Inhibition of Collagen Synthesis With Prolyl 4-Hydroxylase Inhibitor Improves Left Ventricular Function and Alters the Pattern of Left Ventricular Dilatation After Myocardial Infarction
Background Left ventricular (LV) remodeling after myocardial infarction (MI) is associated with fibrosis, dilatation, and dysfunction. We postulated that prevention of fibrosis after MI with a prolyl 4-hydroxylase inhibitor (P4HI) would preserve LV function and attenuate LV enlargement.
Methods and Results Adult female rats (200 to 250 g) had experimental MI and were then randomized to treatment with P4HI (MI-FG041, n=29) or vehicle (MI-control, n=29) 48 hours after MI for 4 weeks in 2 phases. Echocardiograms were performed weekly with a 15-MHz linear transducer, and at 4 weeks, collagen isoform determinations and in vivo hemodynamics were performed. At randomization, the infarct size and LV function and size were similar in MI-FG041 and MI-control but significantly different from shams (n=9). At week 4, the LV function in MI-FG041 was significantly better than in MI-controls (fractional shortening 21% versus 16%, P=0.01; fractional area change 30% versus 19%, P=0.002; ejection fraction 35% versus 23%, P=0.001). In the FG041 group, LV area in systole was less (P<0.05), the dP/dtmax after isoproterenol was higher (P<0.05), and types I and III collagen in noninfarcted LV were less than in MI-control. The hydroxyproline/proline ratio was increased by 64% in MI-control and reduced to the sham value in MI-FG041 rats. In the scar tissue, it was reduced by 24% in MI-FG041.
Conclusions This study demonstrates that prevention of interstitial fibrosis with a P4H inhibitor alters the pattern of LV enlargement and produces partial recovery of LV function after MI.
Received May 29, 2001; revision received July 31, 2001; accepted July 31, 2001.
Left ventricular (LV) remodeling after myocardial infarction (MI) involves disruption of supporting structures, cell slippage, myocyte hypertrophy, and collagen deposition both at the site of infarction and at areas remote from the infarct. This process is adaptive initially but may progress to LV dilatation and dysfunction.1 Interstitial fibrosis of the viable myocardium is believed to be important in ischemic cardiomyopathy2 and may contribute to LV dysfunction after MI in part by causing morphological and functional separation of the myocytes.3,4 Interruption of fibrosis of the viable myocardium may produce favorable effects on cardiac function and arrhythmogenesis.5
Conversely, it has been suggested that disruption of extracellular matrix collagen may be a primary mechanism for LV dilation,6 and inadequate or disrupted fibrillar collagen at the site of infarction may lead to myocardial rupture.7 Of note, however, are studies using ACE inhibitors and angiotensin receptor blockers that showed thinning of the infarcted segment, reduction of fibrosis in the noninfarcted myocardium, attenuation of LV cavity dilatation, and improvement in LV function.8–11 These varied experimental findings suggest the complexity of LV remodeling.
It appears that prevention of fibrosis in the noninfarcted myocardium is feasible, thus allowing an assessment of whether the benefit of preventing fibrosis outweighs the potential adverse effects of collagen scar reduction. Prolyl 4-hydroxylase (P4H) is an essential enzyme in collagen biosynthesis. It catalyzes the hydroxylation of specific proline residues on alpha monomers, resulting in the production of thermally stable triple helical procollagen molecules and their subsequent secretion into the extracellular matrix. Because the final common pathway for collagen formation involves the activity of P4H, inhibition of this enzyme after MI would be expected to prevent interstitial fibrosis.
FG041 is an orally available P4H inhibitor. To test the hypothesis that a P4H inhibitor could prevent cardiac fibrosis and produce favorable effects on LV function and structure after MI, we administered FG041 to rats for 4 weeks beginning 48 hours after experimental MI and evaluated collagen deposition as well as structural and functional changes in the LV.
All studies described in this article were performed according to the guidelines of the American Physiological Society and were approved by the Institutional Animal Care Committee of the University of Illinois at Chicago.
One hundred thirty-five female Wistar rats weighing 200 to 250 g at the onset underwent MI, and 9 rats (shams) underwent sham operation. Of the 135 rats, 74 survived after surgery, and at 48 hours after MI, 2D echocardiography (2DE) was performed. All animals with fractional shortening (FS) <20% and regional wall motion abnormalities (58 rats) were randomized to treatment with FG041 (MI-FG041, n=29) or vehicle (MI-control, n=29), and the rest were excluded from study. Shams were also randomized to FG041 (n=4) or vehicle (n=5). The study was conducted in 2 phases. In phase 1, 14 of 29 MI-FG041 and 12 of 29 MI-control matched rats underwent weekly echocardiography, and at week 4, in vivo LV function and levels of hydroxyproline (Hyp) and proline were determined. In phase 2, 15 of 29 MI-FG041 and 17 of 29 MI-control matched rats underwent weekly 2DE to assess LV area, systolic function, and mitral E-wave velocity, defined as the early mitral filling wave. In each case, treatment was started immediately after randomization and continued for 4 weeks.
Production of MI
An anterior thoracotomy was performed on anesthetized and intubated rats, and the hearts were exteriorized. The left coronary artery was then ligated proximally with a 6-0 silk suture, and the chest wall was closed in layers as previously described.12 Shams underwent identical surgery without ligation of the coronary artery.
Animals lightly anesthetized with ether were placed in a supine or left lateral position with ECG electrodes applied at the paws. A 15-MHz linear transducer and the Acuson Sequoia System (C256) were used to obtain 2DE images. In phase 1, 2DE short-axis images at the mid papillary muscle level were obtained for measurement of LV end-diastolic dimension (LVEDD), LV end-systolic dimension (LVESD), FS, and wall thickness.
In phase 2, short-axis images, apical 4-chamber views, and mitral E-wave velocity were obtained. FS, fractional area change (FAC), LV ejection fraction (LVEF), LV area in diastole (LVAD), LV area in systole (LVAS), and infarct size were determined. For infarct size, short-axis digital images were slowed, and the cursor was placed at the dyskinetic points. The images were then frozen at end diastole and traced. The infarcted segment was expressed as a percentage of the LV area. Measurements were made offline by 2 observers blinded to treatment according to the American Society for Echocardiography leading-edge method.
FG041 was administered by oral gavage twice daily at a dose of 50 mg/kg beginning 48 hours after production of MI and was continued for 4 weeks. This approach was taken to allow for all surgical mortality and to achieve maximal collagen inhibition while at the same time limiting possible adverse effects on wound healing. Serum levels of FG041 were measured.
After 4 weeks of treatment, animals were anesthetized and ventilated. The right carotid artery was cannulated with a Millar ultraminiature pressure transducer (SPR-671). The catheter was then advanced into the LV. After steady state had been established, baseline heart rate, developed pressure, contractile index, systolic and end-diastolic pressures (LVEDP), and maximal rates of pressure rise and fall (±dP/dt) were recorded. Hemodynamic measurements were repeated after a bolus infusion of isoproterenol (0.2 mL of 10−7 mol/L solution over 1 minute) via the femoral vein.
Hearts were excised, weighed, and sliced into 3 transverse sections from the apex to the base. In selected hearts, transverse sections were saved for subsequent histological staining. Otherwise, the scarred tissue was grossly dissected out from the noninfarcted tissue, because the scar was well formed. All the scarred tissue, a section of the right ventricle opposite the septum, and the LV opposite the infarcted region were collected in each of the animals, weighed, and used for Hyp assay. Hyp and proline were determined by the method of Palmerini et al,13 except that l-azetidine-2-carboxylic acid (Aldrich) was substituted for 3,4-dehydroproline as the internal standard.
Western blotting was performed by standard methods. Collagen isoforms were detected by use of primary antibodies against types I and III collagen (Santa Cruz). The bound antibodies were detected by chemiluminescence (Amersham), and the immunoblots were exposed to Hyperfilm (Kodak) for 5 to 30 seconds.
The noninfarcted portion of the LV was dissected free of the scar and was frozen in liquid nitrogen. Subsequently, 5-μm cryostat sections were cut from the basal, mid, and apical regions of the heart. The sections were separately stained with Masson’s trichrome to distinguish areas that represented connective tissue. For image acquisition, slides containing the histological sections for hearts in the untreated and treated MI groups were scanned in color at a resolution of 72 dots per inch with a Nikon LS-3510 film scanner,14 resulting in consistent magnification from image to image. Images were enhanced with Adobe Photoshop 6.0. The areas for each section that suggested connective tissue staining were traced, and the total areas (pixels) were calculated for each section.
Values are given as mean±SEM except where indicated. Differences in Hyp, proline, and echocardiographic and hemodynamic parameters between MI-FG041, MI-control, and sham were initially compared by Levene’s test for equality of variance followed by t test for equality of means with the SPSS statistical program. Values of P<0.05 were considered significant.
At the time of randomization, the MI-FG041 and MI-control animals had similar LV function, internal dimensions, and wall thickness. No differences were detected in the echocardiogram, hemodynamics, and Hyp content between FG041-treated shams and shams that received only the vehicle (sodium methyl cellulose); hence, shams were grouped together in the analysis. Within the period of the study, 7 of 29 MI-control animals (24%) died between days 7 and 28 of MI, and 3 of 29 MI-FG041–treated animals (10%) died between days 14 and 28 of MI. The cause of death was most likely related to progressive heart failure, because autopsy revealed congested lungs, pleural effusions, and markedly dilated right and left ventricles without gross evidence of LV rupture.
In phase 1 of this study, the baseline (value at time of randomization) FS, LVESD, and LVEDD of the MI-FG041 and MI-control animals were similar but significantly different from those of shams (P<0.001, see Table 1). The FS of MI-FG041–treated animals increased from 10% at baseline to 21% at week 4 and was essentially unchanged in the MI-control and sham animals (see Table 1 and Figure 1).
The LVESD was significantly less in MI-FG041 animals at weeks 1 to 3 (P<0.05, see Table 1) but was no longer significant at week 4, even though the LVESD remained unchanged. The LVEDD, conversely, was significantly less only at week 1 (8% versus 17%, P=0.008, see Table 1). Distortion of the LV as a result of the infarction, however, may affect simple internal dimension measurements.
In phase 2, the FAC, LVEF, LVAS, LVAD, infarct size (52% versus 52%), and mitral E-wave peak velocity at randomization were similar in MI-FG041 and MI-control rats. The LVEF was significantly greater in the MI-FG041 than MI-control at all time points (P<0.01, Figure 2). Similarly, the FAC was significantly better in the MI-FG041 than MI-control (P<0.05, Table 2).
At week 4, the LVAS, a more accurate assessment of LV size, was significantly less in the MI-FG041 animals than MI-control in both the short-axis and 4-chamber views (Table 2). Similarly, the LVAD was significantly less in MI-FG041 than MI-control in the short-axis view but was not statistically different in the 4-chamber view (Table 2). Both MI-FG041 and MI-control showed a restrictive filling pattern in the mitral inflow, with the peak E-wave velocity being similar. Little or no detectable A-wave velocity was present. Significant mitral regurgitation, however, was present in most of the MI animals.
The baseline (preisoproterenol) heart rate, developed pressure, contractile index, and systolic blood pressure were similar in sham, MI-control, and MI-FG041–treated animals. The −dP/dt, however, was significantly higher in the sham and MI-FG041–treated animals than in control animals (P<0.05, see Table 3). After isoproterenol infusion, the +dP/dt was significantly increased in the MI-FG041–treated animals compared with MI-control animals (16 830 versus 13 832 mm Hg/s, P<0.05, Table 3). LVEDP (mm Hg) in the control MI group was 17±6 before isoproterenol infusion and 16±6 after. In contrast, in the MI-FG041 group, LVEDP was 9±4 before and 6±4 after (P<0.07, MI-control versus MI-FG041).
Collagen Expression and Histology
There was a 64% increase in the Hyp-to-proline (Hyp/Pro) ratio of the noninfarcted LV in the MI-control relative to shams (see Table 4), whereas there was no difference between MI-FG041–treated animals and shams. This represents a relative 86% reduction (P<0.01) in MI-FG041 relative to MI-control. In the right ventricle, there was a 25% increase in the Hyp/Pro content of the MI-control animals relative to shams, whereas that of MI-FG041–treated animals was the same as in the shams.
The weights of the scar tissue in the infarcted area of both MI-control and MI-FG041 animals were similar (93 versus 91 mg, respectively). There was a significant, 24% reduction of the Hyp/Pro ratio, however, in the scars of the MI-FG041 animals compared with MI-control (P<0.05). Both types I and III collagen isoforms were significantly reduced in the noninfarcted portion of the LV compared with MI-control. A representative gel is shown in Figure 3. Representative histological sections of the hearts of animals with similar infarct size at randomization revealed less ventricular cavity dilatation and better myocardial structure preservation in the MI-FG041–treated animals (see Figure 4, A and B). In addition, a semiquantitative histological analysis of collagen content in MI-control and MI-FG041 animals with matched infarct size showed a marked reduction in collagen in the drug treatment group (Figure 4, C through E).
The mean peak plasma level of FG041 in MI-FG041 animals determined 2 hours after gavage was 7.9 μg/mL. The trough plasma level determined 8 hours after administration was 3.44 μg/mL. These levels were sufficient to inhibit P4H.
This study demonstrates that prevention of interstitial fibrosis in the noninfarcted myocardium alters the pattern of LV enlargement and produces partial recovery of LV function after MI. The effect observed in this study is attributed to reduction of extracellular matrix collagen, because the P4H inhibitor had no afterload-reducing effect.
Collagen Formation After MI
Angiotensin II activation increases mRNA expression of type I and III collagen and synthesis of collagen after myocardial injury.15,16 FG041 treatment reduced collagen expression in the noninfarcted LV to the level found in shams but only by 24% at the site of the infarct. This is not unexpected, because diminished blood supply in the area of the infarct should restrict drug delivery. The result was a relatively normal healing process at the site of the infarct combined with prevention of fibrosis of the noninfarcted myocardium. The scar tissue was of similar weight in both groups, even though there was less Hyp in the MI-FG041 (Table 4). This suggests the presence of underhydroxylated collagen or a relative increase of another protein(s) in the scar tissue.
Because matrix metalloproteinases (MMPs), which degrade collagen, have been reported to increase 4.5-fold at day 2 after MI17 and collagen reportedly increases 3-fold in the first week in rats,18 it appears that a balance between collagen degradation and deposition must occur to prevent rupture of the ventricle. It is possible that FG041, by decreasing collagen deposition, may also have decreased activation of MMPs by indirect negative feedback, thus maintaining a balance between degradation and deposition of collagen. To establish this, however, requires direct measurement of MMP activity at the site of infarction in the presence of FG041.
Experimental MI in Rats
After MI in rats, systolic function declines and the LV cavity dilates significantly, and these changes progress over time. We noted improved systolic performance after P4H inhibition. This improvement was quite dramatic and was sustained from week 1 to week 4, suggesting that fibrosis in the noninfarcted myocardium after MI contributes to LV dysfunction.
The early increase in LV cavity dimension in MI-control animals in this study is consistent with previous studies.19,20 In contrast, the MI-FG041–treated animals had an altered pattern of LV dilatation with less progressive cavity dilatation. This finding should be contrasted with previously published work. Specifically, Zhao and others6,21 suggested that collagen matrix disruption is a primary factor involved in LV dilatation during ischemic cardiomyopathy, whereas Rohde et al22 showed that MMP inhibition, which preserved extracellular collagen, decreased early dilatation of the LV after MI in mice. In addition, a number of studies using ACE inhibitors and angiotensin receptor blockers have also shown reduction of fibrosis at the noninfarcted myocardium, attenuation of LV enlargement, and improved LV function.8–11 This underscores the complexity of LV remodeling after MI. In the present study, early LV enlargement was attenuated despite decreased Hyp; hence, prevention of fibrosis in the viable segments allowed for an overall favorable effect on LV function and geometry that counteracted any potential adverse effect of scar-tissue reduction.
Although the LV was not uniform in shape and we did not perform simultaneous echocardiography and hemodynamic measurements, the estimated LV end-systolic wall stress at the end of 4 weeks23,24 was 35% lower in the MI-FG041–treated animals than the MI-control animals. It is conceivable that the attenuation of LV cavity dimension reduced wall stress and improved LV function.
The systolic thickness of the noninfarcted myocardium was 2.3 mm at week 4 in MI-FG041 compared with 2.0 mm in MI-control, suggesting enhanced contractility in this segment. Enhanced contractility was also evident in the MI-FG041–treated animals after isoproterenol infusion. We postulate that this is a principal mechanism of improved LV function in this setting. By inhibition of fibrosis, myocyte hypertrophy may be unimpeded and there may be less myocyte separation. This could allow for more optimal myocyte performance. This can theoretically be tested by looking at mechanical performance in cardiocytes isolated from treated and untreated cardiac muscle.
Limitations of the Study
Only female rats were used in this study for logistic reasons, and the conclusions should be limited to this sex. Further studies are needed to determine and compare the efficacy of this agent in males, especially because estrogen and progesterone have been shown to inhibit cardiac fibroblast activity and collagen synthesis in the aorta.25,26 The findings apply to animals with large infarcts, and a larger study is needed to confirm the results in animals with smaller infarcts.
There appears to be potential value in the use of agents that inhibit collagen synthesis in the period immediately after MI. Although the drug effect is, in theory, systemic, the major impact appeared to be restricted to the noninfarcted myocardium. The findings in this study are preliminary, and further studies are needed to address mechanisms, establish safety issues, and define the target population.
We wish to acknowledge the support of the program in cardiovascular science at the University of Illinois at Chicago and NIH grants HL62426 and HL63704 (to P.B.). Weihua Zhang provided technical assistance. Robert Folberg, MD, provided invaluable help in the evaluation of tissue histology.
Francis GS, McDonald KM, Cohn JN. Neurohumoral activation in pre-clinical heart failure: remodeling and potential for intervention. Circulation. 1993; 87 (suppl IV): IV-90–IV-96.
Beltrami CA, Finato N, Rocco M, et al. Structural basis of end-stage failure in ischaemic cardiomyopathy in humans. Circulation. 1994; 89: 151–163.
Jugdutt B, Khan M, Jugdutt J, et al. Effect of enalapril on ventricular remodeling and function during healing after anterior myocardial infarction in the dog. Circulation. 1995; 91: 802–812.
De Carvalho-Frimm C, Sun Y, Weber KT. Angiotensin II receptor blockade and myocardial fibrosis of the infarcted rat heart. J Lab Clin Med. 1997; 129: 439–446.
Geenen DL, Malhotra A, Liang D, et al. Ventricular function and contractile proteins in the infarcted overloaded rat heart. Cardiovasc Res. 1991; 25: 330–336.
Weber KT. Hormones and fibrosis: a case for lost reciprocal regulation. News Physiol Sci. 1994; 9: 123–128.
Weber KT. Extracellular matrix remodeling in heart failure: a role for de novo angiotensin II generation. Circulation. 1997; 96: 4065–4082.
Jugdutt BI, Joljart MJ, Khan MI. Rate of collagen deposition during healing and ventricular remodeling after myocardial infarction in rats and dog models. Circulation. 1996; 94: 94–101.
Litwin SE, Katz SE, Morgan JP, et al. Thrombosis/thrombolysis/vasodilator responses: serial echocardiographic assessment of left ventricular geometry and function after large myocardial infarction in rats. Circulation. 1994; 89: 345–354.
Ono K, Matsumori A, Shio T, et al. Cytokine gene expression after myocardial infarction in rat hearts: possible implication in left ventricular remodeling. Circulation. 1998; 92: 149–156.
Pfeffer JM, Pfeffer MA, Fletcher PJ, et al. Progressive ventricular remodeling in the rat with myocardial infarction. Am J Physiol. 1991; 260: H1405–H1414.
Rohde LE, Ducharme A, Arroyo L, et al. Matrix metalloprotein inhibition attenuates early left ventricular enlargement after experimental myocardial infarction in mice. Circulation. 1999; 99): 3063–3070.
Grossman W, Jones D, McLaurin LP. Wall stress and patterns of hypertrophy in the human left ventricle. J Clin Invest. 1975; 56: 56–64.
Dubey RK, Gillespie DG, Jackson EK, et al. 17β-Estradiol, its metabolites and progesterone inhibits cardiac fibroblast growth. Hypertension. 1998; 31 (1 pt 3): 522–528.