INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

PROGRESSIVE WORSENING of left ventricular (LV) systolic and diastolic function is a hallmark characteristic of heart failure (HF) (32, 36). Many biochemical factors have been attributed to this abnormality. Among these factors, Ca²⁺ overloading is one of the important contributors that have been shown to stem from the defect in the sarcoplasmic reticulum (SR) Ca²⁺ uptake and release processes (9, 30). Recently, several studies (1, 3, 14, 23) have reported that cardiac regulatory proteins that control cardiac contractility undergo reversible phosphorylation by protein kinases and phosphatases. The kinases that are involved to phosphorylate cardiac regulatory proteins on a beat-to-beat basis or in the isoproternol-perfused hearts are cAMP-dependent protein kinase A (PKA) and multifunctional Ca²⁺-calmodulin (CaM)-dependent protein kinase type II (CaMKII) (18). In contrast to PKA (5, 37, 38), studies on CaMKII are limited in the pathological state of the heart (6, 15, 19, 27). In the normal (NL) canine LV myocardium, the purified CaMKII migrated as a single band of relative molecular mass (M_r) 55-kDa protein on SDS gel and of a M_r 550,000-kDa protein on native polyacrylamide gel, suggesting a composition of 10 similar subunits (10). This enzyme is predominantly localized in the cytosol (10) but also associated with the membrane of the cardiac cell and has been reported to be involved in cardiomyocyte beating (28), Ca²⁺ infux (21), SR Ca²⁺ uptake and release processes (1, 3, 4, 14, 23), and actomyosin interaction (1, 3, 14, 23). Although the role of CaMKII has been known to regulate the function of normal SR by phosphorylating SR proteins and modulate actomyosin interaction by phosphorylating troponin I, until recently, its role in diseased hearts was controversial (6, 15, 19, 27). To our knowledge, there are only four studies on CaMKII activity in diseased hearts; one study is on the hypertrophied myocardium (6) and the remaining three studies are on failing hearts (15, 19, 27). Studies (15, 19) on failing hearts of humans secondary to dilated cardiomyopathy reported increased CaMKII activity and protein expression, whereas the study (27) of the failing rat myocardium due to myocardial infarction reported reduced CaMKII activity. In vivo, CaMKII activity depends not only on the expression level of the enzyme and the status of CaM but may also depend on the zinc level. CaM is a calcium-binding protein and regulates Ca²⁺-calmodulin-dependent enzymes, including CaMKII enzyme activity. The mRNA level of CaM has been reported to be reduced in the LV tissue of the explanted failed human heart secondary to idiopathic dilated cardiomyopathy (17). However, protein expression level of CaM has yet not been examined in failing hearts. A recent study (2) reported inhibition of cardiac SR-associated CaMKII activity by zinc. The level of zinc has also not been studied in failing hearts. In this study, we have used a canine model of HF produced by intracoronary microembolizations (31). With the use of this canine model of HF, we have reported (11, 12) earlier impaired SR Ca²⁺ uptake and Ca²⁺-ATPase activities, which are associated with reduced expression of phospholamban (PLB) and Ca²⁺-ATPase. With the use of the same dog model of HF, CaMKII activity and expression levels have been examined in the LV myocardium of normal dogs and dogs with chronic HF. In addition, protein levels of CaM, phosphorylated PLB at threonine-17 (PLB-Thr¹⁷) and serine-16 (PLB-Ser¹⁶), and the zinc level were also determined in failing and nonfailing LV myocardium. Our results suggest that CaMKII activity is reduced in the failing LV myocardium, and this abnormality is associated with reduced protein expression level of the enzyme but not due to changes in CaM and zinc levels. Reduced CaMKII activity and phosphorylated PLB-Thr¹⁷ level may be partly responsible for impaired SR function in HF.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Model preparation. A total of 12 dogs of either sex weighing between 19 and 25 kg were used in the study. HF was produced in six dogs, and the remaining six dogs were maintained as NL for comparison purposes. To produce chronic HF, dogs underwent a series of cardiac catherization and intracoronary microembolizations using techniques as previously described (31). Intracoronary microembolizations were performed 1-3 wk apart and discontinued when LV ejection fraction, determined angiographically, was close to 35%. HF dogs were euthanized 4 mo after the last coronary microembolization. With the use of previously published techniques (31, 32), hemodynamic and angiographic measurements were made in HF dogs at baseline before any microembolizations and 4 mo after the last embolization just before they were euthanized. The study was approved by the Care of Experimental Animals Committee and conformed to the "Position of the American Heart Association on Research Animal Use" and the "Guiding Principles in the Care and Use of Animals" of the American Physiological Society. After hemodynamic and angiographic measurements were completed, the animals were euthanized, the chest was opened through a left thoractomy, the pericardium was exposed, and the heart was quickly removed and immersed in ice-cold cardioplegic solution. LV myocardial specimens were cut in 5-mm³ blocks (free of obvious scar or epicardial vessels), quickly frozen in liquid nitrogen, and stored separately at 70°C for biochemical and zinc analysis.

Preparation of homogenate, membrane, and cytosol. From each dog, ~1 g of frozen LV tissue was used to prepare these fractions. All the steps involved in the fraction preparation were performed at 4°C. Tissue specimens were thawed in 10 ml of homogenization buffer consisting of 50 mM Tris · HCl (pH 7.4), 0.5 mM sodium EDTA (pH 7.0), 0.3 M sucrose, and protease inhibitors (0.8 mM of benzamidine, 0.8 mg/l of aprotinin and leupeptin, and 0.4 µg/l of antipain). The thawed tissue was then homogenized as previously described (11, 12). The homogenized tissue was then filtered through four layers of cheesecloths. About 1 ml of filtrate referred to as homogenate was saved, and the remaining was centrifuged at 100,000 g for 40 min. About 1 ml of clear supernatant referred to as cytosol was saved, and the remaining was discarded. The pellet was resuspended in 30 ml homogenization buffer containing 0.6 M KCl and centrifuged at 100,000 g for 40 min. The resulting pellet was again resuspended in 30 ml of homogenization buffer and recentrifuged, and, finally, the resulting pellet was resuspended in 1 ml of medium 1 (20 mM Tris · HCl, pH 7.4; and 0.3 M sucrose) and referred to as membrane. All the fractions were aliquotted into 200-µl portions, immediately frozen in liquid nitrogen, and stored at 70°C until used. Noncollagen protein content of the subcellular fraction was determined by the Lowry method (22) using bovine serum albumin as the standard.

CaMKII activity assay. The enzyme activity was assayed in a total assay volume of 50 µl using CaMKII peptide (281-291) as the enzyme substrate. The assay contained 50 mM Tris · HCl (pH 7.4), 382.7 µM CaMKII substrate (281-291, Calbiochem), 0.3 mM ATP ([-³²P]ATP = 1 µCi per assay), 10 mM magnesium acetate, 0.2 mM ethylene diamine tetraacetic acid, ±5 µM calmodulin, ±250 µM CaCl₂, and an indicated amount of tissue fraction varying protein from 0 to 70 µg. The reaction was initiated by adding enzyme, incubated at 30°C for 10 min, and terminated by adding 20 µl of 300 mM EDTA and ³²P incorporation in the substrate determined as previously described (10). The enzyme activity was expressed as picomoles of ³²P incorporated in the substrate per minute or nanomoles of ³²P incorporated in the substrate per minute per milligram of noncollagen protein.

Western blotting. To determine protein levels, SDS extract of the homogenate, membrane, and cytosol was prepared from LV tissue as previously described (11, 12). To freeze the phosphorylation state of the proteins, LV tissue was homogenized in the presence of the inhibitors of protein kinases (1 mM EDTA and 1 mM EGTA) and protein phosphatases (2 mM sodium pyrophosphate and 10 mM sodium fluoride). Nearly 5 µg or the indicated amount of the SDS extract was separated on 4-20% linear polyacrylamide (Bio-Rad) and transferred electrophoretically on nitrocellulose membrane, and the resulting membrane was incubated with primary antibody as previously described (11, 12). The accuracy of the electrotransfer was confirmed by staining the membrane with 0.1% amido black. Polyclonal antibody (CaMKII, phosphorylated PLB-Thr¹⁷, phosphorylated PLB-Ser¹⁶) or monoclonal antibody (PLB and CaM) was diluted to 500-fold or 2,500-fold, respectively. Primary antibody binding protein was visualized by incubating the blot with a second antibody, a peroxidase-conjugated anti-mouse in the case of monoclonal antibody or anti-rabbit in the case of polyclonal antibody (Sigma), and the enhanced chemiluminescence assay was used as described by supplier (DuPont-NEN). In parallel, calsequestrin, a calcium-binding protein associated with the SR and reported to not be altered during HF (5), was also determined in the LV homogenate and membrane fractions as previously described (11, 12). CaMKII antibody used in this study recognizes all the -, -, and -isoforms of the enzyme because the sequences of the -, -, and -isoforms are very similar in the region to which the immunizing peptides were based. The intensity of the bands was quantified by using a Bio-Rad model GS-670 imaging densitometer. The densitometric unit of measurement was optical density times square millimeters. The density of the phosphorylated PLB-Thr¹⁷ was normalized to the amount of PLB present in LV tissue and that of CaMKII and CaM to the amount of total protein applied on the gel because the expression level of calsequestrin in the homogenate and membrane fractions of LV was not found to be different between NL and HF specimens (data not shown). Before quantitation of the protein expression levels in normal and HF LV tissues, the protein dependency of the immunodetectable bands for all the proteins was established in NL LV tissue. A linear correlation was observed between densitometric units and protein content (<30 µg) for each protein in the study (data not shown).

Elemental analysis. The zinc analysis was performed according to the slight modification of the procedure as previously described (24, 25). The heart tissue (~500 mg) was thawed and placed in a 15-ml precleaned screw-capped Teflon beaker. Five milliliters of concentrated nitric acid were added to the sample and left at room temperature overnight. The beaker was then placed on a hot plate maintained at a temperature of 100°C and then incubated overnight, followed by evaporation to dryness. To the dried mass, 2 ml of HNO₃ followed by 25 drops of H₂O₂ were added drop wise to minimize foaming and then evaporated to near dryness. After three repetitions of the HNO₃-H₂O₂ procedure, the samples were allowed to cool and subsequently made up gravimetrically with 2% HNO₃ to ca 20 g. A VG elemental PlasmaQuad model PQ2 inductively coupled plasma mass spectrometer (ICP-MS) was used for all data acquisition according to the operating conditions described in McGinnis et. al. (24). Single element standard solutions were utilized to prepare calibration and internal standard solutions. Analysis was performed using an external calibration procedure, and cobalt and gallium were used as internal standards for matrix and instrument drift corrections. Procedural blank was analyzed to check for any contribution from the reagents. The standard reference material (SRM)-8414b (bovine muscle) and SRM-1577b (bovine liver) were digested and analyzed to be identical to the unknowns for quality control purposes. The amount of zinc in the LV tissue was expressed as milligram per kilogram of tissue.

Materials. Antibody specific to PLB and phosphorylated PLB-Thr¹⁷ and PLB-Ser¹⁶ was purchased from Phosphoprotein Research (West Yorkshire, UK), antibody specific to CaMKII and CaM was from Upstate Biotechnology (Lake Placid, NY), and antibody specific to calsequestrin was from Dr. Larry R. Jones (Krannert Institute of Cardiology, Indiana University School of Medicine, Indianapolis, IN). Chemicals and supplies for electrophoresis and electrotransfers were purchased from Bio-Rad (San Francisco, CA). Single- element standard solutions were purchased from Inorganic Ventures (Lakewood, NJ). General chemical supplies used in preparation of LV homogenate and measurement of SR Ca²⁺ uptake were obtained from Sigma Chemical (St. Louis, MO).

Statistical analysis. The data shown are means ± SE. Comparisons between failing and nonfailing hearts were based on a t-statistic for two means (unpaired t-test). P < 0.05 was considered statistically significant. The sample size used in this study of six failing and six nonfailing hearts dictated by 80% power to detect a large difference at  = 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Hemodynamic and angigraphic data in dogs with HF. The hemodynamic and angiographic measurements obtained at baseline and 4 mo after last embolizations in dogs with HF are shown in Table 1. LV ejection fraction, LV peak isovolumic contraction and relaxation (+dP/dt and dP/dt, respectively), and stroke volume were significantly decreased in HF compared with baseline (Table 1). These reductions were also associated with a significant increase of LV end-diastolic volume (70 ± 7 vs. 54 ± 6 ml, P < 0.05) and LV end-diastolic pressure (23 ± 3 vs. 6 ± 1 mmHg, P < 0.05). These results are consistent with the presence of LV failure in this canine preparation at the time when the study was conducted.

CaMKII activity. With the use of CaMKII as the peptide, CaMKII activity was determined in the absence and presence of Ca²⁺-CaM. Results are shown in Fig. 1A. The enzyme activity increased at increasing protein concentrations up to 12 µg of LV homogenate. However, the activity at high protein (15-60 µg) concentration of LV homogenate declined. Furthermore, CaMKII activity assayed in the absence of Ca²⁺-CaM was almost 1/100th part of the CaMKII activity assayed in the presence of CaM. Subsequently, CaMKII activity in the presence of Ca²⁺-CaM was determined in 4 µg of LV homogenate from NL and HF dogs (Fig. 1B). In the LV homogenate, CaM-stimulated CaMKII activity was significantly reduced in failing hearts compared with NL hearts (9.22 ± 0.32 vs. 12.08 ± 0.45 nmol ³²P · min¹ · mg protein¹). Because the activity is present in both the membrane and cytosol (16, 17), the CaM-stimulated CaMKII activity was then determined in the isolated membrane and cytosol of LV myocardium of NL dogs and dogs with HF, and the results are depicted in Fig. 2. At all protein concentrations, CaMKII activity declined in the membrane (Fig. 2A), but not in the cytosol (Fig. 2B), of the LV myocardium of dogs with HF compared with NL dogs. In contrast, calmodulin-independent CaMKII activity did not significantly differ in the homogenate (0.12 ± 0.008 vs. 0.12 ± 0.007 nmol ³²P · min¹ · mg¹), membrane (0.016 ± 0.008 vs. 0.017 ± 0.006 nmol ³²P · min¹ · mg¹), and cytosol (0.24 ± 0.02 vs. 0.22 ± 0.02 nmol ³²P · min¹ · mg¹) obtained from the LV myocardium between NL dogs and dogs with HF. In addition to the expression of the activity per milligram of noncollagen protein, CaMKII-stimulated activity was also calculated in whole LV myocardium and compared between NL dogs and dogs with HF. Noncollagen protein (0.86 ± 0.02 vs. 0.78 ± 0.01 mg/mg LV tissue) and LV weight normalized to body weight (4.37 ± 0.07 vs. 4.22 ± 0.08 g/kg) were slightly higher in HF dogs compared with NL dogs, but these increases in dogs with HF were not statistically significant. Total enzyme activity in whole LV myocardium of HF dogs was found to be 7.9 ± 0.4 µmol ³²P · min¹ · mg noncollagen protein¹, which was significantly less than that of NL dogs (9.58 ± 0.4 µmol ³²P · min¹ · mg noncollagen protein¹). These data clearly suggest that CaMKII activity is reduced in dogs with HF regardless of whether the enzyme activity is expressed per milligram of noncollagen protein or in whole LV myocardium.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   A: effect of different concentration of left ventricular (LV) homogenate on Ca2+-calmodulin (CaM)-dependent protein kinase II (CaMKII) activity assayed in the absence (closed circles) and presence (opened circles) of calmodulin. B: CaMKII activity assayed in the presence of calmodulin in LV homogenate of 6 normal (NL) dogs and 6 dogs with heart failure (HF). * P < 0.05 vs. NL indicates statistically significant.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Effect of varying concentrations of membrane and cytosol proteins on CaMKII activity assayed in the presence of calmodulin. Membrane (A) and cytosol (B) were isolated from LV myocardium of 6 NL dogs (filled bar, filled circle) and 6 HF dogs (open bar, open circle). * P < 0.05 vs. normal indicates statistically significant.

Determination of protein expression level of CaMKII. To examine whether reduced enzyme activity is associated with its reduced protein expression level in failing hearts, the expression level of CaMKII was determined in the homogenate, membrane, and cytosol of the LV myocardium of NL and HF dogs. In parallel, the protein expression level of calsequestrin was also determined in the same samples and found not to be changed between NL and HF dogs (data not shown). Figure 3 shows that CaMKII antibody recognized three protein bands of M_r 58, 55, and 47 kDa in the homogenate, only one major protein band of M_r 58 kDa in the membrane, and four protein bands, three bands same as those observed in the homogenate but one more additional band near M_r 72 kDa, in the cytosol of LV tissue. The later band was not visible in the LV homogenate, in part due to a protein upon solubilization by SDS from the membrane blocked the epitope region of the M_r 72-kDa protein present in the cytosol. According to the instructions of the antibody supplier, protein bands corresponding to M_r 50 to M_r 60 kDa are due to the presence of different isoforms of CaMKII. In HF, protein expression levels of CaMKII were reduced in the homogenate (55 and 58 kDa) and membrane (58 kDa) but not in the cytosol compared with NL (Fig. 3). For quantitation, the density of protein bands near 58 to 55 kDa was determined and normalized to the amount of protein loaded on the gel, and the results are shown in Fig. 4. In HF, the protein expression level of CaMKII was reduced by 30% in the homogenate (Fig. 4A) and 52% in the membrane (Fig. 4B), but no significant changes were noted in the cytosol (Fig. 4C) compared with NL.

View larger version (39K): [in this window] [in a new window]   Fig. 3.   Western blot of immunodetectable CaMKII in homogenate (A), membrane (B), and cytosol (C) of LV myocardium of 3 NL and 3 HF dogs. CaMKII bands near molecular mass of 55-58 kDa are reduced in the homogenate and membrane, but not in the cytosol, of LV myocardium of dogs with HF compared with NL dogs.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Densitometric analysis of the immunodetectable CaMKII bands near the molecular mass of 55-58 kDa in a Western blot of the homogenate (A), membrane (B), and cytosol (C) of LV myocardium from 6 NL and 6 HF dogs. * P < 0.05 vs. NL indicates statistically significant.

Expression levels of total PLB, phosphorylated PLB-Thr¹⁷ and PLB-Ser¹⁶, and CaM. Because the amount of the phosphorylated PLB-Thr¹⁷ and PLB-Ser¹⁶ in the LV tissue depends on the amount of PLB present in the tissue, tissue levels of phosphorylated PLB-Thr¹⁷, phosphorylated PLB-Ser¹⁶, and PLB were determined in NL and HF LV specimens. The results of the Western blot of the phosphorylated PLB-Thr¹⁷ and PLB-Ser¹⁶ and total PLB in three NL and three HF dogs are depicted in Fig. 5A, and the densitometric analysis of the phosphorylated PLB-Thr¹⁷ and PLB-Ser¹⁶, normalized to the amount of total PLB, in six NL dogs and six HF dogs are shown in Fig. 5B. As reported earlier (12), LV tissue levels of total PLB (0.22 ± 0.02 vs. 0.31 ± 0.03 densitometric units) were reduced by 30% in HF dogs compared with NL dogs. Tissue levels of phosphorylated PLB-Thr¹⁷ (1.39 ± 0.15 vs. 2.7 ± 0.1 densitometric units normalized to PLB) and PLB-Ser¹⁶ (0.87 ± 0.07 vs. 1.37 ± 0.05 densitometric units normalized to PLB) declined by 48 ± 5% and 36 ± 6%, respectively, in HF dogs compared with NL control dogs (Fig. 5). However, there was no significant difference between the reduction of the tissue level of phosphorylated PLB-Ser¹⁶ and PLB-Thr¹⁷ in HF dogs compared with NL dogs. Next, because CaMKII activity depends on the amount of CaM present in LV tissue, the tissue level of CaM was also determined in the LV homogenate of NL dogs and HF dogs by using monoclonal antibody. The results of Western blots of three NL dogs and three HF dogs and the densitometric analysis of six NL dogs and six HF dogs are shown in Fig. 6. No difference in the CaM level was found to be present between NL and HF dogs (Fig. 6).

View larger version (43K): [in this window] [in a new window]   Fig. 5.   A: Western blot of immunodetectable phosphorylated phospholamban (PLB) at threonine-17 (PLB-Thr) and serine-16 (PLB-Ser) and PLB in the homogenate of LV myocardium from 3 NL and 3 HF dogs. B: densitometric analysis of PLB-Thr and PLB-Ser normalized to the amount of PLB in 6 NL dogs (filled bar) and 6 dogs with HF (open bar). * P < 0.05 vs. NL indicates statistically significant.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   A: Western blot of immunodetectable CaM in LV homogenate of 3 NL and 3 HF dogs. B: densitometric analysis in 6 NL dogs and 6 dogs with HF.

Elemental analysis of zinc. The zinc analysis in LV tissue, determined by ICP-MS, was found to be 16.61 ± 1.01 mg/kg in NL dogs as opposed to 18.20 ± 1.96 mg/kg in HF dogs. These values on zinc content in LV tissues were not statistically different between NL and HF dogs. To ensure that the analyzed zinc values are correct, two reference materials, SRM 1577b (bovine liver) and SRM 8414b (bovine muscle), were used. The reported values for SRM 1557b and SRM 8414b reference materials were 127 ± 16 and 142 ± 14 mg/kg, which were very similar to the analyzed values of 115 ± 16 and 135 ± 16 mg/kg, respectively.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present results demonstrate that the CaMKII activity and protein expression level are reduced in the LV myocardium of dogs with chronic HF. Furthermore, these decreases are observed only in the membrane, but not in the cytosol, of failing LV myocardium. In addition, CaM and zinc levels are unchanged, whereas phosphorylated PLB-Thr¹⁷ and PLB-Ser¹⁶ levels are reduced in failing hearts compared with the normal control. In the same model of HF, we have reported earlier reduced SR Ca²⁺-ATPase activity and expression and reduced SR Ca²⁺ uptake and PLB expression without any changes in the expression level of calsequestrin (11, 12) in the failing LV myocardium. The present results strengthen our previous observations on SR abnormality in failing hearts and also provide a mechanism for reduced SR Ca²⁺ uptake. Thus reduced CaMKII activity and phosphorylated PLB at serine-16 and threonine-17 could, in part, be responsible for reduced SR Ca²⁺ uptake, a characteristic of HF, which has been recognized by many investigators (1, 14, 23).

Phosphorylation of several proteins by PKA and CaMKII controls contractilite function of the heart (18, 29). A number of studies (5, 37, 38) have reported no changes in PKA activity and protein expression in the failing LV myocardium obtained from a different model of HF, and the accepted views are that changes in PKA activity are not involved in pathogenesis of HF (19). In contrast to PKA, very limited studies are reported on CaMKII activity and expression level in the failing LV myocardium. In the NL heart, the most prominent characterized CaMK is multifunctional CaMKII (15). This enzyme was purified to homogeneity from the canine ventricular myocardium and found to be composed of 10 identical subunits of Mr 55 kDa (10). This subunit later on was recognized as the ₃-isoform, the predominant isoform present in the SR of the adult heart, and its mRNA level significantly increased in the failing human LV myocardium due to idiopathic dilated cardiomyopathy (15). Increased mRNA levels for CaMKII were associated with significantly increased CaMKII activity in human failing hearts due to idiopathic but not due to ischemic cardiomyopathy (19). Our results on calmodulin-stimulated CaMKII activity are in contrast to the observations reported in the explanted failed human hearts but agree with that on calmodulin-independent CaMKII activity, which did not differ in the present study between NL and HF dogs. However, our results on calmodulin-stimulated CaMKII activity is very similar to the one reported in rats in which CaMKII activity was also found to be reduced in LV myocardium with HF produced by coronary artery ligation (16). These discordant results between human and animal (dog and rat) failing hearts could, in part, be due to the treatment of terminal congestive HF patients who receive a number of drugs, whereas animals did not receive any therapy. It is, therefore, quite possible that some of these drugs may have impact on the expression level of CaMKII. The second possibility is due to the presence of fibrosis in failing LV tissue. It has been reported that failing hearts have significant interstitial fibrosis (32). But we have used failing LV tissue free from fat and obvious scars and found that the amount of membrane isolated from such failing LV tissue is similar to that from LV of NL hearts, and the amount of calsequestrin has also been found to be similar in the LV tissue between NL and HF dogs (11, 12). In addition to the expression of CaMKII activity per milligram of noncollagen protein, total enzyme activity was also calculated in the whole LV myocardium and was significantly reduced in HF dogs compared with NL dogs. To our surprise, we did not notice any statistically significant changes in normalized LV weight to body weight and in the noncollagen protein of the entire LV myocardium between NL and HF dogs. It may be quite possible because the LV hearts from these HF dogs are undergoing fibrosis, apoptosis, and hypertrophy (32, 33). Despite these conditions, failing LV tissue does show reduced CaMKII activity. Reduced CaMKII protein level in the failing heart then becomes visible even though a commercially available antibody was used, which recognizes almost all the isoforms (, , and ) of CaMKII. Therefore, the observed reduced CaMKII activity and expression level in failing hearts appear to be due to real changes. We further examined whether reduced CaMKII activity is also due to changes in the expression level of CaM and/or the content of zinc. To our knowledge, there is only one study that reported reduced expression of the mRNA encoding CaM in human end-stage HF (17). In that study, the CaM protein expression level was not examined in the explanted failed human hearts. In the present study, CaM protein level was quantitated but not found to be different between NL and HF dogs. Descrepancy of reduced mRNA level but unaltered protein level is not unique and has been reported in other cases, too (5, 34). Next, the content of zinc has not been reported in failing hearts. A recent study (2) reported that zinc in a micromolar concentration can inhibit CaMKII activity. Thus any changes in the zinc level in the failing heart tissue can influence PKM activity. With the use of ICP-MS, no changes in the content of zinc was observed in the LV tissue between NL dogs and dogs with HF. Thus changes in zinc and CaM levels do not occur in failing hearts and are therefore not responsible for the observed reduced CaMKII activity in failing hearts.

CaMKII phosphorylates many physiological substrates in the NL heart, and some of them are localized in the SR (3, 18, 19). These are PLB, Ca²⁺-ATPase, and ryanodine release channels (1, 14, 23). Several investigators reported the status on the phosphorylation of PLB in failing hearts, but their results are discordant (3, 5, 16, 20, 34, 35). In the present study, we have found reduced phosphorylated PLB-Thr¹⁷ and PLB-Ser¹⁶ levels in the LV myocardium of dogs with HF. This finding is in agreement with others, who also reported a diminished level of the phosphorylated PLB at both serine-16 and threonine-17 in failing LV tissue from the explanted failed human heart (34) or the heart of rats with HF produced by coronary artery ligation (16). But this finding is in contrast to our earlier report in which no changes in the activity and protein expression level of CaMKII and the level of phosphorylated PLB-Thr¹⁷ were present in the explanted failed human heart (26). In contrast to human failing hearts, these results suggest that the level of phosphorylated PLB-Thr¹⁷ is reduced in failing LV tissue from animals, and this abnormality is associated with reduced CaMKII activity (27) and increased protein phosphatase (13, 28), two enzymes responsible for maintaining the status of the phosphorylated PLB-Thr¹⁷ in the heart.

In summary, our results are the first to demonstrate in large animals that CaMKII activity is reduced in the membrane, but not in the cytosol, of failing LV myocardium. Furthermore, reduced CaMKII activity in failing hearts may, in part, result from reduced expression level of CaMKII but not due to changes in the CaM and zinc levels. On the basis of these results, we conclude that reduced CaMKII activity and phosphorylated PLB-Thr¹⁷ may be partly responsible for impaired SR function in HF.