Myocardial Stiffness Is Attributed to Alterations in Cross-Linked Collagen Rather Than Total Collagen or Phenotypes in Spontaneously Hypertensive Rats
Background The relative contributions of increases in myocardial collagen, collagen cross-linking, and the ratio of type I to type III collagen to the stiff myocardium in hypertension were determined.
Methods and Results We compared the action of hydralazine (0.07 mmol · kg−1 · d−1) with that of captopril (0.22 mmol · kg−1 · d−1) on the left ventricular end-diastolic (LVED) myocardial stiffness constant, k (g · cm−2) and LV myocardial interstitial characteristics in spontaneously hypertensive rats (SHRs) and Wistar Kyoto (WKY) control rats. LVED k (SHR, 27.9±1; WKY, 19.5±1.2; P<.01), myocardial hydroxyproline concentrations (HPRO; μg/mg dry wt) (SHR, 4.19±0.16; WKY, 3.17±0.09; P<.001), and collagen type I/III ratios (SHR, 7.1±0.7; WKY, 2.1±0.2; P<.001) were increased, whereas the percentage of myocardial collagen extracted after cyanogen bromide digestion (an index of cross-linked collagen) was decreased (SHR, 17±3; WKY, 41±4; P<.001) in SHRs compared with WKY controls. Captopril therapy reduced LVED k, myocardial HPRO, collagen type I/III, and augmented collagen solubility (43±4) in SHRs to values similar to those measured in WKY controls. Hydralazine therapy, despite a favorable effect on LVED k in SHRs (20.±1.6, P<.01 compared with untreated SHRs), failed to influence either myocardial HPRO (4.18±0.18) or collagen type I/III (8±1) but did improve collagen solubility (31±2).
Conclusions An association between alterations in LVED k and collagen solubility but not between changes in LVED k and total collagen or phenotype ratios after antihypertensive therapy in SHRs suggests that myocardial stiffness in hypertension is the consequence of an enhanced myocardial collagen cross-linking rather than of an increase in total collagen or type I phenotype concentrations.
Alterations in LV diastolic performance precede changes in systolic performance in hypertensive patients with LVH.1 A decrease in LV diastolic function is thought to contribute to cardiac failure in hypertensive heart disease.2 A number of myocardial alterations may contribute to a decline in LV diastolic performance in hypertension, including aberrations in the extracellular matrix.3
The interstitial properties that may be involved in modulating diastolic performance in hypertension include an increased myocardial total collagen content4 and the amount of the collagen subtype with a high tensile strength, type I, relative to the other major subtype, type III.5 Despite the evidence in support of a role for these interstitial alterations of collagen contributing to a decrease in LV performance at end diastole,4 there are reports that question the importance of these postulated mechanisms in producing an enhanced myocardial diastolic stiffness in hypertension. Indeed, previous groups have shown that changes in total myocardial collagen6 or collagen phenotypes7 are not always associated with an augmented myocardial stiffness.
The importance of the contribution made by aberrations in myocardial collagen to myocardial stiffness at end diastole has recently been underscored. β-Aminopropionitrile, which inhibits collagen cross-linking and subsequently decreases total myocardial collagen, results in an attenuation of myocardial stiffness constants as measured at end diastole.8 However, it may be argued that a decline in collagen cross-linking, as opposed to myocardial collagen concentration, is responsible for the diminished myocardial stiffness after β-aminopropionitrile administration. Indeed, in volume-overload cardiac disease, an increased myocardial stiffness is associated with a decrement in myocardial collagen solubility (an index of collagen cross-linking) rather than with an enhanced total myocardial collagen or changes in collagen phenotypes.9 Therefore, it is possible that alterations in collagen cross-linking rather than changes in either total collagen or phenotypes contribute to the stiff myocardium of hypertension. Hence, we decided to examine the relative contribution of abnormalities of either total myocardial collagen, collagen phenotypes, or collagen solubility to myocardial stiffness in hypertension.
Certain antihypertensive agents, such as the ACEI, can,10 11 whereas others, including hydralazine, cannot11 attenuate those myocardial extracellular collagen matrix changes described in experimental hypertension. The contrasting effects of these agents on myocardial interstitial characteristics provide an ideal opportunity for examining the importance of myocardial collagen alterations in contributing to a hypertension-induced decrease in diastolic performance. In our present study, we therefore compared the effect of hydralazine and captopril therapy on myocardial stiffness, total collagen, collagen phenotypes, and collagen solubility in SHRs.
Experimental Groups and Noninvasive Blood Pressures
The experimental procedures were approved by the Animal Ethics Committee of the University of the Witwatersrand (certificate number 93/19/2B). Thirty-four male, 23-week-old SHRs (OLAC, Blackthorn, UK) and 35 male, 23-week-old WKYs (Kleintierfarm Madorin Ltd, Germany) were randomly assigned to 6 different groups (3 SHR and 3 WKY). Captopril (SQ 14,225, Squibb) at a dose of 0.22 mmol · kg−1 · d−1 was administered in the drinking water to one group of SHRs (n=10) and one group of WKYs (n=12), and hydralazine HCl (Lennon Pharmaceuticals) at a dose of 0.07 mmol · kg−1 · d−1 was administered in the drinking water to a second group of SHRs (n=11) and WKYs (n=12) for 22 weeks. The third group of SHRs (n=13) and WKYs (n=13) were left untreated for the duration of the experiment. SBP was measured immediately before the commencement of either captopril or hydralazine therapy and thereafter at regular intervals for the 22-week period in all rats as previously described.12 Rats were trained in stocks for 3 days before the first measurement of SBP to allow them to adapt to the procedure. Each reading was taken at midday to avoid diurnal variation.
Cardiac Function and Geometry
At the end of the experimental period, when rats were 45 weeks of age, LV performance was successfully measured in all rats. The surgery, instrumentation, experimental techniques, and calculations used in our present study to measure LVED performance in anesthetized, open-chest, ventilated rats have previously been described and validated.13 14 Briefly, rats were anesthetized with a total of 0.05 mg fentanyl and 2.5 mg droperidol (Janssen Pharmaceuticals). A carotid catheter was inserted for the measurement of blood pressure and heart rate, and positive pressure ventilation was initiated with a constant volume respirator (Harvard Apparatus) before a thoracotomy was performed. After a midline thoracotomy and a parietal pericardectomy, a 21-gauge needle attached to a saline-filled PP25 polyethylene catheter coupled to a Gould P50 pressure transducer was inserted through the apex of the heart for the measurement of LVEDP and LVESP. LVEDDs were measured by use of piezoelectric transducers attached to an apparatus designed and validated in our laboratory.13 14
To examine the accuracy of our pressure measurements, we determined the swept-sine frequency response of the LV catheter coupled to its pressure transducer–dome combination. The frequency responses were recorded against a reference calibrated piezoresistive differential pressure transducer (Honeywell Microswitch 170PC). Identically preamplified (Hellige Servomed) catheter and reference transducer outputs were analyzed for amplitude and phase differences obtained by averaging three repeat measurements over 10 seconds at 1-Hz intervals from DC to 25 Hz (Hewlett Packard Dynamic Signal Analyzer 3562). The amplitude-frequency response was found to be uniform to 10 Hz.
LVEDDs, LVESDs, LVEDPs, and LVESPs were measured over a range of filling volumes during slow and controlled as opposed to instantaneous, complete IVC occlusion. We avoided data collection during instantaneous, complete IVC occlusion because this produced frequent ventricular ectopic beats in untreated SHRs. Before IVC occlusion, a baseline LVEDP of between 10 and 15 mm Hg was obtained by injecting a modified dextran 70 solution12 through the arterial line. If IVC occlusion failed to reduce LVEDP values to 0 mm Hg without resulting in changes in heart rate or inducing ventricular ectopic beats, blood was withdrawn until LVEDP reached a value close to that measured before dextran 70 infusion, and IVC occlusion was repeated. IVC occlusion was repeated at least three times at each baseline filling pressure to ensure reproducibility. Typical schematic examples of short-axis-diameter measurements obtained during rapid, complete IVC occlusion have been published.13 14 Fig 1⇓ shows a portion of an original trace of the LVD and LVDP measurements, obtained at a low chart speed, during slow and controlled IVC occlusion in an untreated SHR in our present study.
After sufficient hemodynamic data had been collected on each rat, the piezoelectric transducer resting on the anterior wall of the heart was sutured to the wall with an opthalmic 7-0 vicryl suture to ensure that the transducer followed the heart closely during ventricular ejection. Because we noted no significant differences in the slope of the relations of LVESP and LVESD (measured during IVC occlusion) obtained before and after suturing in 8 rats in which no change in LV peak systolic pressure occurred during suturing, we were confident that the unsutured anterior wall piezoelectric transducer followed the wall during ventricular ejection. Systolic performance was therefore determined from values obtained before suturing.
LVED myocardial stiffness was calculated from the linearized slope of the relationship between LVED stress and strain.13 14 LVED stress and strain were calculated from LVEDD and LVEDP measurements as follows: Stress=[1.36×LVEDP×(2r)2]/LVEDD2−(2r)2 and strain=(LVEDD−LVEDD0)/LVEDD0, where LVEDD0 is the unstressed LVEDD and internal radius at end diastole (r)= , where Vm (LV wall volume)=0.943×LV wet weight. A thick-walled spherical model of LV geometry was assumed in the calculation of wall stress.15 LV wall thickness (h) values were calculated from the formula h=(LVEDD−2r)/2. Cardiac systolic performance was determined from the slope of the LVES stress versus strain relation (LVES myocardial stiffness).16 End-systolic stress and strain were calculated from the same equations as described for end-diastolic stress and strain except that values for end-systolic pressures and diameters replaced end-diastolic pressures and diameters.
Myocardial Hydroxyproline, Collagen Phenotypes, and Solubility
Samples of LV tissue from all SHRs and untreated WKY controls and a random sample of 8 captopril-treated and 6 hydralazine-treated WKY controls were weighed and stored at −70°C for tissue analysis. Myocardial hydroxyproline content was determined by the method of Stegemann and Stalder.17
Myocardial collagen was extracted and digested with CNBr according to the procedure described by Mukherjee and Sen.5 Polyacrylamide gel electrophoresis was subsequently performed on vertical gels 1.5 mm thick (Min-Protean II, BioRad Laboratories) by stacking and separating gel concentrations of 3% and 12.5%, respectively. Samples of 15 μg of collagen (determined by hydroxyproline measurements of the supernatant obtained after CNBr digestion) were loaded with a microsyringe. When electrophoresis was complete, gels were stained in a fixative (50% water, 40% ethanol, 10% acetic acid) containing 0.1% Coomassie blue (R250) and destained in the fixative alone for several hours. Gels were scanned with a Helena Laboratories EZ scan.
Electrophoretic patterns of myocardial collagen extracts similar to those previously published were obtained, and the bands corresponding to collagen types I and III were identified from collagen type I (Sigma) and III (Calbiochem) standards. The relative amounts of type I:III collagen were determined from the relationship between the quantity of collagen applied to the gel and the relative area under the densitometry curve corresponding to bands G [α1(I)−CB-8] and H [α2(I)−CB-3] (type I)18 and band M [α1(III)−CB-5 plus α1(III)−CB-9] (type III).5 18 Bands G, H, and M were chosen because they contain very little interference from comigrating peptides of other collagen types. To validate the densitometry techniques used to calculate the ratio of type I:III collagen, we correlated the areas measured under the densitometry curves obtained from scanning polyacrylamide gels containing incremental amounts of collagen standards with the amount of collagen loaded onto the gel. The R2 for collagen type I was .92 and for collagen type III, .94.
The percentages of total collagen, type I collagen, and type III collagen extracted after CNBr digestion were expressed as percentages of the collagen extracted after acid hydrolysis alone. We used this as a measure of tissue collagen solubility and subsequently as an index of the degree of collagen cross-linking.19 To differentiate between alterations in collagen cross-linking as opposed to oxidation of methionine residues resulting in changes in the solubility of myocardial collagen to CNBr digestion, we also measured the hydroxyproline content of CNBr-digested tissue after β-mercaptoethanol–induced methionine reduction.20 We chose to examine the degree of collagen cross-linking using CNBr digests as opposed to neutral salt precipitation10 or pepsin digests, because both neutral salt21 and pepsin5 result in low myocardial collagen yields in rats.
Regression analysis was used to determine the lines of best fit for the cardiac function relations. A value of r>.95 was considered to be a good fit. The relations of LVEDP versus LVEDD and LVED stress versus LVED strain were found to best fit exponential functions: LVEDP (or LVED stress)=b×e(m×LVEDD or LVED strain), which were linearized: ln LVEDP (or LVED stress)=ln b+m(LVEDD or LVED strain) for statistical analysis. The LVES stress-versus-strain relation was found to best fit a linear function, LVES stress=m(LVES strain)+b. Multiple comparisons of the slopes of the cardiac function relations, heart weight, h/r, myocardial hydroxyproline concentrations, the ratio of type I to III collagen, and myocardial collagen solubility were made between the groups by a one-way ANOVA followed by a Tukey test. Noninvasive SBP values were compared by a repeated-measures ANOVA followed by the Tukey multiple-comparisons test. To compare SBPs after versus before therapeutic intervention, a repeated-measures ANOVA followed by Dunnett’s test was used. All values in the text are represented as mean±SEM.
Fig 2⇓ illustrates the effect of hydralazine and captopril therapy on SBP as measured noninvasively in SHRs. For clarity, only values obtained at 3-week intervals are represented. Both hydralazine and captopril therapy resulted in SBP values that were no different from WKY control values, decreased in comparison with untreated SHR values, and decreased in comparison with pretreatment values (P<.001) from 3 weeks after initiation of therapy until the end of the study. Not illustrated is the effect of captopril and hydralazine therapy on SBP values in WKY controls. Neither captopril nor hydralazine produced a significant decrease in blood pressure in control rats (SBP at 44 weeks in mm Hg: WKY untreated, 122±4; WKY captopril-treated, 120±4; WKY hydralazine-treated, 119±4).
Table 1⇓ summarizes the results of LV wet and dry weight, ratio of LV wet weight to body weight, and ratio of right ventricular wet weight and LV wall thickness (h) to radius (r) at a calculated end-diastolic internal diameter of 5 mm in treated and untreated SHRs and WKY controls. The increase in LV weight and h/r in SHRs compared with WKY controls indicates the presence of LVH and a concentric LV geometry. Captopril but not hydralazine therapy attenuated LVH and h/r in SHRs. Neither captopril nor hydralazine altered LV weight or h/r in WKY controls.
Fig 3⇓ and Table 2⇓ summarize the results of indexes of LVED performance, as determined in untreated SHRs and WKY controls as well as treated SHRs. LVEDD, as measured across the external surface of the heart, tended to be greater in untreated and hydralazine-treated SHRs compared with captopril-treated SHRs and untreated WKY controls at low filling pressures (Fig 3A⇓ and 3B⇓). The enhanced external diameters in the untreated and hydralazine-treated SHRs compared with the other groups are consistent with a thick-walled rather than a dilated LV, because calculated LVED internal diameters (Fig 3C⇓ and 3D⇓) were similar between the groups over the same range of filling pressures. The relationships between LVEDP and LVED internal or external diameters were steeper in untreated SHRs than in WKY controls, as evidenced by the increased slope (m) and decreased intercept (ln b) of the linearized relations in the former compared with the latter groups (Table 2⇓). These hypertension-induced changes in the slope and intercept of LVEDP–LVED dimension relations suggests an increased LVED stiffness. Both captopril (Fig 3B⇓ and 3D⇓) and hydralazine (Fig 3A⇓ and 3C⇓) therapy administered to SHRs resulted in less precipitous LVEDP-versus–LVED dimension relations compared with untreated SHRs, as evidenced by the decreased slopes (m) and increased intercepts (ln b) of the linearized relations of the treated compared with the untreated SHR groups (Table 2⇓). Both the slopes and the intercepts of the linearized relations of LVEDP versus LVED external and internal diameter of the treated SHR groups were similar to those values obtained in WKY controls (Table 2⇓). Neither captopril nor hydralazine therapy altered either the LVEDP-LVED dimension relations or the slopes and intercepts of the linearized functions of these relations in WKY control rats (not illustrated).
Fig 4⇓ summarizes the results of the LVED stress-strain relations as determined in untreated WKY control rats and SHRs as well as captopril- or hydralazine-treated SHRs. Fig 5⇓ summarizes the effect of hypertension and antihypertensive therapy on the slopes (LVED myocardial stiffness constants, k) of the linearized functions of the LVED stress-strain relations depicted in Fig 4⇓. The left-shifted LVED stress-strain relation and the increased myocardial k noted in untreated SHRs indicate an augmented myocardial diastolic stiffness produced by hypertension. Both captopril (Fig 4⇓, bottom) and hydralazine (Fig 4⇓, top) intervention returned the change in the LVED stress-strain relation and myocardial k (Fig 5⇓) produced by hypertension to levels similar to those noted in either treated or untreated WKY controls. Neither captopril nor hydralazine administration influenced LVED myocardial k in WKY controls (Fig 5⇓). Differences in LVED myocardial k between the untreated SHR group and either the untreated WKY, captopril-treated SHR, or hydralazine-treated SHR groups were not associated with a disparity in heart rates (bpm) at the time of LVEDD and LVEDP measurement (WKY, 362±10; SHR, 383±8; captopril-treated SHR, 366±12; hydralazine-treated SHR, 370±9). Therefore, the hypertension-induced increase and the captopril- and hydralazine-induced attenuation of LVED myocardial k cannot be attributed to alterations in heart rate influencing the extent of myocardial relaxation. The alterations in LV diastolic performance noted in SHRs in our study were not associated with changes in myocardial systolic performance, as determined from the slope of the LVES stress-versus-strain relation (Table 2⇑).
Fig 6⇓ illustrates the effect of hypertension and therapeutic intervention on measures of myocardial interstitial characteristics. An increased myocardial hydroxyproline content and an increased ratio of type I/III collagen were noted in untreated SHRs compared with WKY controls. Captopril but not hydralazine intervention attenuated the hypertension-induced increase in hydroxyproline content and the collagen phenotype changes in SHRs to values similar to those measured in WKY controls. Neither captopril nor hydralazine produced significant effects on the myocardial hydroxyproline content or collagen phenotype ratios in WKY controls.
Fig 7⇓ illustrates the influence of hypertension and therapeutic intervention on myocardial total (top), type I (middle), and type III (bottom) collagen solubility. Reduced total, type I, and type III collagen were measured in CNBr digests of the myocardium of SHRs compared with WKY controls when expressed as a percentage of the total collagen extracted after acid hydrolysis. Thus, collagen solubility was reduced in SHRs compared with WKY controls. The reduced total and type I collagen solubility, as measured in SHRs, was attenuated after therapeutic intervention with either captopril or hydralazine. Alternatively, captopril but not hydralazine therapy returned myocardial type III collagen solubility to values similar to those measured in WKY controls. Neither captopril nor hydralazine produced a significant effect on myocardial collagen solubility in normotensive WKY controls. Differences in collagen solubility after CNBr digestion between untreated SHRs and WKY controls and between treated and untreated SHRs are not attributed to oxidation of methionine residues of collagen in untreated SHRs. Indeed, similar differences in the percentage of myocardial hydroxyproline extracted with CNBr were noted between the groups after β-mercaptoethanol–induced reduction of myocardial collagen methionine residues (WKY, 44±7; untreated SHR, 22±2; captopril-treated SHR, 43±7; hydralazine-treated SHR, 43±5; P<.05 untreated SHR versus the other three groups).
Our results provide new insights into the possible mechanisms that might explain the increase in LVED myocardial stiffness that occurs subsequent to hypertension. We have shown that although the ACEI captopril but not the vasodilator hydralazine produces salutory effects on myocardial hydroxyproline and collagen phenotypes in SHRs, equivalent therapeutic actions on myocardial stiffness were noted with both pharmacological agents. These beneficial effects of therapeutic intervention on LVED myocardial performance are associated with an improved myocardial collagen (mainly type I) solubility, an index of collagen cross-linking, which is attenuated in SHRs.
The enhanced myocardial diastolic stiffness, the increase in myocardial hydroxyproline concentrations, and alterations in myocardial collagen phenotype ratios in SHRs as described by us are findings consistent with other studies.4 10 11 22 Similarly, our results showing a beneficial effect of both ACEIs and hydralazine on diastolic performance in SHRs correspond with previously published data.10 23 The advantageous influence of captopril but the lack of effect of hydralazine on myocardial collagen phenotype ratios that we show also concur with previously published results.11 A range of suggestions may explain the incongruous actions of the ACEI class of agents compared with the nonspecific vasodilators (which include hydralazine), both of which produce equipotent antihypertensive effects on total collagen and collagen phenotype ratios. However, an attractive explanation is that the vasodilators activate rather than inhibit the renin-angiotensin II system. This system has been shown to produce direct cellular effects on collagen expression,24 an action that is likely to be augmented rather than attenuated by antihypertensive agents that enhance plasma renin activity.
The results of our present experiment are not the first to contradict the perception that alterations in myocardial hydroxyproline concentrations and hence total collagen content explain myocardial diastolic stiffness in hypertension. Thiedermann et al6 showed an increased myocardial hydroxyproline content in 40- and 80-week-old SHRs but an enhanced myocardial stiffness in only the 80-week-old group. The interpretation of the results of the study by Thiedermann et al,6 however, is likely to be limited by the presence of normal myocardial diastolic stiffness constants demonstrated in 40-week-old animals by those authors, with a tissue preparation that has subsequently been used to show an increased myocardial stiffness in 48-week-old SHRs.22
It could be argued that our results reflect a unique action of hydralazine that is not shared by captopril on myocardial stiffness at end diastole. Such a hypothetical effect may in turn improve myocardial diastolic performance without influencing those cellular changes, such as alterations in myocardial total collagen or phenotypes, that induce the stiff myocardium in hypertension. Indeed, we have recently demonstrated that the increase in myocardial diastolic stiffness that occurs in the diabetic cardiomyopathy25 is attenuated by habitual exercise through a mechanism unrelated to that thought to be responsible for the stiff myocardium in this disease.26 However, this beneficial effect of habitual exercise on diastolic performance in diabetes mellitus was also noted in euglycemic control animals.13 In our present study, hydralazine administration to WKY controls produced no beneficial action on the myocardial diastolic stiffness constant. We therefore suggest that therapeutic intervention with hydralazine opposes the change in myocardial diastolic performance induced by hypertension by directly or indirectly inhibiting the mechanism responsible for the hypertension-induced increase in myocardial diastolic stiffness.
In our study, the ability to digest myocardial collagen with CNBr, which cleaves collagen at methionine groups,20 was reduced in 45-week-old SHRs exhibiting an increased myocardial diastolic stiffness. Alternatively, the beneficial effect of both captopril and hydralazine on myocardial diastolic stiffness in SHRs was associated with an enhanced CNBr-mediated collagen digestibility, type I in particular. The hypertension-induced decrease in CNBr-mediated digestibility of myocardial collagen and the captopril- and hydralazine-mediated attenuation of the latter effect were noted both before and after methionine residue reduction of collagen molecules of myocardial tissue samples with β-mercaptoethanol. Hence, modifications of myocardial diastolic stiffness that occur as a consequence of hypertension or antihypertensive therapy are proportional to the degree of myocardial collagen cross-linking and not to the degree of oxidation of the methionine residues of collagen.
Captopril but not hydralazine therapy resulted in a decrease in type III myocardial collagen solubility in SHRs, despite a similar beneficial action of both agents on type I solubility. Although we have no explanation for an effect of hydralazine on myocardial type I but not type III collagen cross-linking, it is likely that alterations in type I rather than type III collagen cross-linking will influence myocardial diastolic stiffness, because the type I phenotype is thought to have a higher tensile strength.
The notion that an increased myocardial diastolic stiffness may be attributed to an increment in collagen cross-linking in cardiac disease is supported by findings both in volume-overload states and in the cardiomyopathy of diabetes mellitus. Myocardial collagen cross-linking is augmented in volume-overload conditions, in which neither total myocardial collagen nor myocardial collagen phenotypes are altered but end-diastolic stiffness is increased.9 Similarly, an enhanced myocardial diastolic stiffness in diabetic cardiomyopathy is associated with an increment in myocardial collagen fluorescence (an index of irreversible collagen glycosylation and hence cross-linking) but not with changes in myocardial total collagen content.25
The mechanism by which volume or pressure overload induces an increased myocardial collagen cross-linking has not been determined. However, lysyl oxidase, the enzyme involved in mediating pepsin-sensitive cross-links between collagen molecules, has an enhanced activity in the vasculature (and hence possibly cardiac tissue) in SHRs27 and may therefore augment intermolecular or intramolecular cross-linking of collagen molecules.
Myocardial collagen solubility in response to pepsin digestion is unchanged in SHRs compared with normotensive controls,21 a result that does not contradict our findings of an attenuated collagen solubility to CNBr digestion in untreated SHRs. If an enhanced lysyl oxidase activity is responsible for the increased myocardial collagen cross-linking in SHRs, these sites of cross-linking would be pepsin-sensitive. Thus, in the presence of pepsin, the augmented collagen cross-linking is likely to be disrupted and a similar quantity of collagen will subsequently be extracted from SHR compared with normotensive control tissues. Alternatively, CNBr cleaves collagen molecules between pepsin-sensitive cross-linked sites and hence would plausibly provide a more sensitive measure of the degree of lysyl oxidase–mediated cross-linking in SHRs.
Thirty-week-old SHRs and their normotensive controls have a high degree of CNBr-mediated myocardial collagen solubility.11 Again, the latter results do not contradict our data. Collagen solubility in response to CNBr digestion decreases with age.28 Therefore, the reduced CNBr-mediated myocardial collagen solubility in the untreated SHRs in our study is likely to reflect either a combination of a hypertension- and age-induced effect or an action of prolonged hypertension.
Although we have excluded a possible influence of hypertension-induced changes in total myocardial collagen or phenotypes as contributing to the stiff myocardium that occurs in SHRs, we have not evaluated whether a rearrangement of the cardiac collagen matrix4 contributes to myocardial diastolic stiffness in hypertension. However, an altered myocardial collagen arrangement is usually associated with an increased collagen content.4 As such, it is difficult to argue in favor of an action of hydralazine on the arrangement of the myocardial collagen matrix as being responsible for the attenuated myocardial stiffness after intervention with this vasodilator agent without a finding of a concomitant influence of hydralazine on the hydroxyproline content.
The interstitial properties of the myocardium are not the only factors that need to be considered as potential mediators of an increased myocardial diastolic stiffness. Alterations in coronary perfusion pressure, by changing the erectile properties of the coronary capillaries, may also contribute to myocardial diastolic properties.3 However, even large changes in coronary perfusion pressures produce only small changes in myocardial diastolic stiffness.3 Therefore, it is unlikely that differences in coronary perfusion pressure between the groups during hemodynamic measurement could have influenced our calculated myocardial stiffness constants to a significant degree.
A decrease in the extent of myocardial relaxation also influences the end-diastolic properties of the myocardium.3 Cellular mechanisms such as the activity of the SR Ca2+-ATPase pump may, through alterations in the ability of the SR to sequester Ca2+ during diastole, control the extent of myocardial relaxation. After pressure overload, a decrease in SR Ca2+-ATPase mRNA expression occurs,29 which may result in an accumulation of cytoplasmic Ca2+ and subsequently a stiff myocardium. We have not examined the possible contribution of alterations in the extent of myocardial relaxation to the increased end-diastolic myocardial stiffness noted in SHRs in our study.
LV wall stress calculated from dimension measurements made across one axis of the heart may be underestimated when a spherical chamber geometry is assumed.15 However, we were not interested in comparing diastolic stress values between groups in our study. Rather, our concern was to examine the effect of hypertension and therapeutic intervention on myocardial diastolic stiffness constants. Because the slope of an LVED stress-strain relation calculated from dimension measures made across a single axis of the LV is not influenced by geometric considerations,30 we are confident that the LVED myocardial k values obtained by us reflect true myocardial diastolic stiffness constants.
We also failed to measure indexes of LV relaxation, such as tau or LV −dP/dt. Therefore, we cannot comment on the relaxation properties of the myocardium. However, we deliberately concentrated on measures of late diastolic performance, because it is thought that myocardial interstitial alterations are more likely to influence late as opposed to early diastolic performance.3 In addition, tau and −dP/dt are measures of the rate of rather than the extent of LV relaxation. Because LVED properties are determined by the extent rather than the rate of relaxation, tau and −dP/dt values would not be useful aids to help explain differences in measures of end-diastolic performance between the groups compared in our study.
In conclusion, we have been able to show that the increase in myocardial diastolic stiffness that occurs subsequent to hypertension, although associated with them, is not attributable to alterations in either the myocardial collagen content or collagen phenotypes. Rather, an increased collagen cross-linking, in particular of the type I phenotype, is a more likely explanation for the stiff myocardium of hypertension.
Selected Abbreviations and Acronyms
|IVC||=||inferior vena cava|
|LVD||=||LV external diameter|
|LVDP||=||LV diastolic pressure|
|LVEDD||=||LVED short-axis external diameter|
|LVEDP||=||LV end-diastolic pressure|
|LVESD||=||LVES short-axis external diameter|
|SBP||=||systolic blood pressure|
|SHR||=||spontaneously hypertensive rat|
This research was supported by the Medical Research Council of South Africa, the University of the Witwatersrand Medical Faculty Endowment Fund, the H.E. Griffin Charitable Trust, and the I.E. Hodges Cardiovascular Research Trust. We also acknowledge the kind donation of captopril and hydralazine by Squibb Pharmaceuticals and Lennon Pharmaceuticals, respectively.
Reprint requests to Dr G.R. Norton and Dr A.J. Woodiwiss, Laboratory of Cardiovascular Pathophysiology, Department of Physiology, University of the Witwatersrand Medical School, 7 York Rd, Parktown, 2193, Johannesburg, South Africa.
- Received November 21, 1996.
- Revision received April 9, 1997.
- Accepted April 18, 1997.
- Copyright © 1997 by American Heart Association
Gilbert JC, Glantz SA. Determinants of left ventricular filling and of the diastolic pressure-volume relation. Circ Res. 1989;64:827-851.
Jalil JE, Doering CW, Janicki JS, Pick R, Shroff SG, Weber KT. Fibrillar collagen and myocardial stiffness in the intact hypertrophied rat left ventricle. Circ Res. 1989;64:1041-1050.
Mukherjee D, Sen S. Collagen phenotypes during development and regression of myocardial hypertrophy in spontaneously hypertensive rats. Circ Res. 1990;67:1474-1480.
Weber KT, Janicki JS, Shroff SG, Pick R, Chen RM, Bashey RI. Collagen remodeling of the pressure-overloaded, hypertrophied nonhuman primate myocardium. Circ Res. 1988;62:757-765.
Satoshi K, Spinale FG, Tanaka R, Johnson W, Cooper G IV, Zile MR. Inhibition of collagen cross-linking: effects of fibrillar collagen and ventricular diastolic function. Am J Physiol. 1995;269:H863-H868.
Iimoto DS, Covell JW, Harper E. Increase in cross-linking of type I and type III collagens associated with volume-overload hypertrophy. Circ Res. 1988;63:399-408.
Brilla CG, Janicki JS, Weber KT. Cardioreparative effects of lis-inopril in rats with genetic hypertension and left ventricular hypertrophy. Circulation. 1991;83:1771-1779.
Norton GR, Tsotetsi OJ, Woodiwiss AJ. Hydrochlorothiazide improves ventricular compliance and thus performance without reducing hypertrophy in renal artery stenosis in rats. Hypertension. 1993;21:638-645.
Woodiwiss AJ, Norton GR. Exercise-induced cardiac hypertrophy is associated with an increased myocardial compliance. J Appl Physiol. 1995;78:1303-1311.
Trifunovic B, Norton GR, Duffield MJ, Avraam P, Woodiwiss AJ. An androgenic steroid decreases left ventricular compliance in rats. Am J Physiol. 1995;268:H1096-H1105.
Burns JW, Covell JW, Myers R, Ross J. A comparison of directly measured left ventricular wall stress and stress calculated from geometric reference figures. Circ Res. 1971;28:611-621.
Sagawa K, Maughan L, Suga J, Sunagawa K. Cardiac Contraction and the Pressure-Volume Relationship. New York, NY: Oxford University Press; 1988:367-369.
Laurent GJ, Cockerill P, McAnulty RJ, Hastings JRB. A simplified method for quantification of the relative amounts of type I and type III collagen in small tissue samples. Anal Biochem. 1981;113:300-312.
Brownlee M, Vlassara H, Kooney A, Ulrich P, Cerami A. Ami-noguanidine prevents diabetes-induced arterial wall protein cross-linking. Science. 1986;232:1629-1632.
Conrad CH, Brooks WW, Hayes JA, Sen S, Robinson KG, Bing OHL. Myocardial fibrosis and stiffness with hypertrophy and heart failure in the spontaneously hypertensive rat. Circulation. 1995;91:161-170.
Anderson PG, Bishop SP, Digerness SB. Vascular remodeling and improvement of coronary reserve after hydralazine treatment in spontaneously hypertensive rats. Circulation. 1989;64:1127-1136.
Norton GR, Candy G, Woodiwiss AJ. Aminoguanidine prevents the decreased myocardial compliance produced by streptozotocin-induced diabetes mellitus in rats. Circulation. 1996;93:1905-1912.
Woodiwiss AJ, Kalk WJ, Norton GR. Habitual exercise attenuates the increase in myocardial diastolic stiffness produced by diabetes mellitus in rats. Am J Physiol. 1996;271:H2126-H2133.
Albon J, Karmatowski WSS, Avery N, Easty DL, Duance VC. Changes in the collagenous matrix of the aging human lamina cribrosa. Br J Opthalmol. 1995;79:368-375.
Calderone A, Takahashi N, Izzo NJ Jr, Thaik CM, Colucci WS. Pressure- and volume-induced left ventricular hypertrophies are associated with distinct myocyte phenotypes and differential induction of peptide growth factor mRNAs. Circulation. 1995;92:2385-2390.
Janz RF, Kubert BR, Pate EF, Moriarty TF. Effect of shape on pressure-volume relationships of ellipsoidal shells. Am J Physiol. 1980;238:H917-H926.