J Mel Cell Cardiol22,
1157-I 165 ( 1990)
Enhanced
Deposition of Predominantly Collagen in Myocardial Disease
Type
I
Jill E. Bishop ‘*, Robert Greenbaum 2, Derek G. Gibsonj, Magdi Yacoub4 and Geoffrey J. Laurent’ 1Biochemistry Unit, Department of Thoracic Medicine, National Heart and Lung Institute, University of London, Dovehouse Street, London S W3 6LY, UK. 2Harejield Hospital, Uxbridge, Middlesex, UB9 63H, UK. 3Brompton Hospital, Fulham Road, London SW3 6HP, UK. 4Defiartment of Cardiac Surgery, National Heart and Lung Institute, University of London, Dovehouse Street, London SW3 6L1: UK. (Received 3 August 1989, accepted in revised form 4 May 1990) E. BISHOP, R. GREENBALM, D. C. GIBSON, M. YACOUB AND G. J. LAURENT. Enhanced Deposition of Predominantly Type I Collagen in Myocardial Disease. journal of Moletulur and Cellular Curdio& (1990) 22, 1157-1165. The myocardium consists of a muscle fibre array surrounded and interspersed by a network of connective tissue, principally collagen, which maintains the functional integrity ofthe heart. Changes in collagen composition may therefore contribute to altered ventricular function. Collagen composition was examined in cardiac tissue from 15 patients undergoing orthotopic cardiac transplantation. Of these, 10 had severdy impaired left ventricular function due to coronary artery disease. The remaining five had dilated cardiomyopathy. Normal heart tissue was taken at autopsy from 25 patients who died of causes unrelated to cardiovascular disease. Left ventricular collagen concentration, estimated from hydroxyproline levels, increased from 48.6 & 4.1 mg/g dry weight of tissue in the control group to 95.3 & 9.7 mg/g (P < 0.01) in patients with dilated cardiomyopathy and to 63.5 k-9.8 mg/g in the coronary artery disease group. This increase was attributable to an increase in absolute concentrations of both type I and III collagen, determined by separation of cyanogen bromide peptides by sodium dodecyl sulphate polyacrylamide gel electrophoresis. However. there was a significant decrease in the proportion of type III collagen (compared with type I plus III) from 41.8 _+ l.l?b in controls, to 34.6 + 1.5% (P < 0.01) in the coronary artery disease group and 35.8 _+ 2.89; (P < 0.05) in the dilated cardiomyopathy group. These results suggest that excessive collagen production, with a preponderance of type I, occurs in these forms of myocardial disease, indicative of a remodelling of the collagen matrix, which, by increasing passive myocardial stiffness may contribute to impaired heart function seen in these groups of patients.
J.
KEY
Collagen:
WORDS:
Coronary
artery
disease; Dilated
Introduction The myocardium consists of muscle fibres and blood vessels connected and interspersed by a network of connective tissue. Recent scanning electron microscopy studies in animals have revealed an intimate anatomical relationship between muscle cells and the connective tissue matrix, particularly with respect to the arrangement of collagen fibres (Borg and Caulfield, 1981; Robinson et al., 1987). The collagen matrix forms a scaffolding around the muscle cells and blood vessels playing an im*To whom
all correspondence
0022-2828/90/101157
should
+ 09 $03.00/O
cardiomyopathy
portant role in maintaining the functional integrity of the myocardium (Borg and Caulfield, 1981; Weber et al., 1988; Weber, 1989). Abnormalities in cardiac muscle stiffness, seen in several forms of experimentally induced cardiac hypertrophy, have been correlated with changes in ventricular collagen concentration (Bing et al., 1971; Thiedemann et al., 1983). Although the mechanical measurements were performed on isolated muscles whereas the collagen concentration was estimated in the whole ventricle, more recent
be sent. ii1 1990 Academic
Press Limited
1158
J. E. Bishop
measurements of ventricular stiffness have confirmed the earlier conclusions that myocardial stiffness is, in part, determined by collagen concentration Ualil et al., 1988, 1989; Weber et al., 1988; Doering et al., 1988). This was also the finding in clinical studies by Hess et al. 1984) and Krayenbuehl et al. (1989) in which regression of hypertrophy, 17.5 months following valve replacement in patients with aortic stenosis, was accompanied by an increase in both interstitial fibrosis and left ventricular myocardial stiffness. The relative proportion of the individual collagen types may also affect the physical properties of the collagen matrix. The two most abundant collagen types in human myocardium are types I and III. Type I appears to form chiefly thick fibres and is most abundant in tissues requiring tensile strength, such as bone and tendon (Toole et al., 1972; Hanson and Bentley, 1983). Type III collagen forms thinner fibres and relatively high proportions are found in elastic tissue such as blood vessels and lung (Gay and Miller, 1978; Barnes, 1985; Mays et al., 1988). The relative proportion of these collagens may therefore play an important role in determining the physical properties of the extracellular matrix. In the heart these two collagens appear to, coexist in the matrix, having similar distributions in the collagen struts and pericellular fibres of the endomysium, demonstrated using immunofluorescent techniques (Robinson et al., 1987, 1988). Based on the hypothesis that changes in collagen composition may play a role in the depression of cardiac function seen in myocardial disease in man, the purpose of the present study was to investigate collagen content and the relative proportions of the major collagen types in ventricles of patients with coronary artery disease and dilated cardiomyopathy known to have poor ventricular function. Materials
and
Methods
Collagen content was examined in normal and diseased left ventricular myocardium. All tissue samples were full thickness sections of myocardium with a surface area of approximately 1 cm2, They were generally removed from the diaphragmatic aspect of the left
et al.
ventricle at a level half way between the atrioventricular grove and the apex. Tissues were obtained from three groups. Control tissues were obtained at autopsy from 25 patients (17 male, mean age 57 years, range 19-78 years) without a history of cardiovascular disease, hypertension or exposure to cardiotoxic drugs. Most of these patients died of malignant diseases. A single biopsy was taken from each of the hearts of 15 patients undergoing orthotopic cardiac transplantation for severely impaired cardiac function. Of these, 10 (9 male, mean age 49.3 years, range 34-68 years) had coronary artery disease and 5 (4 male, mean age 3 1.4 years, range 12-49 years) had dilated cardiomyopathy without evidence of significant large vessel coronary artery disease. The coronary artery disease group had a history of at least one full thickness (Q wave) myocardial infarction. Both groups had extremely poor left ventricular function assessed by conventional cineangiography. All biopsies were obtained immediately after the heart was removed from the recipient. In patients with coronary artery disease, regions which were macroscopically replaced by scar tissue were avoided for the purpose of taking biopsies. All the tissues were frozen in liquid nitrogen and stored at - 30°C prior to analysis. Biochemical analysis Collagen concentration was estimated in a full thickness section of tissue following determination of the hydroxyproline levels by the method of Stegemann and Stadler (1967) with more recent modifications (Laurent et al., 1981a). Estimations of the relative amount of types I and III collagen were also made on full thickness sections of tissue as described previously (Laurent et al., 1981 b; Kirk et al., 1984). Briefly, semi-purified collagens were obtained following repeated homogenization of the tissue in phosphate buffered saline (PBS) and 2% sodium dodecyl sulphate (SDS). This material was acetone-dried, then subjected to cyanogen bromide cleavage. The resulting peptides were separated by sodium dodecyl sulphate polyacrylamide gel electrophoresis (unreduced) along with known amounts of purified collagens (with a 3 : 2 ratio of type
Collagen in Heart I : III collagen) and stained (with PAGE blue 83, Fisons plc, Loughborough, England). Gels were destained in 8% (v/v) acetic acid for at least 24 h and each track was scanned using a Gilford 2600 spectrophotometer. The areas beneath the peaks were calculated by computerized integration. The ~1~(I)CB8 and the xl(III)CB5 peptides were used to quantitate types I and III respectively. Standard curves were constructed for these two peptides and the relative proportion of the collagen types were calculated for the tissue samples. Statistical analysis was performed using the Mann Whitney-L’ test (von Fraunhofer and Murray, 1976). However, data are presented as means + S.E.M. Mean values for patient groups were considered significantly different from controls when the test gave a P value of less than 0.05. Results The collagen concentration in the control left ventricle was 48.6 f 4.1 mg/g dry weight of tissue. There was no correlation between age or sex of the patient and collagen concentration (data not shown). The concentration of collagen in the left ventricle of patients with dilated cardiomyopathy was double that of normal ventricles (P < 0.01, see Table 1). In the group with coronary artery disease the concentration increased by about 30% but this was not significant at the 5% level. Figure 1 shows densitometric scans of the cyanogen bromide-derived peptides for standard collagens and collagens from heart tissue following electrophoresis. The peptides used TABLE
1. Collagen concentration left ventricle
and
the relative
Disease
to quantitate types I and III (cri(I)CB8 and ~(1(III)CB5 respectively) are indicated in the figure. They were both sufficiently well separated from adjacent peptides to allow accurate quantitation based on curves constructed from known loadings of collagen standards run on the same gel (see Fig. 1 (a) and Laurent et al., 1981b). Small quantities of collagen types IV (a basement membrane collagen), V and VI will also be present, however their contribution to the total amount of collagen is very small, the distinctive cyanogen bromide peptides for these collagen types described previously (Laurent et al., 1981 b) could not be detected on these gels. Table 1 also shows that the proportion of type III (compared with type I plus III) was lower in both coronary artery disease (by P < 0.01) and cardiomyopathy groups ;lz;1;‘4”,, P < 0.05), compared with controls. This was not due to age or sex differences between groups since no correlation was found between either of these parameters and the proportion of type III collagen in the control tissue samples (see Fig. 2). The estimated concentration of types I and III collagen, derived from the total collagen concentration and the proportion of each type, is also shown in Table 1. The shift in collagen types was apparently achieved by an increase in both type I and III, but there was a disproportionate increase in type I. Discussion The mechanical properties depend both on contractile proportion
Total collagen concentration (mg/g Control
CAD (n = 10) DCM
(n = 5)
dry
wt)
48.6 + 4.1 63.5 f 9.8 95.3 + 9.7 t
1159
of types
I and
III
Type
I
in normal
concentration y. III 41.8 + 1.1 34.6 f 1.5 t 35.8 + 2.8*
(mdg
dry
of the ventricle and connective
wtj
28.3 t 2.5 41.5 + 6.1 61.2 + 9.6t
and disrasrd
Type III concentration (mg/g dry wt: 20.3 + 1.7 22.0 &- 2.4 34.1 &- 4.8t
Values shown arc mean & S.E.M. Total collagen concentration (column I) is expressed as mg/g dry weight of tissue (mg/g dry wt). The proportion of type III (column 2) is expressed as the percentage oftype III collagen compared with the total of type I plus III. Columns 3 and 4 (concentration of the individual types) are derived from the data in the first two columns and do not take into consideration a contribution to the total collagen concentration by types IV, V and VI. CAD = coronary artery disease; DCM = dilated cardiomyopathy; * = P < 0.05 with respect to controls; t = P < 0.01 with respect to controls. For control total collagens, n = 15; for control (I0 III, n = 25.
J. E. Bishop
1160
“‘Lr-
p(I)
et al.
CB8
0
$ 0.9 (b) ,’ 0
o,(I)CB8
5 2
0
20
40
60
80
100
FIGURE I. Densitometric SCIIRJ of cyanogen bromide peptides obtained following separation by SDS-polyocrylamide gel electrophoresis. Figure 1 (a) shows a scan ofpurified collagen standard containing type I and III collagen in a ratio of3 : 2. Also shown are gel tracks ofstandards containing 6-24 pg type I collagen and 4-16 fig type III. The scan was obtained from the third standard (18 pg type I + 12 pg type III). Figure I(b) shows the scan obtained from a sample of myocardium obtained from a patient with dilated cardiomyopathy with the track from which it was obtained. The two bands used to quantitate types I and III collagen, a,(I)CB8 and 011(III)CBS, are indicated.
tissue elements. In general, studies regarding the structure-function relationships have centred on theoretical modelsfor the arrangement of the contractile and elastic elements
50
r
’I
0*
l 0
.
0
40
8
l *
.
l l
l
l 0 0
El “p 2 30 a?
0
0 l
0
0
0
20 I
20
40
I
60
00
Age (years)
FIGURE 2. The relationship between age and Qje III collagen concentration in normal left ventricle. There is no correlation between age of patient and proportion of type III collagen. l = male, 0 = female.
(Sonnenblick, 1964; Spotnitz et al., 1966; Streeter and Hanna, 1973). Anatomical and morphological studies have confirmed the highly ordered nature of ventricular muscle fibre array in man (Greenbaum et al., 1981). The presence of a well developed network of connective tissuein skeletal muscle was highlighted in the 1930’s.Banus and Zeitlin (1938) demonstrated that the passive length/tension curves were dictated in part by the presenceof the connective tissuenetwork. Development of high resolution scanning electron microscopy has facilitated detailed examination of the arrangement of collagen in cardiac muscle. Borg and Caulfield ( 1981) defined the collagen network in three parts. First, a weave ofcollagen bundlessurrounding myocytes, dividing them into groups. Second, collagen struts connecting myocytes within groups, and third, collagen struts connecting myocytes to adjacent capillaries. They suggested that the intermyocyte connections maintain lateral cell to cell alignment especially during diastole when the musclediameter
Collagen
in Heart
decreases. The capillary to myocyte collagen struts may be important in maintaining capillary to myocyte relationships during systole as the latter’s diameter increases, and may allow capillary blood flow to occur during systole in spite of increased wall stress at that time. The concentration of collagen in the human left ventricle, based on hydroxyproline content, was shown in the present study to have increased by a factor of two in dilated cardiomyopathy compared with controls, whilst in coronary artery disease the concentration increased by 3Oq&. Such increases may partly explain altered cardiac function in these groups of patients. It seems unlikely that the values obtained in the control group were influenced by previous treatment since samples were not taken from patients exposed to cardiotoxic agents, and our values are similar to those found in previous studies (Montfort and Perez-Tamayo, 1962; Caspari et al., 1977). In order to try to characterize such changes techniques of non-invasive tissue characterization using ultrasound have been developed (Shaw et al., 1984). The development of fibrotic regions in hypertrophied left ventricles has been demonstrated using a colour coding technique based on variations in the amplitude of reflected ultrasound (Mimbs et al., 1980; Shaw et al., 1984). In patients with left ventricular hypertrophy, increased myocardial echo amplitude correlated strongly with abnormal diastolic function as well as with T wave changes in the ECG (Shapiro et al., 1984). Use of such techniques combined with further studies of collagen composition and distribution in diseaseis likely to increase our understanding of the underlying mechanismsof ventricular disease. Tissue samples without macroscopic evidence of scar tissuewere deliberately selected for our study. Clearly, changesin collagen composition are not restricted to the infarcted regions. Infarct zones have been shown to have a much greater collagen concentration than normal regions (Judgutt and Amy, 1986) and the affected areas have mechanical properties markedly different from viable myocardium (Shaw et al., 1984).Judgutt and Amy i 1986) demonstrated an increase in collagen concentration of more than IO-fold in transmural sections taken at the infarct centre 6
Disease
1161
weeks after experimentally induced myocardial infarction in the dog. At the infarct border there was still a 5-fold greater collagen concentration. However, normal collagen concentrations were found in zones distant from the infarct in these animals. Our results may differ from the animal study for a number of reasons.The extent and duration of the diseaseis likely to be much greater in the human study, and also species variations have been observed in the timing of the remodelling process (Judgutt and Amy, 1986). In the present study, samples were taken from tissuewithout evidence of scarring, assessed at the macroscopic level. The increase in collagen concentration we have demonstrated may be due to a more diffuse interstitial fibrosis, possibly confined to a particular section of myocardium. Since we measured collagen concentration in full thickness sections we have no information on the distribution of the excesscollagen deposited. Further studiesare required to examine the collagen composition in the endocardial, midwall and epicardial regions. An alternative explanation for an increase in collagen concentration in the regions distant from the scar tissueis that the apparently normal myocardium may try to compensate for the lossof viable muscle cells resulting in regional hypertrophy (Roberts et al., 1983; Pasternak et al., 1988). This may be accompanied by an increase in collagen synthesis, since collagen metabolism appears to be stimulated by increasedworkload in the heart (Skosey et al., 1972;Bonnin et al., 1981; Turner et al., 1986). However, this form of hypertrophy is thought to be a volume-overload type (Roberts et al., 1983) which is not normally associated with an increase in collagen concentration (Bartosova et al., 1969; Michel et al., 1986). Further studies are required to determine the mechanismsinvolved in stimulating collagen production in these areas. The changesin collagen concentration, determined biochemically, may not be sufficient to explain fully the collagen related increaseiu myocardial stiffness. Changes in the composition, with regard to the different collagen types, as discussedbelow, may contribute to changes in diastolic function. In addition. rearrangement of the collagen matrix appears to be an important determinant (Weber ~1al.,
1162
J. E. Bishop
1988; Weber, 1989; Doering et al., 1988). An increasein thicknessof the perimysial tendons, weaves and strands was seen in established pressure-overload hypertrophy in the macaque, and newly formed fibres filled the intermuscular spaces(Weber et al., 1988). Borg et al. (1981) observed a correlation between the thickness of the collagen struts in different speciesand the stiffness of isolated papillary muscles.Clearly, morphologic and morphometric studies provide a more detailed description of the remodelling process.This is particularly true where the fibrosis is limited to particular regions of the myocardium. For example, a perivascular fibrosis was observed around small coronary arteries in rat hypertrophied left ventricle (Jalil et al., 1988). In the present study we demonstrated that in myocardial diseasetype III collagen accounts for approximately one third of the total collagen, compared with over 40% in normal humans, and changes in the concentration of types I and III in diseasedmyocardium suggested that a disproportionate amount of newly formed matrix was type I collagen. Comparable changes to those seen here for collagen types I and III have been observed during the development of right ventricular hypertrophy in rabbits (Turner and Laurent, 1986). However, in an animal model of left ventricular hypertrophy caused by gradual onset of hypertension, an increase in type III collagen was observed (Weber et al., 1988). In addition a lower control value was found. A similar trend was observed in copper deficiency-induced cardiac hypertrophy in rats (Dawson et al., 1982). An increasein type III was also seen in experimentally induced hypertrophy in rats following aortic stenosisor renovascular hypertension (Medugorac and Jacob, 1983). There are several explanations for these differences. Firstly the above mentioned studies employed pepsin digestion which, at best, extracts 75% of the total collagen, and this may be substantially reduced in the hypertrophied heart (Iimoto et al., 1988). Love11 et al. (1987) suggested that type III collagen may be less susceptible to pepsin digestion than type I, which might explain why the ratio of type I: type III was lower in control groups in previous studies than described here. In the present study, cyanogen
et al.
bromide digestion was used after first semipurifying the collagen by SDS and PBS washes. This has been shown to extract 80-90% of the total collagen (Laurent et al., 1981b) although we have not examined the extractibility in diseasedhearts. Although differences in the extractibility of the two collagen types has been demonstrated this is not a uniform finding (Chan and Cole, 1984; Maurel et al., 1987). A second explanation for the differences is that increasedtype III collagen may be transient, the normal ratio being restored at a later time point (Weber et al., 1988). In the present study the stage of the diseaseis not known, although it is likely to be advanced, a consequence of which may be an increase in the proportion of type I. Thirdly, the experimental modelsof cardiac hypertrophy cannot be directly compared with the ventricular disease studied here, since changes in the relative proportion of the collagen types may be dependent on the different disease processes.Control values may also differ depending on the speciesunder investigation. The increase in the proportion of type I collagen observed in the apparently nonfibrotic regionsmay further affect the mechanical properties of the ventricle. Although much of the evidence to date is circumstantial, the relative proportion of types I and III in the connective tissuematrix may influence the physical properties of a tissue. These two collagen types appear to colocalize in myocardium (Robinson et al., 1988) and placenta (Amenta et al., 1986). Using immunohistochemistry, type I was seen to form thicker fibres than type III, which were found forming a weave around the type I fibres in placental tissue(Amenta et al., 1986). The nature of the underlying changes in collagen metabolism in human disease remains unclear, but alterations in both synthetic and degradative processesmay be important. Based on the changes in collagen types (Turner and Laurent, 1986) and collagen metabolism (Skosey et al., 1972; Turto, 1977; Lindy et al., 1972; Bonnin et al., 1981; Turner et al., 1986) seenin animal studies of ventricular hypertrophy, it would appear that there is increasedsynthesisof both types I and III, with the increase in type I being greater. It was also suggestedin animal studies that
Collagen
in Heart
changes in breakdown may be important, particularly in the early stages (Turner et al., 1986). This is supported by recent studies on the arrangement of the collagen network in which the thin fibrils disappeared in the evolutionary phase of cardiac hypertrophy (Weber et al., 1988). It is not clear whether the increase in synthesiswas mediated by an increased production by individual cells or an increase in the total number of fibroblasts, although there is evidence for active replication of fibroblastic cells suggestingthe latter mechanism may be important (Morkin and Ashford, 1968; Grove et al., 1969). Skosey et al. ( 1972) demonstrated an increasein collagen production prior to the increase in cell number suggesting also an increasein synthetic activity per cell. In summary, this study has demonstrated enhanced myocardial collagen deposition, particularly of type I collagen, in patients with myocardial disease.The data presented do not
Disease
1163
allow us to conclude that changesin collagen deposition and structure are primary events in the development of impaired ventricular function. Nevertheless,they open up the possibility that the impaired ventricular function seenin these patients may be in part related to a remodelling of connective tissue in what appears macroscopically normal ventricle, rather than due entirely to a biochemical lesion in the contractile cells. Acknowledgements
This work was funded by the British Heart Foundation. Dr Jill Bishop (nee Turner) is currently a British-American Research Fellow of the American Heart Association and British Heart Foundation at the University of Vermont, USA. The authors wish to thank Dr Richard Carter (Royal Marsden Hospital, Sutton, UK) for his co-operation in providing tissuesamples.
References AMENTA PS, GAY, S, VAHERI A, MARTINEZ-HERNANDEZ A. (1986). The extracellular matrix is an integrated unit: Ultrastructural localization of collagen types I, III, IV, V, VI, fibronectin and laminin in human term placenta. Collagen Rel Res 6: 125-152. BANUS MG, ZEITLIN AM (1938) The relation of isomeric tension to length in skeletal muscle. J Cell Comp Physiol 12: &20. BARNES MJ (1985) Collagens in atherosclerosis. Collagen Rel Res 5: 65-97. BARTOSOVA D, CHVAPIL M, KORECKY B, POUPA 0, RAKUSAN K, TIJREK Z, VIZEK M (1969) The growth of the muscular and collagenous parts of the rat heart in various forms of cardiomegaly. J Physiol200: 285-295. BING OHL, MATSUSHITA S, FANBURC BL, LEVINE HJ (1971) Mechanical properties of rat cardiac muscle during experimental hypertrophy. Circ Res 28: 23G-245. BONNIN CM, SPARROW MP, TAYLOR RR (1981) Collagen synthesis and content in right ventricular hypertrophy in the dog. Am J Physiol241: H708-H713. BORG TK, CAULFIELD JB (1981) The collagen matrix of the heart. Fed Proc 40: 2037-2041. BORG TK, RANSON WF, MOSLEHY FA, CAULFIELD JB (1981) Structural basis of ventricular stiffness. Lab Invest 44: 49-54. CASPARI PG, NEWCOME M, GIBSON K, HARRIS P (19771 Collagen in the normal and hypertrophied human ventricle. Cardiovasc Res 11: 554-558. CHAN D, COLE WG (1984) Quantitation of type I and III collagens using electrophoresis of alpha chains and ryanogen bromide peptides. Anal Biochem 139: 322-328. DAWSON R, MILNE G. WILLIAMS RB (1982) Changes in the collagen of rat heart in copper-deficirncy-induced cardiac hypertrophy. Cardiovasc Res 16: 559-565. DOERING CW, JALIL JE, JANICKI JS, PICK R, ACHILI S, ABRAHAMS C, WEBER KT (1988) Collagen network remodrlling and diastolic stiffness of the rat left ventricle with pressure overload hypertrophy. Cardiovasc Res 22: 686-695. GAYS, ~MILLER EJ (1978) In: Collagen in the physiology and pathology ofconnective tissue. S Gay and EJ Miller (Eds). Gustav Fischer Verlag. Stuttgart. New York. GREENBAUM RA, Ho SY, GIBSON DG, BECKER AE, ANDERSON RH (1981) Left ventricular fibre architecture in man. Br Heart J 45: 248-263. GROVE D, NAIR KG. ZAK R (1969) Biochemical correlates of cardiac hypertrophy. III. Changes in DNA content; the relative contribution of polyploidy and mitotic activity. Circ Res 25: 463471. HANSON AN, BENTLEY JP (1983) Quantitation oftype I and type III collagen ratios in small samples of human tendon, blood vessels and atherosclerotic plaque. Analyt Biochem 130: 3240. HESS OM, RITTER M, SCHNEIDER J, GRIMM J, TURINA M, KRAYENBRUEHL HP (1984) Diastolic stiffness and myocardial structure in aortic valve disease before and after valve replacement. Circulation 69: 855-865.
1164
J. E. Bishop et al.
DS, COVELL JW, HARPER E (1988) Increase in cross-linking of type I and type III collagens associated with volume-overload hypertrophy. Circ Res 63: 39%408. JALIL JE, DOERING W, JANICKI JS, PICK R, CWRK WA, ABRAHAMS C, WEBER, KT (1988) Structural vs. contractile protein remodelling and myocardial stiffness in hypertrophied rat left ventricle. J Mol Cell Cardiol20: 11791187. JALIL JE, DOERING CW, JANICIU JS, PICK R, SHROFF SG, WEBER KT (1989) Fibrillar collagen and myocardial stiffness in intact hypertrophied rat left ventricle. Circ Res 61: 104-1050. JUDGUTT BI, AMY RWM (1986) Healing after myocardial infarction in the dog: changes in infarct hydroxyproline and topography. J Am Co11 Cardio17: 91-102. KIRK JME, HEARD BE, KERR I, TURNER-WARWICK M, LAURENT GJ (1984) Q uantitation of types I and III collagen in biopsy lung samples from patients with cryptogenic fibrosing alveolitis. Collagen Rel Res I: 169-182. KRAYENBUEHL HP, Hess OM, MONRAD ES, SCHNEIDER J, MALL G, TUIUNA M (1989) Left ventricular myocardial structure in aortic valve disease before, intermediate, and late after aortic valve replacement. Circulation 79: 744-755. LAURENT GJ, MCANULTY RJ, CORRIN B, COCKERILL P (1981a) Biochemical and histological changes in pulmonary fibrosis induced in rabbits with intratracheal bleomycin. Eur J Clin Invest 11: 441-448. LAURENT GJ, COCKERILL P, MCANULTY RJ, HASTINGS JRB (1981b) A simplified method for quantitation of the relative amounts of type I and type III collagen in small tissue samples. Analyt Biochem 113: 301-312. LINDY S, TURTO H, Urrro J (1972) Protocollagen proline hydroxylase activity in rat heart during experimental cardiac hypertrophy. Circ Res 30: 205-209. LOVELL CR, SMOLENSKI KA, DUANCE JC, LIGHT ND, YOUNG S, DYSON M (1987). Type I and III collagen content and fibre distribution in normal skin during ageing. Br J Dermatol 117: 41%428. MAUREL E, SHUT~LEWORTH CA, Bourssou H (1987 Interstitial collagens and ageing in human aorta. Virchows Arch A. 410: 383-390. MAYS PK, BISHOP JE, LAURENT GJ (1988) Age-related changes in the proportion of types I and III collagen. Mech Ageing Dev 45: 203-2 12. MEDUGORAC I, JACOB R (1983) Characterization of left ventricular collagen in the rat. Cardiovasc Res 17: 15-Z I. MIMBS JW, O’DONNELL M, BAUWENS D, MILLER JM, SOBEL BE (1980) The dependence of ultrasonic attenuation and backscatter on collagen content in dog and rabbit hearts. Circ Res 47: 4958. MICHEL JB, SALZMANN JL, NLOM MO, BRUNEVAL P, BARRES D, CAMILLERI JP (1986) Morphometric analysis of collagen network and plasma perfused capillary bed in the myocardium of rats during evolution ofcardiac hypertrophy. Basic Res Cardiol81: 142-154. MONTFORT TW, PEREZ-TAMAYO R (1962) The muscle ratio in normal and hypertrophic human hearts. Lab Invest 11: 463470. MORKIN E, ASHFORD TP (1968) Myocardial DNA synthesis in experimental cardiac hypertrophy. Am J Physiol 215: 1409-1413. PASTERNAK RC, BRAUNWALD E, SOBEL BE (1988) In: Heart Disease. A textbook of cardiovascular medicine. E. Braunwald (Ed.), W. B. Saunders, Co. pp 122221313. ROBERTS CS, MACLEAN D, BRAUNWALD E, MAROKO PR, KLONER RA (1983) Topographic changes in the left ventricle after experimentally induced myocardial infarction in the rat. Am J Cardiol51: 872876. ROBINSON TF, FACTOR SM, CAPASSO JM, WITTENBERG BA, BLUMFELD 00, SEIFTER S (1987) Morphology, composition and function of struts between cardiac myocytes of rat and hamster. Cell Tiss Res 2* 247-255. ROBINSON TF, COHEN-GOULD L, FACTOR SM, EGHBALI M, BLUMENFELD 00 (1988) Structure and function of connective tissue in cardiac muscle: collagen types I and III in endomysial struts and pericellular fibers. Scanning Microscopy 2: 1005-1015. SHAPIRO LM, MOORE RB, LOGAN-SINCLAIR RB, GIBSON DG (1984) Relation of regional echo amplitude to left ventricular function and the electrocardiogram in left ventricular hypertrophy. Br Heart J 52: 99105. SHAW TRD, LOGAN SINCLAIR RB, SURXN C, MCANULTY RJ, HEART B, LAURENT GJ, GIBSON DG (1984) Relation between regional echo intensity and myocardial connective tissue in chronic left ventricular disease. Br Heart J 51: 46-53. SKOSEY JL, ZAK R, MARTIN AF (1972) Biochemical correlates of cardiac hypertrophy: V. Labeling collagen, myosin and nuclear DNA during experimental myocardial hypertrophy in the rat. Circ Res 31: 145-157. SONNENBLICK EH (1964) Series of elastic and contractile elements in heart muscle: changes in muscle length. Am J Physiol 287: 1330-1338. SPOTNITZ HM, SONNENBLICK EH, SPIRO D (1966) Relation of ultrastructure to function in the intact heart: Sarcomere structure relative to pressure-volume curves of intact left ventricle of dog and cat. Circ Res IS: 44-66. STEGEMANN H, STADLER K (1967) Determination of hydroxyproline. Clin Chim Acta 18: 267-273. STREETER DD, HANNA WT (1973) Engineering mechanics for successive states in canine left ventricular myocardium. 2 Fibre angle and sarcomere length. Circ Res 33: 656664. THIEDEMANN KU, HOLUBARSCH CH, MEDUGORAC I, JACOB R (1983) Connective tissue content and myocardial stiffness in pressure overload hypertrophy. A combined study of morphologic, morphometric, biochemical, and mechanical parameters. Basic Res Cardiol78: 141-155. TOOLE BP, KANG AH, TRELSTAD RL, GROSS J (1972) Collagen heterogeneity within different growth regions of long bones in rachitic and non-rachitic chicks. Biochem J 127: 715-720. TURNER JE, LAURENT GJ (1986) Increased type I collagen compared with type III during right ventricular hypertrophy in rabbits. Biochem Sot Trans II: 1079-1080. IIMOTO
Collagen
in Heart
Disease
1165
TURNER JE, OLIVER MH, GUERREIRO D, LAURENT GJ (1986) Collagen metabolism during right ventricular hypertrophy following induced lung injury. Am J Physiol 251: H915-H919. TURTO H (1977) Collagen metabolism in experimental cardiac hypertrophy in the rat and the effect of digitoxin treatment. Cardiovasc Res 11: 358-366. VON FRAUNWOFER JA, MURRAY JJ (Eds) (1976) In: Stati~fics in Medical, Dental and Biological Studies. Tri-Med Books Ltd, London. pp 77-84. WEBER KT, JANICKI JS, SHROFF SC, F’ICK R, CHEN RM, BAWEY RI (1988) C o II a g en remodeling of the pressureoverloaded, hypertrophied nonhuman primate myocardium. Circ Res 62: 757-765. WEBER KT ( 1989) Cardiac interstitium in health and disease: The fibriljar collagen network. J Am Co11 Cardiol 13: 1637-1652.
1157-I 165 ( 1990)
Enhanced
Deposition of Predominantly Collagen in Myocardial Disease
Type
I
Jill E. Bishop ‘*, Robert Greenbaum 2, Derek G. Gibsonj, Magdi Yacoub4 and Geoffrey J. Laurent’ 1Biochemistry Unit, Department of Thoracic Medicine, National Heart and Lung Institute, University of London, Dovehouse Street, London S W3 6LY, UK. 2Harejield Hospital, Uxbridge, Middlesex, UB9 63H, UK. 3Brompton Hospital, Fulham Road, London SW3 6HP, UK. 4Defiartment of Cardiac Surgery, National Heart and Lung Institute, University of London, Dovehouse Street, London SW3 6L1: UK. (Received 3 August 1989, accepted in revised form 4 May 1990) E. BISHOP, R. GREENBALM, D. C. GIBSON, M. YACOUB AND G. J. LAURENT. Enhanced Deposition of Predominantly Type I Collagen in Myocardial Disease. journal of Moletulur and Cellular Curdio& (1990) 22, 1157-1165. The myocardium consists of a muscle fibre array surrounded and interspersed by a network of connective tissue, principally collagen, which maintains the functional integrity ofthe heart. Changes in collagen composition may therefore contribute to altered ventricular function. Collagen composition was examined in cardiac tissue from 15 patients undergoing orthotopic cardiac transplantation. Of these, 10 had severdy impaired left ventricular function due to coronary artery disease. The remaining five had dilated cardiomyopathy. Normal heart tissue was taken at autopsy from 25 patients who died of causes unrelated to cardiovascular disease. Left ventricular collagen concentration, estimated from hydroxyproline levels, increased from 48.6 & 4.1 mg/g dry weight of tissue in the control group to 95.3 & 9.7 mg/g (P < 0.01) in patients with dilated cardiomyopathy and to 63.5 k-9.8 mg/g in the coronary artery disease group. This increase was attributable to an increase in absolute concentrations of both type I and III collagen, determined by separation of cyanogen bromide peptides by sodium dodecyl sulphate polyacrylamide gel electrophoresis. However. there was a significant decrease in the proportion of type III collagen (compared with type I plus III) from 41.8 _+ l.l?b in controls, to 34.6 + 1.5% (P < 0.01) in the coronary artery disease group and 35.8 _+ 2.89; (P < 0.05) in the dilated cardiomyopathy group. These results suggest that excessive collagen production, with a preponderance of type I, occurs in these forms of myocardial disease, indicative of a remodelling of the collagen matrix, which, by increasing passive myocardial stiffness may contribute to impaired heart function seen in these groups of patients.
J.
KEY
Collagen:
WORDS:
Coronary
artery
disease; Dilated
Introduction The myocardium consists of muscle fibres and blood vessels connected and interspersed by a network of connective tissue. Recent scanning electron microscopy studies in animals have revealed an intimate anatomical relationship between muscle cells and the connective tissue matrix, particularly with respect to the arrangement of collagen fibres (Borg and Caulfield, 1981; Robinson et al., 1987). The collagen matrix forms a scaffolding around the muscle cells and blood vessels playing an im*To whom
all correspondence
0022-2828/90/101157
should
+ 09 $03.00/O
cardiomyopathy
portant role in maintaining the functional integrity of the myocardium (Borg and Caulfield, 1981; Weber et al., 1988; Weber, 1989). Abnormalities in cardiac muscle stiffness, seen in several forms of experimentally induced cardiac hypertrophy, have been correlated with changes in ventricular collagen concentration (Bing et al., 1971; Thiedemann et al., 1983). Although the mechanical measurements were performed on isolated muscles whereas the collagen concentration was estimated in the whole ventricle, more recent
be sent. ii1 1990 Academic
Press Limited
1158
J. E. Bishop
measurements of ventricular stiffness have confirmed the earlier conclusions that myocardial stiffness is, in part, determined by collagen concentration Ualil et al., 1988, 1989; Weber et al., 1988; Doering et al., 1988). This was also the finding in clinical studies by Hess et al. 1984) and Krayenbuehl et al. (1989) in which regression of hypertrophy, 17.5 months following valve replacement in patients with aortic stenosis, was accompanied by an increase in both interstitial fibrosis and left ventricular myocardial stiffness. The relative proportion of the individual collagen types may also affect the physical properties of the collagen matrix. The two most abundant collagen types in human myocardium are types I and III. Type I appears to form chiefly thick fibres and is most abundant in tissues requiring tensile strength, such as bone and tendon (Toole et al., 1972; Hanson and Bentley, 1983). Type III collagen forms thinner fibres and relatively high proportions are found in elastic tissue such as blood vessels and lung (Gay and Miller, 1978; Barnes, 1985; Mays et al., 1988). The relative proportion of these collagens may therefore play an important role in determining the physical properties of the extracellular matrix. In the heart these two collagens appear to, coexist in the matrix, having similar distributions in the collagen struts and pericellular fibres of the endomysium, demonstrated using immunofluorescent techniques (Robinson et al., 1987, 1988). Based on the hypothesis that changes in collagen composition may play a role in the depression of cardiac function seen in myocardial disease in man, the purpose of the present study was to investigate collagen content and the relative proportions of the major collagen types in ventricles of patients with coronary artery disease and dilated cardiomyopathy known to have poor ventricular function. Materials
and
Methods
Collagen content was examined in normal and diseased left ventricular myocardium. All tissue samples were full thickness sections of myocardium with a surface area of approximately 1 cm2, They were generally removed from the diaphragmatic aspect of the left
et al.
ventricle at a level half way between the atrioventricular grove and the apex. Tissues were obtained from three groups. Control tissues were obtained at autopsy from 25 patients (17 male, mean age 57 years, range 19-78 years) without a history of cardiovascular disease, hypertension or exposure to cardiotoxic drugs. Most of these patients died of malignant diseases. A single biopsy was taken from each of the hearts of 15 patients undergoing orthotopic cardiac transplantation for severely impaired cardiac function. Of these, 10 (9 male, mean age 49.3 years, range 34-68 years) had coronary artery disease and 5 (4 male, mean age 3 1.4 years, range 12-49 years) had dilated cardiomyopathy without evidence of significant large vessel coronary artery disease. The coronary artery disease group had a history of at least one full thickness (Q wave) myocardial infarction. Both groups had extremely poor left ventricular function assessed by conventional cineangiography. All biopsies were obtained immediately after the heart was removed from the recipient. In patients with coronary artery disease, regions which were macroscopically replaced by scar tissue were avoided for the purpose of taking biopsies. All the tissues were frozen in liquid nitrogen and stored at - 30°C prior to analysis. Biochemical analysis Collagen concentration was estimated in a full thickness section of tissue following determination of the hydroxyproline levels by the method of Stegemann and Stadler (1967) with more recent modifications (Laurent et al., 1981a). Estimations of the relative amount of types I and III collagen were also made on full thickness sections of tissue as described previously (Laurent et al., 1981 b; Kirk et al., 1984). Briefly, semi-purified collagens were obtained following repeated homogenization of the tissue in phosphate buffered saline (PBS) and 2% sodium dodecyl sulphate (SDS). This material was acetone-dried, then subjected to cyanogen bromide cleavage. The resulting peptides were separated by sodium dodecyl sulphate polyacrylamide gel electrophoresis (unreduced) along with known amounts of purified collagens (with a 3 : 2 ratio of type
Collagen in Heart I : III collagen) and stained (with PAGE blue 83, Fisons plc, Loughborough, England). Gels were destained in 8% (v/v) acetic acid for at least 24 h and each track was scanned using a Gilford 2600 spectrophotometer. The areas beneath the peaks were calculated by computerized integration. The ~1~(I)CB8 and the xl(III)CB5 peptides were used to quantitate types I and III respectively. Standard curves were constructed for these two peptides and the relative proportion of the collagen types were calculated for the tissue samples. Statistical analysis was performed using the Mann Whitney-L’ test (von Fraunhofer and Murray, 1976). However, data are presented as means + S.E.M. Mean values for patient groups were considered significantly different from controls when the test gave a P value of less than 0.05. Results The collagen concentration in the control left ventricle was 48.6 f 4.1 mg/g dry weight of tissue. There was no correlation between age or sex of the patient and collagen concentration (data not shown). The concentration of collagen in the left ventricle of patients with dilated cardiomyopathy was double that of normal ventricles (P < 0.01, see Table 1). In the group with coronary artery disease the concentration increased by about 30% but this was not significant at the 5% level. Figure 1 shows densitometric scans of the cyanogen bromide-derived peptides for standard collagens and collagens from heart tissue following electrophoresis. The peptides used TABLE
1. Collagen concentration left ventricle
and
the relative
Disease
to quantitate types I and III (cri(I)CB8 and ~(1(III)CB5 respectively) are indicated in the figure. They were both sufficiently well separated from adjacent peptides to allow accurate quantitation based on curves constructed from known loadings of collagen standards run on the same gel (see Fig. 1 (a) and Laurent et al., 1981b). Small quantities of collagen types IV (a basement membrane collagen), V and VI will also be present, however their contribution to the total amount of collagen is very small, the distinctive cyanogen bromide peptides for these collagen types described previously (Laurent et al., 1981 b) could not be detected on these gels. Table 1 also shows that the proportion of type III (compared with type I plus III) was lower in both coronary artery disease (by P < 0.01) and cardiomyopathy groups ;lz;1;‘4”,, P < 0.05), compared with controls. This was not due to age or sex differences between groups since no correlation was found between either of these parameters and the proportion of type III collagen in the control tissue samples (see Fig. 2). The estimated concentration of types I and III collagen, derived from the total collagen concentration and the proportion of each type, is also shown in Table 1. The shift in collagen types was apparently achieved by an increase in both type I and III, but there was a disproportionate increase in type I. Discussion The mechanical properties depend both on contractile proportion
Total collagen concentration (mg/g Control
CAD (n = 10) DCM
(n = 5)
dry
wt)
48.6 + 4.1 63.5 f 9.8 95.3 + 9.7 t
1159
of types
I and
III
Type
I
in normal
concentration y. III 41.8 + 1.1 34.6 f 1.5 t 35.8 + 2.8*
(mdg
dry
of the ventricle and connective
wtj
28.3 t 2.5 41.5 + 6.1 61.2 + 9.6t
and disrasrd
Type III concentration (mg/g dry wt: 20.3 + 1.7 22.0 &- 2.4 34.1 &- 4.8t
Values shown arc mean & S.E.M. Total collagen concentration (column I) is expressed as mg/g dry weight of tissue (mg/g dry wt). The proportion of type III (column 2) is expressed as the percentage oftype III collagen compared with the total of type I plus III. Columns 3 and 4 (concentration of the individual types) are derived from the data in the first two columns and do not take into consideration a contribution to the total collagen concentration by types IV, V and VI. CAD = coronary artery disease; DCM = dilated cardiomyopathy; * = P < 0.05 with respect to controls; t = P < 0.01 with respect to controls. For control total collagens, n = 15; for control (I0 III, n = 25.
J. E. Bishop
1160
“‘Lr-
p(I)
et al.
CB8
0
$ 0.9 (b) ,’ 0
o,(I)CB8
5 2
0
20
40
60
80
100
FIGURE I. Densitometric SCIIRJ of cyanogen bromide peptides obtained following separation by SDS-polyocrylamide gel electrophoresis. Figure 1 (a) shows a scan ofpurified collagen standard containing type I and III collagen in a ratio of3 : 2. Also shown are gel tracks ofstandards containing 6-24 pg type I collagen and 4-16 fig type III. The scan was obtained from the third standard (18 pg type I + 12 pg type III). Figure I(b) shows the scan obtained from a sample of myocardium obtained from a patient with dilated cardiomyopathy with the track from which it was obtained. The two bands used to quantitate types I and III collagen, a,(I)CB8 and 011(III)CBS, are indicated.
tissue elements. In general, studies regarding the structure-function relationships have centred on theoretical modelsfor the arrangement of the contractile and elastic elements
50
r
’I
0*
l 0
.
0
40
8
l *
.
l l
l
l 0 0
El “p 2 30 a?
0
0 l
0
0
0
20 I
20
40
I
60
00
Age (years)
FIGURE 2. The relationship between age and Qje III collagen concentration in normal left ventricle. There is no correlation between age of patient and proportion of type III collagen. l = male, 0 = female.
(Sonnenblick, 1964; Spotnitz et al., 1966; Streeter and Hanna, 1973). Anatomical and morphological studies have confirmed the highly ordered nature of ventricular muscle fibre array in man (Greenbaum et al., 1981). The presence of a well developed network of connective tissuein skeletal muscle was highlighted in the 1930’s.Banus and Zeitlin (1938) demonstrated that the passive length/tension curves were dictated in part by the presenceof the connective tissuenetwork. Development of high resolution scanning electron microscopy has facilitated detailed examination of the arrangement of collagen in cardiac muscle. Borg and Caulfield ( 1981) defined the collagen network in three parts. First, a weave ofcollagen bundlessurrounding myocytes, dividing them into groups. Second, collagen struts connecting myocytes within groups, and third, collagen struts connecting myocytes to adjacent capillaries. They suggested that the intermyocyte connections maintain lateral cell to cell alignment especially during diastole when the musclediameter
Collagen
in Heart
decreases. The capillary to myocyte collagen struts may be important in maintaining capillary to myocyte relationships during systole as the latter’s diameter increases, and may allow capillary blood flow to occur during systole in spite of increased wall stress at that time. The concentration of collagen in the human left ventricle, based on hydroxyproline content, was shown in the present study to have increased by a factor of two in dilated cardiomyopathy compared with controls, whilst in coronary artery disease the concentration increased by 3Oq&. Such increases may partly explain altered cardiac function in these groups of patients. It seems unlikely that the values obtained in the control group were influenced by previous treatment since samples were not taken from patients exposed to cardiotoxic agents, and our values are similar to those found in previous studies (Montfort and Perez-Tamayo, 1962; Caspari et al., 1977). In order to try to characterize such changes techniques of non-invasive tissue characterization using ultrasound have been developed (Shaw et al., 1984). The development of fibrotic regions in hypertrophied left ventricles has been demonstrated using a colour coding technique based on variations in the amplitude of reflected ultrasound (Mimbs et al., 1980; Shaw et al., 1984). In patients with left ventricular hypertrophy, increased myocardial echo amplitude correlated strongly with abnormal diastolic function as well as with T wave changes in the ECG (Shapiro et al., 1984). Use of such techniques combined with further studies of collagen composition and distribution in diseaseis likely to increase our understanding of the underlying mechanismsof ventricular disease. Tissue samples without macroscopic evidence of scar tissuewere deliberately selected for our study. Clearly, changesin collagen composition are not restricted to the infarcted regions. Infarct zones have been shown to have a much greater collagen concentration than normal regions (Judgutt and Amy, 1986) and the affected areas have mechanical properties markedly different from viable myocardium (Shaw et al., 1984).Judgutt and Amy i 1986) demonstrated an increase in collagen concentration of more than IO-fold in transmural sections taken at the infarct centre 6
Disease
1161
weeks after experimentally induced myocardial infarction in the dog. At the infarct border there was still a 5-fold greater collagen concentration. However, normal collagen concentrations were found in zones distant from the infarct in these animals. Our results may differ from the animal study for a number of reasons.The extent and duration of the diseaseis likely to be much greater in the human study, and also species variations have been observed in the timing of the remodelling process (Judgutt and Amy, 1986). In the present study, samples were taken from tissuewithout evidence of scarring, assessed at the macroscopic level. The increase in collagen concentration we have demonstrated may be due to a more diffuse interstitial fibrosis, possibly confined to a particular section of myocardium. Since we measured collagen concentration in full thickness sections we have no information on the distribution of the excesscollagen deposited. Further studiesare required to examine the collagen composition in the endocardial, midwall and epicardial regions. An alternative explanation for an increase in collagen concentration in the regions distant from the scar tissueis that the apparently normal myocardium may try to compensate for the lossof viable muscle cells resulting in regional hypertrophy (Roberts et al., 1983; Pasternak et al., 1988). This may be accompanied by an increase in collagen synthesis, since collagen metabolism appears to be stimulated by increasedworkload in the heart (Skosey et al., 1972;Bonnin et al., 1981; Turner et al., 1986). However, this form of hypertrophy is thought to be a volume-overload type (Roberts et al., 1983) which is not normally associated with an increase in collagen concentration (Bartosova et al., 1969; Michel et al., 1986). Further studies are required to determine the mechanismsinvolved in stimulating collagen production in these areas. The changesin collagen concentration, determined biochemically, may not be sufficient to explain fully the collagen related increaseiu myocardial stiffness. Changes in the composition, with regard to the different collagen types, as discussedbelow, may contribute to changes in diastolic function. In addition. rearrangement of the collagen matrix appears to be an important determinant (Weber ~1al.,
1162
J. E. Bishop
1988; Weber, 1989; Doering et al., 1988). An increasein thicknessof the perimysial tendons, weaves and strands was seen in established pressure-overload hypertrophy in the macaque, and newly formed fibres filled the intermuscular spaces(Weber et al., 1988). Borg et al. (1981) observed a correlation between the thickness of the collagen struts in different speciesand the stiffness of isolated papillary muscles.Clearly, morphologic and morphometric studies provide a more detailed description of the remodelling process.This is particularly true where the fibrosis is limited to particular regions of the myocardium. For example, a perivascular fibrosis was observed around small coronary arteries in rat hypertrophied left ventricle (Jalil et al., 1988). In the present study we demonstrated that in myocardial diseasetype III collagen accounts for approximately one third of the total collagen, compared with over 40% in normal humans, and changes in the concentration of types I and III in diseasedmyocardium suggested that a disproportionate amount of newly formed matrix was type I collagen. Comparable changes to those seen here for collagen types I and III have been observed during the development of right ventricular hypertrophy in rabbits (Turner and Laurent, 1986). However, in an animal model of left ventricular hypertrophy caused by gradual onset of hypertension, an increase in type III collagen was observed (Weber et al., 1988). In addition a lower control value was found. A similar trend was observed in copper deficiency-induced cardiac hypertrophy in rats (Dawson et al., 1982). An increasein type III was also seen in experimentally induced hypertrophy in rats following aortic stenosisor renovascular hypertension (Medugorac and Jacob, 1983). There are several explanations for these differences. Firstly the above mentioned studies employed pepsin digestion which, at best, extracts 75% of the total collagen, and this may be substantially reduced in the hypertrophied heart (Iimoto et al., 1988). Love11 et al. (1987) suggested that type III collagen may be less susceptible to pepsin digestion than type I, which might explain why the ratio of type I: type III was lower in control groups in previous studies than described here. In the present study, cyanogen
et al.
bromide digestion was used after first semipurifying the collagen by SDS and PBS washes. This has been shown to extract 80-90% of the total collagen (Laurent et al., 1981b) although we have not examined the extractibility in diseasedhearts. Although differences in the extractibility of the two collagen types has been demonstrated this is not a uniform finding (Chan and Cole, 1984; Maurel et al., 1987). A second explanation for the differences is that increasedtype III collagen may be transient, the normal ratio being restored at a later time point (Weber et al., 1988). In the present study the stage of the diseaseis not known, although it is likely to be advanced, a consequence of which may be an increase in the proportion of type I. Thirdly, the experimental modelsof cardiac hypertrophy cannot be directly compared with the ventricular disease studied here, since changes in the relative proportion of the collagen types may be dependent on the different disease processes.Control values may also differ depending on the speciesunder investigation. The increase in the proportion of type I collagen observed in the apparently nonfibrotic regionsmay further affect the mechanical properties of the ventricle. Although much of the evidence to date is circumstantial, the relative proportion of types I and III in the connective tissuematrix may influence the physical properties of a tissue. These two collagen types appear to colocalize in myocardium (Robinson et al., 1988) and placenta (Amenta et al., 1986). Using immunohistochemistry, type I was seen to form thicker fibres than type III, which were found forming a weave around the type I fibres in placental tissue(Amenta et al., 1986). The nature of the underlying changes in collagen metabolism in human disease remains unclear, but alterations in both synthetic and degradative processesmay be important. Based on the changes in collagen types (Turner and Laurent, 1986) and collagen metabolism (Skosey et al., 1972; Turto, 1977; Lindy et al., 1972; Bonnin et al., 1981; Turner et al., 1986) seenin animal studies of ventricular hypertrophy, it would appear that there is increasedsynthesisof both types I and III, with the increase in type I being greater. It was also suggestedin animal studies that
Collagen
in Heart
changes in breakdown may be important, particularly in the early stages (Turner et al., 1986). This is supported by recent studies on the arrangement of the collagen network in which the thin fibrils disappeared in the evolutionary phase of cardiac hypertrophy (Weber et al., 1988). It is not clear whether the increase in synthesiswas mediated by an increased production by individual cells or an increase in the total number of fibroblasts, although there is evidence for active replication of fibroblastic cells suggestingthe latter mechanism may be important (Morkin and Ashford, 1968; Grove et al., 1969). Skosey et al. ( 1972) demonstrated an increasein collagen production prior to the increase in cell number suggesting also an increasein synthetic activity per cell. In summary, this study has demonstrated enhanced myocardial collagen deposition, particularly of type I collagen, in patients with myocardial disease.The data presented do not
Disease
1163
allow us to conclude that changesin collagen deposition and structure are primary events in the development of impaired ventricular function. Nevertheless,they open up the possibility that the impaired ventricular function seenin these patients may be in part related to a remodelling of connective tissue in what appears macroscopically normal ventricle, rather than due entirely to a biochemical lesion in the contractile cells. Acknowledgements
This work was funded by the British Heart Foundation. Dr Jill Bishop (nee Turner) is currently a British-American Research Fellow of the American Heart Association and British Heart Foundation at the University of Vermont, USA. The authors wish to thank Dr Richard Carter (Royal Marsden Hospital, Sutton, UK) for his co-operation in providing tissuesamples.
References AMENTA PS, GAY, S, VAHERI A, MARTINEZ-HERNANDEZ A. (1986). The extracellular matrix is an integrated unit: Ultrastructural localization of collagen types I, III, IV, V, VI, fibronectin and laminin in human term placenta. Collagen Rel Res 6: 125-152. BANUS MG, ZEITLIN AM (1938) The relation of isomeric tension to length in skeletal muscle. J Cell Comp Physiol 12: &20. BARNES MJ (1985) Collagens in atherosclerosis. Collagen Rel Res 5: 65-97. BARTOSOVA D, CHVAPIL M, KORECKY B, POUPA 0, RAKUSAN K, TIJREK Z, VIZEK M (1969) The growth of the muscular and collagenous parts of the rat heart in various forms of cardiomegaly. J Physiol200: 285-295. BING OHL, MATSUSHITA S, FANBURC BL, LEVINE HJ (1971) Mechanical properties of rat cardiac muscle during experimental hypertrophy. Circ Res 28: 23G-245. BONNIN CM, SPARROW MP, TAYLOR RR (1981) Collagen synthesis and content in right ventricular hypertrophy in the dog. Am J Physiol241: H708-H713. BORG TK, CAULFIELD JB (1981) The collagen matrix of the heart. Fed Proc 40: 2037-2041. BORG TK, RANSON WF, MOSLEHY FA, CAULFIELD JB (1981) Structural basis of ventricular stiffness. Lab Invest 44: 49-54. CASPARI PG, NEWCOME M, GIBSON K, HARRIS P (19771 Collagen in the normal and hypertrophied human ventricle. Cardiovasc Res 11: 554-558. CHAN D, COLE WG (1984) Quantitation of type I and III collagens using electrophoresis of alpha chains and ryanogen bromide peptides. Anal Biochem 139: 322-328. DAWSON R, MILNE G. WILLIAMS RB (1982) Changes in the collagen of rat heart in copper-deficirncy-induced cardiac hypertrophy. Cardiovasc Res 16: 559-565. DOERING CW, JALIL JE, JANICKI JS, PICK R, ACHILI S, ABRAHAMS C, WEBER KT (1988) Collagen network remodrlling and diastolic stiffness of the rat left ventricle with pressure overload hypertrophy. Cardiovasc Res 22: 686-695. GAYS, ~MILLER EJ (1978) In: Collagen in the physiology and pathology ofconnective tissue. S Gay and EJ Miller (Eds). Gustav Fischer Verlag. Stuttgart. New York. GREENBAUM RA, Ho SY, GIBSON DG, BECKER AE, ANDERSON RH (1981) Left ventricular fibre architecture in man. Br Heart J 45: 248-263. GROVE D, NAIR KG. ZAK R (1969) Biochemical correlates of cardiac hypertrophy. III. Changes in DNA content; the relative contribution of polyploidy and mitotic activity. Circ Res 25: 463471. HANSON AN, BENTLEY JP (1983) Quantitation oftype I and type III collagen ratios in small samples of human tendon, blood vessels and atherosclerotic plaque. Analyt Biochem 130: 3240. HESS OM, RITTER M, SCHNEIDER J, GRIMM J, TURINA M, KRAYENBRUEHL HP (1984) Diastolic stiffness and myocardial structure in aortic valve disease before and after valve replacement. Circulation 69: 855-865.
1164
J. E. Bishop et al.
DS, COVELL JW, HARPER E (1988) Increase in cross-linking of type I and type III collagens associated with volume-overload hypertrophy. Circ Res 63: 39%408. JALIL JE, DOERING W, JANICKI JS, PICK R, CWRK WA, ABRAHAMS C, WEBER, KT (1988) Structural vs. contractile protein remodelling and myocardial stiffness in hypertrophied rat left ventricle. J Mol Cell Cardiol20: 11791187. JALIL JE, DOERING CW, JANICIU JS, PICK R, SHROFF SG, WEBER KT (1989) Fibrillar collagen and myocardial stiffness in intact hypertrophied rat left ventricle. Circ Res 61: 104-1050. JUDGUTT BI, AMY RWM (1986) Healing after myocardial infarction in the dog: changes in infarct hydroxyproline and topography. J Am Co11 Cardio17: 91-102. KIRK JME, HEARD BE, KERR I, TURNER-WARWICK M, LAURENT GJ (1984) Q uantitation of types I and III collagen in biopsy lung samples from patients with cryptogenic fibrosing alveolitis. Collagen Rel Res I: 169-182. KRAYENBUEHL HP, Hess OM, MONRAD ES, SCHNEIDER J, MALL G, TUIUNA M (1989) Left ventricular myocardial structure in aortic valve disease before, intermediate, and late after aortic valve replacement. Circulation 79: 744-755. LAURENT GJ, MCANULTY RJ, CORRIN B, COCKERILL P (1981a) Biochemical and histological changes in pulmonary fibrosis induced in rabbits with intratracheal bleomycin. Eur J Clin Invest 11: 441-448. LAURENT GJ, COCKERILL P, MCANULTY RJ, HASTINGS JRB (1981b) A simplified method for quantitation of the relative amounts of type I and type III collagen in small tissue samples. Analyt Biochem 113: 301-312. LINDY S, TURTO H, Urrro J (1972) Protocollagen proline hydroxylase activity in rat heart during experimental cardiac hypertrophy. Circ Res 30: 205-209. LOVELL CR, SMOLENSKI KA, DUANCE JC, LIGHT ND, YOUNG S, DYSON M (1987). Type I and III collagen content and fibre distribution in normal skin during ageing. Br J Dermatol 117: 41%428. MAUREL E, SHUT~LEWORTH CA, Bourssou H (1987 Interstitial collagens and ageing in human aorta. Virchows Arch A. 410: 383-390. MAYS PK, BISHOP JE, LAURENT GJ (1988) Age-related changes in the proportion of types I and III collagen. Mech Ageing Dev 45: 203-2 12. MEDUGORAC I, JACOB R (1983) Characterization of left ventricular collagen in the rat. Cardiovasc Res 17: 15-Z I. MIMBS JW, O’DONNELL M, BAUWENS D, MILLER JM, SOBEL BE (1980) The dependence of ultrasonic attenuation and backscatter on collagen content in dog and rabbit hearts. Circ Res 47: 4958. MICHEL JB, SALZMANN JL, NLOM MO, BRUNEVAL P, BARRES D, CAMILLERI JP (1986) Morphometric analysis of collagen network and plasma perfused capillary bed in the myocardium of rats during evolution ofcardiac hypertrophy. Basic Res Cardiol81: 142-154. MONTFORT TW, PEREZ-TAMAYO R (1962) The muscle ratio in normal and hypertrophic human hearts. Lab Invest 11: 463470. MORKIN E, ASHFORD TP (1968) Myocardial DNA synthesis in experimental cardiac hypertrophy. Am J Physiol 215: 1409-1413. PASTERNAK RC, BRAUNWALD E, SOBEL BE (1988) In: Heart Disease. A textbook of cardiovascular medicine. E. Braunwald (Ed.), W. B. Saunders, Co. pp 122221313. ROBERTS CS, MACLEAN D, BRAUNWALD E, MAROKO PR, KLONER RA (1983) Topographic changes in the left ventricle after experimentally induced myocardial infarction in the rat. Am J Cardiol51: 872876. ROBINSON TF, FACTOR SM, CAPASSO JM, WITTENBERG BA, BLUMFELD 00, SEIFTER S (1987) Morphology, composition and function of struts between cardiac myocytes of rat and hamster. Cell Tiss Res 2* 247-255. ROBINSON TF, COHEN-GOULD L, FACTOR SM, EGHBALI M, BLUMENFELD 00 (1988) Structure and function of connective tissue in cardiac muscle: collagen types I and III in endomysial struts and pericellular fibers. Scanning Microscopy 2: 1005-1015. SHAPIRO LM, MOORE RB, LOGAN-SINCLAIR RB, GIBSON DG (1984) Relation of regional echo amplitude to left ventricular function and the electrocardiogram in left ventricular hypertrophy. Br Heart J 52: 99105. SHAW TRD, LOGAN SINCLAIR RB, SURXN C, MCANULTY RJ, HEART B, LAURENT GJ, GIBSON DG (1984) Relation between regional echo intensity and myocardial connective tissue in chronic left ventricular disease. Br Heart J 51: 46-53. SKOSEY JL, ZAK R, MARTIN AF (1972) Biochemical correlates of cardiac hypertrophy: V. Labeling collagen, myosin and nuclear DNA during experimental myocardial hypertrophy in the rat. Circ Res 31: 145-157. SONNENBLICK EH (1964) Series of elastic and contractile elements in heart muscle: changes in muscle length. Am J Physiol 287: 1330-1338. SPOTNITZ HM, SONNENBLICK EH, SPIRO D (1966) Relation of ultrastructure to function in the intact heart: Sarcomere structure relative to pressure-volume curves of intact left ventricle of dog and cat. Circ Res IS: 44-66. STEGEMANN H, STADLER K (1967) Determination of hydroxyproline. Clin Chim Acta 18: 267-273. STREETER DD, HANNA WT (1973) Engineering mechanics for successive states in canine left ventricular myocardium. 2 Fibre angle and sarcomere length. Circ Res 33: 656664. THIEDEMANN KU, HOLUBARSCH CH, MEDUGORAC I, JACOB R (1983) Connective tissue content and myocardial stiffness in pressure overload hypertrophy. A combined study of morphologic, morphometric, biochemical, and mechanical parameters. Basic Res Cardiol78: 141-155. TOOLE BP, KANG AH, TRELSTAD RL, GROSS J (1972) Collagen heterogeneity within different growth regions of long bones in rachitic and non-rachitic chicks. Biochem J 127: 715-720. TURNER JE, LAURENT GJ (1986) Increased type I collagen compared with type III during right ventricular hypertrophy in rabbits. Biochem Sot Trans II: 1079-1080. IIMOTO
Collagen
in Heart
Disease
1165
TURNER JE, OLIVER MH, GUERREIRO D, LAURENT GJ (1986) Collagen metabolism during right ventricular hypertrophy following induced lung injury. Am J Physiol 251: H915-H919. TURTO H (1977) Collagen metabolism in experimental cardiac hypertrophy in the rat and the effect of digitoxin treatment. Cardiovasc Res 11: 358-366. VON FRAUNWOFER JA, MURRAY JJ (Eds) (1976) In: Stati~fics in Medical, Dental and Biological Studies. Tri-Med Books Ltd, London. pp 77-84. WEBER KT, JANICKI JS, SHROFF SC, F’ICK R, CHEN RM, BAWEY RI (1988) C o II a g en remodeling of the pressureoverloaded, hypertrophied nonhuman primate myocardium. Circ Res 62: 757-765. WEBER KT ( 1989) Cardiac interstitium in health and disease: The fibriljar collagen network. J Am Co11 Cardiol 13: 1637-1652.
