|Home | About | Journals | Submit | Contact Us | Français|
Inclusion body myositis (IBM), the most common muscle disease to afflict the elderly, causes slow but progressive degeneration of skeletal muscle and ultimately paralysis. Hallmark pathological features include T-cell mediated inflammatory infiltrates and aberrant accumulations of proteins, including amyloid-beta (Aβ), tau, ubiquitinated-proteins, apolipoprotein E, and α-synuclein in skeletal muscle. A large body of work indicates that aberrant Aβ accumulation contributes to the myodegeneration. Here we investigated whether active immunization to promote clearance of Aβ from affected skeletal muscle fibers mitigates the IBM-like myopathological features as well as motor impairment in a transgenic mouse model. We report that active immunization markedly reduces intracellular Aβ deposits and attenuates the motor impairment compared to untreated mice. Results from our current study indicate that Aβ oligomers contribute to the myopathy process as they were significantly reduced in the affected skeletal muscle from immunized mice. In addition, the anti-Aβ antibodies produced in the immunized mice blocked the toxicity of the Aβ oligomers in vitro, providing a possible key mechanism for the functional recovery. These findings provide support for the hypothesis that Aβ is one of the key pathogenic components in IBM pathology and subsequent skeletal muscle degeneration.
Prominent amyloid-beta (Aβ) accumulation and the presence of inclusion bodies are pathological hallmarks of several neurological disorders, including Alzheimer disease (AD) and inclusion body myositis (IBM). IBM, the most common skeletal muscle disorder to afflict the elderly, is clinically characterized by proximal and distal skeletal muscle degeneration and accompanied by marked inflammatory infiltration (Dalakas, 2006b). The etiology for the vast majority of IBM cases remains unknown, and no effective treatment has yet been established. Approximately 5–10% cases are hereditary, and can be caused by mutations in several genes, including the myosin heavy chain II, UDP-N-acetylglucosamine-2-epimerase/N-acetylmannosamine kinase (GNE) or in valosin-containing protein (VCP) (Martinsson et al., 2000; Eisenberg et al., 2001; Kayashima et al., 2002; Watts et al., 2004; Askanas and Engel, 2005; Dalakas, 2006a).
IBM-affected skeletal muscle exhibits several distinct myopathological features, which include the accumulation of protein aggregates such as Aβ, hyperphosphorylated tau, and poly-ubiquitinated proteins (Askanas et al., 1991; Askanas et al., 1993; Askanas et al., 1994; Fidzianska and Glinka, 2006). These protein aggregates are more commonly associated with AD, where Aβ is considered to play a central pathogenic role in the neurodegenerative process. In IBM, however, the role of Aβ is still unresolved, but a large body of studies indicates that Aβ may contribute to the muscle degeneration (McFerrin et al., 1998; Querfurth et al., 2001; Wojcik et al., 2006). Moreover, it is well documented that expression of amyloid precursor protein (APP) is significantly elevated, and most notably that Aβ42 preferentially accumulates in affected muscle fibers of IBM patients (Sarkozi et al., 1993; Vattemi et al., 2003; Askanas and Engel, 2008).
Because aberrant accumulation of Aβ appears to play a role in IBM, we generated several mouse models that overexpress human APP or that selectively overproduce Aβ1–42. These models develop key pathological and behavioral features of IBM in an age-dependent fashion, including selective accumulation of Aβ in muscle fibers and myopathological features, including increased CD8+ T-cell infiltrates and hyperphosphorylated tau as well as impaired motor function. Our prior work provides in vivo evidence that Aβ plays a significant role in the development of IBM-like myopathology, in agreement with other reports (Fukuchi et al., 1998; Jin et al., 1998; Moussa et al., 2006).
In the present study, we sought to determine whether the sustained presence of Aβ is necessary to maintain impaired motor function in this mouse model. In other words, once Aβ-initiated motor impairment occurs, is Aβ still required or will its clearance rescue motor function and/or slow the rate of decline. Accordingly, we actively immunized mice against Aβ and found that it markedly attenuated the motor impairment. Moreover, the reduction in insoluble Aβ in skeletal muscle and plasma Aβ levels correlated with the restoration of motor performance. Notably, active immunization reduced oligomeric and fibrillar forms of Aβ in skeletal muscle. To elucidate which Aβ assembly state adversely affects myotubes, we performed in vitro studies and found that that Aβ oligomers were more toxic to differentiated murine myotubes than fibrillar Aβ. Anti-Aβ antibodies isolated from immunized mice significantly blocked the cytotoxicity of Aβ oligomers, highlighting the potent pathological effects these oligomers exert on skeletal muscle, which is comparable to effects observed with neurons.
12-month old MCK-APP/PS1 mice (20 males and 19 females) were used in this study (Kitazawa et al., 2006). Mice were divided into three groups: untreated (5 males and 5 females), sham-treated (7 males and 6 females), and actively immunized (8 males and 8 females).
Aβ1–33–MAP peptide was synthesized on a 4-brached poly-lysine core (Invitrogen, Carlsbad, CA) and used as an antigenic peptide. For the active immunization, 100µg of Aβ1–33–MAP peptide mixed with complete Freund’s adjuvant (CFA; Sigma-Aldrich, St. Louis, MO) was used for the initial injection with subsequent injections using incomplete Freund’s adjuvant (IFA) in total volume of 100 µl adjusted with PBS. The sham-treated group received injections of adjuvant only. The vaccine was delivered intraperitoneally (i.p.) with a two week interval before the first boost, and monthly thereafter. Mice received a total of 5 injections. Blood was collected before the first immunization (pre-bleed) and 10 days after each boost from the retro-orbital sinus into the EDTA-coated tubes. Tubes were centrifuged for 10 min at 4°C, and the plasma were collected as a supernatant and stored at −80°C for the further analysis.
Presence and type of Aβ specific T-cells was analyzed by ELISPOT assay (Cribbs et al., 2003). Production of pro-inflammatory lymphokine IFN-γ (Th1) or anti-inflammatory IL-4 (Th2) was evaluated by restimulation of splenocytes from experimental mice. Briefly, 96-well ELISPOT plates (BD PharMingen, San Diego, CA) were coated with capture IFN-γ or IL-4 specific antibodies. Splenocytes from individual animals were added in tetraplicate wells (2×105 cells/well) and were restimulated with 5 µM Aβ1–33-MAP peptide, 5 µM Aβ1–40 peptide or left without restimulation in a culture medium only. After incubation for 36 h (37°C, 5% CO2), cytokines were detected with biotinylated detection antibodies, followed by avidin-HRP. Substrate AEC (Sigma-Aldrich) was added to develop the reaction. Spots representing cytokine-producing cells were counted using dissecting microscope (Olympus, Tokyo, Japan) by 3 independent investigators. Averaged data are presented as number of cytokine secreting cells per 1×106 splenocytes without re-stimulation or after re-stimulation with peptides Aβ1–33-MAP or Aβ1–40.
Aβ42 peptide was kindly provided by Dr. C. Glabe (UC Irvine Alzheimer’s Disease Research Center). Aβ oligomers or fibrils enriched preparations were made according to previously published method (Kayed et al., 2003; Kayed et al., 2007). Briefly, fibrils were prepared in 10 mM Tris pH7.4 buffer. The samples were stirred with a Teflon coated micro stir bar at 500 rpm at room temperature for 6 days. Fibril formation was monitored by thioflavin T fluorescence. Once fibril formation was complete, the solutions were centrifuged at 14,000 ×g for 20 min, the fibril pellet was washed 3x with the doubly distilled water, and then resuspended in the PBS, pH7.4 buffer. The final peptide concentration was 0.3–0.5 mg/ml.
Soluble oligomers were prepared by dissolving 1.0 mg Aβ42 in 400 µL HFIP for 10–20 min at room temperature. 100 µL of the resulting seedless Aβ solution was added to 900 µL DD H2O in a siliconized tube. After 10–20 min incubation at room temperature, the samples were centrifuged for 15 min at 14,000 ×g, and the supernatant fraction was transferred to a new siliconized tube and subjected to a gentle stream of N2 for 5–10 min to evaporate the HFIP. The samples were then stirred at 500 rpm using a Teflon coated micro stir bar for 24–48 hr at 22 °C. The oligomer formation was assessed by dot blot described below.
C2C12 myoblasts from C3H mouse strain (ATCC, Manassas, VA) were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS), 50 units penicillin and 50 µg/ml streptomycin at 37°C in a humidified atmosphere containing 5% CO2. Cells were plated in 96-well plate at a concentration of 5,000 cells/well for 24 hrs. For differentiating C2C12 cells, culture media were replaced with differentiation media (DMEM with 2% FBS and penicillin/streptomycin) and further incubated for 5 days. Differentiation of C2C12 was confirmed by immunofluorescent staining with muscle creatine kinase (Santa Cruz Biotechnology, Santa Cruz, CA), fast-type skeletal myosin heavy chain or MyoD1 (both from Abcam, Cambridge, MA) with TOTO3 nuclear counter stain. Undifferentiated or differentiated C2C12 cells were exposed to oligomeric or fibrillar Aβ at a concentrating ranging from 0.01 to 10 µM in DMEM containing 1% FBS for additional 24 hrs. For studies involving neutralization with antibodies, 30 µg/ml of purified antibody from immunized mice, 20.1 antibody (Oddo et al., 2006a), or pre-immune mouse IgG was added together with oligomeric Aβ. Cytotoxicity was assessed by lactate dehydrogenase (LDH) assay kit (Sigma-Aldrich).
Skeletal muscle tissue was snap frozen in liquid nitrogen-cooled isopentane and stored at −80°C. Cryosections were cut at 10 µm, placed onto silane-coated slides, and stored at −20°C. Hematoxylin-and-eosin (HE) staining and Gomori trichrome staining were performed to determine the general morphology of the muscle. Serial sections were used to identify oligomeric and amyloid deposits by A11 antibody, OC antibody (generous gift from Dr. C. Glabe), 6E10-biotinylated antibody (Signet Laboratories, Dedham, MA) and specific antibody for x-Aβ40 or x-Aβ42 (generated at IBAD by DHC & VV). A11 antibody recognizes amyloidogenic soluble prefibrillar oligomer species, and OC antibody recognizes Aβ fibrillar oligomers that are distinct from random coil monomers or prefibrillar oligomers (Kayed et al., 2003; Kayed et al., 2007). All double immunofluorescent images were taken and analyzed by z-stack confocal microscopic system (Bio-Rad 2000).
Skeletal muscle tissue was homogenized in T-PER buffer (Pierce) in the presence of protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN) and phosphatase inhibitors (5 mM sodium fluoride and 1 mM sodium orthovanadate). The detergent-soluble fraction was isolated by centrifugation at 100,000 g for 1 hr at 4°C. The resultant pellet was homogenized in 70% formic acid followed by centrifugation at 100,000 g for 1 hr at 4°C to isolate the detergent-insoluble fraction. Equal amounts of protein from each fraction (~50 µg) were resolved by SDS-PAGE (4–12% Bis-Tris gel from Invitrogen, Carlsbad, CA). Primary antibodies used in this study were 6E10 antibody (Pierce, Rockford, IL) for full-length APP and anti-αB-crystallin antibody (Affinity Bioreagents). Membranes were stripped using stripping buffer (GenoTechnology, St. Louis, MO) and re-probed with GAPDH (Santa Cruz Biotechnology, Santa Cruz, CA) to control for protein loading.
Skeletal muscle homogenates (150 µg) were incubated with 2 µg of A11 or 6E10 antibody overnight at 4°C with gentle rocking. Protein A-agarose (for A11) or protein G-agarose (for 6E10) was added and further incubated for 2 hrs. Protein-antibody complex was isolated by serial centrifugation and washing, and subsequently analyzed by immunoblotting.
T-PER extracted skeletal muscle samples (5 µg) were placed onto the nitrocellulose membrane and air-dried for 20 min. The membrane was then blocked with 5% fat-free milk for 2 hrs and incubated with primary antibody, A11 antibody (1:2000) or OC antibody (1:2000), overnight at 4°C. Signals were detected by secondary anti-rabbit-HRP conjugated antibody and Supersignal Pico detection reagent (Pierce).
The titers of anti-Aβ antibodies were measured as previously described with minor modifications (Cribbs et al., 2003). Briefly, wells of 96-well plates (Immulon 2HB, Dynatech Laboratories) were coated with 2.5 µM of Aβ42 in carbonate coating buffer, pH 9.6 (Sigma-Aldrich) and incubated overnight at 4°C. The wells were washed, and blocked with 3% no-fat dry milk for 1 hour at 37°C with shaking. After washing, serial dilutions of plasma from experimental and control mice were added to the wells and the plate was incubated for 1 hour at 37°C with shaking. After washing, HRP-conjugated anti-mouse IgG (H+L) antibodies (Santa-Cruz Biotechnology) were added at 1:4000 dilution. After incubation for 1 hour at 37°C with shaking, wells were washed, and Ultra-TMB ELISA substrate (Pierce) was added to develop the reaction for 15 min. and then 2 N H2SO4 was added to the wells to stop the reaction and the plates were analyzed on a Synergy HT Spectrophotometer (Bio-Tek, Winooski, VT) at 450 nm. Concentration of the antibodies was calculated utilizing 6E10 antibody as a standard on a KD4 Software (Bio-Tek). Isotypes of anti-Aβ antibodies were detected using mouse antibody isotyping kit (Invitrogen) according to manufacture instructions. To detect IgG2ab isotypes, we used biotinylated anti-IgG2ab antibodies (BD Bioscience, San Jose, CA) followed by streptavidin-HRP.
Motor performance was evaluated using the accelerating rotarod (Accuscan Instruments, Columbus, OH) as described previously (Sugarman et al., 2002). Mice were placed on a rotating dowel and required to continuously walk forward to avoid falling off. For measuring balance and coordination, the rod was accelerated over 20 sec to a constant speed of 15 rpm, and each trial was ended at 90 sec. Mice were given 10 training trials per day for two consecutive days, and probe trials were completed on the third day. In the probe trial, each mouse was tested 5 times and time of fall-off was recorded. For the accelerating rotarod, the rod constantly accelerated at a rate of 1 rpm/sec. Mice were tested for 10 trials, and the time of fall-off was recorded.
All data were analyzed using one-way ANOVA with appropriate post-tests, and p<0.05 or lower was considered to be statistically significant.
In this study, we used 12-month old MCK-APP/PS1 mice. MCK-APP/PS1 mice selectively overexpress human APP (about 2–4 fold over the endogenous mouse counterpart) and accumulate Aβ42 in affected muscle fibers, both of which are important features observed in IBM patients (Sarkozi et al., 1993; Vattemi et al., 2003; Askanas and Engel, 2008). Hence, although not a complete model, these mice nevertheless, represent a useful tool for studying the pathogenic role of APP and Aβ on skeletal muscle.
At 12 months, the MCK-APP/PS1 mice start to develop myopathological features as well as motor impairment, which become exacerbated in an age-dependent fashion (Kitazawa et al., 2006). Hence, we sought to investigate whether the sustained presence of Aβ was required for the progression of the pathological phenotype. We actively immunized mice against Aβ to determine whether it would mitigate the disease phenotype. Initial studies revealed that immunizing the MCK-APP/PS1 transgenic mice with full-length fibrillar Aβ1–42 peptide formulated in QuilA or CFA/IFA adjuvant did not produce a detectable anti-Aβ antibody response after three injections, probably due to self-tolerance. Also, it is worth mentioning that C57BL/6 mice are generally a poor antibody responder strain to most antigens and biased towards a Th1 (cellular) type of immune response (Reiner and Locksley, 1995; Gustavsson et al., 1998; Mills et al., 2000) (and our unpublished observations). Consequently, we used a modified immunogen, Aβ1–33-MAP peptide, which demonstrated a higher potency for inducing an antibody response compared to the traditional AN1792 formulation. The new antigen was clearly advantageous, as after two injections, anti-Aβ antibodies were detectable in the actively immunized group, which mainly comprised the IgM isotype (Fig 1a). Subsequent injections significantly augmented antibody production, and we observed antibody class switching from IgM to IgG, which indicates induction of an Aβ-specific T-cell response. The ratio of the IgG/IgM antibody isotype after the second injection was 1:1. At the end of the trial, however, the IgG/IgM ratio was 3:1, consistent with the development of T-cell dependent type of immune response. No detectable anti-Aβ antibody response was observed in untreated or sham-treated animals.
The Aβ specific T-cell response following active immunization was further analyzed by ELISPOT. After 5 injections, only actively immunized animals produced T cells specific to both peptide immunogen Aβ1–33-MAP and Aβ1–40 peptide. Both of the peptides contain a functional T-cell epitope, which is located in the region Aβ16–30 in C57Bl/6 mice (Monsonego et al., 2006). Not surprisingly, we detected a dominant Th1-type of the response with the splenocytes secreting pro-inflammatory cytokine IFN-γ (Fig. 1b) after re-stimulation with Aβ1–33-MAP (204 ± 45 cells) or Aβ1–40 (215 ± 35 cells) peptides, while non-stimulated splenocytes have significantly lower number of IFN-γ secreting cells (29 ± 7 cells). Splenocytes from naïve and adjuvant only groups did not respond for the in vitro restimulation with the Aβ1–33-MAP or Aβ1–40 peptides (Fig. 1b). Moreover, we did not detect splenocytes secreting anti-inflammatory cytokine IL-4 (data not shown), which suggests a pronounced Th1-type of the immune response typical in C57BL/6 mice (Reiner and Locksley, 1995; Mills et al., 2000).
To access the phenotypic efficacy of Aβ immunotherapy, we utilized both the constant speed and accelerating version of the rotarod and evaluated the overall motor performance of each group of mice longitudinally. The constant speed task examines motor coordination and balance, whereas the accelerating version primarily evaluates muscle strength and duration (Boehm et al., 2000; Rustay et al., 2003). The first test was done prior to immunization, and no differences were observed among the three groups (Fig. 2a, d). At this age, all groups reached at average 69–71 sec at 15 rpm constant rotation. Most of the age-matched parental PS1-KI mice (MCK-APP negative) stayed on the rotor for 90 sec throughout the experimental period (Fig. 2). Thus, based on these studies, we determined that the MCK-APP/PS1 mice exhibited impaired motor function before the immunization therapy was initiated. After three injections of the immunogen (2 months after the initial injection), motor function was again assessed, and we found that the Aβ-immunized group performed better than the untreated or sham-treated groups, although the difference was not statistically significant (Fig. 2b, e). After five injections (i.e., 3 months after the initial injection), the disparity in performance among the groups was more robust, and Aβ-immunized group exhibited significantly better motor function than the untreated group (Fig. 2c, f). Interestingly, the sham-treated group showed a trend toward improvement compared with untreated group although it was not statistically significant, suggesting the activation of non-specific immune responses may contribute to attenuating pathological phenotypes as previously reported (Frenkel et al., 2005). Hence, our findings indicate that the impairment of motor function was markedly slowed in the actively immunized mice, whereas motor function grew progressively worse in the control groups. The continuous decline in motor performance in the control or sham-treated groups during the experimental period suggested that the effect of learning the task in this longitudinal study was minimal (Suppl. Fig. 1).
We next determined whether the improved motor function in the transgenic mice was associated with an attenuation of the IBM-like myopathologies. Intramuscular vacuoles are a signature feature of IBM, and histological examination of muscle by hematoxylin-and-eosin (HE) staining revealed that vacuoles (arrowheads) were markedly reduced in muscle fibers from the actively immunized mice (Fig. 3a). At least 20 sections from each mouse were stained with HE, and the extent of vacuoles formed in the skeletal muscle fibers were quantitatively analyzed and found significantly less in the actively immunized group (Fig. 3a). However, it should be noted that these vacuoles were different from rimmed vacuoles observed in human cases, but rather were similar to linear basophilic vacuole-like structures previously described as rimmed-cracks in another mouse model of myopathy (Weihl et al., 2007).
Other histopathological markers such as αB-crystallin and major histocompatibility complex (MHC) class I are elevated in IBM, hence we determined whether levels of these markers were affected by immunization on our mouse model. αB-crystallin is a small heat-shock protein that is abnormally localized and upregulated in skeletal muscle fibers of IBM patients as an indication of cellular stress (Banwell and Engel, 2000; Wojcik et al., 2006). Likewise, MHC class I molecules are also upregulated in muscle fibers, which may be responsible for the infiltration of T-cells (Ferrer et al., 2004). In the untreated or sham-treated MCK-APP/PS1 mice, αB-crystallin was found to be upregulated, whereas their levels were significantly reduced in the immunized mice, suggesting that the reduction of intracellular Aβ achieved by immunization significantly suppressed cellular stress responses (Suppl. Fig. 2a). Conversely, there was no disparity in MHC class I levels among the three groups (Suppl. Fig. 2b).
Aβ levels in skeletal muscle were quantitatively measured by ELISA. Although soluble Aβ40 or Aβ42 was not altered by immunization, insoluble Aβ levels, particularly Aβ42, were significantly reduced in the actively immunized group (Fig. 3c). The steady-state levels of APP or APP processing in skeletal muscle were unaffected by the treatments (Fig. 3b and data not shown). In addition, there was a good correlation between the reduction of Aβ levels and antibody titers in individual mice (Fig. 3d); both insoluble Aβ40 and Aβ42 showed a linear correlation with the antibody titer with a correlation coefficient of 0.48 and 0.49, respectively. Similarly, free Aβ levels in plasma were also decreased in the immunized mice with higher anti-Aβ antibody titers (Suppl. Fig. 3). Our data suggest that the antibody generated by the active immunization effectively sequestered Aβ in the plasma and lowered free plasma Aβ, which is consistent with previous findings of passive immunization on a mouse model of AD (DeMattos et al., 2001).
We next addressed the mechanism by which active immunization led to the clearance of Aβ in skeletal muscle fibers. It has recently been reported that antibodies are capable of being internalized and facilitating the clearance of the intracellular protein aggregates in a model of Parkinson’s disease (Masliah et al., 2005). To determine if a similar mechanism participated in the reduction of Aβ in skeletal muscle following immunization, we double labeled skeletal muscle sections with anti-mouse IgG and an antibody called A11, which detects oligomeric Aβ species (Kayed et al., 2003). In the immunized mice, we obtained evidence showing that some muscle fibers clearly contained antibodies that were detected by anti-mouse IgG, and these antibodies were in the vicinity of A11-positive deposits (Fig. 4). Notably, in untreated mice, we did not find any evidence of mouse antibodies in the fibers harboring A11-positive deposits (Fig. 4). Similarly, mouse IgG surrounded Aβ40- and Aβ42-immunoreactive deposits in actively immunized mice but not in untreated or sham-treated mice, further confirming the high affinity of antibodies to Aβ deposits (Fig. 4). Therefore, this analysis revealed that some of the mouse endogenous antibodies were able to penetrate within the skeletal muscle fibers. Based on these data, we suggest that some portion of antibody is able to enter the muscle fibers and directly bind to the amyloid deposits and mediate its degradation and clearance.
Recent evidence from in vitro and in vivo studies indicates that Aβ oligomers play a crucial role in mediating neuronal toxicity and subsequent neuronal dysfunction (Walsh et al., 2002b; Lesne et al., 2006; Malaplate-Armand et al., 2006; Oddo et al., 2006b). Hence, we examined the levels of Aβ oligomers in the muscle fibers and evaluated whether active immunization reduced their levels. To quantify Aβ oligomers, we used a dot blot assay with two oligomer-specific antibodies, A11 and OC, which detect amyloidogenic pre-fibrillar oligomers and fibrillar Aβ oligomers, respectively (Kayed et al., 2003; Kayed et al., 2007). In the immunized mice, we observed a significant reduction in the A11 signal in the dot blots compared to the untreated and sham groups (p<0.05; Fig. 5a). Likewise, the biochemical analysis also indicated that the OC-positive species were also significantly reduced in the immunized versus control groups (Fig. 5a). Finally, we immunohistochemically labeled muscle sections with antibodies A11 and OC from mice that underwent the various treatments and found that active immunization markedly reduced intracellular oligomers in skeletal muscle (Fig. 5b). Double immunofluorescent staining of skeletal muscle with antibodies 6E10 and A11 showed an overlap in staining, indicating that some immunoreactive deposits contained Aβ oligomers (Suppl. Fig. 5a). A few amyloid deposits that were not A11-positive were also found in skeletal muscle (Suppl. Fig. 4a). Furthermore, immunoprecipitation and immunoblot analysis also demonstrated that A11- or OC-positive oligomers were Aβ (Suppl. Fig. 4b). Therefore, immunotherapy was not only able to reduce Aβ levels in skeletal muscle, but more importantly, it was able to significantly reduce Aβ oligomers, which are generally considered to be the most cytotoxic assembly state of Aβ (Walsh et al., 2002a).
Relatively little is known about the pathophysiological role of Aβ oligomers in muscle cells. Hence, we used both undifferentiated and differentiated murine myoblast C2C12 cells to assess cytotoxicity of oligomeric Aβ species using the LDH assay. The differentiation of C2C12 cells was confirmed by the expression of muscle creatine kinase, myosin heavy chain and MyoD1, as well as the presence of multi-nucleated cells (Fig. 6e). Two solutions containing predominantly Aβ oligomers or Aβ fibrils were prepared (Suppl. Fig. 5a). We found that Aβ oligomers induced cytotoxicity in both undifferentiated and differentiated C2C12 cells in a dose-dependent manner after 24 hrs of exposure (Fig. 6a, c, Suppl. Fig. 5b, c). Our studies also indicated that Aβ oligomers were more toxic than fibrillar Aβ even at low concentrations in these cells.
Because anti-Aβ antibodies provided therapeutic benefit in mice, we sought to determine whether the oligomer-induced cytotoxicity could be neutralized by using antibodies isolated from the actively immunized mice. Undifferentiated C2C12 cells were exposed to various doses of Aβ oligomers for 24 hrs either in the presence or absence of affinity purified antibodies (30 µg/ml) isolated from the immunized animals. Notably, we found that sera harvested from the actively immunized mice and affinity purified on an Aβ1–15 column was able to significantly block the oligomer-mediated toxicity (Fig. 6b, Suppl. Fig. 5b). As a positive control, we also found that mouse monoclonal N-terminal specific anti-Aβ antibody (20.1 antibody (Oddo et al., 2006a)) also effectively blocked toxicity (data not shown). In contrast, the protective effect was not due to non-specific co-treatment with IgG as mouse monoclonal IgG failed to block oligomer cytotoxicity (Fig. 6b). Likewise, the toxicity of Aβ oligomers was significantly neutralized by co-treatment with an anti-Aβ antibody (20.1) in differentiated C2C12 cells (Fig. 6d). These in vitro findings demonstrate that Aβ oligomers are potent toxic mediators that may play a key role in the progression of the muscle degeneration in a mouse model of IBM. Our in vivo studies suggest anti-Aβ antibodies generated by active immunization can neutralize oligomer toxicity in muscle.
In this study, we determined whether the sustained presence of Aβ in skeletal muscle is required to maintain motor dysfunction in the MCK-APP/PS1 transgenic mice. We actively immunized against Aβ, which successfully reduced Aβ levels and accumulation of insoluble forms in skeletal muscle. We found the immunogen consisting of Aβ1–33 peptide linked to a lysine backbone effectively elicited an immune response and generated Aβ-specific antibodies as early as after the second injection. After 3 months of immunization, insoluble Aβ levels were significantly reduced in the skeletal muscle, as were histopathological features such as vacuoles and αB-crystallin. The motor performance of the immunized mice was markedly better as determined by the constant speed and accelerating rotarod tests. These results provide strong evidence that reducing Aβ levels in muscle can mitigate the motor impairments.
Notably, our study also shows that active immunization lowers Aβ oligomers in skeletal muscle. Aβ oligomers are believed to play a critical role in the pathogenesis of AD and are more neurotoxic than monomers or fibrils (Caughey and Lansbury, 2003; Gong et al., 2003; Demuro et al., 2005). In addition, Aβ oligomers are detected in muscle fibers from sporadic IBM patients, and these oligomers interact with αB-crystallin to potentially induce cellular stress and subsequent degenerative mechanisms (Wojcik et al., 2006). In agreement with these reports, we show that Aβ oligomers are highly toxic to both undifferentiated and differentiated C2C12 myoblast cells, and the cytotoxicity is effectively blocked by co-treatment with anti-Aβ antibodies collected from the immunized mice. These results suggest that Aβ oligomers may be key mediators of toxicity in skeletal muscle.
Although numerous in vitro, in vivo and clinical studies point to a toxic role for Aβ in muscle, its role in the pathogenesis of IBM remains poorly understood. Some investigators argue that IBM is a combined proteinopathy and inflammatory disease, and these two processes can interact and influence each other (Dalakas, 2006b; Needham and Mastaglia, 2008). Recent findings showing that inflammation mediates Aβ or tau pathology in skeletal muscle supports this view (Kitazawa et al., 2008; Schmidt et al., 2008). On the other hand, others suggest Aβ accumulation is a downstream event, because all known genetic mutations for the hereditary form of IBM are not associated with APP processing or degradation. Interestingly, recent mouse models of hereditary IBM harboring GNE or VCP mutations exhibit intracellular Aβ-like deposits in the skeletal muscle (Malicdan et al., 2007; Weihl et al., 2007), indicating that although these mutations do not have a direct relationship with APP, they modulate APP processing or Aβ degradation, causing the accumulation of Aβ in affected skeletal muscle fibers. It is important, however, to note that mouse endogenous Aβ does not aggregate as readily as human Aβ, and differs by three amino acids at position 5, 10 and 13; several studies have shown that these amino acids in the human peptide are particularly important for controlling the aggregation property of Aβ and are presumed to have toxic properties (Fraser et al., 1992; Dyrks et al., 1993; Otvos et al., 1993; Atwood et al., 1998).
The anti-Aβ antibodies induced in the MCK-APP/PS1 transgenic mice by active immunization effectively removed various assembly states of Aβ including fibrils and oligomers. Further experiments will be required to determine the exact mechanism(s) by which the antibodies facilitate clearance of intracellular Aβ. However, several studies have demonstrated that intracellular deposits of Aβ or α-synuclein can be cleared from neurons following immunization (Oddo et al., 2004; Billings et al., 2005; Masliah et al., 2005).
In the case of active immunization against Aβ in AD, the remarkable successes in animal models, largely free of adverse events, did not translate to the clinical trial in AD patients, as approximately 6% of the patients developed meningoencephalitis (Schenk et al., 2005). Currently, more than ten new approaches to active and passive immunotherapy are in various stages of clinical trials with the goal of improving the safety and efficacy anti-Aβ immunotherapy approaches (Wisniewski and Konietzko, 2008). As a consequence of this continued improvement in anti-Aβ therapeutic strategies, there may be real opportunities to safely pursue a clinical trial in IBM patients.
The progression of skeletal muscle weakness and atrophy in IBM is relatively slow but irreversible. Ultimately, patients are forced to rely on assisting devices and/or caregivers. Currently, no effective treatment for IBM exists; a number of immunosuppressive or immunomodulatory therapies that have worked well on other inflammatory myopathies have shown only marginal improvements in IBM patients (Walter et al., 2000; Dalakas et al., 2001; Badrising et al., 2002; Rutkove et al., 2002; Tawil, 2004). Hence, there is a critical need to develop novel approaches to treating IBM. In this study, we provide evidence that Aβ is one of the key pathological components of IBM and promoting clearance of Aβ from skeletal muscle may be a valid therapeutic strategy to pursue. Additional studies using different animal models would further advance the understanding of this devastating age-associated myopathy.
This study was supported by grants from the National Institutes of Health (NIH): NIH/NIA R01AG20335 (F.M.L.), R01AG20241 (D.H.C.), P01AG00538 (D.H.C) and NIH/NIAMS K99AR054695 (M.K.). Aβ peptides and antibodies were provided by the UCI Alzheimer’s Disease Research Center (ADRC) funded by NIH/NIA grant P50AG16573 and the Institute for Brain Aging and Dementia (IBAD) funded by the NIH Program Project Grant, AG00538.