INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MYOCARDIAL COLLAGEN CONCENTRATION is elevated in chronic arterial hypertension, aortic stenosis, experimental renovascular hypertension, and genetic hypertension (2, 6, 17, 28). In view of the mechanical strength and inextensibility of collagen (19), an increased concentration of this material within the myocardium would be expected to have a significant influence on left ventricular (LV) chamber and myocardial stiffness. Studies in spontaneously hypertensive rats (SHR), which show a marked increase in myocardial stiffness and fibrosis, appear to suggest that a change in intrinsic myocardial function may be caused at least in part by alterations in the extracellular matrix (5). However, significant hypertrophy also occurs in these various models of LV pressure overload, and one could argue that myocyte enlargement also contributes to the abnormal stiffness.

Two studies designed to determine the separate influences of hypertrophy and abnormal collagen concentration on myocardial stiffness have resulted in conflicting conclusions (23, 25). Narayan et al. (23) assumed a spherical LV to calculate myocardial stiffness from LV pressure and volume data and concluded that increased collagen accumulation, but not hypertrophy, was responsible for an abnormal diastolic stiffness in the SHR. Schraeger et al. (25) used ventricular strips from SHR with and without hypertrophy to obtain tension-length curves and reported that increased collagen concentration does not affect muscle stiffness.

Others suggested that increased connective tissue would be responsible for the increased passive stiffness of hypertrophied trabecular and papillary muscles; however, they did not experimentally rule out the potential contribution of muscle hypertrophy (4, 14). It was further suggested that myocardial fibrosis may restrict myofibrillar motion and thereby impair systolic and diastolic function (29). Conrad et al. (12) observed in SHR failing hearts a reduction in tension development in association with an increased LV hydroxyproline concentration, but they did not conclude whether the myocardial dysfunction was caused by fibrosis or by a relative reduction in the number of myocytes.

On the other hand, few studies have addressed the effects of decreased collagen content without ischemia on myocardial function. Caulfield et al. (10) observed that the loss of collagen struts that interconnect myocytes had no effect on either myocyte contractility or force delivery to the ventricle. However, they did find this loss to cause a marked dilation of the ventricle and increased distensibility. Thus the purpose of this study was to analyze the relationship between LV myocardial collagen content and papillary muscle passive and active stiffness. To this end, LV papillary muscles from groups of rats with different amounts of myocardial collagen were studied.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental procedure. Thirty-three male Wistar rats were used in the study. Their care and use conformed with National Institutes of Health guidelines and the protocol was approved by the University Animal Care and Use Committee. In the first group, 10 rats (6 wk old) were anesthetized with pentobarbital sodium (50 mg/kg ip), and renal hypertension was produced by placing a silver clip around the left renal artery to constrict it to an external diameter of 0.25 mm; the contralateral kidney remained normally perfused. After a 5-wk follow-up, the rats were treated for 3 wk with the angiotensin-converting enzyme (ACE) inhibitor ramipril (20 mg · kg¹ · day¹ in drinking water; RHTR group). In the second group (GSSG group, n = 14), myocardial collagen degradation was induced using the method described by Caulfield and Wolkowicz (11). Briefly, 10 wk-old rats were anesthetized and received two intravenous infusions over 3 h (0.11 ml/min), 1 wk apart, of 20 ml of a 2 mM solution of oxidized glutathione. The animals were killed 3 wk after the second infusion, when myocardial hydroxyproline is expected to be at a minimum (11). A third group (control, n = 9) consisted of unoperated and untreated normotensive rats that were the same age as the other two groups at the end of the experiment (i.e., 14 wk old). All rats were housed in a temperature-controlled room (24°C) with 12-h light:dark cycles, and food and water were supplied ad libitum. At the end of the experiment, tail cuff systolic arterial pressure (SAP) was measured in all rats.

Isolated papillary muscle study. The animals were anesthetized with pentobarbital sodium (50 mg/kg ip), and the body weight (BW) was recorded at the time of death. The chest was opened by median sternotomy, and the heart was removed and placed in oxygenated Krebs-Henseleit solution at 28°C. The LV and septal wall were separated from the right ventricle, and their weights were determined. One papillary muscle was dissected from the LV, mounted between two spring clips, and placed vertically in a bathing chamber. The lower spring clip was attached to a Kyowa model 120T-20B force transducer by a thin (1/15,000 in.) steel wire. The upper spring clip was connected by a thin wire to a rigid lever arm above which was mounted an adjustable micrometer stop for the adjustment of unstimulated muscle length. Oxygenated (95% O₂-5% CO₂) bathing medium consisted of (in mM) 118.5 NaCl, 4.69 KCl, 2.52 CaCl₂, 1.16 MgSO₄, 1.18 KH₂PO₄, 5.50 glucose, and 25.88 NaHCO₃ dissolved in deionized water. The temperature of the bathing medium was maintained at 28°C.

2

1

2

1

Biochemical study. It has been demonstrated that hydroxyproline concentration in the LV free wall is similar to that in the papillary muscle (15). Therefore, we assumed that the hydroxyproline observed in the apex of the LV is representative of that in the entire ventricle, including the papillary muscle. We measured hydroxyproline in tissue obtained from the LV apex according to the method described by Switzer (27). Briefly, the tissue was dried for 4 h using a SpeedVac Concentrator SC 100 attached to a refrigerated condensation trap TR 100 and vacuum pump VP 100 (Savant Instruments, Farmingdale, NY). Tissue dry weight was determined, and the samples were hydrolyzed overnight at 110°C with 6 N HCl (1 ml/10 mg dry tissue). An aliquot of 50 µl of hydrolysate was transferred to an Eppendorf tube and dried in the SpeedVac Concentrator. One milliliter of deionized water was added, and the sample was transferred to a tube. One milliliter of potassium borate buffer (pH 8.7) was added to maintain stable pH, and the sample was oxidized with 0.3 ml of chloramine T solution at room temperature for exactly 20 min. The oxidation was stopped by the addition of 1 ml of 3.6 M sodium thiosulfate with thorough mixing for 10 s. The solution was then saturated with 1.5 g of KCl, and the tubes were capped and heated in boiling water for 20 min. After the tubes cooled to room temperature, 2.5 ml of toluene were added and the tubes were shaken over 5 min. The tubes were briefly centrifuged at low speed, and 1 ml of toluene extract was transferred to a 12 × 75 mm test tube. In the next step, 0.4 ml of Ehrlich's reagent was added to allow the color to develop for 30 min. Absorbencies were read at 565 nm with a double-beam spectrophotometer (A-160 spectrophotometer, Shimadzu) against a reagent blank. Deionized water and 20 µg/ml hydroxyproline were used as blank and standard, respectively.

Histology and morphometry. Transverse sections of LV were fixed in 10% buffered Formalin and embedded in paraffin. Five-micrometer-thick sections were cut from the blocked tissue and stained with hematoxylin-eosin and with the collagen-specific stain picrosirius red (Sirius red F3BA in aqueous saturated picric acid). Myocyte CSA (MA) was determined for at least 100 myocytes per slide stained with hematoxylin-eosin. The measurements were performed using a Leica microscope (×40 magnification lens) attached to a video camera and connected to a personal computer equipped with image analyzer software (Image-Pro Plus 3.0, Media Cybernetics, Silver Spring, MD). MA was measured with a digitizing pad, and the selected cells were transversely cut with the nucleus clearly identified in the center of the myocyte. Interstitial collagen volume fraction (CVF) was determined for the entire section of the heart stained with picrosirius red using an automated image analyzer (Image-Pro Plus 3.0, Media Cybernetics). The components of the cardiac tissue were identified according to their color level: red for collagen fibers, yellow for myocytes, and white for interstitial space. The digitized profiles were sent to a computer that calculated collagen volume fraction as the sum of all connective tissue areas divided by the sum of all connective tissue and myocyte areas. On the average, 35 microscopic fields were analyzed with a ×20 lens. Perivascular collagen was excluded from this analysis.

Statistics. All grouped data were expressed as means ± SD and compared by one-way ANOVA and post hoc Tukeys test. Statistical analyses were performed with SigmaStat statistical software (Jandel Scientific Software, San Rafael, CA). Differences with P  0.05 were considered significant. Straight lines were fit to the systolic tension-length relations using linear regression analysis (22). The resulting slopes corresponded to AS, and the means among the groups were compared by ANOVA. Before the diastolic tension-length relationship was compared for the three groups, the resting tension at the muscle length corresponding to 90% of L_max (L₉₀) was subtracted from all subsequent tension data in each experiment to have all tension-length curves intercepting the y-axis origin at L₉₀.

(L  L0)

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Average group values for BW, LV weight (LVW), right ventricular weight, papillary muscle CSA, SAP, and LVW normalized to BW (LVW/BW) are shown in Table 1. In the RHTR group, treatment with an ACE inhibitor for 3 wk significantly reduced systolic blood pressure from an average value of 202 ± 31 mmHg to 111 ± 11 mmHg (P < 0.001) and regressed LVW to a value comparable to the control and GSSG groups. MA was significantly higher in the GSSG group compared with control and RHTR groups (Fig. 1). CVF and hydroxyproline (Fig. 2) were statistically higher in RHTR than in the other two groups. The difference between GSSG and control groups reached a level of significance of 10% (Fig. 2A), whereas hydroxyproline was statistically lower in the GSSG compared with the control group (Fig. 2B).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Myocyte cross-sectional area in control, oxidized glutathione (GSSG), and ramipril-treated rat (RHTR) groups. Data are means ± SD analyzed by one-way ANOVA with Tukey's posttest procedure.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Collagen volume fraction (A) and hydroxyproline concentration (B) in control, GSSG, and RHTR groups. Data are means ± SD analyzed by one-way ANOVA with Tukey's posttest procedure.

The isolated papillary muscle functional parameters RT, L_max, AT at L_max, AT at L₉₀, +dT/dt, dT/dt, TPT, TR_1/2, and AS are shown in Table 2. RT was significantly higher in the RHTR group (0.64 ± 0.08 g/mm²) compared with control (0.47 ± 0.14 g/mm²) and GSSG (0.35 ± 0.10 g/mm²) groups. AT at L₉₀ and at L_max were not different among the groups.

In all experiments the relation between peak developed active tension and muscle length was linear, as evidenced by the coefficient of determination (r²), which was typically >0.94. This finding means that at least 94% of the sum of squares of deviations of AT values about their means is attributable to the linear relation between AT and muscle length (22). The slope of these linear regressions corresponds to the myocardial AS, which was significantly increased in the GSSG group compared with the control group (5.86 ± 1.14 vs. 3.96 ± 1.33 g · mm² · %L_max¹; P = 0.008). The differences between GSSG and RHTR groups and between control and RHTR groups were not statistically significant (Fig. 3).

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Left ventricular papillary muscle active and passive tension-length curves obtained from control, GSSG, and RHTR groups. Results are presented as means ± SD. The active stiffness (AS) obtained from the GSSG group was statistically higher than that from controls (P = 0.008). Statistically, no differences were observed between GSSG and RHTR groups (P = 0.493) and between control and RHTR groups (P = 0.085). AS was analyzed by ANOVA and post hoc Tukey's test. The RHTR passive tension-length curve was shifted upwards compared with either control (F = 14.25; P < 0.01) or GSSG (F = 38.8; P < 0.01) groups. The curve from the GSSG group was shifted downwards compared with the control group (F = 9.95; P < 0.01). The passive tension-length curves were fitted to a monoexponential relation, and comparisons were made by constructing an F ratio from the residual sum of squares. Statistical significance was taken to be P  0.05/k where k is the number of comparisons.

The passive tension-length curve from the RHTR group was shifted upward from that of the control group (F = 14.25; P < 0.01) and that of the GSSG group (F = 38.8; P < 0.01), reflecting an increased passive stiffness. The GSSG curve was shifted downwards from the control group (F = 9.95; P < 0.01), indicating decreased passive stiffness (Fig. 3).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In a previous study (21), we showed that renovascular hypertension induces marked myocardial hypertrophy and interstitial fibrosis. Treatment with ramipril for 3 wk did not reverse perivascular and interstitial fibrosis but fully treated the arterial hypertension and promoted regression of myocardial hypertrophy. Therefore, we used that experimental model to study myocardial function in papillary muscle from rat heart with increased collagen concentration without myocardial hypertrophy. In the present study, collagen amount was measured with CVF and hydroxyproline. It has been shown that total volume fraction is closely related to hydroxyproline concentration in the LV (28), and in our study both measurements indicated that the interstitial collagen was altered in the treated groups relative to the control rats. However, we observed that the CVF measurement was associated with a greater variation than the measurement of hydroxyproline, and, consequently, the decrease in CVF in the GSSG group came close but did not reach the level of statistical significance. The variability of CVF might be caused in part by the measurement method used. In the present investigation we used a ×20 microscope objective to obtain a large field. This magnification would detect only large perimysial collagen fibers, thereby decreasing the sensitivity of the measurement. Even so, a power analysis indicated that the difference between the GSSG and control groups would have reached the level of significance if but a few additional histological samples were available.

Despite similar LVW, the papillary muscles were significantly stiffer in the group with greater collagen concentration. This result is similar to that obtained by Narayan et al. (23) using hydralazine to prevent myocyte hypertrophy but not abnormal collagen accumulation in SHR. The collagen excess resulted in abnormally elevated passive myocardial stiffness. In contrast, Schraeger and co-workers (25) concluded that ACE inhibitor-induced regression of LV hypertrophy in SHR significantly decreased the passive stiffness of skinned trabecular muscle despite abnormally elevated hydroxyproline levels. It is not clear to what extent, if any, the 48-h incubation in the skinning solution at 0°C influenced their observations. The discrepancies observed between the studies may be caused by the animal strains as well as by the different experimental models used to produce hypertrophy and fibrosis.

In our study, using the model of presumed collagenase activation by oxidized glutathione described by Caulfield and Wolkowicz (11), it was possible to induce, in vivo, an 11% reduction of myocardial collagen concentration measured by hydroxyproline concentration and a 19% reduction in the interstitial CVF. These results are less expressive than the 30-35% reduction in collagen as reported by Caulfield et al. (10). The authors have shown that the double infusion of GSSG resulted in no visible myocyte damage at any time as examined by light microscopy and scanning electron microscope (SEM). The collagen matrix alteration was not visible by light microscopy; SEM revealed damage to the endomysium with loss of the weave that surrounds groups of myocytes and the struts that interconnect myocyte to myocyte and myocyte to adjacent capillaries, with no change in coiled perimysial fibers. These changes in the fibrillar collagen network resulted in increased ventricular volume and compliance, suggesting that damage to the intermyocyte struts and to the weave complex might be more important than the decrease in myocardial collagen. Other studies have shown that the double infusion of GSSG in rats promotes a reduction in CVF, ventricular dilatation, and a shift to the right of the diastolic pressure-volume curve of the entire LV (18, 20). However, a similar effect in the papillary muscle preparation has not been studied previously. The main advantage of this preparation is that the muscle force and length are directly measured and that the mathematical assumptions required when myocardial mechanical characteristics are evaluated in the LV chamber are unnecessary.

The study of cardiac function in the whole heart is based on the pressure-volume and stress-strain relationships. In that condition, myocardial stiffness is derived from chamber measurements using mathematical models and assumptions regarding LV shape. If the LV is assumed to be a thick-walled sphere, the stress will be underestimated (30), whereas the assumption of an ellipsoid shape would result in an overestimated wall stress (7). Therefore, isolated muscle experiments provide descriptions of myocardial behavior without the influence of chamber and wall geometry. In our study, the diastolic tension-length relations obtained for the three groups were different from each other, showing that the changes in collagen content, measured by hydroxyproline and CVF, are associated with myocardial passive properties. Compared with the control group, the diastolic tension-length curves were significantly shifted upwards and to the left in the RHTR group and downwards and to the right in the GSSG group. Therefore, our results allow us to conclude that the decreased passive stiffness in the GSSG group strongly correlates with the fibrillar collagen loss and that increased collagen content strongly correlates with the elevated passive stiffness observed in the RHTR group. Previous studies have suggested that collagen cross-linking (9) may affect myocardial stiffness, regardless of collagen amount. In addition to the effect of altered collagen amounts, it is important to be mindful of the effects of collagen crosslink density, as well as collagen type (type I or III) and collagen distribution. At present, we cannot rule out that changes in the collagen characteristics might also have influenced myocardial stiffness in the present study. Nevertheless, the results clearly indicate that alterations in collagen concentration and papillary muscle function are correlated.

Ventricular elastance and myocardial stiffness are indexes of contractility of the ventricular chamber and myocardium, respectively (8, 26). Elastance is the ratio of the change in peak isovolumetric pressure for a given change in volume, and stiffness is defined as the ratio of the change in active force related to change in muscle length (8). Myocardial contractility is a very complex property of the heart that is difficult to measure directly. During the last two decades it has been proposed that an ideal index of myocardial contractility must be able to measure the ability of the myocardium to generate force independently of loading condition. The slope of the linear pressure-volume relationship in the isolated canine heart has been shown to be relatively independent of preload and afterload and therefore has been used as an index of contractility (26).

Using the slope of active tension-length (active stiffness) as an index of myocardial contractility, we have shown an enhancement of active stiffness when the muscle is stretched from 90% to 100% of L_max in the GSSG group. The mechanisms underlying the association between decreased myocardial collagen and enhanced active stiffness are not well established, and the results presented in this study do not answer all the questions concerning this matter. When collagen is reduced, ventricular dilatation occurs (10) and myocyte hypertrophy takes place in response to alterations in the loading state of the ventricle (17). Therefore, myocyte hypertrophy might play an important role in the improvement in contractility observed in the GSSG group. Another explanation would be related to the intracellular glutathione metabolisms. The glutathione level in the heart is ~1.2 µM/g (16), mainly in the reduced form, GSH, because of the high activity of GSSG reductase (13). That means that, inside the cell, most of the infused GSSG was rapidly converted to GSH. The action of excess GSH or GSSG in the heart is not completely elucidated. Bauer et al. (3), working on fiber bundles from papillary muscle of porcine right ventricle, observed an increased sensitivity of contractile protein to calcium and, consequently, an increased force development in the presence of GSH. In our study, considering that the half-life of the glutathione is only a few minutes (1), it is doubtful that the double infusion of oxidized glutathione might increase the glutathione level in the cardiac tissue after 3 wk. Nevertheless, this is a very complex matter that requires further study. Active tension and active stiffness in papillary muscles from RHTR rats were similar to those in the control rats, suggesting that regression of hypertrophy by treatment with an ACE inhibitor is associated with preserved myocardial contractility.

We conclude that decreased collagen content induced by GSSG is associated with myocyte hypertrophy, decreased passive stiffness, and increased active stiffness. Abnormally high collagen concentration correlates with myocardial diastolic dysfunction and has no relation with systolic function.