Research ArticleMuscular Dystrophy

Cystamine Suppresses Polyalanine Toxicity in a Mouse Model of Oculopharyngeal Muscular Dystrophy

See allHide authors and affiliations

Science Translational Medicine  02 Jun 2010:
Vol. 2, Issue 34, pp. 34ra40
DOI: 10.1126/scitranslmed.3000723

Abstract

Oculopharyngeal muscular dystrophy (OPMD) is caused by a trinucleotide repeat expansion mutation in the coding region of the gene encoding PABPN1 (polyadenylate-binding protein nuclear 1). Mutant PABPN1 with a polyalanine tract expansion forms aggregates within the nuclei of skeletal muscle fibers. There is currently no effective treatment. We have developed cell and mouse models of OPMD and have identified the aggregation of mutant PABPN1 and apoptosis as therapeutic targets. Here, we show that transglutaminase activity is increased in muscle from OPMD model mice. Elevated transglutaminase 2 expression enhances, whereas TG2 knockdown suppresses, the toxicity and aggregation of mutant PABPN1 in cells. Cystamine protects against the toxicity of mutant PABPN1 and exerts its effect via the inhibition of transglutaminase 2, as cystamine treatment is unable to further reduce the protective effect of transglutaminase 2 knockdown on mutant PABPN1 toxicity in cells. Cystamine also reduces the aggregation and toxicity of mutant PABPN1 in human cells. In a mouse model of OPMD, cystamine treatment reduced the elevated transglutaminase activity, attenuated muscle weakness, reduced aggregate load, and decreased apoptotic markers in muscle. Therefore, inhibitors of transglutaminase 2 should be considered as possible therapeutics for OPMD.

Introduction

Oculopharyngeal muscular dystrophy (OPMD) is a late-onset muscular dystrophy that typically occurs in the fifth or sixth decade; the first signs are drooping eyelids and difficulty in swallowing (1). OPMD is a progressive myopathy, and the disease can eventually affect all voluntary muscles, leading to proximal muscle weakness. OPMD is generally inherited as an autosomal dominant trait, although recessive forms of the disease do rarely occur and these have a more severe disease course. The mutation that causes OPMD was identified in 1998 and is an abnormal expansion of a (GCG)n trinucleotide repeat in the coding region of the gene encoding PABPN1 (polyadenylate-binding protein nuclear 1) (2). In unaffected individuals, (GCG)6 codes for the first six alanines in a homopolymeric stretch of 10 alanines. In most patients, this (GCG)6 repeat is expanded to (GCG)8–13, leading to a stretch of 12 to 17 alanines in mutant PABPN1. PABPN1 with an expanded polyalanine tract forms aggregates consisting of tubular filaments within the nuclei of skeletal muscle fibers (24). There is currently no effective treatment for OPMD.

Overexpression of mutant PABPN1 in COS-7 cells results in the formation of PABPN1 aggregates and a concomitant increase in cell death, as shown by increased abnormal, apoptotic nuclei and elevated apoptotic markers (57). Mice expressing a mutant PABPN1 transgene (with a 17-alanine repeat; A17 mice) develop progressive muscle weakness, aggregates in skeletal myocyte nuclei, an increase in nuclei labeled by TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling), and elevated concentrations of the proapoptotic protein BAX (7). Increased numbers of apoptotic, TUNEL-labeled nuclei were also observed in another mouse model of OPMD in which a PABPN1 transgene is expressed highly in all tissues under control of the CAG (cytomegalovirus enhancer and chicken β-actin) promoter (8). We recently showed that wild-type PABPN1 overexpression can reduce mutant PABPN1 toxicity in both cell and mouse models of OPMD, likely through an antiapoptotic mechanism (9), reinforcing the importance of apoptosis in the disease.

Reduction of mutant PABPN1 aggregation in cells by molecular and chemical chaperones (including chaperones that do not affect susceptibility to proapoptotic insults) correlates with reduced cell death (5, 6). Similarly, inhibition of mutant PABPN1 oligomerization or aggregation by deletions in its oligomerization domains decreased toxicity in cells (10). We subsequently showed that in our transgenic mouse model of OPMD, the known antiaggregation compounds doxycycline and trehalose delayed the onset of muscle weakness, accompanied by a reduction in aggregate load in skeletal muscle (7, 11). Although the role of aggregates in protein conformation disorders such as OPMD is controversial, a widely accepted view is that aggregates, the aggregation process, or early oligomeric species are toxic and pivotal to OPMD pathology. Aggregates themselves may not be toxic (12). However, a reduction in aggregate load may actually reflect a decrease in the abundance of a toxic folding intermediate or toxic oligomer of PABPN1. Thus, compounds that both decrease the aggregation of mutant PABPN1 and are antiapoptotic are candidate therapeutics for OPMD.

The codon reiteration and protein aggregation of OPMD is also a feature of the nine other diseases caused by polyglutamine tract expansion, so strategies that show promise in polyglutamine diseases such as Huntington’s disease may be of value in OPMD. Cystamine ameliorates the pathogenic phenotype in mouse (13, 14) and Drosophila models of Huntington’s disease (15) and decreases aggregate load and cell death in cell models (15, 16). Cystamine may protect against polyglutamine-induced toxicity in several ways, including the direct inhibition of caspase-3 (17), acting as an antioxidant by increasing l-cysteine (18) and glutathione (19) and by inhibiting the transamidating activity of transglutaminase 2 (TG2 or tissue transglutaminase) (13, 14). TG2 is a thiol- and Ca2+-dependent acyl transferase that catalyzes the formation of covalent bonds between γ-carboxamide groups of glutamine residues and various primary amines including the ε-amino group of lysine (20). Its action can result in the posttranslational cross-linking of proteins via ε-(γ-glutamyl)lysine linkages, resulting in products that are often insoluble and resistant to proteolysis and degradation. Cystamine is thought to inhibit TG2 activity by forming a mixed disulfide with a critical thiol group of its active site (17), and the reduced form of cystamine (cysteamine) is a competitive inhibitor of TG2 (21). TG2 cross-links mutant huntingtin in vitro (22, 23), and ε-(γ-glutamyl)lysine bonds are present in aggregates in Huntington’s disease postmortem brains and mouse models of Huntington’s disease (13, 24). Similarly, it has been proposed that transglutaminase-catalyzed cross-linking contributes to the formation of insoluble neurofibrillary tangles from soluble tau monomers (2527) and that TG2 can cross-link β-amyloid (28, 29). TG2 also modulates apoptosis (3033). We therefore tested whether TG2 inhibition and cystamine treatment could be beneficial in OPMD.

Results

Transglutaminase activity is elevated in OPMD mice

We assayed transglutaminase activity in protein lysates from biceps muscle and found that it was markedly elevated in OPMD transgenic (A17) mice relative to nontransgenic littermates (Fig. 1A). Western blots of the same lysate showed a single band of TG2 in nontransgenic mice and two bands in A17 mice (Fig. 1B), the lower one being the same molecular mass as the single band, 62 kD. The predominant isoform of mouse TG2 has an apparent molecular mass of 62 kD, with very little or no full-length TG2 (82 kD) detected (34).

Fig. 1

Transglutaminase activity is elevated in muscle from A17 mice. (A) Relative transglutaminase activity in biceps muscle lysates from 8-month-old nontransgenic (NT) and A17 mice. Activity in nontransgenic mice is set to 1. (B) Western blot of biceps muscle lysates from 8-month-old nontransgenic or A17 mice, probed with an antibody to TG2; tubulin was used as a loading control. (C) Western blot of biceps muscle lysates from 6-month-old nontransgenic or A17 mice and the same lysates pretreated with CIP. Nonadjacent lanes from the same gel are separated by a black line. (D) Transglutaminase activity in gastrocnemius muscle lysates from 6-month-old nontransgenic and A17 mice. The same lysates were pretreated with CIP. (E) Immunohistochemistry of biceps muscle sections from 6-month-old A17 mice shows colabeling of PABPN1 aggregates (green) with an antibody against ε-(γ-glutamyl) isopeptide (red). Myocyte nuclei were visualized with DAPI (blue). (F) High–molecular weight forms of mutant PABPN1 were immunoprecipitated from biceps muscle lysates from A17 but not nontransgenic mice with Sepharose beads conjugated to ε-(γ-glutamyl) isopeptide antibody (X). As a control, lysates were also incubated with Sepharose beads conjugated to protein G (G). *P < 0.05; ***P < 0.0001. NS, nonsignificant. Error bars represent SEM.

We tested whether the upper TG2-immunoreactive protein band (Fig. 1B) was a phosphorylated form of TG2. Treatment of the lysates from A17 mice with calf intestinal phosphatase (CIP) abolished the upper band (Fig. 1C), suggesting that it is a phosphorylated form of TG2. We incubated muscle lysates from A17 mice with CIP before performing a transglutaminase activity assay and found that transglutaminase activity was reduced in A17 mice and that this effect was not seen in nontransgenic mice (Fig. 1D). Although CIP treatment causes a global dephosphorylation of proteins in muscle lysates, this experiment suggests that the increased phosphorylation of TG2 is associated with enhanced activity. Although we cannot formally exclude that the phosphorylation of another protein may regulate transglutaminase activity, this is very unlikely as CIP had no effect on this activity in nontransgenic mice.

We were interested to test whether the elevated transglutaminase activity seen in our mouse model of OPMD could contribute to pathogenesis. TG2 is known to cross-link aggregate-prone proteins such as tau (2527), polyglutamine-containing proteins (2224), and β-amyloid (28, 29), making them insoluble and more resistant to degradation and proteolysis. To test whether mutant PABPN1 was cross-linked by transamidation linkages, we performed immunohistochemistry with an antibody raised to ε-(γ-glutamyl) isopeptide (Fig. 1E). This antibody specifically recognizes ε-(γ-glutamyl) bonds (25) and labeled PABPN1-containing aggregates. In addition, high–molecular weight forms of PABPN1 were immunoprecipitated with a Sepharose-linked ε-(γ-glutamyl) isopeptide antibody from skeletal muscle lysates from A17 mice but not from nontransgenic littermates (Fig. 1F). The smear at the top of the blot (Fig. 1F) suggests that mutant PABPN1 is covalently cross-linked by ε-(γ-glutamyl) linkages, contributing to the formation of high–molecular weight oligomeric species.

Elevation of TG2 activity is toxic and inhibition of TG2 activity is protective in cells

To assess the effect of elevated TG2 activity on mutant PABPN1 toxicity, we used a cell model of OPMD. Differentiated C2C12 myotubes transfected with enhanced green fluorescent protein (EGFP)–tagged mutant PABPN1 (A17) had intranuclear aggregates and increased level of cell death (as assessed by number of abnormal or apoptotic nuclei) relative to cells transfected with EGFP alone (EGFPC1) (Fig. 2A). Cotransfection of an expression plasmid encoding full-length TG2 had no effect on cell death in differentiated C2C12 cells transiently expressing EGFPC1 (Fig. 2A). In contrast, TG2 coexpression increased toxicity in differentiated C2C12 cells expressing A17 (Fig. 2A) and this was associated with an increased number of cells with A17 aggregates. This increase in cell death was not seen when A17 was cotransfected with mutant TG2 without transamidating activity (C277S) (Fig. 2A) (32, 35). Therefore, toxicity is dependent on the transamidating activity of TG2.

Fig. 2

Inhibition of TG2 activity reduces A17 toxicity. (A) Differentiated C2C12 cells were cotransfected with A17 or EGFP alone (C1), and either empty vector, a plasmid encoding full-length TG2, or TG2 with no transglutaminase activity (C277S). The number of cells with A17 aggregates and abnormal nuclei was recorded. (B) Western blot of TG2 concentrations in HEK293 cells transfected with control or TG2 siRNA. Actin is a loading control. TG2 is present as three bands representing the three different splice isoforms in HEK293 cells. (C) HEK293 cells were transfected with either control or TG2 siRNA and A17. The number of cells containing aggregates was recorded. Cells were also treated with cystamine to test whether cystamine has effects on aggregation in addition to those due to inhibition of TG2 activity. (D) Cells from (C) were scored for cell death (abnormal nuclei). (E) Aggregation and cell death (abnormal nuclei) in COS-7 cells transfected with A17 and treated with cystamine (A17 + cystamine) or left untreated (A17). (F) Differentiated C2C12 cells were transfected with EGFP alone (C1), A10, or A17. A17-transfected cells were left untreated or treated with cystamine, trehalose, or doxycycline. (G) Cultured primary myoblasts isolated from unaffected individuals were transiently transfected with EGFP alone (C1), A10, or A17 and induced to differentiate into myotubes. A17-transfected cells were left untreated or treated with cystamine. Myotubes were scored for aggregation and death. (H) Aggregation and cell death (abnormal nuclei) in the human rhabdomyosarcoma cell line RD transiently transfected with EGFPC1 (C1) or A17 and either left untreated or treated with cystamine. (I) Human rhabdomyosarcoma cell line TE671 was transfected, treated, and scored as in (H). *P < 0.05; **P < 0.001; ***P < 0.0001. Error bars represent SEM.

As increased expression of TG2 and elevated transamidating activity enhanced the toxicity of A17 (Fig. 2A), we tested the effect of TG2 inhibition on A17 toxicity. We used small interfering RNA (siRNA) to reduce protein concentrations of TG2 in human embryonic kidney 293 (HEK293) cells (Fig. 2B), as they are easy to transiently transfect with siRNA and expression plasmids. A17-transfected HEK293 cells have high levels of aggregate formation and cell death, which is easy to score. Knockdown of TG2 reduced the formation of A17 aggregates and cell death; HEK293 cells cotransfected with A17 and siRNA directed against TG2 had fewer aggregates and a decreased apoptosis relative to cells cotransfected with A17 and control siRNA (Fig. 2, C and D).

Cystamine protects against toxicity caused by expression of the polyglutamine expansion mutation through a variety of mechanisms, including the inhibition of TG2 (13, 14, 16, 18, 36, 37). We therefore wanted to see whether cystamine could protect against mutant PABPN1 toxicity in cell models of OPMD. Indeed, treatment with cystamine decreased both aggregation and cell death in COS-7 cells transiently transfected with A17 (Fig. 2E). In addition, cystamine reduced A17 aggregation and rescued toxicity caused by expression of mutant PABPN1 in differentiated C2C12 cells to similar extents as trehalose and doxycycline, two compounds that have beneficial effects in our OPMD mouse model (Fig. 2F) (7, 11). It is likely that cystamine is exerting its protective effect on A17 toxicity via the inhibition of TG2, as cystamine treatment was unable to further rescue the toxicity of mutant PABPN1 in HEK293 cells transiently transfected with A17 and siRNA against TG2 (Fig. 2, C and D). If cystamine were not acting predominantly via the inhibition of TG2, one would expect a greater effect of cystamine treatment and siRNA knockdown combined relative to TG2 knockdown alone. This is not seen.

Cystamine treatment also decreased the aggregation and toxicity of A17 in human muscle cell models of OPMD (Fig. 2, G to I). To generate a human cell model of OPMD, we transiently transfected PABPN1 expression constructs into cultured myoblasts isolated from unaffected individuals and induced these to differentiate and form myotubes. Differentiated human primary skeletal myotubes show a high level of cell death and the formation of aggregates when transiently transfected with A17 (Fig. 2G and fig. S1), and cell death and the formation of aggregates were reduced by treatment with cystamine (Fig. 2G). Cystamine also reduced the formation of aggregates and the number of apoptotic nuclei in A17-transfected human rhabdomyosarcoma cell lines RD (Fig. 2H) and TE671 (Fig. 2I and fig. S1).

Our data suggested that the inhibition of TG2 activity may reduce the toxicity of mutant PABPN1 by decreasing aggregation (Fig. 2). However, TG2 has also been shown to modulate apoptosis (3033), and cystamine may have an antiapoptotic effect (independent of the inhibition of TG2) by inhibiting the activation of caspase-3 (17, 38). This may apply to OPMD, as it is likely that apoptosis contributes to myofiber loss (7). It is possible that whether TG2 has proapoptotic or protective effects depends on the experimental paradigm being studied and cellular context (3032). We tested whether decreasing TG2 expression and inhibiting TG2 activity was antiapoptotic in our cell models. Cystamine pretreatment protected COS-7 cells against apoptosis induced chemically by staurosporine treatment and genetically by the overexpression of EGFP-tagged BAX (Fig. 3, A and B). We also tested whether the inhibition of TG2 activity by siRNA knockdown was antiapoptotic. We found decreased cell death in HEK293 cells transfected with siRNA against TG2 compared to cells transfected with control siRNA after staurosporine treatment, as shown by scoring abnormal apoptotic nuclei (Fig. 3C) and active caspase-3 (Fig. 3, D and E). TG2 knockdown also protected against apoptosis due to the transient overexpression of BAX in HEK293 cells (Fig. 3F). The combination of cystamine and TG2 siRNA did not reduce staurosporine-induced toxicity more than TG2 siRNA alone (Fig. 3C). Therefore, in cell models of OPMD, TG2 inhibition may be acting to reduce mutant PABPN1 toxicity both by reducing the aggregation of mutant PABPN1 and by indirect antiapoptotic effects.

Fig. 3

TG2 knockdown and cystamine protect against proapoptotic insults. (A) COS-7 cells were treated with cystamine or left untreated (control) before treatment with staurosporine (ST). Cells were scored for death (abnormal nuclei). (B) Cystamine-treated or control COS-7 cells transfected with EGFP-tagged BAX. EGFP-positive cells were scored for death (abnormal nuclei). (C) Cell death in HEK293 cells transfected with control siRNA before treatment with staurosporine (control siRNA + ST), transfected with siRNA directed against TG2 and treated with staurosporine (TG2 siRNA + ST), or transfected with siRNA directed against TG2 and pretreated with cystamine before incubation with staurosporine (TG2 siRNA + ST + cystamine). (D) Western blot of lysates from cells treated as in (C) showing levels of active caspase-3 and full-length (fl) caspase-3. Actin was used as a loading control. (E) Densitometric quantification of three separate experiments performed as in (D). (F) HEK293 cells were transfected with either control or TG2 siRNA and incubated before transfection with EGFP-tagged BAX. Sixteen hours later, cells were fixed and EGFP-positive cells were scored for abnormal apoptotic nuclei. **P < 0.001; ***P < 0.0001. Error bars represent SEM.

Cystamine attenuates muscle weakness in a mouse model of OPMD

As cystamine is protective in cell models of OPMD, we next tested its efficacy in our mouse model of OPMD. Mice expressing A17 develop progressive muscle weakness starting at ~4 months of age, accompanied by the formation of aggregates and increased TUNEL-labeled nuclei in skeletal muscle (7). This phenotype is not seen in mice expressing a wild-type PABPN1 transgene (10 alanines) at similar levels (7). We treated A17 mice with cystamine starting at 6 weeks of age and continued until the end of the study. Cystamine delayed the onset of muscle weakness in A17 mice and significantly improved forelimb grip strength as well as grip strength of all paws in A17 mice relative to nontransgenic littermates from 4 months of age until the end of our study (at an age of 10 months) (Fig. 4A). Cystamine-treated OPMD transgenic mice also performed significantly better in the wire maneuver task than placebo-treated mice from 6 to 9 months of age (Fig. 4B).

Fig. 4

Cystamine attenuates muscle weakness in a mouse model of OPMD. (A to D) Male A17 mice were given drinking water containing cystamine (900 mg/liter) (5 weeks, n = 25; 4, 5, 6, and 7 months, n = 22; 8 months, n = 21; 9 and 10 months, n = 19) or normal drinking water (5 weeks and 4 and 5 months, n = 25; 6 and 7 months, n = 24; 8, 9, and 10 months, n = 23) from 6 weeks of age. (A) Forelimb grip strength and grip strength from all limbs were improved in A17 mice treated with cystamine compared to untreated mice at all time points tested (individual t tests of each time point all, P < 0.0001; repeated-measures ANOVAs for determination of the overall effect from all treatment time points, P < 0.0001). No significant difference was seen between the different groups before treatment at 5 weeks of age. Error bars represent SD. (B) Cystamine treatment improved the overall performance of A17 mice at the wire maneuver task between 5 and 10 months of age (P < 0.0001). Score: 0, active grip with hind legs; 1, difficulty grasping with hind legs; 2, unable to lift hind legs; 3, falls within 30 s; 4, falls immediately. (C) Pelvic elevation was improved in cystamine-treated A17 mice compared to control A17 mice between the ages of 7 and 10 months (overall effect P = 0.0006). Score: 0, normal elevation; 1, barely touches; 2, markedly flattened. (D) Cystamine-treated A17 males performed significantly better at the vertical gripping test at 8, 9, and 10 months of age than untreated control A17 males. Score: 0, grips the grid; 1, falls off the grid. *P < 0.05; **P < 0.001; ***P < 0.0001.

Fewer cystamine-treated A17 mice dragged their pelvis when walking (a measure of ability to lift body weight) than untreated mice from 8 to 10 months of age (Fig. 4C), and in the vertical gripping test, cystamine-treated A17 males performed better than untreated males from 8 to 10 months of age (Fig. 4D). Vertical gripping and pelvic elevation phenotypes only become evident in A17 mice from 7 months of age. There was no difference in performance at any of these tasks between the two groups before the start of the treatment, at 5 weeks of age (Fig. 4, A to D).

There was no difference in the performance of untreated and cystamine-treated nontransgenic mice in the vertical gripping task and pelvic elevation (fig. S2). Cystamine treatment did not affect the grip strength of nontransgenic mice, except to decrease the grip strength of all paws at 7 and 8 months of age (fig. S2). Consistent with other studies (36), we observed an effect of cystamine on the body weight of nontransgenic mice; cystamine-treated mice weighed less than untreated nontransgenic mice (fig. S2), and this was significant between the ages of 5 and 8 months. Because heavier mice generate greater force on grip strength, this probably accounts for the decreased grip strength caused by cystamine in the nontransgenic mice. This effect on weight was probably also responsible for the observed differences between cystamine-treated and cystamine-untreated nontransgenic mice at the wire maneuver task, which was significant from 6 to 8 months of age (fig. S2), because lighter mice may be able to pull themselves up on the wire more easily. There was no difference between the weight of untreated and cystamine-treated A17 mice (fig. S2), so weight was unlikely to have influenced the results of grip strength and wire maneuver in A17 mice.

Cystamine may be acting through a variety of mechanisms to reduce the toxicity of mutant PABPN1, and many of these may be via the inhibition of TG2. Indeed, transglutaminase activity was reduced in muscle lysates from 6-month-old A17 mice treated with cystamine compared to untreated A17 mice (Fig. 5A). However, this reduction was not associated with a reduction in the phosphorylated form of TG2, as shown by Western blot (fig. S2), suggesting that cystamine and cysteamine are directly inhibiting TG2 (17, 21). In addition to attenuating muscle weakness in A17 mice, cystamine treatment also improved pathology. The number of TUNEL-positive nuclei was reduced in muscle sections from cystamine-treated compared to untreated A17 mice at 6 months (Fig. 5B). Cystamine also reduced the number of myocyte nuclei containing PABPN1 aggregates (Fig. 5C), and of these aggregates, fewer were labeled with an antibody raised against ε-(γ-glutamyl) cross-links (Fig. 5D).

Fig. 5

Cystamine reduces transglutaminase activity and attenuates pathology in a mouse model of OPMD. (A) Transglutaminase activity in biceps muscle lysates from 6-month-old A17 mice treated with cystamine (A17 cystamine) and A17 mice given normal drinking water (A17 control). (B) Biceps muscle sections from 6-month-old control or cystamine-treated A17 mice were TUNEL-labeled, and the number of positive nuclei was scored. (C) Quantification of aggregate-containing nuclei in sections of biceps muscles from 6-month-old control or cystamine-treated A17 mice. (D) Biceps sections were treated as in (C) but colabeled with antibodies against PABPN1 and ε-(γ-glutamyl) cross-links. The number of PABPN1 aggregates that were immunopositive for ε-(γ-glutamyl) cross-links was scored. *P < 0.05; **P < 0.001. Error bars represent SEM.

Apoptotic markers are elevated in muscle from OPMD mice and reduced by cystamine

We have previously speculated that apoptosis may contribute to the pathogenesis of OPMD (7). We therefore tested whether other apoptotic markers could be detected in muscle sections from A17 mice. Indeed, we detected muscle fibers immunopositive for active caspase-3 (Fig. 6A) and with a diffuse pattern of cytochrome c labeling (as opposed to mitochondrial localization, corresponding to cytochrome c release from the mitochondria; Fig. 6B) in muscle sections from A17 mice. Furthermore, an increased number of active caspase-3–immunoreactive myofibers were seen in A17 mice relative to nontransgenic littermates at 6 and 12 months of age (Fig. 6C). Likewise, the number of myofibers with a cytosolic distribution of cytochrome c was elevated in A17 mice (Fig. 6D). Cystamine treatment decreased both the number of active caspase-3–immunopositive myofibers and myofibers with a cytosolic distribution of cytochrome c at both 6 and 12 months (Fig. 6, C and D). However, cystamine treatment did not affect concentrations of BAX, which were unaltered in muscle lysates from 6-month-old untreated and cystamine-treated A17 mice (Fig. 6E).

Fig. 6

Apoptotic markers are elevated in muscle from a mouse model of OPMD and reduced by cystamine treatment. (A) Representative images of active caspase-3–immunoreactive myofibers (arrow) in biceps muscle sections from 12-month-old A17 mice. For comparison, unlabeled, nonactive caspase-3–immunoreactive fibers are marked with an asterisk. (B) Biceps muscle sections from 12-month-old nontransgenic and A17 mice were immunolabeled with an antibody against cytochrome c (red). Cytochrome c is mainly localized to mitochondria at the myofiber periphery but can also have a diffuse pattern across entire myofibers corresponding to its release from the mitochondria and indicating apoptotic myofibers (arrow). (C) Biceps muscle sections from 6- and 12-month-old nontransgenic, A17 control untreated, and A17 cystamine-treated mice were labeled with an active caspase-3 antibody [as in (A)], and the number of immunopositive myofibers was scored. (D) Quantification of the number of myofibers with a diffuse pattern of cytochrome c labeling (corresponding to cytochrome c release from the mitochondria to cytosol) in sections treated as in (B). (E) Western blot of biceps muscle lysates from A17 mice treated with cystamine (Cyst) or untreated A17 mice (C), probed with an antibody against BAX. Tubulin was used as a loading control. *P < 0.05; **P < 0.001; ***P < 0.0001. Error bars represent SEM.

Discussion

We have shown that the inhibition of TG2 reduces both apoptosis and aggregate formation caused by mutant PABPN1 expression. Our data show that transglutaminase activity is elevated in muscles from a mouse model of OPMD. An increase in transglutaminase activity is likely to contribute to OPMD pathogenesis, as the coexpression of full-length TG2 with A17 in C2C12 cells enhanced aggregate formation and cell death compared to cells expressing A17 alone. This increase in both aggregation and cell death is likely to be dependent on the transamidating activity of TG2, as the cotransfection of mutated full-length TG2 with no transamidating activity (C277S) did not increase A17 toxicity or aggregates.

Our data raise the possibility that TG2-catalyzed cross-linking may contribute to the formation of mutant PABPN1 aggregates. TG2 cross-links mutant huntingtin and is involved in the formation of aggregates (22, 23). Mutant huntingtin is an obvious substrate for TG2, as it contains a large number of glutamine residues. However, TG2 can also cross-link aggregate-prone proteins with a lower glutamine content, such as tau and β-amyloid. The human tau isoforms tau23 and tau40 contain 8 glutamine residues and 10 lysine residues capable of participating in isopeptide bond formation (39), and the amyloid βA4 (1–40) peptide contains 1 glutamine and 2 lysine residues (40). Similarly, PABPN1 contains 4 glutamine and 10 lysine residues (fig. S3). We cannot exclude the possibility that mutant PABPN1 aggregation may be mediated by another tightly associated, transglutaminase cross-linked protein. The mechanism by which TG2 activity contributes to the aggregation of PABPN1 remains an important issue to explore in future studies.

The knockdown of TG2 was protective against A17 toxicity in cells. We also observed that cystamine could rescue toxicity of mutant PABPN1 in cell models of OPMD, including human muscle cell models of OPMD. Cystamine is most likely exerting its effect via the inhibition of TG2, as cystamine treatment is unable to further reduce the protective effect of TG2 knockdown on mutant PABPN1 toxicity in cells. If cystamine were not acting predominantly via the inhibition of TG2, one would expect a greater effect of cystamine treatment and siRNA knockdown combined relative to TG2 knockdown alone, and this is not seen. In addition to providing protection by decreasing the aggregation of mutant PABPN1, cystamine and the inhibition of TG2 may be directly inhibiting apoptosis. Both cystamine treatment and TG2 knockdown can protect cells from apoptosis induced chemically by staurosporine or genetically by the overexpression of BAX.

Cystamine treatment rescued muscle weakness in a mouse model of OPMD, and this was associated with a decrease in transglutaminase activity, aggregate load, and TUNEL-labeled nuclei. We have previously proposed that apoptosis may contribute to OPMD pathology (7, 9), a hypothesis consistent with the data presented here. We show elevated apoptotic markers (active caspase-3 and cytosolic distribution of cytochrome c) in muscle from A17 mice. Furthermore, cystamine treatment reduced the number of muscle fibers with apoptotic markers in A17 mice. Thus, cystamine and TG2 knockdown may be protecting from mutant PABPN1 toxicity in part via an antiapoptotic mechanism. The role of apoptosis in muscle fiber death and muscular dystrophy is unclear and is likely to vary among muscular dystrophies. There is activation of apoptosis in some muscular dystrophies, and antiapoptotic strategies, such as the genetic overexpression of Bcl2 or inactivation of BAX, are protective in some mouse models of muscular dystrophy (41, 42).

In summary, we found that TG2 activity was elevated in a mouse model of OPMD. This is likely to be deleterious, as TG2 overexpression enhanced the toxicity and aggregation of mutant PABPN1 in cell culture models. Genetic knockdown and chemical inhibition of TG2 with cystamine were protective in cells, and the beneficial effects of cystamine in an OPMD mouse model suggest that this drug may be a suitable therapeutic for this disease.

Materials and Methods

Cell studies

C2C12 (mouse myoblast), COS-7 (African green monkey kidney), HEK293 RD, and TE671 (human rhabdomyosarcoma) cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Sigma-Aldrich) supplemented with 10% fetal bovine serum, penicillin-streptomycin (100 U/ml), 2 mM l-glutamine, and 1 mM sodium pyruvate at 37°C in 5% carbon dioxide. Cells were transiently transfected with pEGFPC1-PABPN1-A17 (5), pEGFPC1-PABPN1-A10 (5), EGFPC1 (Clontech), wild-type full-length TG2, and TG2 with no transamidating activity (TG2 and C277S) (32, 35) using Lipofectamine reagent (Invitrogen). Cotransfections were carried out with plasmids at a molar ratio of 1:1. C2C12 cells were transfected 24 hours after the addition of differentiation medium [DMEM supplemented with 2% horse serum, penicillin-streptomycin (100 U/ml), 2 mM l-glutamine, and 1 mM sodium pyruvate] and harvested 48 hours after transfection.

Primary human myoblasts were obtained from the Muscle Tissue Culture Collection of the EuroBioBank. Myoblasts were cultured in proliferation medium, comprising DMEM supplemented with 20% fetal bovine serum, penicillin-streptomycin (100 U/ml), 2 mM l-glutamine, insulin (10 μg/ml), human basic fibroblast growth factor (25 ng/ml), and epidermal growth factor (10 ng/ml), at 37°C in 5% carbon dioxide. Transient transfections of plasmid DNA were performed using Lipofectamine 2000 (Invitrogen), and myoblasts were induced to differentiate 24 hours after transfection by incubating in differentiation medium [DMEM supplemented with 5% horse serum, penicillin-streptomycin (100 U/ml), and insulin (10 μg/ml)].

Small interfering RNA against human TG2 (siGENOME SMARTpool siRNA; Dharmacon) or control siRNA (siCONTROL Non-Targeting pool; Dharmacon) was transfected into cells at 100 nM using Lipofectamine 2000 following the manufacturer’s instructions.

To induce apoptosis, we added staurosporine (3 μM; Sigma-Aldrich) to the cell media for the final 6 hours after transfection. In an alternative model of cell death, GFP-BAX (43) was transfected into cells 16 hours before collection.

Transgenic mice

OPMD transgenic mice were described in (7). All studies and procedures were carried out following UK Home Office regulations, and animals were caged under standard conditions (12-hour light, 12-hour dark; food and water available ad libitum). Male A17 mice and nontransgenic littermates were treated orally from 6 weeks of age with cystamine (900 mg/liter) in drinking water. Mice in the control groups received normal drinking water, and the solutions were changed twice a week. Mice were given alphanumeric identities that provided no clue to genotype.

Supplementary Material

www.sciencetranslationalmedicine.org/cgi/content/full/2/34/34ra40/DC1

Materials and Methods

Fig. S1. Human cell models of OPMD.

Fig. S2. Effects of cystamine treatment on nontransgenic mice.

Fig. S3. Amino acid sequence of bovine, human, and mouse PABPN1.

Footnotes

  • * These authors contributed equally to this work.

  • Citation: J. E. Davies, C. Rose, S. Sarkar, D. C. Rubinsztein, Cystamine suppresses polyalanine toxicity in a mouse model of oculopharyngealmuscular dystrophy. Sci. Transl. Med. 2, 34ra40 (2010).

References and Notes

  1. Acknowledgments: We thank O. Sadiq for technical assistance, G. Johnson Voll for TG2 overexpression constructs, E. Wahle for PABPN1 bacterial expression construct, R. O’Connor and A. Marx for rhabdomyosarcoma cell lines, B. Underwood and B. Ravikumar for helpful comments, and the Muscle Tissue Culture Collection for providing control human myoblasts. Funding: This work was funded by the Wellcome Trust (Senior Fellowship to D.C.R.) and the Muscular Dystrophy Campaign, UK. The Muscle Tissue Culture Collection is part of the German network on muscular dystrophies (MD-NET, service structure S1, 01GM0601) funded by the German Ministry of Education and Research (Bonn, Germany). The Muscle Tissue Culture Collection is a partner of EuroBioBank (http://www.eurobiobank.org) and TREAT-NMD (http://www.treat-nmd.eu). Author contributions: D.C.R. and J.E.D. designed the experiments; J.E.D., C.R., and S.S. performed the experiments and analyzed the data with D.C.R. J.E.D., C.R., and D.C.R. wrote the manuscript. D.C.R. secured funding. Competing interests: The authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article