INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

TUMOR NECROSIS FACTOR- (TNF-) is a pleiotropic peptide that is elevated in the sera and myocardium of patients with heart failure (HF). Its importance in the pathophysiology of left ventricular (LV) dysfunction was first recognized in studies demonstrating acute negative inotropic effects on single cardiac myocytes and isolated heart preparations (1, 14, 16, 26, 61). Furthermore, rats subjected to prolonged (15 day) intraperitoneal infusion of TNF- developed reversible LV dilatation and dysfunction associated with a reduced contractility of isolated LV myocytes (5). To further elucidate the role of TNF- in the development of HF, we created transgenic (TG) mice with cardiac-restricted overexpression of TNF- (TNF1.6 mice) (29). The TNF1.6 mice (29) and similar TNF- overexpressor mice (7) exhibited a HF phenotype with ventricular hypertrophy, fibrosis, extracellular matrix remodeling, dilatation, diminished LV function, loss of -adrenergic responsiveness, activation of fetal gene programs, and premature death. The TNF1.6 mice showed a striking gender-dependent difference in the severity of LV dysfunction and mortality between young (<20 wk of age) male (TG-M) and female (TG-F) TNF1.6 mice, despite identical levels of TNF- expression (24).

Changes in LV structure and function in HF result from several processes, including fibrotic remodeling of the extracellular matrix, realignment (slippage) of myocytes within the LV wall, and/or alterations in the structure and contractility of individual LV myocytes (6, 22, 38). Both altered intracellular Ca²⁺ (Ca) homeostasis (26, 61) and reduced myofilament responsiveness to Ca (14, 16) have been suggested to underlie the acute negative inotropic effect of TNF- in vitro. However, it is unclear whether the diminished ventricular function effected by chronic exposure to TNF- in vivo is attributable to the intrinsic changes at the single myocyte level or to extracellular matrix remodeling and changes in LV geometry. The present study was designed to determine whether and to what extent alterations in LV structure and function in mice overexpressing TNF- are related to changes in the morphometry, Ca handling, and/or contractile function in single LV myocytes.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals. Twelve-week-old TG mice (TNF1.6) and their wild-type (WT) littermates (29) were utilized under protocols approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh. Twelve-week-old mice were utilized in these studies because gender-dependent differences were most marked at that age (24, 32).

Echocardiography. Echocardiographic studies were performed on 12-wk-old TNF1.6 mice and their WT littermates with an ultrasonograph system (Sequoia 512, Acuson), using methods described in detail previously (24, 32).

Myocyte isolation. LV myocytes were isolated using the method described by Xu et al. (59). Briefly, hearts were perfused by the Langendorff method with HEPES-buffered Earle's balanced salt solution (GIBCO-BRL) supplemented with 6 mM glucose, amino acids, and vitamins (buffer A) and then with buffer A containing 0.8 mg/ml collagenase B (Boehringer-Mannheim) and 10 µM CaCl₂ at 35°C. The enzyme solution was filtered (2-µm pores) and recirculated through the heart until the flow rate doubled (12-20 min); the LV was then removed and minced in collagenase-containing buffer A. Thereafter, the tissue pieces were transferred to fresh (enzyme free) buffer A supplemented with 1.25 mg/ml taurine, 5 mg/ml BSA (Sigma), and 150 µM CaCl₂ and mechanically dissociated by gentle trituration. The resulting suspension was filtered, and isolated cells were obtained by sedimentation. The myocytes were kept in buffer B containing 500 µM CaCl₂ until used, usually within 5-7 h.

Measurement of cell size and shortening. The morphological and mechanical properties of the myocytes and Ca activity were measured using the methods described by Wolska and Solaro (58). Briefly, in each heart, the myocyte length, width, and contractile function (unloaded shortening) were examined in a portion of the myocytes not loaded with fura 2 and placed in temperature-controlled chamber (25°C) mounted on the stage of an inverted microscope (Nikon Diaphot). The cells were bathed with modified Tyrode solution, which contained (in mM) 137 NaCl, 15 glucose, 5.4 KCl, 1.3 MgSO₄, 2.0 CaCl₂, and 20 HEPES (pH 7.4), and stimulated with 5-ms pulses at 0.5 Hz. A video motion edge detector (Crescent Electronics) and IonOptics (Milton, MA) software were used to analyze and calculate the myocyte shortening fraction and the maximal rates of the cell shortening and relengthening. Calibration of the voltage output of the edge detector was carried out using coverslips with an etched micrometer grid (Cell VU). The cell length and width were assessed in 11-15 myocytes/heart. To avoid bias, these measurements were carried out both randomly and systematically, i.e., in each rod-shaped myocyte encountered upon moving the microscope objective from the top left to bottom right of the perfusion chamber.

Measurement of Ca activity. For Ca measurements, the myocytes were incubated with 5 µM fura 2 for 30 min in the physiological buffer with 250 µM Ca²⁺ at room temperature (23-24°C). After being loaded, the cells were washed and resuspended in physiological buffer with 500 µM Ca²⁺. Fluorescence emissions were detected between 480 and 520 nm with a dual-excitation fluorescence multiplier system (IonOptics) after cells were illuminated at 360 nm, at the beginning and the end of the recording, and at 380 nm for the duration of the recording. The ratio of fluorescence at 360 nm to fluorescence at 380 nm was calculated as a function of Ca after the background fluorescence was subtracted. In each heart, the background fluorescence was estimated by measuring and averaging the autofluorescence of randomly picked 9-12 myocytes not loaded with fura 2. Recordings were performed under basal conditions and after a 1-min exposure to 1 µM isoproterenol (Iso). Steady-state contractions or Ca transients (5-10) were averaged for each cell and used to analyze the measured parameters.

Analysis of transcript levels. For measurements of mRNA levels of -myosin heavy chain (MHC), -MHC, atrial natriuretic factor (ANF), cardiac sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA) 2, and phospholamban (PLB), hearts were removed from euthanized mice and flash frozen. RNA isolation, Northern blot hybridizations, preparation of cDNA and oligonucleotide probes, and quantitative analysis of hybridized filters were performed as previously reported (28). Hybridization signals were normalized to that of the 18S probe to correct for differences in RNA mass and efficiency of transfer. For evaluation with respect to both genotype and gender, the data were normalized to the mean of the male and female WT samples combined, arbitrarily set at 100%.

Analysis of protein expression. For measurements of protein expression of SERCA 2 and PLB using Western blots, frozen hearts were homogenized [50 mM KHPO₄ (pH 7.0), 0.3 M sucrose, 10 mM NaF, 1 mM EDTA, and 0.5 M dithiothreitol] and briefly centrifuged to remove gross debris. Protein levels were determined (Bio-Rad) using bovine IgG as a standard. Proteins were subjected to gel electrophoresis, transferred to nitrocellulose membranes, interacted with antibodies, washed, developed by chemiluminescence, and quantified as previously described (41). Antibodies included anti-PLB monoclonal antibody (1 µg/ml, Upstate Biotechnology) or rabbit polyclonal anti-SERCA antibody (5 µg/ml, Zymed). Filters were stripped and reinteracted with anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH; diluted 1:1,000, Research Diagnostics) with the resultant GAPDH signal serving to normalize the protein mass loaded and transferred. These results, representing an average of four experiments, were in turn normalized to the mean of the male and female WT samples combined, arbitrarily set at 100%.

Statistics. Results are presented as means ± SE; n indicates the number of mice. For experiments in isolated myocytes, 4-8 and 11-15 cells from individual hearts were used for functional and morphometric assessments, respectively, and each heart was treated as a single n. Statistical significance of differences was determined using ANOVA with a post hoc Student-Newman-Keuls test in morphometric and functional measurements and using ANOVA and Tukey's post hoc test in molecular studies. Values of P < 0.05 were considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

LV structure and function. At 12 wk of age, male and female WT mice were not different with respect to the parameters assessed using echocardiography and gravimetric measurements in explanted hearts (Table 1). Compared with WT mice, TG-F mice showed a significant increase in LV weight and the LV weight-to-body weight ratio. Cardiac systolic performance, indexed by fractional shortening (FS), was only slightly reduced in TG-F versus WT mice at baseline, whereas the Iso-dependent increase in FS was significantly smaller in TG-F mice (Table 1). In contrast, TG-M littermates exhibited a significantly larger increase in the LV weight-to-body weight relative relative to TG-F mice and LV dilation relative to both WT and TG-F mice. Also, baseline LV systolic dimension and FS were significantly reduced in TG-M mice compared with both WT and TG-F mice and failed to appreciably increase after Iso challenge. Heart rate, measured under anesthesia, and body weight, assessed gravimetrically, were similar in all compared groups (Table 1). The present echocardiographic results support those reported previously in TNF1.6 mice (23, 31).

Gene expression at transcript levels. Northern blot analysis showed that relative to WT hearts, transcripts encoding -MHC, SERCA 2, and PLB were significantly reduced, whereas transcripts encoding -MHC and ANF were significantly increased, in TG hearts (Fig. 1, A and B). Additional patterns of gene expression emerged upon categorization of the hearts by both genotype and gender (Fig. 1, C-G). Specifically, whereas both TG-M and TG-F hearts showed significantly elevated expression of ANF and -MHC transcripts compared with WT littermates, a relative increase in these two transcript was approximately twofold larger in TG-M versus TG-F hearts (Fig. 1, C and D). Message levels of -MHC, SERCA 2, and PLB were significantly reduced in TG versus WT hearts when assessed with respect to both genotype and gender (Fig. 1, E-G). Figure 1, E-G, also shows higher message levels of -MHC, SERCA 2, and PLB in male versus female WT hearts and a lack of significant gender differences of these transcripts in TG hearts.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Northern blot analysis of gene expression. A: representative Northern blot images in wild-type (WT) and transgenic (TG) hearts. MHC, myosin heavy chain; SERCA, sarco(endo)plasmic reticulum Ca2+-ATPase; ANF, atrial natriuretic factor; PLB, phospholamban. B: quantitative Northern blot analysis of combined male + female (M+F) hearts. Mean of M+F WT = 1.0 (n = 19-20 mice/group). * P < 0.001 vs. WT hearts. C-G: quantitative differences in gene expression analyzed by gender. WT-M, WT males; WT-F, WT females; TG-M, TG males; TG-F, TG females. * P < 0.05. NS, not significant.

Protein expression of SERCA 2 and PLB. An impaired Ca²⁺ handling in TG myocytes (see Figs. 3 and 4 and Table 2) may be consequent to an altered protein expression and/or function of SERCA-2 or PLB (6, 22). To directly address this issue, we examined the protein expression of SERCA 2 and PLB proteins using Western blotting. Protein levels of SERCA 2 or PLB were not different in TG versus WT hearts (Fig. 2) or in male versus female mice in both groups (data not shown).

View this table: [in this window] [in a new window]   Table 2.   Myocyte morphometry, contractile function, and Ca kinetics in WT and TG TNF 1.6 mice

View larger version (32K): [in this window] [in a new window]   Fig. 2.   Western blot analysis of PLB and SERCA protein expression. A: representative Western blot image of PLB, SERCA, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) protein expression in WT and TG hearts. B: quantitative analysis (n = 10 WT and TG hearts; males = 50% in each).

Structural and functional properties of isolated LV myocytes. At 12 wk of age, there were no appreciable differences between myocytes isolated from WT male and female mice, and both groups were combined and denoted as WT mice. Figures 3 and 4A show representative examples of basal contractile function and Ca activity in electrically stimulated ventricular myocytes isolated from the hearts of 12-wk-old WT, TG-F, and TG-M mice. Pooled data from these experiments, and the morphometry of rested myocytes, are summarized in Table 2. Compared with WT myocytes, the resting cell length of TG-F and TG-M myocytes was similarly increased (by 20%, P < 0.001), whereas the cell width was not different among the compared groups. The length-to-width ratio was the same in TG-F and TG-M myocytes and was significantly increased relative to WT myocytes (Table 2). During steady-state stimulation at 0.5 Hz, Ca activity and contractile parameters were comparable in WT and TG-F myocytes except a longer time to peak Ca and time to 50% decline (TD_50%) of the Ca transient in TG-F myocytes (Figs. 3 and 4A and Table 2). In contrast, compared with WT myocytes, TG-M myocytes displayed a significant reduction in the maximal rate of development of Ca transients (+dCa/dt_max), the amplitude, and the maximal rate of decline of Ca transients (dCa/dt_max) of the electrically stimulated Ca transients and twitch contractions and a prolongation of TD_50% of the Ca transient. Furthermore, the amplitude of the Ca transients and contractions and the +dCa/dt_max were also significantly reduced in TG-M versus TG-F myocytes (Figs. 3 and 4A and Table 2).

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Effect of cardiac-specific overexpression of TNF- on contractile function in single murine left ventricular (LV) myocytes. Myocytes were isolated from the hearts of 12-wk-old WT mice and TNF1.6 TG-F and TG-M littermates. The cells were electrically stimulated at 0.5 Hz in the presence of 2 mM extracellular [Ca2+] at 25°C. Positive and negative deflections of dL/dt (top), maximal rate of development and decline of unloaded cell shortening; %L (bottom), percent change from diastolic length. Note the similar %L but slower kinetics of contractions in TG-F vs. WT myocytes and the marked reduction of all contractile parameters in TG-M myocytes compared with both WT and TG-F myocytes.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Effect of cardiac-specific overexpression of TNF- on Ca2+ regulation in single murine LV myocytes in the absence and presence of -adrenergic stimulation. Original recordings show intracellular Ca2+ (Ca) concentration transients indexed by the fura 2 380-to-360-nm fluorescence ratio (bottom) and their first derivative (top) in WT, TG-F, and TG-M ventricular myocytes stimulated at 0.5 Hz before (A) and after a brief (1 min) exposure to 1 µmol/l isoproterenol (Iso; B).

Effect of -adrenergic stimulation on Ca regulation in LV myocytes. Figure 4 shows representative examples of steady-state Ca transients measured in WT, TG-F, and TG-M myocytes before (A) and after a brief (1 min) exposure to Iso (B). Pooled data from these experiments are illustrated in Fig. 5. In all compared groups, the Iso-dependent increase in diastolic Ca averaged 5-7%, not different from control (P = not significant). Potentiation of the amplitude of the Ca transients and their +dCa/dt_max and dCa/dt_max were similar in WT and TG-F myocytes exposed to Iso (Figs. 4 and 5). In contrast, Iso-dependent augmentation of the amplitude and the +dCa/dt_max of the Ca transients were significantly reduced in TG-M myocytes relative to both WT and TG-F myocytes. Also, an increase in the dCa/dt_max in response to Iso was significantly smaller in TG-M than WT myocytes (Figs. 4 and 5).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Average data from experiments illustrated in Fig. 4 showing the effect of Iso on diastolic Ca, amplitude (systolic  diastolic) of Ca transients (Ca2+i), and maximal rates of development (+dCa/dtmax) and decline (dCa/dtmax) of Ca transients in ventricular myocytes isolated from the hearts of WT (n = 6), TG-F (n = 4), and TG-M (n = 4) mice. Values are means ± SE; n = number of hearts, with 4-5 myocytes/heart.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our findings in single myocytes, combined with the assessments of changes in the structure and function of age-matched hearts in vivo, provide several new observations relevant to cardiac pathophysiology consequent to overexpression of TNF-. First, we found that that LV myocytes isolated from TNF1.6 male mice display significant abnormalities in Ca handling, contractile function, and -adrenergic response, whereas myocytes isolated from TNF1.6 females differ from WT myocytes only in the slower kinetics of the Ca transients. Second, similarities in functional alterations in single LV myocytes isolated from TNF1.6 mice and LV function in vivo suggest that the latter is determined, in part at least, at the myocyte level. Third, our results show that isolated LV myocytes exhibit a significant hypertrophy, associated with an increase in the cell length but not width, changes that are independent of gender or myocyte function. Also, whereas both male and female TNF1.6 hearts show a similar hypertrophy, the former display a marked LV dilatation and a larger increase in gene transcripts of ANF and -MHC. Finally, we find that an impaired Ca handling in TNF1.6 myocytes does not result from changes in protein expression of SERCA 2A or PLB.

Molecular remodeling in TNF1.6 mice. Reactivation of the embryonic gene program is a characteristic feature of cardiac hypertrophy and failure (8, 25). Consistently, transcripts of ANF and -MHC were significantly increased in TG versus WT hearts (Fig. 2, A-D). Interestingly, whereas male and female TG hearts displayed a similar extent of cardiac hypertrophy (indexed by the LV mass-to-body weight ratio; Table 1) and cellular hypertrophy (Table 2), the increase in ANF and -MHC mRNA was approximately twofold larger in TG males than in females (Fig. 1, C and D). This difference may reflect, in part, increased transmural tension/myocyte stretch, a potent modulator of gene expression in cardiac myocytes (25), consequent to a significant dilatation of TG male hearts (Table 1). The pattern of changes in -MHC transcripts (Fig. 1, A, B, and E) was opposite to that of -MHC transcripts, consistent with a partial replacement of -MHC with -MHC in hypertrophied and failing hearts (35, 47). Compared with WT hearts, transcripts of SERCA 2 and PLB were significantly reduced in TG hearts (Fig. 1, A, B, F, and G). These findings are compatible with reports in human failing hearts (33, 42) and experimental hypertrophy (9, 46, 56). We also observed relatively higher message levels of -MHC, SERCA 2, and PLB in WT males versus females (Fig. 1, E-G). However, these differences were not reflected in the function of WT hearts or isolated myocytes (Tables 1 and 2) or protein expression of SERCA 2 or PLB.

Structural remodeling in TNF1.6 mice. Structural remodeling of the heart (i.e., changes in mass and/or geometry) is consequent to several processes, including fibrotic remodeling of the extracellular matrix, structural alterations of individual cardiac myocytes, their displacement within the cardiac wall, and/or alternations in the abundance of nonmyocyte cardiac cells (2, 6). Morphometry of single LV myocytes isolated from TNF1.6 mice showed that compared with WT myocytes, the length of TG-F and TG-M myocytes was similarly, significantly increased (by 20%), whereas the cell width was virtually the same in all experimental groups (Table 2). These observations are consistent with morphometric changes in ventricular myocytes isolated from human failing hearts (15) and with gravimetric measurements in TNF1.6 hearts (Table 1). Hypertrophy of the LV chamber and single myocytes (Table 1 and 2) may arise, in part, from a direct hypertrophic effect of TNF- on cardiac myocytes. Such an effect has been suggested based on a substantial increase in protein synthesis (including sarcomeric proteins) in transiently cultured adult cardiac myocytes stimulated with TNF- (61). Also, a moderate (11%) increase in the myocyte cross-sectional area and a 20% increase in the cell volume has been previously shown to occur after a 15-day TNF- infusion (intraperitoneally) in rats (5). Interestingly, the degree of cardiac and cellular hypertrophy was nearly identical in TG-F and TG-M mice, despite marked differences in the contractile function of hearts in situ and isolated LV myocytes (Table 1 and 2). These findings suggest that structural remodeling of LV myocytes may precede the myocyte (or cardiac) contractile dysfunction. Hypertrophy consequent primarily to increased length, as observed in 12-wk-old TNF1.6 mice, is thought to lead to an earlier LV dilatation and contractile dysfunction compared with transverse myocyte growth (6, 50). This type of myocyte remodeling is not unique to cardiac hypertrophy and/or failure consequent to overexpression of TNF-. For example, recent studies in a hypertensive rat model of HF have shown that LV myocytes exhibit a maladaptive lengthening long before LV pump failure and are apparently devoid of an adaptive ability to normalize wall stress through transverse growth (50). More importantly, the present results show that TG overexpression of TNF- affects myocyte hypertrophy and function through different pathways because, in contrast to functional differences, the extent of hypertrophy in TNF1.6 mice was not gender dependent (Tables 1 and 2).

Functional remodeling in TNF1.6 mice. One of the important findings of the present study is that in both TG-M and TG-F hearts, changes in the functional properties of single myocytes closely corresponded to functional changes in age-matched hearts in situ. Thus these results suggest that cardiac function in young TNF1.6 mice is determined, at least in part, at the myocyte level. Specifically, compared with WT littermates, 12-wk-old TG-F mice displayed a near normal basal cardiac function, as determined by echocardiographic assessments of basal FS, but a blunted increase of FS in response to Iso (Table 1). In contrast, age-matched TNF1.6 males displayed cardiac function consistent with HF, as evidenced by a significantly reduced basal FS and virtually no increase of FS in response to Iso (Table 1). Compared with WT myocytes, TG-M myocytes exhibited a significant reduction in the amplitude and kinetics of Ca transients and contractions (Figs. 3 and 4A and Table 2). In particular, +dCa/dt_max, a sensitive index of SR Ca²⁺ release flux (57), was reduced relative to both WT and TG-F myocytes. Also, dCa/dt_max and TD_50% of the Ca transient were reduced in TG-M versus WT myocytes (Table 2). The dCa/dt_max and TD_50% of the Ca transient are thought to index the rate of SR Ca²⁺ reuptake (34), which may reflect both changes in Ca²⁺ pumping by SERCA 2 and Ca²⁺ "leak" from the SR due to defective RYR function. Relative to WT myocytes, TG-M myocytes also showed an approximately twofold reduction in the inotropic and lusitropic effects of Iso on Ca transients (Figs. 4 and 5). These changes in the configuration of the Ca transient and contractions closely correspond to those reported previously in myocytes isolated from failing human hearts (3, 10) and in experimental models of HF (21, 27).

Gender differences. Consistent with the present findings, we recently reported a more severe cardiac dysfunction and markedly higher mortality in young (<20 wk) TNF1.6 males versus females (23, 24). Whereas other laboratories have also generated mice that display HF consequent to cardiac-specific overexpression of TNF (7, 55), they did not report the sex of mice employed or whether gender differences were actually considered. Thus it is not clear whether the gender differences we observed (Refs. 23 and 24 and the present results) are unique to this line of TNF- overexpressor mice. Interestingly, a more rapid onset and/or a greater severity of cardiac dysfunction in male versus female hearts have been recently reported in other transgenic models of HF (18, 49) and in a rat model of cardiac hypertrophy (12, 56). Our previous study (24) suggested that gender differences in TNF1.6 mice may be related, in part, to the higher expression of TNF receptor transcripts and higher production of ceramide, a TNF--activated second messenger, in male hearts. Other recent studies have indicated that alterations in cardiac telomerase activity (30) or replicative capacity of cardiac fibroblasts in response to hypoxia (17) may also contribute to the more severe cardiac dysfunction in males versus females. There is also strong, although indirect, evidence that sex hormones may affect the translation and/or function of proteins implicated in cardiac structure, contractile function, and Ca²⁺ homeostasis (37, 53, 56). Nonetheless, gender differences in the context of HF remain incompletely understood. Thus the gender differences in structural and functional remodeling of hearts and myocytes characterized in the TNF1.6 mouse line suggests its potential use as a model to elucidate specific mechanisms underlying gender-dependent differences in the onset and severity of HF.