Research ArticleMYOPATHY

Targeting protein homeostasis in sporadic inclusion body myositis

See allHide authors and affiliations

Science Translational Medicine  23 Mar 2016:
Vol. 8, Issue 331, pp. 331ra41
DOI: 10.1126/scitranslmed.aad4583

Targeting protein dyshomeostasis in myopathy

Sporadic inclusion body myositis (sIBM) is a debilitating adult myopathy that is difficult to treat. Although both inflammation and protein dyshomeostasis have been implicated in sIBM pathogenesis, treatments have only targeted the inflammatory component, and all have failed in clinical trials. In a new study, Ahmed et al. tested the effects of targeting protein dyshomeostasis using arimoclomol, a co-inducer of the heat shock response. In rat myoblast cell culture, arimoclomol reduced key pathological features of IBM. In mutant valosin-containing protein (VCP) mice, which develop an inclusion body myopathy, treatment with arimoclomol ameliorated disease pathology and improved muscle function. The authors then treated a small number of sIBM patients with arimoclomol and showed that it was safe and well tolerated.

Abstract

Sporadic inclusion body myositis (sIBM) is the commonest severe myopathy in patients more than 50 years of age. Previous therapeutic trials have targeted the inflammatory features of sIBM but all have failed. Because protein dyshomeostasis may also play a role in sIBM, we tested the effects of targeting this feature of the disease. Using rat myoblast cultures, we found that up-regulation of the heat shock response with arimoclomol reduced key pathological markers of sIBM in vitro. Furthermore, in mutant valosin-containing protein (VCP) mice, which develop an inclusion body myopathy, treatment with arimoclomol ameliorated disease pathology and improved muscle function. We therefore evaluated arimoclomol in an investigator-led, randomized, double-blind, placebo-controlled, proof-of-concept trial in sIBM patients and showed that arimoclomol was safe and well tolerated. Although arimoclomol improved some IBM-like pathology in the mutant VCP mouse, we did not see statistically significant evidence of efficacy in the proof-of-concept patient trial.

INTRODUCTION

Sporadic inclusion body myositis (sIBM) is the commonest idiopathic inflammatory myopathy (IIM) occurring in patients more than 50 years of age (17), but no treatment is available. The prevalence of sIBM differs between different populations and ranges between 1 and 71 individuals per million (813). sIBM is distinguished from other IIMs by early asymmetric finger flexor and knee extensor weakness, which leads to loss of hand function and propensity to fall, and resistance to immunosuppressive therapy. However, any skeletal muscle may be affected including esophageal and pharyngeal muscles. Late-stage disease is characterized by significant morbidity, including motor disability, swallowing failure, and poor quality of life. Death in IBM is related to malnutrition, cachexia, aspiration, and respiratory failure (17).

Although the etiology of sIBM remains uncertain, the varied pathological findings observed in patient muscles have driven a number of hypotheses, including viral infection, accumulation of toxic proteins, autoimmune attack, myonuclear degeneration, endoplasmic reticulum (ER) stress, and impairment of autophagy and proteasome function (1418). Muscle biopsies from sIBM patients typically show several pathological features broadly described as either inflammatory or degenerative. Inflammatory features include endomysial infiltration by mononuclear cells, which surround and invade nonnecrotic muscle fibers and overexpression of major histocompatibility complex class I (MHC-I), which is not constitutively expressed by skeletal muscle. The degenerative features of sIBM include the formation of rimmed vacuoles and inclusion bodies containing a range of proteins including β-amyloid precursor protein (β-APP), heat shock proteins (HSPs), phosphorylated tau (p-Tau), p62, and the cytoplasmic mislocalization of RNA-binding proteins including transactive response DNA binding protein 43 (TDP-43), heterogeneous nuclear ribonucleoprotein A1 (hnRNPA1), and hnRNPA2B1 (1922).

Despite lack of any experimental evidence, protein accumulation could theoretically play a role in triggering inflammation in IBM muscle. For example, amyloid oligomers can induce key components of sIBM pathology, at least in vitro, so that β-APP overexpression and consequent accumulation impairs innervation by cocultured neurons (23), causes mitochondrial dysfunction (24), and in vivo, results in calcium dyshomeostasis (25). Conversely, inflammation may induce the onset of degenerative features through exposure to inflammatory cytokines and increased nitric oxide production (26). In muscle biopsies of sIBM patients, expression of APP mRNA correlates with inflammation and expression of mRNAs of chemokines and interferon-γ (IFN-γ). Furthermore, unlike muscles of inflammatory myopathies, in sIBM muscles, inflammatory mediators colocalize with β-amyloid deposits within myofibers. In vitro, exposure of human myotubes to interleukin-1β (IL-1β) causes up-regulation of APP with subsequent aggregation of β-amyloid. These findings suggest that there is a link between the production of proinflammatory mediators and β-amyloid–associated degeneration in sIBM muscle (27).

Previous clinical trials in sIBM have only tested agents directed at the inflammatory component of pathology, and all were ineffective (3, 4, 2831). Whether the degenerative aspect of IBM is primary to the pathogenesis or not, it very likely plays a role in the deleterious effects in muscle and may be a potential therapeutic target. Protein homeostasis (proteostasis) is essential for normal cellular functions (32), and under conditions of cellular stress, proteins can become unfolded or misfolded, leading to their aberrant aggregation (33). Protein misfolding is normally controlled by endogenous chaperone proteins (34) that prevent aberrant protein-protein interactions and promote correct protein folding. HSPs are a family of ubiquitously expressed protein chaperones, which are up-regulated after stress-induced activation of the heat shock response, an endogenous cytoprotective mechanism. Because the heat shock response declines with advanced age (35), up-regulation of the heat shock response in disorders in which there is evidence of protein mishandling, such as sIBM, may be of therapeutic value.

Arimoclomol is a pharmacological agent that co-induces the heat shock response (36) by prolonging the activation of heat shock factor 1 (HSF-1) (32), the main transcription factor that controls HSP expression, thus augmenting HSP levels (3639). Arimoclomol only acts on cells under stress in which HSF-1 is already activated and does not induce the heat shock response in unstressed cells (40), thereby avoiding the side effects associated with widespread, nontargeted heat shock response activation. Indeed, previous examination of the effects of arimoclomol in mice has shown that treatment with arimoclomol has no detectable effect on nonstressed cells and there was no increase in HSP expression in any tissue studied (including spleen, heart, brain, spinal cord, nerve, and muscle).

Here, we took a three-step translational approach to test whether heat shock response augmentation may be a potential therapy for sIBM. We first established robust in vitro models representing the degenerative and inflammatory components of sIBM. Primary rodent myotube cultures were induced to overexpress human β-APP by transfection of the wild-type human gene, or cultures were exposed to the inflammatory cytokines IL-1β and tumor necrosis factor–α (TNFα). In these models, muscle cells developed several key IBM-like pathological features, which were used as outcome measures to examine the therapeutic potential of arimoclomol in IBM. Having established the beneficial effects of arimoclomol in vitro, we next tested its ability to ameliorate sIBM-like pathology in mice overexpressing mutant valosin-containing protein (VCP; p97 in mouse). These transgenic mice model the degenerative disorder multisystem proteinopathy (MSP), which has a phenotypic spectrum that includes inclusion body myopathy, Paget’s disease of the bone, and frontotemporal dementia (41). Although these mice are a model of the genetic rather than sporadic form of IBM, they do develop progressive muscle weakness and pathological hallmarks of inclusion body myopathy. In addition, we also undertook a randomized, double-blind, placebo-controlled, proof-of-concept trial of arimoclomol versus placebo for the treatment of sIBM in human patients. The aim of this trial was to evaluate the safety and tolerability of arimoclomol and to gather exploratory efficacy data.

RESULTS

Arimoclomol improves sIBM-like pathology in myoblast cultures

We first established accurate in vitro models of sIBM-like pathology in which primary rat myoblast cultures were either transfected with β-APP or exposed to inflammatory cytokines. For both cell culture models, we first undertook a series of titration experiments to determine the optimal conditions; the effects of different DNA and Lipofectamine ratios or different cytokine concentrations on β-APP and Aβ40 and Aβ42 expression were examined at different stages in vitro.

Transfection of primary rat myocytes with full-length human β-APP increased the expression of β-APP and its toxic cleavage product amyloid β-42 (Aβ42) and led to the formation of cytoplasmic inclusion bodies, a hallmark of sIBM, containing β-APP, ubiquitin, and TDP-43 among other proteins (Fig. 1, A to C, and fig. S1, A to E). Treatment with arimoclomol after transfection significantly reduced the formation of cytoplasmic inclusions (P < 0.01; n = 3; Fig. 1C). Additionally, β-APP overexpression increased the expression of MHC-I, a characteristic feature of sIBM muscle, and the proapoptotic protein caspase-3 (fig. S1, F and G). Exposure of cultured rat myocytes to IL-1β and TNFα was also associated with elevated expression of β-APP and Aβ42, although in the absence of inclusion body formation. Increased MHC-I expression was also reproduced in the cytokine-exposed cultures (fig. S1G). Similar results were obtained with IFN-γ.

Fig. 1. β-APP overexpression or exposure to inflammatory mediators induces sIBM-like pathology in cultured rat myocytes that is abrogated by arimoclomol.

(A and B) Cytoplasmic inclusion bodies (white arrows) in rat myocytes transfected with full-length human β-APP that are immunoreactive for β-APP and ubiquitin, and TDP-43 and ubiquitin. (C) The number of rat myocytes containing ubiquitinated inclusion bodies as a percentage of the total number of myocytes present [n = 3; *P < 0.05, one-way analysis of variance (ANOVA)]. Ari, arimoclomol. (D and E) Expression of TDP-43 (green) after empty vector (EV) or β-APP transfection and arimoclomol treatment and the number of rat myocytes with cytoplasmic mislocalization of TDP-43 (n = 3; *P < 0.001, unpaired t test). (F and G) TDP-43 expression (green) after exposure to inflammatory mediators and arimoclomol and quantification of TDP-43 mislocalization in cytokine-treated cultures (n = 3; *P < 0.05, unpaired t test). (H) Western blot analysis of TDP-43 expression in rat myocyte cultures exposed to cytokines in the presence and absence of arimoclomol. (I and J) Images show the expression of NF-κB subunit p65 (green) in β-APP–transfected cultures [4′,6-diamidino-2-phenylindole (DAPI)–labeled nuclei in blue] and cultures exposed to cytokines in the presence and absence of arimoclomol. (K) The number of rat myocytes with the nuclear subunit of NF-κB, p65 as a percentage of the total number of myocytes present (n = 3; *P < 0.05, one-way ANOVA). Error bars represent SEM. Scale bars, 10 μm (A and B); 20 μm (D, F, I, and J).

Overexpression of β-APP or exposure to inflammatory mediators resulted in cytoplasmic mislocalization of the C terminus of TDP-43 from the nucleus (Fig. 1, D to G), which is a key pathological feature of sIBM patient muscle. This mislocalization was almost completely prevented by treatment with arimoclomol (Fig. 1, D to G). Thus, although only 1.4 ± 0.6% of rat myocytes showed mislocalized TDP-43 in control cultures, overexpression of β-APP induced TDP-43 mislocalization in 52.2 ± 5.3% of myocytes, which was reduced to only 2.4 ± 0.53% after treatment with arimoclomol for 24 hours after transfection (P < 0.001; n = 3; Fig. 1, D and E).

Treatment with inflammatory mediators also induced TDP-43 mislocalization in rat myocytes, which was reduced by arimoclomol treatment, from 47.4 ± 2.8% to 24.8 ± 2.0% in myocyte cultures exposed to IL-1β (P < 0.05) and from 59.9 ± 9.0% to 29.8 ± 2.1% in TNFα-exposed cultures (P < 0.05; Fig. 1, F and G). In addition, exposure to inflammatory cytokines also increased overall TDP-43 expression, and this was also reduced by treatment with arimoclomol (Fig. 1H). The N terminus of TDP-43 remains in its nuclear location under all conditions studied (fig. S1H).

Although the extent of TDP-43 mislocalization induced by β-APP overexpression and exposure to inflammatory mediators was similar, arimoclomol was more effective in preventing TDP-43 mislocalization in the β-APP overexpression model. This difference most likely reflects the underlying cause of TDP-43 mislocalization in the two models, where stress induced by protein misfolding is the key pathological trigger in β-APP overexpressing cells. Arimoclomol directly reduces the levels of misfolded proteins and cell stress by up-regulating the heat shock response. The mode of action of arimoclomol in the inflammatory models is less direct and may involve a general reduction in cell stress and consequently nuclear factor κB (NF-κB) activation, as shown in Fig. 1 (I to K).

β-APP overexpression also activated the NF-κB cascade, most likely as part of a nonspecific stress response, as shown by nuclear translocation of the NF-κB subunit p65, observed in 43.1 ± 6.2% of myocytes in β-APP–transfected cultures compared to 8.8 ± 1.2% in control cultures (P < 0.01; Fig. 1, I and K). Treatment with arimoclomol reduced this to 23.6 ± 4.2% (P < 0.05). Unsurprisingly, exposure to inflammatory cytokines also activated the NF-κB cascade, from 4.6 ± 1.3% and 9.1 ± 1.4% in control cultures to 88.1 ± 6.3% and 38.9 ± 2.9% (P < 0.001) in IL-1β–treated and TNFα-treated cultures, respectively (Fig. 1, J and K). Arimoclomol reduced this effect to 41.2 ± 6.8% (P < 0.05) and 24.3 ± 3.1% (P < 0.05), respectively.

Arimoclomol augments HSP70 expression and improves cell survival

Overexpression of β-APP and exposure to inflammatory mediators resulted in a significant increase in HSP70 expression, which increased by 3.7- and 2.3-fold of control after β-APP overexpression and treatment with IL-1β, respectively (P < 0.05; n = 3). Arimoclomol further increased HSP70 expression by 2.4- and 2.1-fold in β-APP and IL-1β–treated cultures, respectively (Fig. 2, A to C). There was no difference in HSP70 levels in untreated controls and empty vector–treated cultures (fig. S1I). β-APP overexpression and exposure to inflammatory mediators also resulted in significant cell death, which was reduced by treatment with arimoclomol (P < 0.05; n = 3; Fig. 2, D and E).

Fig. 2. Arimoclomol augments HSP70 expression and improves survival of cultured rat myocytes.

(A) HSP70 expression (red) in cultured rat myocytes after empty vector (EV) transfection (left-hand panel), β-APP overexpression (middle panel), or β-APP overexpression plus treatment with arimoclomol (right-hand panel). (B) HSP70 expression (red) in cultured rat myocytes that were untreated (control; left-hand panel) or exposed to IL-1β alone (middle panel) or IL-1β plus arimoclomol (right-hand panel). (C) Western blot analysis of HSP70 expression in rat myocyte cultures exposed to inflammatory mediators or after β-APP overexpression in the presence or absence of arimoclomol. (D) Cytotoxicity in β-APP–overexpressing rat myocyte cultures as a percentage of that in control cultures (n = 3; *P < 0.02, one-way ANOVA), as assessed by a lactate dehydrogenase (LDH) assay. (E) Cytotoxicity after exposure to inflammatory mediators in the presence or absence of arimoclomol (n = 3; *P < 0.05, one-way ANOVA), as assessed by an MTT [3-(4, 5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide] assay. Error bars represent SEM. Scale bars, 10 μm.

Arimoclomol attenuates ER stress and protein mishandling

β-APP overexpression and exposure to inflammatory mediators increased ER stress in cultured rat myocytes, as determined by measurement of intracellular calcium ions [Ca2+], where reduced cytosolic [Ca2+] was indicative of ER stress. In β-APP–overexpressing cells exposed to the ER stressor thapsigargin, cytosolic [Ca2+] was significantly lower than in controls (P < 0.05; n = 3; Fig. 3A). β-APP overexpression also induced up-regulation of expression of the ER stress mediator CHOP (Fig. 3B). This β-APP–induced disruption in ER calcium ion handling was prevented by treatment with arimoclomol, where cytosolic [Ca2+] was significantly increased to the levels observed in control cultures (P < 0.05; n = 3; Fig. 3A), accompanied by a decrease in CHOP expression (Fig. 3B).

Fig. 3. Disruption of protein homeostasis in rat myocyte cultures is prevented by arimoclomol.

(A) The cytosolic calcium ion response induced by the ER stressor thapsigargin (an indicator of ER stress) in β-APP–overexpressing rat myocyte cultures and in cell cultures exposed to inflammatory mediators (n = 3; *P < 0.05, one-way ANOVA). (B) Expression of the ER stress mediator CHOP determined by Western blot analysis in rat myocyte cultures overexpressing β-APP or exposed to inflammatory mediators, in the presence or absence of arimoclomol. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. (C) The images show the presence of a cytoplasmic aggregate immunoreactive for p62 (red; white arrow) in a desmin-positive rat myocyte after β-APP transfection. (D) The images show the expression of the autophagic protein LC3-II (red) in β-APP–transfected desmin-positive rat myocytes after empty vector transfection (left-hand panel), β-APP overexpression (middle panel), or β-APP overexpression plus treatment with arimoclomol (right-hand panel). The white arrow indicates punctate LC3-II staining of autophagosomes. (E) The chymotrypsin-like proteasome activity assessed 48 hours after β-APP transfection in the presence or absence of arimoclomol (n = 3; *P < 0.05, one-way ANOVA). (F) The expression of the autophagosome marker LC3-II in β-APP–transfected rat myocytes with or without arimoclomol (n = 3; *P < 0.05, one-way ANOVA). EV, empty vector. Error bars represent SEM. Scale bars, 10 μm.

Exposure to the inflammatory cytokines IL-1β and TNFα also resulted in a significant reduction in cytosolic [Ca2+] (Fig. 3A) and an increase in the expression of CHOP (Fig. 3B). Treatment with arimoclomol restored cytosolic [Ca2+] to control levels and reduced the expression of CHOP (P < 0.05; n = 3; Fig. 3, A and B).

Aberrant expression of the shuttle protein p62 (sequestosome 1) was observed in β-APP–overexpressing cells (Fig. 3C). This protein targets misfolded proteins for degradation and was observed in cytoplasmic aggregates in cultured rat myocytes, suggesting altered protein handling. Assessment of proteasome function demonstrated a significant decrease in proteasome activity after β-APP overexpression (P < 0.05; n = 3); however, this was not improved by treatment with arimoclomol (Fig. 3E). Autophagic degradation was assessed by examination of the autophagosome marker LC3-II, which increased by 1.28-fold in β-APP–overexpressing cells compared to controls. However, this was reduced to 0.74-fold the expression level of controls in arimoclomol-treated cultures, indicating a reduction in autophagic degradation, possibly due to a reduction in misfolded proteins (P < 0.05; n = 3; Fig. 3, D and F).

Arimoclomol ameliorates IBM-like pathology in mutant VCP mice

Arimoclomol has clear beneficial effects in cellular models of sIBM. We therefore next tested the efficacy of arimoclomol in vivo in mutant VCP mice. These mice model MSP and develop characteristic hallmarks of inclusion body myopathy (IBM). Mutant VCP mice were treated with arimoclomol from 4 to 14 months of age. By 14 months, there was a significant decrease in muscle force in mutant VCP mice (P < 0.0001; n = 10), as assessed by longitudinal analysis of grip strength as well as acute in vivo physiological assessment of isometric muscle force (Fig. 4, A to C). The loss of muscle force observed at 14 months was prevented by treatment with arimoclomol. Grip strength was significantly greater in treated mutant VCP mice than untreated mice (P = 0.0175; n = 10; Fig. 4, A to C). There was also an improvement in the toe-spreading reflex observed in arimoclomol-treated mutant VCP mice (fig. S2, A and B). However, arimoclomol had no effect on the abnormal increase in body weight observed in mutant VCP mice (fig. S2C). Histopathological analysis of hindlimb muscles of mutant VCP mice confirmed the presence of significant IBM-like pathology including myofiber atrophy, increased endomysial connective tissue, and the presence of degenerating fibers and vacuoles as well as hypertrophic fibers (Fig. 4D). All of these pathological findings were reduced in arimoclomol-treated mutant VCP mice (Fig. 4, D to I, and movie S1). In addition, muscles of mutant VCP mice showed evidence of inflammatory cell infiltration as well as up-regulation of MHC-I molecules and the phosphorylated form of the NF-κB substrate, inhibitor of nuclear factor κB α (IκBα) (Fig. 4, E to G). In contrast, there was no evidence of inflammatory infiltrates in muscles of arimoclomol-treated mutant VCP mice, and there was a reduction in expression of MHC-I and IκBα (Fig. 4, E to G). In particular, treatment with arimoclomol markedly improved mitochondrial morphology and reduced sarcoplasmic reticulum swelling and vacuole number in mutant VCP mouse muscle fibers compared to controls (Fig. 4H and movies S2 to S4). Furthermore, the macrophage infiltration observed in mutant VCP mouse muscle was also reduced in arimoclomol-treated mice (fig. S2, D to F). Arimoclomol-treated mutant VCP mouse muscles also showed a decrease in muscle fiber diameter compared to untreated mutant mouse muscle (P < 0.0001; n = 3; Fig. 4I). These beneficial effects of arimoclomol were accompanied by an increase in HSP70 expression in mutant VCP mouse muscles (Fig. 4J), a decrease in the expression of ubiquitin (Fig. 5, A to C), and a reduction in cytoplasmic mislocalization of TDP-43 (P = 0.01; n = 3; Fig. 5, D to F).

Fig. 4. Arimoclomol treatment improves muscle strength, muscle contractile characteristics, and IBM-like pathology in mutant VCP mice.

(A) The change in grip strength in wild-type VCP (WT-VCP), mutant VCP (mVCP), and arimoclomol-treated mutant VCP (mVCP + Ari) mice at 4 and 14 months (M) of age (n = 10; *P < 0.0001, unpaired t test). (B) Typical traces of muscle twitch and maximum tetanic force of extensor digitorum longus (EDL) muscles in untreated and arimoclomol-treated mutant VCP mice. (C) Mean maximum force of extensor digitorum longus muscles of WT-VCP, mutant VCP, and arimoclomol-treated mutant VCP mice (n = 10; *P < 0.04, one-way ANOVA). (D) Hematoxylin and eosin (H&E) staining of tibialis anterior (TA) muscles of mice in each experimental group (white arrow, atrophied fiber; black arrowheads, hypertrophic fibers). (E) H&E staining showing clear inflammatory cell infiltration in mutant VCP tibialis anterior muscle. (F and G) Western blots show MHC-I and phospho-IκBα expression in tibialis anterior muscles of mice in each experimental group. (H) Transmission electron microscopy of tibialis anterior muscles of mice in each experimental group. (I) Muscle fiber diameter in tibialis anterior muscles from untreated and arimoclomol-treated mutant VCP mice compared to WT-VCP mice (n = 3; *P < 0.0001, one-way ANOVA). (J) Western blot analysis of HSP70 expression in tibialis anterior muscles from mice in each experimental group. Bar chart shows mean relative optical density (n = 3; *P = 0.01, unpaired t test). Error bars represent SEM. Scale bars, 50 μm (D and E); 2 μm (H).

Fig. 5. Arimoclomol abrogates IBM-like pathology in mutant VCP mouse muscle.

(A to F) Cross sections of tibialis anterior muscle from WT-VCP, mutant VCP, and arimoclomol-treated mutant VCP mice immunostained for ubiquitin (red) and TDP-43 (green); nuclei stained with DAPI (blue). Scale bars, 50 μm; 25 μm (insets).

Clinical trial of safety and tolerability of arimoclomol for the treatment of sIBM

We next undertook a randomized, double-blind, placebo-controlled, proof-of-concept trial of arimoclomol versus placebo for the treatment of sIBM. The primary aim of the trial was to evaluate the safety and tolerability of arimoclomol, but exploratory efficacy data were also gathered. Sixteen sIBM patients were randomized to receive arimoclomol and eight to receive placebo. The duration of the treatment period was 4 months, and the follow-up continued for a further 8 months after treatment ceased, with an overall trial duration of 12 months. At 4 months (end of the treatment phase), all sIBM patients were still participating in the trial. At 8 months, two patients had discontinued the study (because of travel difficulties), but one of these patients returned for the final assessment at 12 months (fig. S3). Baseline clinical and demographic characteristics were similar between groups (table S1).

There were no significant differences between treatment groups regarding the rate, type, and severity of adverse events (table S2). There were 8 adverse events in the placebo group and 14 in the arimoclomol group, the most common being gastrointestinal (see table S2 for details). In the arimoclomol group, one serious adverse event was reported as a result of persistent high blood pressure requiring overnight hospitalization in a patient with known poorly controlled hypertension, in whom the first trial muscle biopsy was identified as a stressful event. Blood pressure normalized after adjustment of the patient’s antihypertensive medication and remained within the normal range throughout the remainder of the trial. Hypertensive episodes were also observed in two placebo patients, under similar circumstances, although these cases did not require hospitalization. Two cases of hyponatremia and one case of high thyroxine levels were observed in the arimoclomol group; however, these were transient, asymptomatic, and did not require treatment. The episode of hematuria in the arimoclomol group was also limited and did not require treatment. All infections resolved with standard treatments, with or without antibiotics, and did not require hospitalization. Ocular toxicity and arrhythmia were not observed in any study subjects.

Clinical trial secondary outcomes

Overall, there was no statistically significant difference in the secondary outcome measures favoring arimoclomol. However, there were trends in favor of arimoclomol, but these will require a formal large-scale patient trial powered for efficacy to be assessed further. Physical function and muscle strength decline rates over time were numerically higher in the placebo group compared to the arimoclomol group (Fig. 6, A to C, and table S3). At 8 months, there was a trend favoring the arimoclomol group, with P values of 0.055 for change in the IBM functional rating scale (IBMFRS) score (−0.68 ± 1.58 versus −2.50 ± 3.31; mean ± SD), 0.147 for change in the average manual muscle testing (MMT) score (−0.12 ± 0.22 versus −0.26 ± 0.27), and 0.064 for change in the right-grip maximum voluntary isometric contraction testing (MVICT) score (1.26 ± 2.63 versus −0.54 ± 1.86). No differences were seen for changes in the other MVICT scores, changes in dual-energy x-ray absorptiometry (DEXA) fat-free mass percentage, or changes in myosin-adjusted HSP70 expression in muscle (table S3).

Fig. 6. Clinical trial secondary outcomes (efficacy measures).

The change from baseline to end point on three different clinical scales assessed at 4, 8, and 12 months (means ± SEM) in sIBM patients treated with arimoclomol for 4 months. (A to C) IBMFRS score, MMT average score, and right-hand grip (HGR) MVICT score. The IBMFRS is a disease-specific functional questionnaire for patients with sIBM and measures physical function/disability. MMT is a measure of muscle strength scored by the physician on the basis of the clinical assessment. MVICT is a measure of muscle strength performed using a quantitative muscle assessment (QMA) system that uses an adjustable cuff to attach the patient’s arm or leg to an inelastic strap that is connected to a force transducer. Error bars represent SEM. No statistically significant clinical efficacy measures were observed.

DISCUSSION

Here, we examined the effects of targeting the heat shock response in sIBM. In sIBM patients, arimoclomol was found to be safe and well tolerated, with a trend of a slower decline in muscle strength and physical function compared with placebo-controlled sIBM patients, although the trend was not significant. Although we did not observe statistically significant clinical efficacy or significant morphological changes in the repeat muscle biopsies taken from arimoclomol-treated patients, studies both in cellular models in vitro and in an in vivo mouse model, which recapitulates many features of sIBM in muscle, showed that the pathological and functional deficits associated with sIBM were ameliorated by arimoclomol in these model systems.

In vitro, both β-APP overexpression and exposure to inflammatory cytokines induced degenerative sIBM-like pathology in rat myocytes, including an increase in the formation of ubiquitinated inclusion bodies. Treatment with arimoclomol reduced inclusion body formation, indicating an improvement in protein handling. This was most likely due to up-regulation of the heat shock response, in particular enhanced HSP70 expression. Both models also recapitulated the increase in mislocalized TDP-43 observed in sIBM patient myofibers. TDP-43 is cleaved by caspase-3 (42), allowing the C terminus to leave the nucleus, thus linking its translocation to increased cell stress. Arimoclomol also reduced TDP-43 mislocalization and TDP-43 expression in cultures exposed to inflammatory mediators, suggesting that the drug reduced levels of cell stress. Indeed, both β-APP overexpression and inflammatory mediators induced cell death, which was reduced by arimoclomol.

Nuclear translocation of the NF-κB subunit p65, which was observed after exposure to IL-1β and TNFα, was examined as an indication of NF-κB activation. β-APP overexpression also activated NF-κB as has been observed in cellular models of Alzheimer’s disease in which Aβ42 peptide activates the inflammatory cascade (43), most likely as part of a nonspecific stress response. Arimoclomol had an inhibitory effect on NF-κB activation in both the β-APP and the inflammatory cell models, which likely reflects up-regulation of the heat shock response and down-regulation of the NF-κB signaling cascade by HSPs (44).

Arimoclomol also decreased the disruption in ER calcium ion homeostasis induced in both the β-APP and inflammatory cellular models of sIBM and consequently reduced the ER stress response, a key mechanism triggered by aberrantly folded proteins. This finding was reflected by restoration of cytosolic calcium ion concentrations. Indeed, in TNFα-treated rat myotube cultures, the reduction in cytosolic calcium ion concentration was prevented by arimoclomol, as was the expression of the ER stress mediator CHOP. This protective effect of arimoclomol may reflect, at least in part, the chaperone activity due to augmented expression of HSP70, which serves to improve the efficiency of handling and degradation of misfolded proteins.

Examination of the two major protein degradation pathways in the cell culture models revealed disruption in both proteasome function and increased autophagy. However, arimoclomol had no effect on proteasome function, although formation of mature autophagosomes was reduced, suggesting a reduced misfolded protein load in the lysosomal pathway.

In vivo, treatment of mutant VCP mice with arimoclomol improved the pathological features and functional deficits characteristic of the inclusion body myopathy associated with MSP. Thus, in arimoclomol-treated mutant VCP mice, we observed an increase in muscle strength and a reduction in key pathological hallmarks of IBM, including decreased ubiquitin expression, a reduction in cytoplasmic mislocalization of TDP-43, and a marked improvement in mitochondrial morphology accompanied by an increase in HSP expression.

These findings provide evidence for the beneficial effects of arimoclomol in experimental models of sIBM and support clinical assessment of the effects of arimoclomol in sIBM patients. The primary end point of our randomized, double-blind, placebo-controlled, proof-of-concept trial in sIBM patients was met, with results showing that arimoclomol was both safe and well tolerated by sIBM patients. Regarding the secondary end points (efficacy measures), there were no statistically significant differences between the treatment groups. However, this finding was not surprising because this first experimental study of a compound targeting the heat shock response in sIBM had several limitations. These included the small sample size, which had been advised by our ethics committees and the U.S. Food and Drug Administration (FDA) because arimoclomol had never been given to sIBM patients before and this study was intended as a proof-of-concept/safety study that was not powered for efficacy. The short duration of treatment (only 4 months) was also mandated by regulatory agencies. In a slowly progressing disease like sIBM, longer treatment periods are likely to be required to be able to detect changes in efficacy outcome measures. Whether arimoclomol can ameliorate sIBM pathology will only be determined by performing adequately powered clinical trials of longer duration. A 12-month study powered for efficacy that will enroll 150 IBM patients has now been approved and fully funded and enrollment will commence in late 2016.

Finally, regarding the measurement of muscle HSP70 expression in sIBM patients, this readout also had several limitations, namely, the fact that muscle HSP70 expression is extremely sensitive to multiple factors, including disease stage, physical activity, age, and gender (4547), and the fact that muscles on the opposite sides of the body are biopsied at baseline and after treatment.

Arimoclomol has previously been shown to be safe in both animal models and humans (36, 39, 48) and has been indicated to be of potential therapeutic benefit in several neurological disorders including diabetic peripheral neuropathy and retinopathy (49). Arimoclomol has also been found to be beneficial in animal models of neurodegeneration, including models of acute injury–induced neuronal death (37) and has been shown to have therapeutic value in mouse models of motor neuron disease, including amyotrophic lateral sclerosis (ALS) (36, 50) and Kennedy’s disease (51). Arimoclomol has been through seven phase 1 clinical trials in healthy volunteers to assess its safety, tolerability, and pharmacokinetic properties, as well as a small-scale phase 2 trial in ALS patients (48). A phase 2a dose-ranging trial in ALS has shown arimoclomol to be safe and well tolerated up to 100 mg three times daily (52), and a phase 2/3 randomized, double-blind, placebo-controlled trial is currently under way in familial superoxide dismutase 1 (SOD1) ALS patients (NCT00706147).

Our results show that arimoclomol is safe and well tolerated in sIBM patients and ameliorates key degenerative and inflammatory features of IBM pathology in experimental cellular and animal models. Together, these findings support further investigation of arimoclomol for the treatment of sIBM.

MATERIALS AND METHODS

Study design

Here, we took a three-step translational approach to test whether augmentation of the heat shock response may be a potential therapy for sIBM, by examining the effects of arimoclomol in rat muscle cells in vitro, in a preclinical study in mutant VCP mice, and in sIBM patients in a safety and tolerability trial. In vitro, the effects of arimoclomol were examined in primary rat muscle cultures overexpressing β-APP or treated with inflammatory cytokines by assessing the effects on IBM-relevant histopathological characteristics. All experiments were repeated in at least three independent cultures. In the preclinical efficacy study, mutant VCP mice were randomized to arimoclomol or vehicle treatment arms, and the effects on muscle force and histopathological characteristics were set as the study end points. The experimenter was blind to genotype and treatment group throughout. In sIBM patients, we undertook a randomized, double-blind, placebo-controlled, proof-of-concept trial of arimoclomol versus placebo to evaluate the safety and tolerability of arimoclomol and to gather exploratory efficacy data by testing muscle strength and HSP70 expression. This trial was an exploratory study conducted without prior knowledge of effect size of arimoclomol in sIBM. The sample size was chosen on the basis of feasibility. Randomization was performed centrally for both study sites.

In vitro model of IBM

Primary muscle cultures were used as an in vitro model of IBM by extracting satellite cells that lie under the basal lamina of myofibers. Muscle cells were induced to model either the degenerative features of sIBM by overexpression of β-APP or the inflammatory characteristics by treatment with inflammatory mediators (see Supplementary Methods for details of muscle culture preparation and treatment).

In these in vitro models of sIBM, the following features of sIBM pathology were examined and the effects of treatment with 10 μM arimoclomol investigated: (i) inclusion body formation, (ii) HSP expression, (iii) cytoplasmic translocation of TDP-43, (iv) cell survival, (v) NF-κB activation, and (vi) ER stress (for details, see the Supplementary Materials).

Breeding and maintenance of mutant VCP mice

All experimental procedures were carried out under license from the UK Home Office (Scientific Procedures Act 1986) and after approval by the University College London (UCL) Institute of Neurology’s Animal Welfare and Ethical Review Board. Mice overexpressing the wild-type or mutant (A232E) human VCP gene under the cytomegalovirus-enhanced chicken β-actin promoter [see Custer et al. (41)] were bred and maintained at the UCL Institute of Neurology. Transgenic female mice carrying the wild-type or mutant gene were mated with wild-type C57BL/6J male mice to generate transgenic and nontransgenic littermates. Only male offspring were used in this study to prevent gender differences. The mice were genotyped by polymerase chain reaction (PCR) amplification of ear notches and analyzed using agarose gel electrophoresis and visualized using GelRed stain (Sigma-Aldrich). All mice used in this study were housed in a controlled temperature and humidity environment with a 12-hour light/dark cycle and had access to drinking water and food ad libitum.

Treatment of mutant VCP mice with arimoclomol

After genotyping, male mice were randomly assigned to a treatment or vehicle arm of the study. From 4 months of age to the time of examination at 14 months, mutant VCP mice were treated with arimoclomol (120 mg/kg per day) dissolved in drinking water (n = 10) or water alone (vehicle; n = 10). The body weight of all mice was recorded fortnightly. Arimoclomol was obtained from Orphazyme ApS. All experiments were undertaken blinded to genotype and treatment.

Longitudinal assessment of grip strength and body weight

Grip strength was assessed fortnightly in all mice from 4 to 14 months of age using a Bioseb force gauge according to the manufacturer’s instructions. An average of four maximum readings was obtained. Body mass was recorded at the same time as grip strength. The ratio of grip strength to body mass was determined for individual animals and was pooled by each genotype.

In vivo analysis of isometric muscle force

The mice were prepared for in vivo assessment of muscle function [see Kieran et al. (36)]. Briefly, mice were deeply anesthetized [inhalation of 1.5 to 2.0% isoflurane in oxygen delivered through a Fortec vaporizer (Vet Tech Solutions Ltd.)]. The distal tendons of the tibialis anterior and extensor digitorum longus muscles in both hindlimbs were dissected free and attached to isometric force transducers (Dynamometer UFI Devices). The sciatic nerve was exposed and sectioned, and all branches were cut except for the deep peroneal nerve that innervates the tibialis anterior and extensor digitorum longus muscles. Muscle length was adjusted for maximum twitch tension. Isometric contractions were elicited by stimulating the nerve to the muscles using square-wave pulses of 0.02-ms duration at supramaximal intensity using silver wire electrodes. Contractions were elicited by trains of stimuli at frequencies of 40, 80, and 100 Hz for 450 ms, and the maximum twitch and tetanic tension were measured using force transducers connected to a PicoScope 3423 oscilloscope (Pico Technology) and subsequently analyzed using PicoScope software v5.16.2 (Pico Technology). Muscle histology, immunohistochemistry, Western blot, and electron microscopy protocols are described in the Supplementary Materials.

Clinical trial design and patient population

In this investigator-initiated, double-blind, placebo-controlled study, 24 patients meeting the Griggs diagnostic criteria for definite or probable sIBM (53, 54) were randomized to arimoclomol (100 mg) or placebo (2:1), three times a day for 4 months (treatment phase). This restricted 4-month treatment period, mandated by the FDA, was followed by an 8-month blinded assessment phase. The study was conducted from August 2008 to May 2012, at two centers in two countries (12 patients per center): University of Kansas Medical Center (KUMC) and Medical Research Council (MRC) Centre for Neuromuscular Diseases. Detailed inclusion and exclusion criteria can be found in Supplementary Methods.

Randomization was performed centrally for both study sites at KUMC. This was done by a General Clinical Research Center (GCRC) statistician, who sent the randomization codes, created using a random number generator table, to the respective research pharmacies. All personnel involved in the conduct of the trial were blind to the identity of the treatment assignments, except for the unblinded statistician and the pharmacist at each site who labeled the study medication using codes provided by the unblinded statistician. The appearance of the placebo was identical to that of arimoclomol.

The study was conducted according to the ethical principles of the Declaration of Helsinki and approved by the Independent Ethics Committee or Institutional Review Board for each center. Informed consent was obtained from each patient before randomization. The study was registered with ClinicalTrials.gov (NCT00769860) and with International Standard Randomized Controlled Trial Number Register (ISRCTN80057573).

Clinical trial: Primary outcomes

The safety and tolerability of arimoclomol compared to placebo were the primary outcome of the trial. Participants were seen for assessment of adverse events at every study visit. Unscheduled visits to evaluate potential adverse events could occur at any time. All serious adverse events were reported to the sponsor and regulatory authorities according to standard operating procedures. The trial was monitored by an Independent Safety Monitoring Committee.

During the first 4 months (treatment phase), all participants completed a study medication diary. Pill bottles were brought to each visit for a count by a research team member to check on whether participants were taking the study medication in the appropriate dosages. At screening and months 1, 2, 3, 4, and 12, the study participants had full safety laboratory analyses done, which included full blood count with differential, prothrombin time, activated partial thromboplastin time, urea, creatinine, electrolytes, glucose, phosphate and calcium, alanine transaminase, aspartate transaminase, total bilirubin, albumin, full urine analysis, and 24-hour urine protein content and creatinine clearance. At months 0.5, 1.5, 2.5, and 3.5, participants had partial safety laboratory analyses done including serum creatinine, urea, electrolytes, glucose, phosphate, calcium, and a full urine analysis. An electrocardiogram was performed at screening and months 1 and 3. An ophthalmic examination was performed at screening and month 10. The electrocardiogram and ophthalmic examination were introduced as safety measures following results from an animal model study that raised concerns of potentially accelerated cataract formation and apparent risk of sudden, unexplained death with arimoclomol at very high doses and with arimoclomol used in conjunction with riluzole (CytRx Corporation; arimoclomol Investigator’s Brochure version 7.0, September 2010).

Clinical trial: Secondary outcomes

Physical function was measured using the IBMFRS that is intended only for patients with sIBM. It includes 10 measures (swallowing, handwriting, cutting food and handling utensils, fine motor tasks, dressing, hygiene, turning in bed and adjusting covers, changing position from sitting to standing, walking, and climbing stairs) graded on a Likert scale from 0 (being unable to perform) to 4 (normal). The sum of the 10 items gives a value between 0 and 40, with a higher score representing less functional limitation. The IBMFRS is a sensitive and reliable tool for assessing activities of daily living in patients with sIBM and is quickly administered (see Supplementary Methods) (7, 30, 5557).

Muscle strength was assessed by MMT and MVICT using the QMA system designed by Computer Source (30, 57) (see Supplementary Methods). Each muscle was tested twice, and the maximum force generated by the patient from the two trials was recorded for each muscle group. The total summed score of strength in kilograms was computed for each patient. MVICT has been shown to be reliable and valid in several neuromuscular disorders, including sIBM (30, 57, 58). Body composition was obtained using a standard DEXA whole-body scan to assess total body fat-free mass. DEXA has been used to measure lean body mass in previous neuromuscular disease studies, including sIBM (30, 57). Patient muscle HSP70 expression was determined as described in Supplementary Methods.

Statistical methods

In the preclinical cellular and mutant VCP mouse models, analysis of normally distributed results was performed using an unpaired t test or, for comparison of greater than two groups, a one-way ANOVA. Otherwise, the nonparametric Mann-Whitney U test was used.

Data management and statistical analysis were performed by GCRC informatics staff and a GCRC statistician. Data were entered blindly to a GCRC password-protected electronic database, and analyses were performed after database lock. Descriptive statistics were used to summarize subject disposition and adverse events by treatment group. To reduce measurement error, baseline scores were computed by calculating the average of visit 1 (screening visit) and visit 2 (baseline visit), which had to be less than 21 days apart. Continuous variables were compared between treatment groups using the Mann-Whitney U test. Categorical variables were compared between treatment groups using χ2 or Fisher’s exact test, as appropriate. Treatment groups were compared at baseline as well as regarding changes in several outcome measures at 4 months (IBMFRS, MMT, MVICT, DEXA, and HSP70 content), 8 months (IBMFRS, MMT, and MVICT), and 12 months (IBMFRS, MMT, MVICT, and DEXA). Statistical analyses were performed using STATA v10 and in all analyses, statistical significance was set at the two-sided 5% level. Error bars represent SEM unless otherwise stated.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/8/331/331ra41/DC1

Materials and Methods

Fig. S1. Overexpression of β-APP and exposure to inflammatory mediators induce sIBM-like pathology in cultured myocytes.

Fig. S2. Mutant VCP mice show further signs of pathology.

Fig. S3. Consort diagram of the patients participating in the clinical trial.

Table S1. Baseline characteristics of the study population.

Table S2. Summary of all adverse events over the course of 1 year.

Table S3. Mean changes (±SD) in secondary outcome measures throughout the study period.

Movie S1. Low-magnification serial block-face scanning electron microscopy of arimoclomol-treated and untreated mutant VCP mouse muscle.

Movie S2. High-magnification serial block-face scanning electron microscopy of wild-type VCP mouse muscle.

Movie S3. High-magnification serial block-face scanning electron microscopy of untreated mutant VCP mouse muscle.

Movie S4. High-magnification serial block-face scanning electron microscopy of arimoclomol-treated mutant VCP mouse muscle.

References (59, 60)

REFERENCES AND NOTES

  1. Acknowledgments: We thank the participants who enrolled in this intensive study and the KIT (Keep In Touch) and Myositis UK support groups. The following clinicians, research nurses, coordinators, and technicians contributed substantially to the trial: M. Walsh, M. Michaels, F. Raja, A. Dick, K. Latinis, L. Dewar, G. Barreto, and I. Skorupinska. We thank Orphazyme ApS and CytRx for providing drug and placebo. Funding: This work was funded by Arthritis Research UK (ref. no. 19255), MRC Centre for Neuromuscular Diseases grant (G0601943), Kansas University Neurology Department Ziegler grant, Kansas University GCRC CReFF (General Clinical Research Centre Clinical Research Feasibility Funding) grant, and the Wellcome Trust (107116/Z/15/Z) (G. Schiavo). L.G. is the Graham Watts Senior Research Fellow, supported by the Brain Research Trust, Rosetrees Trust, The Stoneygate Trust, and funded by the European Community’s Seventh Framework Programme (FP7/2007–2013). C.S. is the recipient of an MRC Studentship. L.C. and A.W. are funded by Cancer Research UK, the MRC, BBSRC (UK Biotechnology and Biological Sciences Research Council), and EPSRC (Engineering and Physical Sciences Research Council) under grant award MR/K01580X/1 to L.C. and P. O’Toole (York University). P.M.M. reports funding from the National Institute for Health Research (NIHR) Rare Diseases Translational Research Collaboration (RD TRC) and from the NIHR University College London Hospitals (UCLH) Biomedical Research Centre (BRC). The views expressed are those of the author and not necessarily those of the UK National Health Service (NHS), the NIHR, or the Department of Health. This project was also supported by an Institutional Clinical and Translational Science Award and NIH/NCATS (National Center for Advancing Translational Sciences) grant (UL1TR000001). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH. Author contributions: M.A., A.M., and L.G. designed the in vitro experiments. M.A. and L.G. designed the in vivo experiments, and M.A. and C.S. performed the experiments and the data analyses. J.P.T. generated and characterized the VCP mice. P.M.M., A.M., L.H., Y.W., A.L.M., M.P., P.G., J.S., S.B., M.P., J.H., M.G.H., R.J.B., M.M.D., and L.G. designed the clinical trial. P.M.M. performed the clinical trial assessments in the UK. L.H., Y.W., A.L.M., M.P., R.J.B., and M.M.D. performed the clinical trial assessments in the United States. J.H. and J.N. conducted data analyses of the trial data. M.M.D., A.L.M., H.S., and G. Samandouras performed the trial muscle biopsies. J.L.H. performed muscle histopathology of patient biopsies from the trial. C.-H.L. and B.K. performed the HSP measurements in the trial muscle tissue. A.W. and L.C. carried out the electron microscopy analyses, and G. Schiavo undertook their interpretation. M.G.H., R.J.B., and M.M.D. supervised the clinical trial. L.G. supervised the in vitro cell models and in vivo preclinical trial experiments and HSP measurements in the muscle tissue. M.A., P.M.M., C.S., and L.G. wrote the drafts of the manuscript and all authors provided scientific input to the manuscript. All authors read and approved the final version of the manuscript. Competing interests: L.G. became an unpaid consultant to Orphazyme ApS (the owner of arimoclomol) after completing this study. The other authors declare that they have no competing interests.
View Abstract

Navigate This Article