Calcium Dyshomeostasis in β-Amyloid and Tau-bearing Skeletal Myotubes*

The relative scarcity of inclusion-affected muscle cells or markers of cell death in inclusion body myositis (IBM) is in distinction to the specific and early intracellular deposition of several Alzheimer's Disease (AD)-related proteins. The current study examined the possible correlation between myotube β-amyloid and/or Tau accumulations and a widespread mishandling of intracellular muscle calcium concentration that could potentially account for the unrelenting weakness in affected patients. Cultured myogenic cells (C2C12) expressed β-amyloid-42 (Aβ42) and fetal Tau peptides, as human transgenes encoded by herpes simplex virus, either individually or concurrently. Co-expression of Aβ42 in C2C12 myotubes resulted in hyperphosphorylation of Tau protein that was not observed when Tau was expressed alone. Resting calcium concentration and agonist-induced RyR-mediated Ca2+ release were examined using calcium-specific microelectrodes and Fluo-4 epifluorescence, respectively. Co-expression of Aβ42 and Tau cooperatively elevated basal levels of myoplasmic-free calcium, an effect that was accompanied by depolarization of the plasma membrane. Sarcoplasmic reticulum (SR) calcium release, induced by KCl depolarization, was not affected by Aβ42 or Tau. In contrast, expression of Aβ42, Tau, or Aβ42 together with Tau resulted in enhanced sensitivity of ryanodine receptors to activation by caffeine. Notably, expression of β-amyloid, alone, was sufficient to result in an increased sensitivity to direct activation by caffeine. Current results indicate that amyloid proteins cooperate to raise resting calcium levels and that these effects are associated with a passive SR Ca2+ leak and Tau hyperphosphorylation in skeletal muscle.

Both cortical neurons and skeletal muscle are electrically specialized, post-mitotic tissues endowed with cholinergic synapses and internal calcium release channels and are uniquely disordered in Alzheimer's disease (AD) 1 and inclusion body myositis (IBM), respectively. Deposits of AD-associated proteins, such as ␤-amyloid peptide (A␤), hyperphosphorylated Tau, and other shared misfolded proteins, such as, ␤-amyloid precursor protein (␤APP), ubiquitin, as well as various chaperones and kinases are thought to play pathologic roles in the cognitive decline (1) and muscular failure (2), which define these incurable human conditions. Ca 2ϩ ions play essential roles in numerous cellular processes. The release of Ca 2ϩ from the intracellular stores and the influx of Ca 2ϩ through the surface membrane ion channels and transporters are critical to both memory formation and contraction in their cognate tissues (3). The prolonged elevation of cellular Ca 2ϩ has been shown to be deleterious to neuronal and muscle function (4).
There is an extensive literature base devoted to the concept that neural calcium dyshomeostasis contributes to the pathogenesis of Alzheimer's disease (5,6). As pertains to the ␤-amyloid hypothesis, published results indicate that application of A␤ 42 peptide potentiates excitotoxic mechanisms through oxidative damage, increase in Ca 2ϩ influx and reduction of Ca 2ϩ removal (7,8). These alterations in Ca 2ϩ homeostasis could be linked to inhibition of numerous intracellular and plasma membrane ion-motive ATPases (9) or it could arise from the ability of A␤ to form selective calcium permeant pores with conductances following electrochemical gradients (10). Calcium dysregulation may also involve aberrant storage or release of calcium from internal stores involving the endoplasmic reticulum (ER) and its calcium release channels (11,12). This hypothesis has been supported by correlation of the changes in [ 3 H]ryanodine binding, with the load of both neurofibrillary and A␤ pathologies in AD entorhinal/hippocampal cortex (13), thereby suggesting an alteration of activity of the ryanodine receptor Ca 2ϩ release channels (RyR). One possible mechanism for AD progression is governed by a pathologic cycle of Ca 2ϩ dysregulation and cell damage, whereby, the effect (increased [Ca 2ϩ ]) can reciprocate to further enhance the cause (A␤ production) (14 -16).
IBM is the most common acquired muscle disorder affecting the elderly. The etiology of IBM is unknown. While sporadic IBM is clearly an inflammatory condition for which evidence exists to support an immune basis, there are also unique clinical and pathological features pointing to a primary degenerative cause (17). IBM disorder shares several common features with the AD. In IBM, intracellular inclusions bearing ADassociated proteins (rimmed vacuoles) occur in scattered, non-necrotic, atrophic, angulated myofibers (18). Their sparse occurrence, however, make it likely that clinical muscle weakness arises from a more widespread or generalized metabolic defect, for instance related to oxidative stress (19) that can possibly lead to mitochondrial failure, calcium dyshomeostasis and/or a disorder of filament function and cytoskeletal protein interactions. Calcium dysregulation for example, could be limited to myofibers affected with inclusions or also involve normal appearing myofibers, which harbor undetectable but nevertheless toxic levels of intracellular soluble oligomeric ␤-amyloid species (see Ref. 20).
Most of the previous work in the field of Ca 2ϩ dysregulation in AD has utilized numerous mutant presenilin (PS)-based models. The introduction of PS-1 mutations into cultured neuronal cells, has been shown to increase the levels of expression of the RyR that lead to a potentiation of caffeine-induced calcium release (21) and heightened sensitivity to the application of A␤ peptides and other injurious stimuli, resulting in enhanced apoptosis (21,22). The mechanism of abnormal Ca 2ϩ signaling in cells harboring presenilin mutations has been shown to involve an overload of ER Ca 2ϩ stores (22)(23)(24). It has also been found that the C-terminal domain of ␤APP (AICD) has nuclear transcriptional activities and modulates calcium handling, suggesting that a signal transduction mechanism underlies the buildup of ER calcium stores (25). However, unresolved issues with this approach include the specific role of A␤ 42 levels and the possible extent to which RyR-mediated Ca 2ϩ release is altered.
To date, little attention has been given to the possibility that dysregulation of intracellular calcium in diseased skeletal myofibers may have a significant role in IBM pathogenesis. Since the pathogenesis of AD and IBM share several hallmarks, and neuronal and skeletal muscle have some functional similarities, IBM could serve as a suitable model for studying the effects A␤ and Tau proteins as they pertain to development of AD. In the current study we utilized this model to study Ca 2ϩ dysregulation in response to direct overexpression of exogenous ␤-amyloid and Tau peptides in skeletal muscle myotubes. The results of this study indicate that in addition to significant depolarization of the plasma membrane, the resting intracellular Ca 2ϩ levels in AD protein-expressing muscle cells were significantly elevated. This finding was associated with the enhancement of Ca 2ϩ -induced Ca 2ϩ release (CICR); supporting the concept that chronic degeneration in IBM may represent a decline in non-activity dependent muscle function. It was also examined whether depolarization-induced Ca 2ϩ release from the sarcoplasmic reticulum of cultured skeletal muscle cells was affected in a way that could account for the elevation in basal Ca 2ϩ levels.

MATERIALS AND METHODS
Herpes Simplex Virus Synthesis-␤-amyloid 42 and the fetal isoform (0N3R) of human Tau (HFT) sequences were cloned into a Herpes simplex virus amplicon for expression in muscle cells. A 17 amino acid signal peptide ER localization sequence of ␤-amyloid precursor protein precedes the A␤ 42 sequence, as previously reported (26). The Tau sequence was digested from the pBR322-based p19 Tau plasmid (27) and inserted into pHSV-PrPUC (gift from R. Neve, McLean Hospital, Boston, MA) as shown in Fig. 1A. The amplicons were transfected into Vero cells with LipofectAMINE plus. Sixteen hours after transfection 1 ϫ 10 6 plaque-forming units of ␣8 helper strain were added, and Herpes viruses were subjected to three rounds of amplification. Viral particles were purified on a sucrose gradient according to Ref. 28. Titers ranged from 2.4 ϫ 10 7 to 1.5 ϫ 10 8 for HSV-A␤ 42 and 4.3 to 5.3 ϫ 10 7 for HSV-HFT. Control Herpes amplicon coding for ␤-galactosidase (HSV-LacZ) was described (26).
Cell Culture Conditions-C 2 C 12 cells were expanded in growth medium, which consisted of Dulbecco's modified essential medium supplemented with 20% fetal calf serum (Invitrogen, Carlsbad, CA). Upon reaching full confluency, growth medium was replaced with differenti-ation media, which consisted of DMEM supplemented with 2% heatinactivated horse serum. The cells were differentiated for 3-7 days, until most of the myocytes had fused, and infected with either 0.7 multiplicity of infection (moi) HSV-sigA␤ 42 or 0.2 moi HSV-HFT or both. From previous work, similar moi resulted in maximal transgene expression without cell loss at the viral exposure times used for Western analysis (see below). Control cells were infected with 0.9 moi of HSV-LacZ. Viral load was always equalized between treatments by the addition of the appropriate amounts of HSV-LacZ. Cells were continuously exposed to virus for 9 or 21 h following infection, depending on the type of experiment, and prior to harvesting for Western blot (21 h), fixation for immunostaining (9 h), or loading for Ca 2ϩ imaging (9 h).
Western Blot Analysis-2 ϫ 10 5 cells were harvested 21 h after being infected with the herpes simplex viral constructs. Cells were washed once with ice-cold phosphate-buffered saline then lysed in 50 l of radioimmune precipitation assay buffer (50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 5 mM EDTA, 50 mM sodium fluoride, 1 mM sodium orthovanadate, 1mM phenylmethylsulfonyl fluoride, 2 mM sodium pyrophosphate, 1 g/ml pepstatin A, 100 g/ml leupeptin, and one protease inhibitor mixture tablet/50 ml (Roche Applied Science). Cells were immediately scraped from each well, incubated on ice for 20 min, vortexed, and centrifuged at 14,000 ϫ g for 15 min. The supernatant was transferred to a new tube and frozen at Ϫ80°C until analyzed for protein content using the Bio-Rad protein assay and bovine IgG for the standards. 12.5 g of total protein was separated on a 4 -12% Bis/Tris SDS-PAGE gel and transferred to polyvinylidene difluoride (PVDF) membrane. PVDF membranes were washed once in Tris-buffered saline containing 0.5% Tween-20 (TBST), pretreated with 15% H 2 0 2 for 15 min and followed by 6 ϫ 5 min washes. Blocking took place in TBST containing 2.5% bovine serum albumin and 2.5% nonfat milk for 45 min at 25°C. Membranes were incubated with primary antibodies in TBST containing 5% bovine serum albumin overnight at 4°C. The following day membranes were washed three times for 10 min in TBST, followed by 45 min of incubation with secondary antibody (e.g. horseradish peroxidase-conjugated goat anti-mouse (Amersham Biosciences) at 1:1000) in TBST at 25°C. Membranes were washed three times for 10 min in TBST before signal detection by enhanced chemiluminescence according to the manufacturer's protocol (Amersham Biosciences). For multiple detections, membranes were stripped with sodium dodecyl sulfate and ␤-mercaptoethanol for 30 min at 50°C, and blocked in 5% nonfat milk before incubation with the next primary antibody. Primary antibodies (mc, monoclonal; rab, rabbit and dilutions; see Immunostaining of C 2 C 12 Cells-Cells were grown in an 8-chamber slide to a density of 2 ϫ 10 4 cells and were differentiated in Dulbecco's modified Eagle's medium containing 2% horse serum for 2-4 days. The cells were infected with herpes simplex viruses as described above for 9 h. Cells were washed once with ice-cold phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde in phosphate buffer for 15 min at 25°C. Each well was washed three times for 5 min prior to permeabilization with 3% Triton X-100 in PBS for 3 min and further washed three times for 5 min. Nonspecific sites were blocked by incubating with 10% normal goat serum in PBS. Cells were incubated overnight at 4°C with combinations of primary antibodies: mc anti-␤-galactosidase 1:50 (Roche Applied Science), mc anti-Tau-5 1:50 (Signet), mc anti-6E10 1:100, rab anti-Tau phosphothreonine 231 1:100 (gift from Dr. Koichi Ishiguro), and/or rab R1282 1:300 (gift from Dr. Dennis Selkoe) in PBS containing 10% normal goat serum and 0.1% Triton X-100 in PBS. The following day slides were washed three times for 5 min with PBS. The HSV-LacZ well was incubated with horseradish peroxidase-conjugated goat anti-mouse IgG at 37°C for 45 min. Cells were washed, and the color was developed by incubation in NBT. The other chambers were washed and incubated with either Cy2-conjugated goat anti-mouse IgG or Cy3-conjugated goat anti-rabbit IgG (1:300) in 10% normal goat serum and 0.1% Triton X-100 in PBS for 45 min at 37°C. Each chamber was washed three times for 5 min with PBS and mounted in Npropylgalate. Cells were visualized with a Nikon Diaphot microscope.
Fluorescence Measurements-Fluorescence measurements were performed as described previously by (29). C 2 C 12 myotubes were continuously infected with HSV particles bearing transgenes for 9 h. Prior to loading, differentiation media were removed, and myotubes were washed with the Ringer Solution containing in mM: 125 NaCl, 5 KCl, 1.2 MgSO 4 , 6 glucose, 25 HEPES, 2 CaCl 2 , pH 7.4. Solutions used for depolarization contained a constant product of NaCl and KCl at 130 mM. Cells were then loaded for 30 min in Ringer supplemented with Ca 2ϩ indicator dye (Fluo-4AM, 5 M). To terminate loading, myotubes were washed several times with the Ringer. Whole cell fluorescence changes were detected using PTI delta-RAM as the light source with a 12-bit digital intensified CCD (Stanford Photonics) interfaced with an inverted microscope equipped with an Olympus Uapo/340 ϫ40 oil immersion objective. Intracellular [Ca 2ϩ ] transients were characterized as changes in Fluo-4 fluorescence intensity. All experiments were conducted in Ringer at room temperature (22°C). Solution exchange within each well was achieved via pressure-controlled perfusion system (Automate Scientific Inc., Berkley, CA). Changes in fluorescence from the regions of interest within each cell were analyzed using QED imaging software (QED Software, Pittsburgh, PA).
Membrane Potential and Resting [Ca 2ϩ ] i Recording-Intracellular resting [Ca 2ϩ ] i , and the plasma membrane potentials were recorded simultaneously using double-barreled Ca 2ϩ -selective microelectrodes that were prepared from thin-walled borosilicate glass capillaries (WPI PB150F-4, Sarasota, FL). Prior to pulling, all capillaries were washed with HCl followed by a rinse with distilled water and then dried at 150°C for 3 h. Capillaries were pulled to make short taper microelectrodes with an outside tip diameter being ϳ0.6 m. The larger barrel (1.5-mm outside diameter) was silanized by exposure to dimethyldichlorosilane vapor. Twenty-four hours later the tip was backfilled with the neutral carrier ETH129 (Fluka, Ronkontioma, NY). The remainder of the barrel was backfilled with pCa7 solution 24 h later. The smaller barrel (0.84-mm outside diameter) was backfilled with 3 M KCl (tip resistance 10 -15 ⍀M). All Ca 2ϩ -selective microelectrodes were calibrated at 22°C in solutions of known [Ca 2ϩ ] (Lopez et al. 1983) before and after the Ca 2ϩ measurements. To better mimic the intracellular ionic conditions all calibration solutions were supplemented with 1 mM Mg 2ϩ . Only those Ca 2ϩ microelectrodes that provided a Nernstian response between pCa3 and pCa7 (29.5 mV/pCa unit at 22°C) were used in this study. No more then 10 measurements were performed with each microelectrode. The potentials from the 3 M KCl barrel (V m ) and the Ca 2ϩ barrel (V Ca ) were recorded via high impedance amplifier (WPI FD-223, Sarasota, FL). The V m was subtracted electronically from V Cae , to produce a differential Ca 2ϩ -specific potential (V Ca ) that represents the resting myoplasmic Ca 2ϩ concentration. The V m and V Ca potentials were filtered  to improve the signal to noise ratio.

RESULTS
To assess the influence of Alzheimer's disease and IBMassociated proteins on skeletal muscle calcium homeostasis we examined the expression patterns of viral-encoded ␤-amyloid and Tau human transgene proteins, alone or in combination, as well as their effects on basal and stimulated cytosolic Ca 2ϩ levels in differentiated murine C 2 C 12 myotubes in culture.
Western Blot Analysis and Immunocytochemistry-In Fig.  2A, ␤-amyloid expression was not detected in whole cell lysate from C 2 C 12 myotubes that were either uninfected or infected with HSV-LacZ (lanes 2 and 3, respectively) or Tau alone (lane 5). Expression of monomeric, 4 kDa ␤-amyloid protein in C 2 C 12 myotubes infected with either HSV-A␤ 42 alone (lane 4) or in combination with HSV-human fetal Tau (HFT, lane 6) was readily detected with the 6E10 antibody. Additionally, cells that were infected with A␤ 42 also exhibited formation of various soluble oligomeric species. HFT co-infection did not affect A␤ oligomerization.
In analogous experiments using the phosphorylationspecific PHF-1 and anti-S262 and non-phosphorylationdependent Tau-5 probes, endogenous Tau was not detected in soluble extracts from uninfected, cells infected with ␤-galactosidase (Fig. 2B, lanes 1 and 2) or with A␤ 42 transgene alone (lane 4). When C 2 C 12 cells were infected with Tau alone or in combination with A␤ 42 an appearance of closely migrating bands was observed, ranging in molecular mass from ϳ52 to 60 kDa (lanes 3 and 5). This is consistent with both, the transgene expression of the short Tau isoform (30) and lack of endogenous "small Tau" expression in rat muscle fibers (31,32). When endogenous Tau degradation is inhibited, a complex of closely spaced bands, similar to ours, appears in rat muscle. These are regarded as various modifications of small Tau (33). Tau appeared more highly phosphorylated at Ser 262 in cells that were infected with both Tau and A␤ 42 when compared with those infected with Tau alone (lanes 5 versus 3). PHF-1 reactivity, specific to phosphoepitopes in the flanking region C-terminal to the microtubule binding repeat domains, was also increased in Tau-positive cells that were also expressing A␤ (lane 5 versus 3).
To examine cellular expression at times corresponding to physiological recordings below, differentiated C 2 C 12 myotubes were infected with HSV particles encoding either LacZ, Tau, A␤ 42 , or simultaneously with HSV-Tau and HSV-A␤ 42 (A␤). Nine hours after control infections in expression time course experiments, ϳ70 -90% of the myotubes infected with HSV-LacZ stained positive for ␤-galactosidase (results not shown).
The earliest morphological changes were observed in myotubes co-infected with A␤ and Tau. These changes consisted of rounding (Fig. 3, panel G) and granulovacuolarization (Fig. 3, panel J). Consistent with our previous studies (34), exogenously expressed A␤ 42 was retained intracellularly (Fig. 3, panels B, H, and K). When infected individually with HSV-A␤ 42 , or HSV-HFT, myotubes were positive for their cognate transgene (panels B and E, respectively), but were negative for the phosphorylated Thr 231 Tau epitope (panels C and F). As expected, myotubes expressing only Tau showed no reactivity with A␤detecting antibodies (not shown). In co-infections, both transgenes were expressed (panels H and I). Only in co-infections was phosphorylation of Tau on Thr 231 observed (panel L), indicating that the presence of A␤ had a hyperphosphorylating effect on Tau. Conversely, co-infections with Tau and A␤ showed a moderate decrease in the number and signal intensity of myotubes labeled using an antibody (Tau-1) specific for dephosphorylated Tau residues Ser 198 , Ser 199 , Ser 202 , and Thr 205 , relative to cells infected with Tau-encoding virus alone (data not shown). These results are also consistent with the complementary phosphorylation at Ser 262 noted in Fig. 2.
Tau, A␤, and TauϩA␤ Alter Resting Membrane Potential and Resting [Ca 2ϩ ] i -As demonstrated in Fig. 3, overexpression of AD-related proteins produced distinct alterations in the myotube morphology. Since C 2 C 12 cells are excitable myogenic cells it is feasible that the observed alterations in the cellular morphology could be accompanied by changes in the resting membrane potential (V m ). Microelectrode recordings were conducted to determine if the PM of myotubes overexpressing either one of these proteins exhibited aberrant V m with respect to the uninfected cells (Fig. 4). Fig. 4A shows representative membrane potential records of a myotube that was co-infected with Tau and ␤-amyloid (A␤) and that which was not (NV). The impalement and the removal of the microelectrode were accompanied by a downward and an upward deflection, respectively. As described previously (35) the V m of the myotubes is typically in the range between Ϫ60 mV and Ϫ70 mV, which was confirmed in the uninfected cells (Fig. 4, A and B). As demonstrated by the upper record in Fig. 4A and in Fig. 4B, an overexpression of the A␤ was accompanied by a substantial depolarization of the plasma membrane (Ϫ52 ϩ 1.9 mV). Similar to A␤, overexpression of either Tau or A␤ alone also appeared to produce a depolarizing effect on the PM (Fig. 4B,  records not shown). However, unlike in the presence of A␤, the mean magnitude of depolarization was smaller in cells expressing individual Tau or A␤ proteins (Fig. 4B). Expression of ␤-galactosidase (HSV-LacZ) did not result in changes in V m (Ϫ62 Ϯ 0.74 mV).
As described under "Materials and Methods," the use of double barrel microelectrodes allows for simultaneous recording of the V m and resting intracellular [Ca 2ϩ ] i . Fig. 4C presents intracellular Ca 2ϩ potential (V Ca 2ϩ) records (see "Materials and Methods") from the same two cells described in Fig. 4A. It is evident in cells overexpressing A␤ (Fig. 4A), Tau or A␤ alone (records not shown), that in addition to the changes in the V m , there is a significant change in the intracellular Ca 2ϩ potential, which is proportional to the changes in the resting myoplasmic Ca 2ϩ concentration ([Ca 2ϩ ] i ). Determined [Ca 2ϩ ] i for each tested condition are summarized in Fig. 4D. Consistent with previous reports (36), the mean resting [Ca 2ϩ ] i in the uninfected myotubes was 117 Ϯ 1.45 nM, and 116 Ϯ 12 in LacZexpressing cells, which served as a negative control. However, overexpression of either Tau, A␤, and especially A␤ proteins in these cells resulted in a significant increase in the resting [Ca 2ϩ ] i , 306 Ϯ 17.4 nM, 281 Ϯ 10 nM and 405 Ϯ 12.7 nM, respectively (Fig. 4D). Moreover, these results were specific to the transgenes expressed and not to the viral delivery system employed since infection with a tetracycline-regulated adenovirus, which encodes the same A␤42 (37) resulted in the same increase in resting [Ca 2ϩ ] i , in the presence of doxycycline as did the HSV-A␤42 vector (results not shown). To determine if the increase in resting [Ca 2ϩ ] i was dependent on extracellular Ca 2ϩ , control, and A␤-expressing cells were bathed in Ca 2ϩdepleted Ringer solution. The removal of Ca 2ϩ was accomplished by replacement with same molar amount of Mg 2ϩ and addition of 1 mM EGTA. One hour incubation of A␤-expressing cells in Ca 2ϩ -free medium resulted in similar alterations in both, the membrane potential (53.2 Ϯ 0.45mV) and elevation in resting [Ca 2ϩ ] i (404 Ϯ 23 nM). Fig. 4 it could be inferred that overexpression of Tau, A␤, or A␤ in myotubes lead to significant dysregulation of the Ca 2ϩ homeostasis and plasma membrane potential. Since the ability to respond to depolarization (i.e. the initial step in excitation-contraction coupling) is one of the key functional features of skeletal muscle, it is important to determine whether this function is preserved to the same extent in cells expressing either of these proteins. To address this question, PM of the myotubes was depolarized by perfusion with solutions containing varied [KCl] o (see "Materials and Methods"). As described previously, the degree of PM depolarization is proportional to changes in [K ϩ ] o (38) . Fig. 5, A and B presents cellular responses to 80 mM [KCl] o , which is expected to generate a near maximal detectable release of calcium from its stores. Fig. 5A shows representative time courses of RyR-mediated Ca 2ϩ release from an uninfected cell and from a cell that was expressing A␤. The relative amount of released Ca 2ϩ (in arbitrary units of fluorescence, a.u.), as calculated by integrating the total area of the Ca 2ϩ transients, was not significantly different (p Ͼ 0.05) in (13.8 Ϯ 1.8), A␤ (16.8 Ϯ 2.6), or A␤ (17.5 Ϯ 2.2) expressing myotubes when compared with untreated (16.3 Ϯ 2.2) or LacZ-expressing cells (15.0 Ϯ 3.1) (Fig.  5B, data from three independent experiments). Additionally, there were no significant changes in Ca 2ϩ release between uninfected and A␤-expressing cells at each of the lower tested [KCl] o values (Mann Whitney test, p Ͼ 0.05) (Fig. 5C). It has to be noted that upon withdrawal of the stimulant, the fluorescence returned back to its baseline in each tested condition, as seen in Fig. 5A, indicating that the response was highly voltagedependent and that Ca 2ϩ removal apparatus was still functional. Furthermore, during the sequential stimulations, the magnitude of the transients remained unaltered and the same baselines were achieved in all tested conditions (not shown), attesting to the efficiency of the Ca 2ϩ reuptake.

Tau, A␤, and A␤ Do Not Affect Depolarization-induced Ca 2ϩ Release-From
Effects of Tau, A␤, and A␤ on Caffeine Sensitivity-It has been previously proposed that intracellular Ca 2ϩ is involved in a process known as CICR. (For detailed review of skeletal muscle EC-coupling see Ref. 39). Since the resting [Ca 2ϩ ] i is increased when the above mentioned peptides are overexpressed, there is a distinct possibility that this increase in Ca 2ϩ leads to enhancement of the CICR in the transgene-bearing cells. To investigate whether the increase in resting [Ca 2ϩ ] i is accompanied by heightened CICR, we conducted measurements of intracellular Ca 2ϩ fluorescence in response to increasing concentrations of the agonist caffeine (40,41) (Fig. 6, A and  B). Each caffeine solution was perfused over the cells for 10 s and rapidly washed out of the well for 60 s to assure sufficient removal of caffeine and to allow for the normal reuptake of Ca 2ϩ into SR as judged by the complete return to baseline. The fitted and normalized sigmoidal dose response curves from uninfected and A␤-infected cells are shown in Fig. 6A. For technical reasons these experiments required a much larger number of cells to be plated into each well than the depolarization experiments, making it difficult to unequivocally identify whether the tested cells expressed the transgenes. As a result, a direct comparison of the dose-response curves was not attempted, because the EC 50 values obtained from such curves would incorporate a significant number of uninfected cells. Instead, we chose to determine EC 50 values for caffeine in each tested cell. The assumption for this approach was that the EC 50 values that were significantly different then the lowest or highest values obtained in the uninfected cells would likely represent the values from the cells expressing the transgenes of interest. Fig. 6, B and C present the mean and the distribution of individual EC 50 values obtained from each tested condition, respectively. Fig. 6C provides a visual representation of distribution within each population of cells, confirming a non-normal distribution in cells exposed to viral-encoded transgenes as compared with uninfected cells. The Kruskal-Wallis nonparametric statistical evaluation demonstrated that the mean EC 50 values from the A␤ and to a lesser degree A␤expressing cells, were significantly different (p Ͻ 0.05) from the values obtained from the uninfected cells. Although, statistical analysis did not reveal a significant difference between control and Tau-alone expressing cells (p Ͼ 0.05), the skewed distribution of EC 50 values in the presence of Tau resembled those in the presence of A␤ and A␤, thus suggesting the presence of a population of cells exhibiting a higher sensitivity to caffeine, a population that was not present in the uninfected cells. From these experiments it can be concluded that expression of A␤ transgene alone is sufficient to make myotubes more susceptible to caffeine-induced Ca 2ϩ release.

DISCUSSION
The goal of the present study was to investigate the additive influence of human A␤ and Tau proteins on calcium homeostasis, when these peptides are expressed heterologously in skeletal muscle cell culture. Using murine myotubes, which lack detectible endogenous Tau, we have shown that in the presence of A␤ 42 , exogenously expressed human wild-type Tau is hyperphosphorylated at several AD-relevant epitopes. We also show for the first time that intracellular coexpression of A␤ 42 and fetal Tau isoform additively induced the depolarization of the plasma membrane, elevated the level of resting free myoplasmic calcium and enhanced susceptibility of the RyR to Ca 2ϩinduced Ca 2ϩ release.
Rationale for Studying AD Pathogenesis through an IBM Model-This investigation was inspired by human Inclusion Body Myositis, a skeletal muscle condition that is biochemically analogous to Alzheimer's disease. Since skeletal muscle and neurons share several functional characteristics, while acknowledging their obvious differences, the findings of this study may also be relevant to calcium dysregulation in neurons under combined cellular A␤ and Tau stress.
The 42-amino acid version of A␤ is particularly relevant to the current study of intracellular processes, over shorter forms, particularly A␤ 40 , since the latter is preferentially secreted and less prone to nucleation (42). The intraneuronal expression of A␤ 42 in transgenic mice has been shown to be synaptotoxic and to precede extracellular plaque formation (43)(44)(45)(46). Additionally, synthetic A␤ 42 species were found to be especially toxic to long term potentiation (LTP) in the rat hippocampus (47,48). It was previously demonstrated that only a relatively small amount of viral-encoded A␤ 42 peptide is actually secreted whereas the large majority of it is retained, in both, the soluble and insoluble fractions (34), in agreement with other studies (49). Consistent with inclusion formation in IBM, long incubation times with HSV-A␤ 42 eventually lead to formation of intracellular deposits of A␤ and myotube attrition (26). However, the current study was designed to temporally reflect the preinclusion stages of the disorder; therefore, our findings are unlikely to result from the formation of amyloid inclusions or cell death.
In AD as well as in IBM, insoluble aggregates of abnormally hyperphosphorylated Tau and neurofilaments have the ultra-

FIG. 4. Plasma membrane potential and resting intracellular calcium concentration [Ca 2؉ ] i in AD protein-bearing myotubes.
Double-barreled microelectrode recordings in C 2 C 12 myotubes continuously infected (9 h) with HSV particles coding for either A␤ or Tau () singly or concurrently (A␤). Controls are no virus (NV) or HSV-LacZ additions. A, representative membrane potential (V m ) records from uninfected and A␤-infected cells. B, summary of mean (ϮS.E.) V m for each tested condition. For each point, the records from 10 -36 cells, over three separate experiments were averaged. Compared with single transgene infections or no virus/LacZ conditions, double A␤-expressing myotubes were significantly depolarized at rest (p Ͻ 0.05 and Ͻ0.005, respectively). C, representative Ca 2ϩ potential (V Ca ) records from uninfected and A␤-infected cells from the cells described in A. D, similar to panel B, the corresponding intracellular calcium concentration calculated from V Ca was significantly higher in double-expressing cells (p Ͻ 0.05 versus single infections). All data are presented as mean Ϯ S.E. structural appearance of paired helical filaments (PHF). Studies with animal models expressing the same human fetal Tau isoform used here, have shown the development of insoluble hyperphosphorylated Tau protein, a reduction in the number of microtubules, and diminished axonal transport (30). These characteristics recapitulate the properties of PHF-Tau as found in the human tauopathies, but not those of the other Tau isoforms of the normal adult. For instance, the fetal Tau isoform, used in these studies, shares many of the same phosphorylation sites with the PHF form (50), of which most were proline-directed (51). It also appears to be phosphorylated to a greater degree than the longer 4-repeat adult Tau isoforms (52). Additionally, similar to PHF-Tau, fetal Tau was found to have a greater intrinsic reduction in microtubule binding and stabilization in vitro (53)(54)(55). Here we chose to examine the phosphorylation status of 4 motifs, each of which is also affected in PHF-Tau formation. These motifs are either positioned in the flanking targeting areas (Thr 231 and Ser 396 / Ser 404 ; NЈ-and CЈ to the repeat domains, respectively) or in a non-proline-directed motif within repeat no. 1 (Ser 262 ) (51,56). Importantly, the phosphorylations at Ser 396 , Ser 404 , Thr 231 , and Ser 262 shown here, also factored highly in mice made transgenic for wild-type human fetal (30).
Our immunocytological and immunoblot data suggest cooperation between cellular A␤ and Tau. We find it plausible that the phosphorylation status and possibly aggregation state of Tau, is strongly linked with calcium homeostasis. We also show that intracellular A␤ expression in a cell culture model may incite Tau phosphorylation. Since the virus dosages and transgene expression levels were submaximal and neither transgene altered the expression level of the other, the effects of their combination on calcium homeostasis are cooperative and not duplicative. Numerous other cell culture studies have shown that extracellular application of synthetic A␤ fibrils can push the balance in favor of hyperphosphorylation (57)(58)(59). Additionally, two recent studies using transgenic mice with the FTD-(P301L) Tau mutation had more than the expected burden of PHF neurofibrillary tangles and phosphorylation of Tau in the presence of secreted or injected A␤ (60,61).
We now show an association of AD-related proteins with Ca 2ϩ dyshomeostasis in myotubes. Whether the calcium load is actually required for these phosphorylations to proceed or vice versa, or whether A␤ directly or indirectly stimulates the responsible kinases remains to be determined (34,62). For instance, an increase in [Ca 2ϩ ] i through leak and/or influx is one mechanism by which Tau kinases could become activated and lead to NFT changes in AD and IBM (8, 10) (63). Specifically, elevated resting calcium and A␤42 may activate calpain (64) and lead to Tau hyperphosphorylation through constitutive stimulation of cdk5 A␤ 42 (65)(66)(67). In our skeletal muscle system, the presence of Tau clearly aggravated the A␤-induced mismetabolism of calcium. Further experimentation is required to answer whether relative degrees of non-phosphorylated versus phosphorylated Tau, independent of A␤, have functional effects on calcium homeostasis.
Effects of AD-associated Proteins on Ca 2ϩ Homeostasis-The most significant finding reported here is that the expression of either of the AD-associated peptides leads to a large elevation in resting [Ca 2ϩ ] i , and a significant depolarization of the plasma membrane (Fig. 4). It also has to be noted that coexpression of Tau and A␤ proteins exhibits an additive effect with respect to the magnitude of change in resting [Ca 2ϩ ] i (288.3 nM) and the magnitude of the membrane depolarization (ϳ15 mV). The intrinsic and additive importance of Tau in clinical and cytological toxicity is underscored by several AD animal models where the interaction of A␤ and mutant Tau was investigated using single, double, and triple transgenic approaches (60,61,68,69).
In contrast to the AD transgene-expressing cells which exhibited changes in the V m , the depolarization-induced responses appeared to follow the same pattern under all conditions. Voltage-dependent Ca 2ϩ release in skeletal muscle involves activation of the voltage sensors in the T-tubules of the PM, which in turn activate Ca 2ϩ release via a direct mechanical interaction with the RyRs of the SR (39). In the experiment described in Fig. 5, voltage-dependent release was achieved by increasing the extracellular concentration of K ϩ , thereby clamping the plasma membrane at a specific voltage. Although, it has been hypothesized that CICR plays a part during the voltage activation of the RyR, the specifics of this contribution are currently unresolved (70). The current study did not at-tempt to address this subject with respect to the amyloid and Tau proteins. Instead, we sought to determine if the depolarization-induced Ca 2ϩ release was still operative in a voltagedependent manner. Experiments presented in Fig. 5 show that expression of A␤ and Tau does not perturb the voltage dependence of Ca 2ϩ release and suggests that the threshold for this mechanism of activation is likely unaltered.
On the other hand, the elevations in intracellular Ca 2ϩ described above are expected to enhance the ability of caffeine to induce RyR-mediated Ca 2ϩ release (40,41). As expected, expression of either Tau, A␤, or A␤ have resulted in a significant reduction EC 50 for activation of Ca 2ϩ release by caffeine. Although, the direct mechanism by which the resting Ca 2ϩ was elevated remains to be determined, the reduction of the EC 50 can be explained by the fact that increased local [Ca 2ϩ ] rendered the RyRs more susceptible to CICR and therefore more readily activated by caffeine (71). Alternatively, RyRs have been shown to be modulated by numerous proteins and peptides (72)(73)(74). It is therefore also possible that each of the tested peptides could interact directly with the RyRs and modulate their function in a way that could render them leaky in the resting state or as demonstrated where A␤ alone is apparently sufficient, more susceptible to activation by caffeine.
Resting myoplasmic free calcium levels are normally maintained by a complex equilibrium between the activities of calcium release channels, SR-ATPase reuptake pumps, Ca 2ϩ /Na ϩ exchangers, various intracellular calcium buffers and surface membrane channels. Although, the source of the increased Ca 2ϩ remains to be identified, there are several intracellular components that are likely to play a significant role in the establishment of this phenomenon. Direct inhibition of re-uptake of Ca 2ϩ through SR Ca 2ϩ ATPase (SERCA) by A␤/ could be expected to raise intracellular [Ca 2ϩ ] or at least to prolong the time to baseline after stimulation (75). As demonstrated by these results, depolarization (Fig. 5A) and caffeine-elicited (not shown) Ca 2ϩ transients consistently returned back to the baseline at the end of each treatment, thereby suggesting that the Ca 2ϩ removal and reuptake apparatus was relatively unaltered. However, since this study did not address the effects of the AD peptides on the individual components of the Ca 2ϩ removal, we cannot state with any certainty that their function was not altered. Overloaded SR Ca 2ϩ stores have been described to induce Ca 2ϩ leaks and to increase the magnitude of Ca 2ϩ release (76). Although Fluo-4 is not an ideal fluorophore for the quantitative determination of the amounts of the released Ca 2ϩ , it can serve a good indicator of any gross changes that might arise. Current results show that in response to 40 mM caffeine, a concentration in which maximal detectable Ca 2ϩ release can be achieved, the magnitude of the Ca 2ϩ transients were essentially the same in all tested conditions (control: 17.2 Ϯ 2.3, : 12.7 Ϯ 1.7, A␤: 16.3 Ϯ 1.7, and A␤: 15.9 Ϯ 1.3). Taken together, these results argue against the possibility of SR Ca 2ϩ store overload.
Since A␤ has been demonstrated to be localized to the myoplasm, associated with the ER membrane (26,37,49), it is tempting to speculate that the elevation of resting Ca 2ϩ in A␤-expressing cells is produced by a direct leak involving the SR through A␤-formed pores (10) or other damage to membrane bilayers. Our data using calcium-depleted media makes a leaky influx at the level of the PM less attractive an hypothesis to explain elevated resting calcium levels. Alternatively, A␤ could affect the functional state of plasma membrane Ca 2ϩ channels, such as the store operated CCE channels and L-type Ca 2ϩ channels (77).
In conclusion, results of the current study provide a framework for further investigations into the cooperation between A␤ and Tau, as they relate to the disruption of Ca 2ϩ -dependent cellular functions in Alzheimer's Disease and inclusion body myositis.