Anthracyclines Induce Calpain-dependent Titin Proteolysis and Necrosis in Cardiomyocytes*

Titin, the largest myofilament protein, serves as a template for sarcomere assembly and acts as a molecular spring to contribute to diastolic function. Titin is known to be extremely susceptible to calcium-dependent protease degradation in vitro. We hypothesized that titin degradation is an early event in doxorubicin-induced cardiac injury and that titin degradation occurs by activation of the calcium-dependent proteases, the calpains. Treatment of cultured adult rat cardiomyocytes with 1 or 3 μmol/liter doxorubicin for 24 h resulted in degradation of titin in myocyte lysates, which was confirmed by a reduction in immunostaining of an antibody to the spring-like (PEVK) domain of titin at the I-band of the sarcomere. The elastic domain of titin appears to be most susceptible to proteolysis because co-immunostaining with an antibody to titin at the M-line was preserved, suggesting targeted proteolysis of the spring-like domain of titin. Doxorubicin treatment for 1 h resulted in ∼3-fold increase in calpain activity, which remained elevated at 48 h. Co-treatment with calpain inhibitors resulted in preservation of titin, reduction in myofibrillar disarray, and attenuation of cardiomyocyte necrosis but not apoptosis. Co-treatment with a caspase inhibitor did not prevent the degradation of titin, which precludes caspase-3 as an early mechanism of titin proteolysis. We conclude that calpain activation is an early event after doxorubicin treatment in cardiomyocytes and appears to target the degradation of titin. Proteolysis of the spring-like domain of titin may predispose cardiomyocytes to diastolic dysfunction, myofilament instability, and cell death by necrosis.

Titin is the largest known protein and is an integral part of the myofilament system in vertebrate striated muscle. Full-length titin extends the half-sarcomere from the Z-disk to the M-line, with the I-band region comprised of distinct elastic segments that account for the extensibility of titin. Titin has been implicated in the process of sarcomerogenesis, where it is thought to serve as a framework for ordered assembly of myofilament proteins (1)(2)(3). The contribution of the elastic domain of titin to passive and restoring forces is well established (4 -6), and a physiological role is emerging where titin is an important determinant of the Frank-Starling relationship of the heart (7)(8)(9). Given the central role of titin in sarcomere formation as well as myocyte mechanics, pathological conditions leading to enhanced titin degradation may have profound consequences for myofilament stability and function. Indeed, there is evidence for disorganization and loss of titin in ischemic and failing human hearts (10 -12). The mechanisms, however, for cardiac titin degradation under normal and pathological conditions have not been examined.
Titin is known to be extremely sensitive to proteolysis, in vitro, which is preventable with calcium chelation in combination with leupeptin protease inhibition, conditions known to be necessary for inhibition of the calcium-dependent proteases, the calpains (13). The calpains are a family of calcium-dependent cysteine proteases that are believed to participate in basic calcium-mediated intracellular processes (14). Of the calpain family, calpain isoenzymes I (-calpain) and II (m-calpain) are ubiquitously expressed and are activated in vitro in the presence of micro-and millimolar concentrations of calcium, respectively. Although the exact physiological role of calpains in the myocardium is not known, the ability of calpains to cleave cytoskeletal and myofilament proteins desmin, fodrin, filamin, C-protein, tropomyosin, troponin T, troponin I, nebulin, gelsolin, and vinculin in a variety of cell types, in vitro, suggest a regulatory role for calpains in remodeling of the myofibril (15)(16)(17). Calpain activity is increased in a wide variety of pathological conditions associated with calcium overload including Alzheimer's disease (18,19), cataracts (20), oxidative stress (21), and ischemia reperfusion injury (22). In post-ischemic myocardium, the proteolytic activity of calpain may be linked to degradation of sarcomeric proteins troponin I and desmin (17,23). Interestingly, skeletal muscle-specific calpain-III (also known as p94) has been shown to bind to titin in skeletal muscle, and mutations in p94 have been associated with a severe form of limb-girdle muscular dystrophy (24 -26). The association between p94 and titin is of particular interest to researchers in food science, where the level of calpain activity and the degree of titin degradation have been related to postmortem meat tenderness (27,28). The extreme sensitivity of titin to calciumdependent proteolysis as well as the known association between titin and p94 in skeletal muscle make titin a plausible target of the calpains in cardiac muscle.
Doxorubicin cardiomyopathy is typically associated with myofibrillar deterioration; however, the early mechanisms responsible for this degenerative phenotype remain elusive (for review see Ref. 29). In addition to myofilament damage, doxoru-bicin is also known to cause intracellular calcium overload (29). These two aspects of doxorubicin cardiotoxicity may well be related, where a pathological increase in intracellular calcium may trigger indiscriminate activation of calcium-dependent proteases resulting in degradation of key myofibrillar proteins. We therefore hypothesized that titin degradation is an early process in doxorubicin-induced myofilament injury, with consequences for myofilament stability and perhaps cell viability. We further hypothesized that titin proteolysis occurs by the action of the calciumdependent proteases, the calpains. To test the prospective role for calpains in pathologic titin degradation, we used a cell culture model of anthracycline-induced myofilament injury. Our results support a role for calpain in titin proteolysis early after doxorubicin treatment. The implications of these findings for doxorubicin and other forms of myocardial injury are discussed.
Neonatal rat ventricular myocytes were isolated as described previously (31). Cells were cultured for 48 h in Dulbecco's modified Eagle's medium supplemented with 7% fetal calf serum and penicillin/streptomycin prior to treatment. Neonatal myocytes were treated with 0.5 mol/liter doxorubicin without or with protease inhibitors (ALLN or BDFMK) and incubated for 24 h.
Gel Electrophoresis-Following treatment, myocytes were solubilized by adding 1 volume of cell suspension to 9 volumes of solubilization buffer (50 mmol/liter Tris-HCl, 5% SDS, 10% glycerol, 80 mmol/liter dithiothreitol, pH 6.8 at 25°C) preheated to 90 -95°C. The samples were solubilized for 60 s and allowed to cool down to room temperature. Protein concentration was determined by the Lowry method. Bromphenol blue was added (0.025%), and equal amounts of sample (200 g) were analyzed by SDS-PAGE using a 2% acrylamide gel strengthened with 0.5% agarose (32). Gels were run at 15 mA for 4 h at room temperature and fixed in 50% methanol and 7% acetic acid. Gels were stained with Coomassie overnight, destained with 50% methanol, 7% acetic acid followed by 5% methanol, 7% acetic acid. Wet gels were scanned at 400 or 900 dpi using a commercially available scanner (Molecular Analyst, Bio-Rad). T1 titin (both N2B and N2BA isoforms), titin degradation product T2, total titin (T1 isoforms ϩ T2), and myosin heavy chain from the same samples were quantitated by densitometry. Total titin, T1, and T2 were normalized relative to myosin heavy chain to account for potential inaccuracies in protein loading.
Calpain Assay-Calpain activity was determined using a fluorogenic assay as described by Sasaki et al. (33). At the end of the experimental protocol, myocytes were harvested and sonicated in calcium-free assay buffer containing 63.2 mmol/liter imidazole-HCl, 20 mmol/liter EGTA, 25 mmol/liter EDTA, and 10 mmol/liter 2-mercaptoethanol, pH 7.3. Samples were diluted to 1 mg/ml, and 25 l of sample was combined with either 75 l of calcium-free assay buffer or 75 l of calcium buffer (containing 63.2 mmol/liter imidazole-HCl, 1.25 mmol/liter CaCl 2 , and FIG. 1. Effect of doxorubicin on titin integrity. Titin degradation was assessed in Coomassie-stained 2% polyacrylamide gels. A, control rat myocytes contained T1 titin isoforms, with a predominant N2B titin band and a minor N2BA band. Incubation of myocytes with 1 mol/liter doxorubicin for 24 h resulted in partial degradation of titin as shown by the appearance of titin degradation product T2 (arrowhead), and almost complete disappearance of titin following 48 h of incubation (B). A-C, pretreatment of myocytes with calpain inhibitors ALLN (100 mol/liter), calpeptin (CP, 5 mol/liter), or leupeptin (LP, 50 mol/liter) for 1 h prior to 1 or 3 mol/liter doxorubicin (Doxo) treatment for 24 h preserved titin similar to control levels. Pretreatment with a broad spectrum caspase inhibitor BDFMK (100 mol/ liter), on the other hand, did not prevent the doxorubicin-induced titin degradation. Results are from four to five independent experiments.
10 mmol/liter 2-mercaptoethanol, pH 7.3) in a 96-well fluorescent plate. Samples were incubated at 37°C for 10 min, and N-succinyl-Leu-Leu-Val-Tyr-7-amido-4-methylcoumarin was added to the samples for a final concentration of 50 mol/liter. After an additional 30 min of incubation, fluorescence was detected at 355 nm excitation and 460 nm emission. Calpain activity was determined as the difference between the calcium-dependent and the calcium-independent fluorescence. All experiments were run in duplicate.
Caspase-3 Assay-The caspase-3 activity within the cells was measured with a caspase-3 fluorescent assay kit (Molecular Probes). Briefly, the caspase-3-specific substrate acetyl-Asp-Glu-Val-Asp-amido-4-methylcoumarin was incubated with 20 g of myocyte lysates at room temperature for 30 min, and fluorescence of the cleaved product was detected at 355 nm excitation and 460 nm emission. All experiments were run in duplicate. The specificity of the cleavage reaction was verified by including 10 mol/liter acetyl-Asp-Glu-Val-Asp-CHO (caspase-3 inhibitor) in the incubation, and Ͻ10% of activity was detected.
Fura-2 Measurement-Calcium transients were measured using the fluorescent calcium indicator fura-2, as described previously (34). Briefly, redifferentiated cardiomyocytes treated with 1 mol/liter doxorubicin at the indicated time points were incubated for 15 min in Tyrode's buffer (in mmol/liter: NaCl 137, KCl 5.4, CaCl 2 1.2, MgCl 2 0.5, HEPES 10, and glucose 10, pH 7.4, 37°C) containing 1 mol/liter membrane-permeant fura-2/AM (Molecular Probes) and 500 mol/liter probenicid (Sigma) to prevent leakage of fura-2 from cells. Fura-2 loaded cardiomyocytes were perfused in a cell chamber with Tyrode's buffer maintained at 37°C and field-stimulated at 3 Hz. Calcium transients were measured by dual excitation (360 and 380 nm) of the fura-2-loaded cells, and emission fluorescence was recorded at 510 nm (IonOptix Inc.). The fluorescence ratio (F 360 /F 380 ) and time constant () of the calcium transient decay reflect the intracellular calcium concentration, and the rate of cytosolic calcium removal, respectively. The area under the calcium transient reflects the total amount of cytosolic calcium throughout the contractile cycle.
Immunofluorescence Microscopy-Cell cultures were fixed in 4% paraformaldehyde and incubated with primary antibodies to ␣-actinin, myomesin (clone B4, a kind gift from Dr. Jean-Claude Perriard), titin PEVK domain (9D10; developed by Dr. Marion Greaser at the Development Studies Hybridoma Bank at the University of Iowa), and titin M-line domain (M8; courtesy of Dr. Mathias Gautel). Myocytes were subsequently incubated with rhodamine-phalloidin or fluorescein isothiocyanate-phalloidin for staining of filamentous actin, and secondary antibodies were conjugated to fluorescein isothiocyanate, TRITC, or cyan-5 (Molecular Probes). Stained preparations were analyzed with a Leica confocal scanner TCS NT on an inverted microscope Leica DMIRB-E. Myofibrillar disarray was assessed by an investigator blinded to treatments using a Nikon Diaphot 300 fluorescence microscope equipped with a 40ϫ oil immersion objective. A total of 100 -150 myocytes were counted for each experimental condition in each experiment.
Trypan Blue Staining and Creatine Kinase Assay-Following treat-ment, media were collected for each experiment and frozen at Ϫ80°C for determination of creatine kinase activity. Myocytes from the same experiments were subsequently stained with 0.4% trypan blue as described previously (35). Creatine kinase activity in the media samples was determined from a diagnostic kit according to manufacturer's instructions (CK-10, Sigma). Time-lapse Video Microscopy-Cardiomyocytes were maintained in an incubation and perfusion system with CO 2 and humidity controller on a heated microscope stage (Carl Zeiss Co.) during acquisition of video time-lapse images. Cardiomyocytes were exposed to doxorubicin for 48 h, and images were acquired every 15 min using an automated shutter and video camera (Kappa Opto-Electronics Inc.). Following the video time-lapse, the same area was fixed in 4% paraformaldehyde and immunostained for myomesin and actin for confocal microscopy.
Flow Cytometry-Apoptotic neonatal myocytes were quantified by flow cytometric analysis as previously described (31). Briefly, following treatment both detached and trypsinized adherent cells were pelleted, resuspended in PBS (Invitrogen), fixed in 70% ethanol/PBS, and stored at Ϫ20°C until use. Cells were rinsed once in PBS and resuspended in FIG. 2. The effect of doxorubicin on calpain and caspase activity in adult rat ventricular myocytes. A, treatment with 1 mol/ liter doxorubicin resulted in an ϳ3-fold increase in calpain activity, which occurred as early as 1 h and remained elevated at 48 h. B, caspase-3 activity was significantly increased at 24 and 48 h. Pretreatment of myocytes with 100 mol/liter ALLN prevented the doxorubicininduced increase in calpain activity at 24 h (A) but had no effect on caspase-3 activity (B). Conversely, pretreatment with 100 mol/liter BDFMK prevented the increase in caspase-3 activity at 24 h (B) with no effect on calpain activity (A). *, p Ͻ 0.05 versus 0 h; †, p Ͻ 0.05 versus 24 h. Shown are mean Ϯ S.E. from four to seven independent experiments. PBS containing 20 g/ml propidium iodide (Sigma) and 0.1 mg/ml RNase A (5 kilounits/ml, Sigma). After a 2-h incubation period at room temperature, 10,000 cells from each sample were counted by FACScan (BD Biosciences). Gating was performed to exclude small debris with Ͼ2 logs weaker propidium iodide fluorescence than G 0 cells. Apoptosis was calculated as the percent of cells in the sub-G 1 /G 0 peak. Statistics-Data are presented as mean Ϯ S.E., unless indicated otherwise. Where appropriate, results were analyzed by Student's t test or analysis of variance with a post-hoc test of least significant differences. p Ͻ 0.05 was considered statistically significant.

Doxorubicin Degrades Titin in Intact Cardiomyocytes-We
examined the effect of doxorubicin on titin using 2% Coomassie-stained SDS-polyacrylamide gels (Fig. 1). As reported previously (36), control rat myocytes express a predominant N2B titin band with a minor N2BA band compared with human and pig myocardium (see calibration gel). Incubation of myocytes with 1 mol/liter doxorubicin resulted in partial degradation of titin, as evidenced by the appearance of titin degradation product T2 (Fig. 1, A and B). Table I shows that compared with the untreated control group, the T2/MHC ratio increased with doxorubicin incubation time and that pretreatment with calpain inhibitor ALLN, but not caspase inhibitor BDFMK, prevented the increase in the T2/MHC ratio. Interestingly, there was no difference in the total titin (T1 isoforms ϩ T2)/MHC ratio, suggesting a very specific and limited degradation of titin to its T2 fragment at the time points studied (Table I). Higher concentrations of doxorubicin and longer incubation times, on the other hand, resulted in almost complete degradation of titin ( Fig. 1, B and C). Pretreatment of myocytes with calpain inhibitors ALLN (100 mol/liter), calpeptin (5 mol/liter), or leupeptin (50 mol/liter) for 1 h prior to doxorubicin treatment preserved titin similar to control levels, whereas caspase inhibitor BDFMK (100 mol/liter) had no effect (Fig. 1, A-C).
Western blot analysis and wet gel scans were performed to determine the effect of doxorubicin on other sarcomeric proteins. Treatment with 1 mol/liter doxorubicin up to 24 h had no effect on protein levels of ␣-actinin, actin, desmin, or myosin heavy chain (Table I). Thus, doxorubicin treatment of myocytes results in early degradation of titin, which is preventable with calpain but not caspase inhibition.
Doxorubicin Activates the Calpains and Caspase-3-The effect of doxorubicin on protease activity was examined in adult rat ventricular myocytes in culture at 1, 6, 24, and 48 h. Treatment with 1 mol/liter doxorubicin resulted in a significant increase in calpain activity, which occurred as early as 1 h and remained elevated at 48 h ( Fig. 2A). The increase in calpain activity was similar to that in cardiomyocytes treated with the calcium ionophore, ionomycin (data not shown). Furthermore, the doxorubicin-induced increase in calpain activity coincided with significant prolongation of the calcium transient and increase in total cytosolic calcium levels, assessed by time constant of the fura-2 transient decay and the area under the fura-2 transient curve, respectively (Table II). Caspase-3 activ-ity, on the other hand, did not increase until 24 h of doxorubicin treatment (Fig. 2B). Pretreatment of myocytes with 100 mol/ liter ALLN prevented the doxorubicin-induced increase in calpain activity at 24 h but had no effect on caspase-3 activity. Conversely, pretreatment with 100 mol/liter BDFMK prevented the increase in caspase-3 activity at 24 h but did not inhibit calpain activation.
Calpain-Calpastatin Ratio Is Increased with Doxorubicin-We examined the effect of 24 and 48 h of doxorubicin (1 mol/liter) treatment on protein levels of calpains-I, -II, and calpastatin. By Western blot, calpain-I appeared as a single 80-kDa band that was significantly decreased at 24 and 48 h of doxorubicin (Fig. 3A). Doxorubicin treatment for 24 h resulted in significant autolysis of the 80-kDa calpain-II band to a higher mobility 78-kDa band, which is widely believed to be the more active form of calpain-II (for review see Ref. 37). Calpastatin was significantly decreased at 24 and 48 h of doxorubicin treatment (Fig. 3A). Because the proteolytic activity of the calpains is negatively regulated by calpastatin, the calpain/ calpastatin ratio provides a gauge to assess the proteolytic potential of the calpains (38,39). The ratio of calpain-II (combined 80-and 78-kDa products) to calpastatin was significantly elevated and the increase occurred earlier in comparison to the calpain-I/calpastatin ratio (Fig. 3B).
Calpain Degrades Titin in Skinned Cardiomyocytes-To determine whether calpain can directly degrade titin, we incubated Triton-permeabilized myocytes with 10 units/ml of calpain-I (Calbiochem) with or without calpain inhibitor for 30 min. The skinning procedure resulted in some degradation of titin as shown by the T2 band in the control lane (Fig. 4A). Incubation with calpain-I resulted in a more prominent T2 band compared with control, and co-treatment with calpain inhibitor preserved titin to control levels. This effect of calpain on titin integrity was also visualized directly in skinned myocytes immunostained with 9D10, a monoclonal antibody to I-band titin at the PEVK domain (40). Immunofluorescence microscopy revealed that 9D10 immunostaining was reduced in skinned myocytes treated with calpain compared with control myocytes, and this effect was prevented by calpain inhibition (Fig. 4B).
Calpain Inhibition Prevents Doxorubicin-induced Myofibrillar Disarray-We examined the pattern of titin staining with 9D10 in myocytes treated with doxorubicin. High resolution images of myocytes stained with antibody 9D10 demonstrated a doublet pattern consistent with staining of the I-bands on either side of the Z-line (Fig. 5A). Myocytes treated with doxorubicin showed an overall decrease in 9D10 staining, along with loss of the doublet pattern (Fig. 5B). Concurrent treatment with a calpain inhibitor preserved the immunostaining of the PEVK domain of titin (Fig. 5C). To examine further the effect of doxorubicin on titin structure, we co-immunostained 9D10 with M8, an antibody to the M-line of titin (Fig. 6A). Most interesting, we were able to observe loss of 9D10 staining with rela- tively preserved M8 staining in some myocytes treated with doxorubicin (Fig. 6B). This pattern was heterogeneous with some myocytes showing preservation of both 9D10 and M8 staining. The effect of doxorubicin on myofilament structure was assessed by evaluating the immunostaining pattern of filamentous actin and myomesin or ␣-actinin. Consistent with our previous report (30), myocytes treated for 24 h with 1 mol/liter doxorubicin show significant myofilament disarray when compared with nontreated cells (Fig. 7, A, B, and D). Pretreatment with 100 mol/liter ALLN or 4 mol/liter calpastatin (the biological inhibitor of the calpains) significantly reduced the doxorubicin-induced myofibrillar disarray (Fig. 7, C and D).
Effect of Doxorubicin on Cell Death-Anthracyclines have been shown to induce dose-dependent apoptosis and necrosis (31), and both caspases and calpains have been implicated in cell death. We therefore examined the putative roles of the caspases and calpains in doxorubicin-mediated apoptotic and necrotic cell death. Despite a nearly 4-fold increase in caspase-3 activity at 24 and 48 h of doxorubicin exposure, we did not see a significant increase in apoptotic cell death by terminal dUTP nick-end labeling at 48 h in our long term adult cardiomyocyte culture system (control ϭ 1.7 Ϯ 0.4% versus doxorubicin ϭ 2.2 Ϯ 0.6%; p ϭ not significant). We did observe cell loss, with an increase in trypan blue uptake and CK release following doxorubicin exposure suggesting loss of cell membrane integrity, which is indicative of necrotic cell death. Doxo-rubicin exposure (1 mol/liter) caused a time-dependent increase in the percentage of trypan blue-positive cells from a base-line value of ϳ8 to 12% following 48 h of doxorubicin (Fig.  8A). The loss of myocytes is clearly illustrated in Fig. 8C showing time-lapse images of cardiomyocytes, exposed to doxorubicin for 48 h, undergoing changes in morphological features consistent with necrosis (also see video demonstration in Supplemental Material). Co-treatment with calpain inhibitor for 48 h (100 mol/liter ALLN) completely prevented the increase in doxorubicin-induced trypan blue staining, although caspase inhibition (100 mol/liter BDFMK) had no effect (Fig. 8A). These data were closely paralleled by the CK release results, with an increase in CK activity from a control value of ϳ3 to 5 units/liter after 48 h of doxorubicin, which was prevented with calpain but not caspase inhibition (Fig. 8B).
By using a neonatal cardiomyocyte culture system, which we have shown previously to undergo apoptosis following low dose anthracycline exposure (31), we were able to observe caspasedependent apoptosis that does not involve the calpains. Neonatal cardiomyocytes exposed to 0.5 mol/liter doxorubicin for 24 h exhibited a nearly 3-fold increase in the percentage of apoptotic cells as quantified by flow cytometry (Fig. 9, A and B). Caspase inhibition (100 mol/liter BDFMK) significantly reduced the frequency of doxorubicin-induced cardiomyocyte apoptosis. Calpain inhibition (100 mol/liter ALLN), on the other hand, did not prevent the doxorubicin-induced increase in ap- optotic cell death. Neither calpain nor caspase inhibition alone, i.e. in the absence of doxorubicin, had any effect on cardiomyocyte apoptosis (data not shown).

DISCUSSION
In this report we demonstrate that cardiotoxic anthracyclines can induce titin degradation via activation of calpains, and that ultimately this contributes to a spectrum of cell injury including myofilament disorganization and necrotic cell death. Titin is an essential component of the sarcomere, making important contributions to sarcomere organization and myocyte compliance. Previous studies (10 -12) have reported disorganization and loss of titin in ischemic and failing human hearts. The present study shows for the first time that titin proteolysis is an early process in myocyte injury, in this case by doxorubicin, and that titin proteolysis is mediated by the activation of calpains. Moreover, these results implicate titin proteolysis as an initial step leading to both myofibrillar disarray and cardiomyocyte necrosis in the setting of anthracycline cardiotoxicity, and conceivably other pathological conditions associated with calcium overload.
In this study, we used the dedifferentiated adult myocyte culture as it provides a stable model for studying myofilament protein turnover (41). Treatment of dedifferentiated myocytes with doxorubicin results in myofibrillar disarray as seen by a misalignment of the myofibrils and contraction bands (30,42,43). Myofilament disarray is also seen in the intact heart of animals treated with anthracyclines, and therefore may in part be responsible for the chronic cardiotoxicity of anthracyclines (42). Our present work suggests that the mechanism for this disarray involves titin proteolysis, perhaps as an initial step. The appearance of a T2 degradation product of titin by gel electrophoresis suggests preferential cleavage of the springlike PEVK domain of titin as reported previously (5,44). The degradation of titin seen on gels was experimentally corroborated by a reduction in immunostaining of 9D10 (antibody raised against the PEVK domain of titin) in doxorubicin-treated cells. It is interesting that when we co-immunostained 9D10 with M8 (antibody to titin at the M-line), we observed a reduction in 9D10 but a normal M8 striated pattern in the FIG. 4. Calpain-induced degradation of titin. Triton-permeabilized myocytes were treated with 10 units/ml calpain-I with or without calpain inhibitor for 30 min. Titin degradation was assessed in Coomassie-stained 2% polyacrylamide gels. A, calpain-I treatment induced degradation of titin as evidenced by an increase in the T2 band (arrowhead) compared with control myocytes, which was prevented by co-treatment with 100 mol/liter ALLN. B, immunofluorescence microscopy revealed that 9D10 immunostaining was reduced in skinned myocytes treated with calpain-I and that co-treatment with ALLN was able to preserve 9D10 similar to control levels. same myocyte, suggesting targeted proteolysis of the elastic domain of titin. Although doxorubicin treatment was clearly associated with myofibrillar disarray marked by disorganized immunostaining of myomesin, ␣-actinin, and filamentous actin, we were unable to detect any changes in the levels of these proteins as well as desmin and myosin heavy chain. Furthermore, the decrease in 9D10 staining was sometimes observed in myocytes with apparently normal myofilament ultrastructure. These observations suggest that doxorubicin-induced titin fragmentation precedes degradation of other sarcomeric proteins, with adverse consequences for the stability of the sarcomere. Titin is critical for sarcomere assembly and stability (45)(46)(47); hence degradation of titin could lead to destabilization of the myofilament as well as disrupt incorporation of newly synthesized proteins into the myofilament lattice, ultimately leading to sarcomere disarray.
Our data suggest that the mechanism by which doxorubicin induces calpain activation likely involves impaired intracellular calcium homeostasis coupled with a decrease in the expression of the biological calpain inhibitor, calpastatin. Doxorubicin is known to generate reactive oxygen species through redox cycling of the quinone moiety, leading to lipid peroxidation and membrane instability of vital organelles including the sarcoplasmic reticulum (SR) (48). In addition, doxorubicin has been reported to increase the L-type calcium current and directly bind the SR ryanodine receptors to induce calcium release from SR vesicles (49 -51). Oxidative damage to the SR and activation of the sarcolemmal L-type calcium channels or the SR ryanodine receptors would result in calcium accumulation in the cytosol, which alone may be sufficient for activation of calpains (14). Decreased calpastatin, on the other hand, could further potentiate the calpain proteolytic system. Although some reports have attributed a pathological decrease in calpastatin to proteolytic degradation by the calpains or caspases (39,(52)(53)(54), the mechanism for anthracycline-mediated calpastatin downregulation remains to be elucidated.
The observed link between titin proteolysis and calpain activation in doxorubicin-treated myocytes is similar to earlier work showing the extreme susceptibility of titin to degradation in the presence of calcium and the association of calpain-III (p94) with titin in skeletal muscle (13,24,25). As calpain-III is not expressed in the heart, presumably other members of the calpain family, notably the ubiquitous calpains-I and -II, are responsible for cardiac titin proteolysis (55). Whereas total calpain activity was elevated following doxorubicin exposure, calpain-I expression seemed paradoxically decreased and calpain-II was increased only at 24 h. Calpastatin, on the other hand, was down-regulated with doxorubicin exposure. By using the calpain/calpastatin ratio as a gauge for the proteolytic potential of the calpains, this ratio is increased for both calpains. Because both calpains-I and -II appear to target the same substrates (for review see Ref. 37), in the absence of specific calpain isoenzyme inhibitors it is difficult to determine which one of these is responsible for the degradation of titin. Collectively, however, our observations that doxorubicin decreased the expression of calpain-I, induced autolysis of calpain-II to a more proteolytically active form, and that the calpain-II/calpastatin ratio was elevated early relative to calpain-I lead us to speculate that calpain-II may be predominantly responsible for the doxorubicin-induced increase in total calpain activity. This observation is similar to what has been observed early after ischemic cardiac injury, where increases in calpain-II activity, mRNA, and protein levels were much greater compared with those in calpain-I (56,57). Interestingly, a reduction in calpain-I activity and in the rate of calpain-I autolysis has been reported in a neuroblastoma cell line treated with doxorubicin (58). Further studies are needed, however, to delineate the precise roles of the ubiquitous calpains in anthracycline-mediated cardiac injury.
Using several distinct calpain inhibitors we were able to preserve titin protein levels, PEVK immunostaining, and myofibrillar structure in doxorubicin-treated cells. Although the calpain inhibitor ALLN also inhibits the proteasome and the lysosomal cathepsins, several lines of evidence argue against these proteases as being involved in the initiating step of titin degradation. The proteasome is unable to degrade complexed myofilament proteins or intact myofibrils, presumably due to the inability of the proteasome core to accommodate multicomponent or large proteins such as titin (59). The lysosomal cathepsins are minimally activated in cardiac and skeletal muscle following acute anthracycline treatment (60,61). Our finding that calpain-specific inhibitors calpeptin and calpastatin prevent doxorubicin-induced titin degradation and sarcomere disarray implicate the calpains in the initial step of titin degradation. Because the calpains do not degrade proteins to their constituent amino acids, complete proteolysis of titin must require the participation of one or more additional proteases, most likely the proteasome (59).
Alternative mechanism(s) by which the anthracyclines cause cardiotoxicity may involve cell death leading to permanent loss of contractile function. In adult cardiomyocytes in long term culture, doxorubicin induced morphological and biochemical features in cardiomyocytes that are characteristic of necrotic cell death. This may contribute, in part, to the early cytotoxic effects of doxorubicin exposure. Most important, the doxorubicin-induced cardiomyocyte necrosis appears to be mediated by the calpains, with the caspases having little or no role in this form of cell death. Despite significant caspase-3 activation at 24 and 48 h, we did not observe an increase in apoptosis in adult ventricular myocytes under our experimental conditions. At present, we do not know the reasons for the relative resist- . Some, but not all, cardiomyocytes exposed to doxorubicin for 48 h undergo hypercontracture and subsequent rounding up (open arrows), morphological changes consistent with necrosis. Following the video time-lapse, the same area was fixed in 4% paraformaldehyde and immunostained for myomesin (green) and filamentous actin (red). Note that the upper right rounded-up cardiomyocyte (see arrowhead) was lost during the fixation process.
ance of the adult myocyte in long term culture to doxorubicininduced apoptosis. Whereas caspase-3 has been linked to degradation of small myofilament proteins (62), we did not find a role for caspase-3 in doxorubicin-induced titin degradation. Consistent with our results, Chen et al. (63) reported a reduction in CK release and a decrease in infarct size following calpain but not caspase inhibition in the ischemic reperfused heart, although no clear distinction was made between necrotic and apoptotic cell death.
We did find caspase-dependent apoptosis in neonatal cardiomyocytes exposed to doxorubicin, which is consistent with studies in other cell types where doxorubicin-induced apoptosis was suppressed following caspase inhibition (64,65). Calpain inhibition, on the other hand, had no effect on neonatal cardiomyocyte apoptosis suggesting a calpain-independent apoptotic cell death program. In addition to apoptosis and necrosis, a recent report (66) suggests that cardiomyocyte loss can also occur through autophagy associated with accumulation of ubiquitinated proteins. In our experiments, however, doxorubicin did not induce an increase in ubiquitinated proteins 2 ; hence, we believe that autophagy is not a major factor during early an-thracycline cardiotoxicity. Taken together, our data suggest that the calpains and the caspases serve distinct roles in anthracycline-induced cardiomyocyte injury, with the calpains being the predominant contributor to necrotic and the caspases to apoptotic cell death pathways.
Myofilament protein turnover is a dynamic and highly regulated process with protein degradation and synthesis exquisitely balanced to maintain normal sarcomere integrity (67). Doxorubicin accelerates the degradation of titin, and this appears to be an early requisite step in the degeneration of myofibrils. Other events, however, such as the suppression of myofilament transcription, are likely required for the generation of myofibrillar disarray in a given myocyte. Previous reports (42,68) have shown a decrease in myofilament gene transcripts, in vivo and in vitro, following doxorubicin exposure. In failing human hearts, titin mRNA levels were significantly reduced, which correlated with a decrease in titin immunostaining (11). In addition, doxorubicin has been shown to inhibit assembly of myofilament proteins (69). Thus, the overall effect of doxorubicin on myofilament structure is likely a composite of increased titin degradation along with an impairment in protein synthesis and assembly.
Our findings reveal a spectrum of anthracycline-induced in- jury resulting from calpain-dependent proteolysis of titin. At one end of the spectrum, calpain-mediated proteolysis of the elastic domain of titin may have acute physiological consequences. The extensible segment of titin underlies the passive and restoring forces of the cardiomyocyte and regulates myofilament calcium sensitivity, thus titin contributes to the Frank-Starling mechanism and helps maximize myocardial efficiency (4 -9). Proteolysis of the elastic domain of titin will predictably lead to impaired diastolic as well as systolic function, both of which can occur acutely after doxorubicin treatment (29). At the other end of the spectrum is cardiomyocyte necrosis, with an intermediate phenotype being myofibrillar disarray. The reasons for the wide ranging response of cardiomyocytes to injury is not known, although we can speculate that there is heterogeneity in the magnitude of oxidative stress and defense capabilities, calcium homeostasis, or activity of other systems that modulate stress responses, which will ultimately determine the fate of the cell. Although the exact physiological function of the calpains in the myocardium is still unclear, our study supports a model where calpain-mediated proteolysis of titin catalyzes the disassembly of the myofilament lattice during the process of normal sarcomere turnover. Presumably, titin degradation will lead to disassembly of the myofilament complex, requiring de novo synthesis of intact titin molecules around which recycled or newly synthesized components can reorganize. Hence, the energetic cost associated with accelerated titin degradation is large and likely requires significant time for recovery with adverse consequences for myocyte structure, function, and survival. Strategies that prevent titin damage and degradation during injury may therefore offer a novel therapeutic approach for treatment of pathologies associated with myocardial calcium overload and oxidative stress.