The effects of an ActRIIb receptor Fc fusion protein ligand trap in juvenile simian immunodeficiency virus-infected rhesus macaques

Abstract

There are no approved therapies for muscle wasting in children infected with human immunodeficiency virus (HIV), which portends poor disease outcomes. To determine whether a soluble ActRIIb receptor Fc fusion protein (ActRIIB.Fc), a ligand trap for TGF-β/activin family members including myostatin, can prevent or restore loss of lean body mass and body weight in simian immunodeficiency virus (SIV)-infected juvenile rhesus macaques (Macaca mulatta). Fourteen pair-housed, juvenile male rhesus macaques were inoculated with SIVmac239 and, 4 wk postinoculation (WPI) treated with intramuscular injections of 10 mg ⋅ kg⁻¹ ⋅ wk⁻¹ ActRIIB.Fc or saline placebo. Body weight, lean body mass, SIV titers, and somatometric measurements were assessed monthly for 16 wk. Age-matched SIV-infected rhesus macaques were injected with saline. Intervention groups did not differ at baseline. Gains in lean mass were significantly greater in the ActRIIB.Fc group than in the placebo group (P < 0.001). Administration of ActRIIB.Fc was associated with greater gains in body weight (P = 0.01) and upper arm circumference than placebo. Serum CD4⁺ T-lymphocyte counts and SIV copy numbers did not differ between groups. Administration of ActRIIB.Fc was associated with higher muscle expression of myostatin than placebo. ActRIIB.Fc effectively blocked and reversed loss of body weight, lean mass, and fat mass in juvenile SIV-infected rhesus macaques.—O’Connell, K. E., Guo, W., Serra, C., Beck, M., Wachtman, L., Hoggatt, A., Xia, D., Pearson, C., Knight, H., O’Connell, M., Miller, A. D., Westmoreland, S. V., Bhasin, S. The effects of an ActRIIb receptor Fc fusion protein ligand trap in juvenile simian immunodeficiency virus-infected rhesus macaques.

Loss of body weight and muscle wasting are common sequelae of many chronic illnesses, including cancer, heart failure, chronic obstructive lung disease, end-stage renal disease, and HIV infection (1–5). Loss of metabolically active lean muscle tissue has been linked to adverse disease outcomes, including more rapid disease progression and increased risk of opportunistic infections, hospitalizations, and mortality in HIV-infected patients (6). Furthermore, loss of muscle mass often leads to impaired physical function, increasing the risk of dependence and disability (7). Although the prevalence of acquired immunodeficiency syndrome (AIDS) wasting has decreased in the developed countries due to the widespread availability and use of antiretroviral drugs, it continues to be a significant problem worldwide especially in Africa, Asia, and even in the United States (5, 8, 9).

Currently, the only approved therapies for the treatment of AIDS wasting are orexigenic drugs such as megestrol acetate or recombinant human growth hormone (10–12). Orexigenic agents such as megestrol acetate can promote weight gain, but they do not promote lean mass accretion (13). In fact, due to its antiandrogenic effects, the administration of megestrol acetate is associated with loss of lean body mass (13, 14). Recombinant human growth hormone promotes lean mass accretion, but it has not been shown to improve muscle strength or function (15). Consequently, the past decade has witnessed considerable investment in the development of novel pharmacologic therapies for the treatment of muscle wasting associated with HIV and other chronic diseases. Among these novel pharmacologic therapies, myostatin antagonists and selective androgen receptor modulators are the farthest along in development.

Myostatin, also known as growth and differentiation factor 8 (GDF-8), is a highly conserved member of the TGF-β superfamily and is expressed in adult skeletal muscle (16–18), as well as in adipose and cardiac muscle (19). Spontaneous mutations in the myostatin gene are associated with hypermuscularity in a number of mammalian and nonmammalian species (19–23). Conversely, overexpression of myostatin is associated with loss of muscle mass (24, 25). Myostatin has been postulated to induce muscle atrophy by several mechanisms, including inhibition of myoblast proliferation, induction of the ubiquitin-proteasome pathway, and inhibition of the IGF1-Akt pathway (26). Consequently, a number of strategies have been developed to inhibit myostatin action, which have been shown to increase muscle mass in postnatal life in preclinical models and in a limited number of human trials (27–29).

Here, we tested the hypothesis that ActRIIB.Fc-mediated inhibition of myostatin and other TGF-β/activin family members would reverse the muscle wasting that occurs during HIV infection. Accordingly, we conducted a placebo-controlled, proof-of-concept trial in a nonhuman primate model of HIV to determine the efficacy of ActRIIB.Fc, a ligand trap that blocks the actions of myostatin and other members of the TGF-β/activin family, which contribute to muscle loss and cachexia associated with illness, in reversing weight loss associated with SIV infection and in promoting lean mass accretion. Muscle wasting and weight loss are a feature of simian acquired immunodeficiency syndrome (SAIDS) in rhesus macaques infected with the SIV. Moreover, juvenile macaques experience a failure to thrive when inoculated with SIV similar to that observed in HIV-infected children (30). The SIV model closely parallels the human disease and therefore was utilized to study the effects of ActRIIB.Fc on muscle wasting in HIV/AIDS.

MATERIALS AND METHODS

The rhesus macaques were pair-housed in a centralized ABSL-2 facility at the New England Primate Research Center and were maintained in accordance with the Guide for the Care and Use of Laboratory Animals (ILAR, 8th edition, 2011). The study protocol was approved by Harvard Medical School’s Standing Committee on Animals. Macaques were fed a certified commercial primate diet (8714; Teklad, Harlan Laboratories, South Easton, MA, USA) and provided fresh water ad libitum.

Animal assignment

Fourteen male juvenile (2.5–3 yr old) rhesus macaques (M. mulatta) were studied. Animals were randomized based on age and major histocompatibility complex status. All macaques were intravenously inoculated with SIVmac239 (50 ng of p27 viral-antigen equivalent). Seven macaques were treated with ActRIIB.Fc (10 mg ⋅ kg⁻¹ ⋅ wk⁻¹), an experimental-grade ligand trap kindly provided by Dr. Carl Morris (Pfizer, Incorporated, Cambridge, MA, USA) by intramuscular injection weekly for 12 wk starting 4 wk after SIV inoculation. The remaining 7 were administered a saline placebo injection intramuscularly weekly for 12 wk.

Body composition

Body composition was determined using dual-energy X-ray absorptiometry (DXA) and by somatometrics at baseline and 4, 8, 12, and 16 WPI. Animals were sedated with either ketamine HCl (10 mg/kg for procedures requiring short sedation) or tiletamine (5mg/kg for procedures requiring anesthesia for a longer duration) intramuscularly. The same technicians performed these measurements. A standard tape measure was used with consistent tightness. The following measurements were taken: crown-heel length (crown to heel of straightened leg), crown-rump length (crown to base of tail), upper arm circumference (midway between shoulder and elbow), upper leg circumference (midway between hip and knee), chest circumference (across nipples), and head circumference (above ears, across brow.)

DXA scans were performed using a total body scanner (Lunar; GE Healthcare, Westborough, MA, USA), by generating X-rays at 2 energy levels (40 and 70 kVp) as described previously (31). Scans were performed in conjunction with anthropomorphic measurements and did not require additional sedation. A series of transverse scans was made from head to toe, at 1-cm intervals. Data were collected for ∼120 pixel elements/transverse scan, with a pixel size of 5 × 10 mm. Body composition was derived using a computer algorithms provided by the manufacturer.

SIV copy number PCR

Plasma viral loads were determined using procedures detailed in Cline et al. (32).

Muscle histomorphometry and immunohistochemistry

Immunofluorescent staining was carried out on 8 μm thick cryosections of the triceps muscle. Sections were rehydrated in PBS, fixed in 4% paraformaldehyde (PFA) for 10 min at 4°C, then permeabilized with 0.2% Triton X-100 for 12 min and blocked in 5% normal goat serum for 1 h at room temperature. Sections were incubated with the primary antibodies overnight at 4°C in 1% bovine serum albumin, then with secondary antibodies for 1 h at room temperature. Primary antibodies used were as follows: mouse monoclonal anti-laminin α2 (MAB1922, 1:250; Millipore, Billerica, MA, USA) for cross-sectional area (CSA) staining, mouse monoclonal anti-slow myosin heavy chain (MyHC) (M8421, 1:1500; Sigma-Aldrich, St. Louis, MO, USA), mouse monoclonal anti–fast MyHC (A4335, 1:100; Sigma-Aldrich), and mouse monoclonal anti–neural cell adhesion molecule-1 (NCAM1) (MEM-188, 1:80; Abnova, Taipei City, Taiwan) rabbit polyclonal anti-laminin (ab11575, 1:30; Abcam, Cambridge, MA, USA) for the MyHC counterstaining. Secondary antibodies used were: goat polyclonal to mouse Cy3-conjugated (115-165-062, 1:300; Jackson Immunoresearch Laboratories, West Grove, PA, USA), and goat polyclonal to rabbit FITC-conjugated (111-095-144, 1:200; Jackson Immunoresearch Laboratories). Nuclei were counterstained with DAPI. Images were acquired using a Nikon Eclipse TE2000-E microscope (Nikon Instruments Incorporated, Melville, NY, USA) and analyzed using SPOT advanced software (Diagnostic Instruments Incorporated, Sterling Heights, MI, USA). Muscle fiber CSA was calculated using Vision Assistant Software LabView (National Instruments Corporation, Austin, TX, USA). For the staining of slow and fast MyHC, fixation in 4% PFA was omitted.

Western blot analysis

Pulverized triceps muscle was homogenized in 1× cell lysis buffer (#9803; Cell Signaling Technology, Beverly, MA, USA) supplemented with 0.1% SDS and 1 mM PMSF. Primary antibodies were obtained from Cell Signaling Technology (Phospho-Akt, #4060; Akt, #9272; phospho-JNK, #9251; JNK, #9258; phospho-FOXO3a, #13129); Bethyl (Foxo3a, A300-453A); R&D (myostatin, AF788); Santa Cruz Biotechnology (ActRIIB, sc-5665; β-actin, sc-47778; follistatin, sc-30194, Murf1, sc-32920; Santa Cruz, CA, USA). Proteins were separated by SDS-PAGE, transferred to PVDF membrane, blocked in 5% milk, and incubated with the first antibody overnight at 4°C with gentle shaking. Membrane was washed and incubated in second antibody at room temperature for 1 h. Proteins were detected by exposure to X-ray film after incubation with LumiGLO (#7003; Cell Signaling Technology) and quantified by densitometry using NIH ImageJ program.

RT-qPCR

Pulverized triceps muscle was homogenized in Trizol (#15596018; Invitrogen, Carlsbad, CA, USA). RNA extraction was completed using the RNeasy plus mini kit (#74134; Qiagen, Valencia, CA, USA). Single-strand cDNA was synthesized using ProtoScript First Strand cDNA Synthesis Kit (#E6300S; New England Biolabs, Ipswich, MA, USA) following the manufacturer’s instruction. RT-PCR was performed using SYBR master mix on a 7500 Sequence Detection System (Thermo Fisher Scientific, Waltham, MA, USA). The primer sequences are: myostatin forward: 5′-CCCGTCGAGACTCCTACAA-3′, myostatin reverse: 5′-CAGTGCCTGGGTTCATGTCA-3′, β-actin forward: 5′-GGACTTCGAGCAGGAGATGG-3′, β actin reverse: 5′-GCTCGTTGCCAATGGTGATG-3′. Primer sequences for additional targets are listed in Supplemental Table 1.

Proteasomal activity

Muscle sample was prepared in buffer containing 50 mM Tris, 0.5 mM EDTA, 2 mM ATP, 1 mM DTT, 0.025% digitonin, 0.5% triton. Proteasomal activity was measured using a commercial assay kit (#K245-100; Biovision, Milpitas, CA, USA), following the manufacturer’s instruction.

Clinical pathology

Phlebotomy was performed via femoral vein in conjunction with anthropomorphic measurements and did not require additional sedation. At time points where anthropomorphic measurements were not performed, sedation was achieved with ketamine HCl (10 mg/ml) intramuscularly. A Vacutainer system (BD, Franklin Lakes, NJ, USA) with a 22 gauge needle was used to collect blood samples for hematology and flow cytometric analysis at 0, 1, 2, 3, 4, 8, 12, and 16 WPI. Hematologic analysis was performed within 4 h on a Hemavet 1700 Analyzer (Drew Scientific, Waterbury, CT, USA). Manual review of blood smears was performed on Wright-Giemsa stained blood films. Blood chemistries were analyzed in a commercial veterinary laboratory (IDEXX, North Grafton, MA, USA).

Flow cytometry and ELISAs

Blood lymphocyte subsets were determined using fluorochrome-conjugated monoclonal antibodies against phenotypic cell surface antigens, followed by flow cytometric analysis. Specifically, the following antibodies were used: CD3-FITC, CD8-PerCP, and CD4-APC (BD Pharmingen, Franklin Lakes, NJ, USA).

A MILLIPLEX Map Kit (Millipore) was used to measure IFN-γ, IL-10, IL-15, IL-1β, IL-6, IL-8, TNF-α. Resistin and insulin were measured using ELISA kits (Millipore) and read on a Dynex plate reader (MRX revelation 4.25; Chantilly, VA, USA). For the Milliplex assays, plates were run on a Luminex 200 and analyzed for median fluorescence intensity using a 5-parameter logistic method.

Euthanasia and pathology

Animals were sedated with ketamine HCl (10 mg/kg) intramuscularly and desired antemortem samples collected. Beuthanasia-D (sodium pentobarbital and phenytoin sodium) was administered in either the saphenous or cephalic vein. Respiratory arrest was immediately followed by cardiac arrest. Death was confirmed by thoracic auscultation by a veterinarian. Complete necropsies were performed on all macaques, including collection of triceps, biceps, quadriceps, pectorals, gastrocnemius, and diaphragm for histomorphometry and molecular analyses.

Statistical analyses

All statistics were performed using Prism (version 4.0c; GraphPad Software Incorporated, La Jolla, CA, USA). Repeated measures analysis of variance with an intervention factor (ActRIIB.Fc or saline) and a time-in-treatment factor was performed to assess the effects of intervention over time in somatometric, DXA, clinical pathology, and flow cytometry parameters. When F values were significant, Bonferroni’s post hoc test was used. Nonparametric t tests were used to compare end point data sets and changes from baseline. All data in graphs are presented as means ± sem. Tables 1 and 2 show data as means ± sd. Effect sizes to estimate the magnitude of change were calculated for total lean mass, total fat mass, and body weight.

TABLE 1.

Baseline characteristics of the animals

Variable	ActRIIB.Fc	Placebo	Variable	ActRIIB.Fc	Placebo
Somatometrics			Chemistry
Body weight (kg)	4.2 ± 0.4	4.0 ± 0.5	Glucose (mg/dl)	56.8 ± 8.3	63.3 ± 13
Crown-rump length (cm)	42 ± 2.1	40.1 ± 1.9	Albumin (g/dl)	3.7 ± 0.7	3.9 ± 0.9
Upper arm circumference (cm)	12.7 ± 0.5	12.1 ± 1.0	Total protein (g/dl)	5.5 ± 0.9	6.0 ± 1.4
Upper leg circumference (cm)	17.1 ± 2.0	16.9 ± 1.0	Creatinine (mg/dl)	0.46 ± 0.11	0.57 ± 0.20
Chest circumference (cm)	29.5 ± 1.6	29.3 ± 1.3	Cholesterol (mg/dl)	119 ± 20	137 ± 26
Head circumference (cm)	26.7 ± 1.0	26.1 ± 0.3	Triglyceride (mg/dl)	28.4 ± 6.3	32.7 ± 10.6
BMI	24 ± 2.8	24.6 ± 2.4
			DXA
Hematology			Whole body lean (g)	3608 ± 284	3319 ± 424
WBC (k/μl)	7.2 ± 1.5	8.6 ± 1.5	Whole body fat (g)	374 ± 165	429 ± 126
lymph (k/μl)	2.4 ± 1.1	2.4 ± 1.3	Trunk lean (g)	1843 ± 177	1699 ± 216
RBC (M/μl)	6.2 ± 0.5	5.9 ± 0.5	Trunk fat (g)	156 ± 92	195 ± 73
HB (g/dl)	14.1 ± 1	13.2 ± 0.9	Arms lean (g)	479 ± 29	442 ± 74
HCT (%)	45 ± 3.7	42.3 ± 3.6	Arms fat (g)	68.5 ± 21.5	72.9 ± 16.6
RDW (%)	15.6 ± 0.6	15.9 ± 1.2	Legs lean (g)	903 ± 96	836 ± 119
PLT (k/μl)	536 ± 83	496 ± 97	Legs fat (g)	118 ± 37	123 ± 28

The data are means ± sd. n = 6 in the group assigned to ActRIIB.Fc and n = 7 in the group assigned to placebo. BMI, body mass index; HB, hemoglobin; HCT, hematocrit; PLT, platelets.

TABLE 2.

Hematologic and immunologic characteristics of the animals at 16 wk postinfection

Variable	ActRIIB.Fc	Placebo	P (from baseline)
WBC × 103/μl	7.0 ± 1.5	8.5 ± 4.6	0.840
HCT %	36 ± 3.1	38 ± 4.5	0.290
RDW %	19.3 ± 3.5	15.4 ± 1.3	0.030*
CD4+ cells/μl	42.4 ± 14.3	38.7 ± 15.1	0.840
CD8+ cells/μl	85.9 ± 50.1	87.5 ± 26.9	0.790
CD4/CD8 ratio	0.59 ± 0.3	0.48 ± 0.3	0.630
SIV particles/ml	24.3 ± 3.9	10.8 ± 1.3	>0.999

The data are means ± sd. n = 6 in the group assigned to ActRIIB.Fc and n = 7 in the group assigned to placebo with P values based on absolute change from baseline. HCT, hematocrit. *The change in RDW% from baseline was statistically significant.

RESULTS

Baseline characteristics and overall health were similar between groups

The animals in the 2 intervention arms did not differ in their baseline characteristics (Table 1). Specifically, plasma SIV copy numbers, CD4⁺ and CD8⁺ T-lymphocyte counts, body weight, and composition were similar between the 2 groups. The macaques were monitored daily for general behavior, attitude, and appetite. With the exception of 1 animal demonstrating a rapid progression to SAIDs in the active intervention arm that was euthanized at 14 wk after SIV inoculation and excluded from analysis, no animals showed any signs of clinical illness. Mild gingival hyperplasia (3 in the ActRIIB.Fc group and 2 in the placebo group) and epistaxis (2 in ActRIIB.Fc arm and 3 in placebo arm) were noted in several animals. Diarrhea was noted in all animals. Epistaxis and diarrhea were attributed to bacterial infections and were either self-limiting or clinically responsive to antibiotic therapy.

Whole body and appendicular lean and fat mass increased during ActRIIB.Fc administration

Whole body and appendicular lean and fat mass were measured using DXA scans at baseline and at 4, 8, and 16 wk (Fig. 1). Whole body lean mass remained unchanged in animals assigned to the placebo group, but increased significantly over time in those treated with ActRIIB.Fc (Fig. 1A). There was a significant time-treatment interaction (P = 0.004); the gain in whole body lean mass from baseline was numerically greater in the ActRIIB.Fc group (320 ± 166 g) than in the placebo group (68 ± 344 g, P = 0.051; Fig. 1B), although the difference was not statistically significant. It is noteworthy that the effect size of the intervention for total lean mass (1.09 sd units) was large. Likewise, changes in arm and leg lean mass demonstrated significant time-treatment interaction (arms: P = 0.008, Fig. 1E; legs: P = 0.011, Fig. 1I); the gains in lean mass in the arms (P = 0.051) were numerically greater in ActRIIB.Fc-treated animals than in those treated with placebo (Fig. 1F), although the gains in leg lean mass were not significantly different (Fig. 1J). Whole body fat mass showed a significant time-treatment interaction (P = 0.006, Fig. 1C) and was increased in ActRIIB.Fc-treated animals during the intervention period (374 ± 165 g to 572 ± 262 g) but decreased in the placebo-treated group (429 ± 126 g to 362 ± 166 g, Fig. 1D). The intervention’s effect size of 1.26 sd units was large. The change in arm fat mass over time was greater in the ActRIIB.Fc treatment group than in the placebo group (time-treatment interaction P = 0.04, Fig. 1G, H). There was no significant change in leg fat mass over time (Fig. 1K), and the change from baseline in leg fat mass did not differ between groups (Fig 1L).

Figure 1.

Effects of ActRIIB.Fc or placebo on whole body lean and fat mass and arm and leg lean mass in SIV-infected juvenile rhesus macaques. Effects of ActRIIB.Fc or placebo on whole body lean (A) and fat (C) mass, and arm and leg lean (E, I) and fat (G, K) mass in SIV-infected juvenile rhesus macaques. A) The changes from baseline in whole body (B), arm (F), and leg (J) lean mass (g), and whole body (D), arm (H), and leg (L) fat mass are also shown. For longitudinal analyses, data are shown as means ± sem and P values indicate time-treatment interactions. For endpoint comparisons, error bars indicate means ± sem and P values indicate t tests for comparison of absolute changes from baseline. n = 6 in the group assigned to ActRIIB.Fc and n = 7 in the group assigned to placebo.

Loss of body weight is reversed by ActRIIB.Fc therapy

Body weight showed significant interaction between time and treatment (P < 0.001, Fig. 2A). The gain in body weight from baseline was greater in animals treated with ActRIIB.Fc than in those treated with placebo (P = 0.01, Fig. 2B). The intervention’s effect size was large at 2.33 sd units. Similarly, body mass index showed significant interaction between time and treatment (P < 0.001, Fig. 2C). The increase in body mass index from baseline in the ActRIIB.Fc group (1.5 ± 2.6 kg/cm²) was greater than that in the placebo group (−3.7 ± 2.3 kg/cm², P = 0.01, Fig. 2D). The change in upper arm circumference over time was numerically greater in the ActRIIB.Fc group (0.9 ± 1.2 cm) than in the placebo group (0.3 ± 0.5 cm) but did not reach statistical significance (P = 0.09, Fig. 2E, F). Crown-rump length increased with time but the changes were similar in the 2 intervention arms (P = 0.14), indicative of the continued growth in juveniles (data not shown). Chest circumference increased over the course of study in the treatment group with a significant time-treatment interaction (P = 0.008, Fig. 2G) and change from baseline (P = 0.04, Fig 2H). Increases in leg circumference were similar between groups (P = 0.6, data not shown).

Figure 2.

Body weight and other somatometric measures in animals randomized to ActRIIB.Fc or placebo. Effects of ActRIIB.Fc or placebo on body weight (A), body mass index (BMI; C), mean upper arm circumference (E), and chest circumference (H). For longitudinal analyses, data are shown as means ± sem and P values indicate time-treatment interactions. The changes from baseline in body weight (B), BMI (D), upper arm circumference (F), and chest circumference (I) are also shown. For end point comparisons, error bars indicate means ± sem and P values indicate t tests of absolute changes from baseline. n = 6 in the group assigned to ActRIIB.Fc and n = 7 in the group assigned to placebo.

Necropsy findings

There were no significant differences in organ weights at necropsy. All animals had typical histologic evidence of SIV infection characterized by lymphoproliferation in multiple organs. Although they did not have any grossly detectable signs of progression, 2 animals in the intervention group had SAIDs-defining microscopic lesions including 1 opportunistic infection with enteropathogenic Escherichia coli and 1 giant cell pneumonia. The rapid progressor (euthanized at 14 WPI and excluded) had Mycobacterium avium infection in the gastrointestinal tract and lymph nodes.

Muscle fiber cross-sectional area, muscle fiber typing, and satellite cell number

To evaluate if the increase in lean body mass in the animals treated with ActRIIB.Fc was associated with changes in muscle fiber size, we quantified the muscle fibers CSA of the triceps muscle. As shown in Fig. 3A, the ActRIIB.Fc-treated animals showed a small (5.4%) numerical increase in fiber CSA in examined sections, although this increase was not statistically significant. In addition, the triceps muscle of the ActRIIB. Fc-treated animals showed a reduced number of slow muscle fibers (slow MyHC⁺, Fig. 3B), and a parallel significant increase (P = 0.04) of fast fibers (fast MyHC⁺, Fig. 3C). Number of satellite cells (NCAM1⁺) was numerically greater in the triceps muscle of the ActRIIB.Fc-treated animals than in saline-treated animals, although this difference did not reach statistical significance (Fig. 3D).

Figure 3.

Effect of ActRIIB.Fc on muscle histomorphometry. The muscle fiber CSA (A), relative proportion of slow fibers (B), relative proportion of fast fibers (C), and satellite cell number (D) are shown. Images were processed and analyzed using the SPOT advanced software and fiber CSA was calculated using the Vision Assistant Software LabView. Error bars indicate means ± sem. n = 6 in the group assigned to ActRIIB.Fc and n = 7 in the group assigned to placebo.

Treatment with ActRIIB.Fc increases myostatin expression

The ActRIIB.Fc administration was associated with an increased level of myostatin mRNA expression (P = 0.04), as would be expected after effective blockade of signaling through ActRIIB receptor, and consistent with target engagement by the drug (Fig. 4A). However, no difference was detected in the level of full length or mature form of myostatin protein (Supplemental Fig. 1A, B). Similarly, we found no difference in muscle expression of follistatin, ActRIIB, or Smad2/3 proteins (Supplemental Fig. 1C–E). Consistent with these findings, we found no significant difference in the steady-state expression of the myostatin/TGF-β target genes, such as Smad7, ATF4, and decorin, between the 2 groups (Supplemental Fig. 2C, D, G). Expression of activin A, a low-affinity ligand for ActRIIB, did not differ between the 2 groups (Supplemental Fig. 2A). Similarly, treatment with ActRIIB.Fc did not affect muscle expression of other important growth factors and cytokines (Supplemental Figs. 2 and 3) and genes involved in energy metabolism, anabolic, and catabolic activities, including myoD, myogenin, IGF1, IRS1, glycogen synthase, ribosomal protein S5, cathepsin B, calpain 2, and proapoptotic protein Bnip3 (Supplemental Figs. 4–6).

Figure 4.

Effect of ActRIIB.Fc on muscle gene expression, protein expression, and proteasomal activity in the triceps muscle of SIV-infected juvenile male rhesus macaques. Mst gene expression (A), proteosomal activity (B), phosphorylated foxo3a (C), and phosphorylated JNK1 (D) are shown. Error bars indicate means ± sem. n = 6 in the group assigned to ActRIIB.Fc and n = 7 in the group assigned to placebo.

The proteasomal activity was numerically lower in animals treated with ActRIIB.Fc than in those administered placebo (P = 0.051, Fig. 4B). Skeletal muscle hypertrophy and atrophy are both largely regulated by insulin/IGF1 signaling, through Akt-mediated activation of protein synthesis and inactivation of protein degradation. Therefore, we compared the intramuscular expression of activated Akt and its downstream target Foxo3a, a transcription factor that regulates muscle atrogen Murf1 and MAFbx1. Unexpectedly, the expression of phospho-Akt and Akt were not different between groups, each with large variations (Supplemental Fig. 1F, G). A near 2-fold increase of phospho-Foxo3a:Foxo3a ratio was detected in the ActRIIB.Fc group, although the difference did not reach statistical significance due to the large individual variations and limited sample size (Fig. 4C and Supplemental Fig. 1H–I). No difference was detected in the Foxo3a target genes, including MurF1 and MAFbx1 (Supplemental Figs. 1J and 4C, D). Similarly, the animals treated with ActRIIB.Fc showed a near 2-fold decrease in the expression of phospho-JNK1 (Fig. 4D and Supplemental Fig. 1K–L), a stress-activated protein kinase, compared with placebo-treated animals, although the difference did not reach statistical significance.

Hematocrit decrease is greater in ActRIIB.Fc-treated macaques compared with saline-treated controls

Total white blood cell (WBC) counts, neutrophil and lymphocyte counts, and platelet counts did not differ significantly between the 2 groups (Table 2). There was a decrease in WBC count over time in both groups (P = 0.015) but the change in WBC count from baseline did not differ between groups (P = 0.8). The hematocrit decreased over time in both groups (P < 0.0001), but change in hematocrit from baseline between the 2 groups was not significantly different (P = 0.29, Table 2). Because the decrease in hematocrit was numerically greater in animals assigned to ActRIIB.Fc group than in those assigned to placebo group, we checked the stool specimens for occult blood loss using the hemoccult test. Only 1 animal in the treatment group had a positive hemoccult test for occult blood. Similarly, the number of animals with epistaxis did not differ between groups (2 in the ActRIIB.Fc group and 3 in the placebo group). The red cell distribution width (RDW) increased significantly from baseline (P = 0.03, Table 2) in the treated group over the course of study (15.6 ± 0.6% to 19.3 ± 3.5%) with a significant time-treatment interaction (P = 0.0014).

Clinical chemistries, including albumin, alanine aminotransferase, aspartate aminotransferase, bilirubin, blood urea nitrogen (BUN), glucose, cholesterol, triglycerides, creatinine, calcium, phosphorus, and alkaline phosphatase at 0, 4, 8, 12, and 16 WPI did not differ between groups. The BUN at 16 WPI (P = 0.08) was slightly higher in the placebo group than in the ActRIIB.Fc group, but was within the normal reference range for rhesus macaques (data not shown).

CD4⁺/CD8⁺ lymphocytes, SIV copy numbers, and cytokines are similar between ActRIIB.Fc-treated macaques and saline-treated controls

The CD4⁺ T lymphocyte counts and CD4⁺/CD8⁺ lymphocyte count ratios changed significantly over time (P < 0.0001), but the counts did not differ between groups (Table 2). SIV copy numbers changed over time, but did not differ significantly between groups (Table 2). Circulating levels of several cytokines including IL-10, IL-6, and TNF-α, IL-15, IFN-γ, IL-8, and IL-1β were measured in 3–6 animals in the ActRIIB.Fc group and 5–7 animals in the placebo group at 0, 4, 8, 12, and 16 WPI, but did not differ between groups. Insulin, glucose, and resistin levels were determined by ELISA in all animals at 0, 4, 8, 12, and 16 WPI and did not differ between groups (data not shown).

Discussion

These data constitute the first demonstration that administration of an ActRIIB.Fc ligand trap effectively reverses the weight loss associated with SIV infection in a nonhuman primate model of HIV-associated weight loss (Fig. 2). The weight gain induced by administration of the ActRIIB.Fc ligand trap was composed of gains in both the lean and fat compartments (Fig. 1). Furthermore, the gains in lean mass were distributed in both the upper and lower extremities (Fig. 1). The changes in body composition assessed by DXA were generally confirmed by concordant changes in the anthropometric measures (Figs. 1 and 2). Thus, the intervention was efficacious in promoting body weight gains and lean mass accretion in SIV-infected macaques.

The intervention was well tolerated in the monkeys. One animal in the treatment group exhibited rapid disease progression requiring euthanasia, which is consistent with the reported incidence of 5% rapid progressors following inoculation with SIVmac239 (33). Two additional animals in the active intervention group had histologic findings consistent with SAIDS but no clinical signs. Changes in CD4⁺ T lymphocytes and SIV copy numbers did not differ between groups (Table 2). Some of the animals experienced diarrhea in a manner typical of SIV-infected macaques (30). Gingival hyperplasia and epistaxis were noted in several animals but were not associated with group assignment. Both are common findings in rhesus macaques within this colony and are not likely associated with the therapeutic intervention. Although the hematocrit was numerically lower in animals treated with ActRIIB antagonist, we did not find evidence of bleeding in these animals. Only 1 animal had occult blood in the stool. Neither the sample size nor the study duration was sufficient for a comprehensive assessment of safety.

The mechanisms by which ActRIIB.Fc ligand trap reverses SIV-associated loss of body weight and muscle loss remain poorly understood and need further investigation. Here we show that administration of ActRIIB.Fc engaged the target and was associated with trends toward reduced proteasomal activity (Fig. 4). These data suggest that treatment with ActRIIB.Fc likely blocks muscle protein breakdown. In experimental models of cachexia in mice bearing tumor or chronic kidney disease, the administration of myostatin inhibitors has been reported to suppress Smad3 signaling and improve insulin signaling to inhibit Foxo3a activity and down-regulate the expression of muscle atrogenes MurF1 and MAFbx1 (34, 35). Other rodent studies using similar strategies have also reported improvement of muscle wasting but found no effect on Murf1 or MAFbx1 expression (36). The reversal of weight loss and muscle loss by ActRIIB.Fc in this primate model of HIV was not associated with the previously reported effects on atrogene expression. We also did not find significant changes in genes involved in other signaling pathways, including myoD, myogenin, IGF1, IRS1, glycogen synthase, and ribosomal protein S5 (Supplemental Data). The expression of other genes involved in protein degradation, including cathepsin B, calpain 2, and proapoptotic protein Bnip3, also did not differ between groups (Supplemental Data). However, we only assessed steady-state gene expression levels at the study end point; it is possible that initial changes at earlier time points in the regulatory pathways that mediated the observed increase in muscle mass could have been missed. Also, the SIV-infected juvenile macaques display milder weight loss and failure to thrive rather than severe wasting observed in tumor-bearing mice. Furthermore, species differences in the mechanisms of cancer-related cachexia between primates and mice have been well recognized (1).

Skeletal muscle inflammation has been shown to play a role in insulin resistance (37, 38) as well as age- and disease-related muscle loss (39–41). However, serum cytokine analysis revealed no difference in common markers of systemic inflammation. We also found no difference in muscle expression of TNF-α and IL-6, both of which were expressed at very low level (Supplemental Data). Muscle expression of other NFkB downstream genes, including iNOS and MMP3, were below the limit of detection. The ActRIIB.Fc-treated animals displayed lower expression of phospho-JNK1/2, a stress-activated protein kinase (Fig. 4 and Supplemental Data). Myostatin has been reported to activate JNK1 in muscle of aging mice (42). Interestingly, JNK activation is involved in up-regulation of proteasomal activity (43) and increased proteasomal protein degradation (44). This is in line with our findings of the reduced proteasomal activity in skeletal muscle of ActRIIB.Fc-treated animals.

It is remarkable that the animals in the ActRIIB.Fc group gained substantially greater amounts of lean as well as fat mass than those assigned to the placebo group (Fig. 1). These changes were associated with an increase in type 2 fibers (Fig. 3). This last observation indicates that the ActRIIB.Fc induces a shift of the muscle metabolic profiling, consistent with what has been observed in myostatin null mouse (45). The effects of myostatin inhibitors on fat mass have varied in different contexts. The genetic disruption of myostatin is associated with resistance to fat accumulation (46). Loss of fat is noted during aging of myostatin null mice, especially in females (47). Postnatal administration of myostatin inhibitors to mice has been associated with increased muscle mass and strength, as well as modest loss of fat mass (48–50). Of note, weight loss in HIV-infected individuals and in other states characterized by cachexia comprises loss of lean as well as fat mass (27, 51). Similarly, restoration of health by administration of effective antiretroviral therapy in patients with AIDS wasting restores both lean and fat mass (52). Therefore, our observation of increased fat mass in response to ActRIIB.Fc treatment in the SIV-infected animal may be a marker of an improvement of general health condition instead of a direct effect of ActRIIB.Fc.

The dose and regimen of ActRIIB.Fc used in the trial was effective in restoring body weight and lean body mass. The gains in body weight and lean body mass associated with administration of the ActRIIB.Fc were substantial and averaged 10–15%; the intervention effect was quite large averaging 1.09, 1.26, and 2.33 sd units for lean mass, fat mass, and body weight, respectively. As a reference for perspective, the gains in lean body mass reported in randomized trials of testosterone and selective androgen receptor modulators have typically averaged ∼2.5–3% (53). Although we do not know what the clinical meaningful gains in body weight and lean body mass are in rhesus macaque, the minimal clinically important difference (MCID) in lean mass in human trials of function promoting therapies has been deemed to be 1.5 kg (∼3%) (54). The observed improvements in body weight and lean body mass compare favorably with the effects observed in human trials of other function promoting therapies and the consensus MCID estimates.

Our study also has some limitations. The number of animals was relatively small, which did not allow comprehensive assessment of safety in this proof-of-concept efficacy trial. Similarly, the intervention duration was insufficient to allow assessment of effects on survival or disease progression. A longer study period may have also allowed for some of our outcome parameters to reach statistical significance. In addition to myostatin, ActRIIB.Fc ligand trap blocks the actions of other members of the TGF-β/activin family, including activin. Activins are potent negative regulators of muscle mass and are increased in cancer cachexia in mice (55). Lee (56, 57) shown that genetic disruption of other members of the TGF-β/activin family, such as Activin A, in addition to myostatin, leads to more pronounced gains in muscle mass than disruption of myostatin alone. Follistatin has also been proposed as a therapy for a mouse model of muscle degeneration (58). In addition to myostatin, ActRIIB.Fc also blocks the action of GDF-11 (59) but this has been previously shown to not impact muscle size (60).

Weight loss in HIV-infection portends poor disease outcome. The loss of lean tissue in HIV affects muscle strength, physical performance, and immune function and is associated with increased morbidity and mortality (61). Our data demonstrate that ActRIIB.Fc ligand trap can reverse the cachexia associated with the immunodeficiency virus. Although weight gain associated with effective antiretroviral therapy is associated with improved survival, we do not know whether reversal of weight loss and muscle wasting by ActRIIB.Fc would also be associated with improved disease outcomes. In mice with cancer cachexia, the administration of an ActRIIB.Fc ligand trap attenuated muscle wasting and increased survival (27). The data from this proof-of-concept efficacy trial in nonhuman primate model of SIV-wasting provide strong rationale for clinical trials in HIV-infected patients with weight loss to determine the effects of myostatin inhibition on patient-important disease outcomes.

Glossary

AIDS

acquired immunodeficiency syndrome

BUN

blood urea nitrogen

DXA

dual-energy X-ray absorptiometry

GDF-8

growth and differentiation factor 8

HIV

human immunodeficiency virus

NCAM1

neural cell adhesion molecule-1

RDW

red cell distribution width

SAIDS

simian acquired immunodeficiency syndrome

SIV

simian immunodeficiency virus

WBC

white blood cell count

WPI

weeks postinoculation