor-bearing vehicle-treated controls from Tseng et al. intersected with genes differentially regulated in quadriceps muscle from both severe and moderately wasting C-26 tumor-bearing mice relative to tumor-free controls from Bonetto et al. B) Canonical pathway analysis using GSEA of the 147 overlapping genes from A. Appendix Figure S7. Western blot analysis of phospho(p)STAT3 in gastrocnemius muscles from representative mice treated in Study 1. This data is a replicate blot generated from the samples used in Figure 6B. tSTAT3, total STAT3. Appendix Figure S8. Principle Component Analysis plots of Study 1 RNA-seq samples utilized for subsequent analyses.
Appendix Figure S12. A) Genes differentially regulated 4-fold or greater in gastrocnemius muscle by 50 mg/kg AR-42 treatment relative to tumor-bearing vehicle-treated controls from Tseng et al. intersected with genes differentially regulated 2-fold or greater in gastrocnemius muscle by 10 mg/kg AR-42 treatment relative to tumor-bearing controls from Study 1. B) Scatter plot of the 209 overlapping genes identified in (A). C) Canonical pathway analysis using the 147 overlapping genes identified in (A).
Appendix Figure S13. Standard volcano plot from RNA-seq analyses of Study 1 gastrocnemius muscles showing DEGs for GTx-024-treated tumor-free controls versus vehicle-treated tumor-free controls (colored red). Log2-transformed fold change (FC) in expression is plotted on the x-axis and -log10 transformed Benjamini-Hochburg adjusted p-values are plotted on the y-axis.

Appendix Tables
Appendix Table S1A-D (related to Figure 1): p-values from statistical comparisons. One way ANOVA followed by Tukey's multiple comparison test.    Appendix Table S10 (related to Figure EV2): p-values from statistical comparisons. One way ANOVA followed by Tukey's multiple comparison test.
A. Luteinizing Hormone ( Figure EV2A) B. Spleen Mass ( Figure EV2C) Appendix Table S11 (related to Figure EV3): p-values from statistical comparisons. One way ANOVA followed by Tukey's multiple comparison test.