Collagen Rel. Res. Vol. 5/1985, pp. 337-347

Collagen Types in Neocartilage Tissue Resulting from Rib Perichondrial Graft in an Articular Defect - a Rapid Semi-Quantitative Methodology

DAVID AMIEL, FRED L. HARWOOD, MARK F. ABEL and WAYNE H. AKESON

Division of Orthopaedics and Rehabilitation, M-030 University of California at San Diego, La Jolla, CA 92093, USA.

Abstract

A method for estimating type II to type I collagen ratios in small tissue samples has been developed. The cyanogen bromide peptides of the tissue collagens were analyzed by SDS-gel electrophoresis. Marker peptides representative of each collagen type were established and their relative amounts determined by integration of the stained peptide bands following gel scans. Marker peptide ratios were then computed for each of several standard type II1type I mixtures and these peptide ratios were mathematically correlated with the corresponding type II1type I collagen ratios. A linear relationship between marker peptide ratio and collagen type ratio was established. This relationship was applied to the analyses of type II1type I ratios in sampies of rib perichondrium and neocartilage derived from perichondrial graft repairs of full thickness femoral condyle defects. The results indicated that perichondrial grafts synthesize both types II and I collagens and that the proportion of type II increases with increasing post-transplant time.

Key words: cartilage repair, collagen typing, perichondrial graft.

Introduction

Alterations in collagen types and their metabolism in connective tissues are now recognized to be significant in a wide variety of clinical conditions ranging from newly identified heritable disorders to the conditions where metabolie responses to infection, foreign bodies, tissue repair and neoplasia occur (Gay et al., 1978; and Nimni and Deshmukh, 1973). It has been reported, for example, that while type II collagen is synthesized in normal articular cartilage, type I collagen may be additionally synthesiz­ed by osteoarthrotic articular cartilage (Deshmukh and Hemrick, 1976; and Nimni and Deshmukh, 1973), suggesting a partial change to a collagen type more typical of fibrillar structures. Indeed, type II collagen, which comprises the vast majority of the collagen in hyaline cartilage, is found mixed with larger amounts of type I collagen in complex fibrillar structures such as annular fibrosus (Eyre and Muir, 1975).

Experiments using transplanted rib perichondrium as a source for repair of full­thickness defects in the femoral condyle of a rabbit have been carried out in this

338 D. AmieI, F. L. Harwood, M. F. Abel and W. H. Akeson

laboratory. These experiments have resulted in the growth of neocartilage that grossly and histologically resembles hyaline cartilage. Because of the probable importance of collagen type to tissue function, it was necessary to develop a method for estimating type II1type I collagen ratios ip sm all neocartilage tissue sampIes.

Various methods have been used to identify and quantify the different collagen types making up collagenous tissues. Differential solubility in neutral salt solution has been used to selectively salt out and weigh specific collagen types (Bailey et al., 1975). Other methods have been developed which involve partially digesting collagen with cyanogen bromide (CNBr) and separating the resultant peptides by a combination of ion-ex­change and molecular sieve chromatography (Chung et al., 1974; Epstein, 1974; and Eyre and Muir, 1975). .

Both of these approaches, however, have the disadvantage of requiring relatively large amount (>50 mg) of tissue in order to be successfully applied to collagen type ratio estimation. In light of this, several investigators have employed immunohistoche­mical techniques to localize and estimate collagen type content in tissues (Gay et al., 1976; and Lane et al., 1982). This has become a powerful tool and supplement to biochemical studies. The success and reproducibility of histological approaches, howe­ver, depend in large part on the specificity and quality of antibodies employed. The use of insufficiently purified and characterized antibodies could therefore leadto contro­versial resu1ts.

Semi-quantitative analysis of the electrophoretic patterns of the CNBr collagen pep­tides provides a complementary method of collagen type estimation, requiring only small amounts (5 mg) of tissue. Because of molecular weight differences, certain of the larger CNBr peptides can be separated from one another and used as markers for collagen types. SDS gel elctrophoresis of CNBr peptides has been applied in serveral variations to the estimation of type I1type III ratios in tisslle (Reiser and Last, 1980; Laurent et al., 1981; and Hansen and Bentley, 1983). At this time, however, semi­quantitative electrophoretic analysis of type I1type 11 ratios in tissue has not yet been described.

This communication focuses on the separation by electrophoresis of these type­specific (marker) peptides in a rabbit model and describes the relationship between the ratio of the type I/type 11 marker peptides quantitated on the gels and the ratio of collagen types yielding those peptides. A method is thus proposed for estimating the ratios of collagen types present in limited amounts of tissue based on a semi-quantita­tive assessment of the electrophoretic patterns of the CNBr peptides on SDS-10 % polyacrylamide.

This proposed method is enabling us to investigate the use of perichondrial grafts for biological resurfacing, and to characterize the newly formed cartilage in a rabbit model which utilizes rib perichondrium to repair full thickness defects in the femoral condyle.

Materials and Methods

Surgical Model

Male adult New Zealand white rabbits weighing 2.5-3.0 kg were utilized as describ­ed previously (Coutts et al., 1984a). The cartilaginous portion of the left lower rib was removed and was lIsed to harvest the perichondrial layer yielding a thin rectanglilar strip measuring 1 cm X 0.5 cm. The femoral condyle was exposed and a hollow cylindrical drill bit was used to obtain a bone core. This co re was prepared for grafting

Rib Perichondrial Graft in Articular Ddect 339

by removing the cartilaginous surface with a burr down to cancellous bone. The graft was draped over the core, with the surface facing the rib cartilage exposed to the joint, and sutures of 6-0 monofilament nylon were used to fix the graft. The perichondrium­covered core was replaced in the defect so that the surface was recessed approximately 1 mm below the surrounding cartilage. The rabbits were allowed cage activity for two study periods: 6 and 12 weeks post-surgery.

Experimental Tissue Preparation

Specimens of normal perichondrium and neocartilage dissected from the 6 and 12 weeks postoperative groups (Fig. 1) were obtained immediately after sacrifice and extracted for 24 h at 22 oe with 4 M guanidine hydrochloride to remove noncollage­nous proteins. The insoluble residues were washed exhaustively with H2ü and freeze­dried.

Preparation of Types land II Collagen Standards

Hyaline cartilage was extracted with 4 M guanidine hydrochloride to remove non­collagenous proteins and standard type II collagen was isolated from the residue by limited solubilization in pepsin and precipitation in neutral salt solution (Eyre and

Fig. 1. Neocartilage tissue (arrows) resulting from perichondrial grafts of femoral condyle defects. A. 6 weeks post transplant B. 12 weeks post transplant

340 D. Amiel, F. L. Harwood, M. F. Abel and W. H. Akeson

Muir, 1975). Standard type I collagen was isolated from pepsin digests of rabbit skin by repeated differential salt precipitation (Epstein, 1974).

CNBr Cleavage of the Collagens

Sampies of standard types I or II collagen (10 to 100 mg) and experimental tissues (3-5 mg, after removal of non-collagenous proteins) were digested with CNBr in 70% formic acid (Epstein et al., 1971). The duration of CNBr digestion on neocartilage tissues was increased to 12 h in order to improve the percentage of tissue solubilized. The digests were centrifuged to remove any undissolved tissue and the supernatants were then lyophilyzed. Hydroxyproline analyses of both the supernatants and undi­gested residues of several CNBr-treated neocartilage sampies were performed in order to determine the efficiency and recovery of CNBr digestion. The peptides resulting from this treatment were preliminarily separated from contaminating materials by chromatography on a column of Bio-Gel P-2 using 0.1 M acetic acid as eluting buffer. The peptide digests were then lyophilized and stored.

CM-Cellulose Chromatography of Standard CNBr Peptides

50 to 200 mg each of CNBr peptide digests from purified types I or II collagen were dissolved at 45°C in 10 to 50 ml of starting buffer: .03 M Na3C6HsÜ7 . 2HzÜ, .01 M NaCl, pH 3.6. The solution was filtered to remove insoluble matter and de-aerated prior to chromatography on a column (1.0 X 10 cm or 1.6 X 12 cm) of CM-cellulose (Whatman, CM-32). Elution of the peptides was achieved using a linear salt gradient from .01 M to .16 M NaCl. Adopting the conventional method, the peptides were numbered in the order of their elution from CM-cellulose.

SDS-Polyacrylamide Gel Electrophoresis

The CNBr digests of purified types land II collagen were accurately and separately weighed into two test tubes. An exact volume of .01 M sodium phosphate, 0.5% SDS, pH 7.2 was added so that the peptide concentration in each tube was 2.00 mg/mI. The tubes were then incubated at 40°C for one hour. The incubated CNBr peptides of types land II collagens were mixed precisely in five different ratios (II : I) ranging from 0.10 to 3.00. Each standard mixture was prepared in triplicate. A 40 1-11 aliquot con­taining a total of 80 I-Ig peptides from each mixture was layered separately onto 10% polyacrylamide disk gels, and electrophoresis was performed for 5 to 6 h under a constant current of 7 mA per gel.

Electrophoresis of the CNBr digests of experimental tissues was performed under identical conditions to those used for type I and type II standards. Additionally "elec­trophoretic recovery" of the CNBr peptides was assessed by introducing into neocartil­age digests an internal standard. Type I-type II mixtures of known proportions were thus added to some sampies, and electrophoresis was then performed. The sampies to which an internal standard was added were also analyzed without such a standard, and the results were compared. All gels stained 12 h in 50% methanol containing 0.25% Coomassie Brilliant Blue-R and 9.2% acetic acid. Background color was removed by suspending the gels for 12 h in a continuously stirred solution of 7% acetic acid containing 40% methanol followed by 18 h in 7% acetic acid containing 25% methanol. Gels were then transferred to tubes containing 7% acetic acid. As soon as

Rib Perichondrial Graft in Articular Oefect 341

swelling was camplete, the gels w ere scanned by linear transport at 590 ml-l in order to quantitate the relative optical d ensities of the stained peptide bands. The recorded scans were electronically copied, and integration of the peptide bands was achieved by a digitizer linked to a PDP- I1103 computer.

Results

CM-Cellulose Chromatography of Standard Types l and II CNB r Peptides

Identification of the larger C NBr p eptides ( molecular weight > 4000) was achieved by collecting the material under each peak and performing SDS-gel electrophoresis o n 10% polyacrylamide (Fig. 2). The relative migration positions for each of the CNBr

.6

~ .5 ::1

a,-CBB A E

0 .4 A '" ~ Q,(I)·C B1 + CBS ;; a,(/j·C86 'Cii .3 o.,j1rCB3 c:

'" 0 C; .2 .2 Q2 - CB3.5 ä. 1A 2A 3A 4A 5' .. 0 .1

80 160 240 Effluent Volume (mi)

.7 af - CB9 ... '0

.6 B :i B E .5

0

'" ~ ;;.4 O,III}CB8 .;;; c:

'" 0 .3 C; .2 '. 2. 3. ,. ä. .2 0

.1

80 160 240 320 Effluent Volume (mi)

Fig. 2. CM-cellulose chrornatography a nd subsequent SOS-gel electrophoresis o f the CNBr peptides frorn standard type I and type II collagen. A - left: CM-Cellulose chrornatograrn of t ype I peptides. A - right: SOS-gel electrophoresis of the CM-Cellulose t ype I p eptides.

* lane 1A: Type I CNBr peptides before CM-Cellulose chrornatography. B - left: CM-Cellulose chrornatograrn of t ype II peptides. B - right: SOS gel electrophoresis of the CM-Cellulose resolved t ype II peptides.

,." lane 1B: Type II CNBr peptides before CM-Cellulose chrornatography.

342 D. Amiel, F. L. Harwood, M. F. Abel and W. H. Akeson

peptides were noted and compared with those previously published for types land II collagen (Click and Bornstein, 1970; Epstein et al., 1971; Fietzek and Piez, 1969; and Miller, 1971). From these results it was determined that each collagen type has at least one CNBr peptide band of a molecular weight (gel position) specific enough to be used as a marker. Thus, for a mixture of types land II collagens, either al (l)-CB6 (Mr = 17,000) or a2(l)-CB3,5 (M r = 60,000) potentially serves as a marker for type I colla­gen, while al(II)-CBI0 (M r = 30,000) denotes the presence of type II collagen.

Electrophoresis and Quantitation of Collagen Types

For quantitation of collagen types it was decided to explore the correlation between marker peptide ratios (measured on SDS-polyacrylamide gels) and parent collagen type ratios. In order to test the hypothesis that the relative amounts (ratios) of marker CNBr peptides are proportional to the relative weights (ratios) of the parent collagen types yielding those peptides, known mixtures of type I-type 11 collagen CNBr peptides were prepared quantitatively as described in Methods. SDS-gel electrophoresis was per­formed and the gels were densitometrically scanned. The appropriate marker peptide bands were integrated and their ratios were computed for each gel. A least squares plot of marker peptide ratio vs. collagen type ratio is illustrated in Fig. 3. Ir is seen that the

- a ,(II)-CB10 ~ o 5

1111 11/1

Fig.3. Least squares plots of marker peptide ratios vs. type lIItype I collagen ratios.

left:

right:

al(IIl-CBlO a2(I) -CB3,S

al(ll)-CBlO al(l) -CB6

vs. ~ Type I

vs. ~ Type I

middle: SDS-gel electrophoresis of the five different type II1type I CNBr peptide ratios. Gel 1 0.11

2 0.33 3 0.67 4 1.50 5 3.00

Rib Perichondrial Graft in Articular Defect 343

ratio of al(II)-CB10 to a2(1)-CB3,5 varies linearly with the ratio of type II to type I collagen (r == 0.99). Likewise, the ratio of a1(II)-CB10 to a1(I)-CB6 also varies linearly with the ratio of types II to I (r == 0.99). Using the equation of a straight line, a relationship between marker peptide ratio and collagen type ratio was demonstrated as folIows:

(1)

(2)

[a1(II)-CB10]" [a2(1)-CB3,5] "

1.98" ". x wt. type II wt. type I

[al (II)-CB10]'" == 1.44"" x wt. type II [a2(1)-CB6]" wt. type I

* brackets indicate peptide peak areas. *,' slope t y-intercept

+ 0.18t

+ 0.07t

The respective slopes in equations (1) and (2) do not equal one since the absolute color yields for equimolar amounts of the marker peptides are not the same. This is to be expected, since their molecular weights are unequal. Also the color yield of a2(I)­CB3,5 is expected to be less since it is derived from the a2(I) chain, which constitutes just 1/3 of the type I molecule. Nevertheless, these equations demonstrate that the ratios of marker peptides are proportional to the ratio of collagen types, with the constants of proportionality represented by the slopes.

The technique described above was applied to the estimation of type II to type I ratio in a perichondrial transplant model for regeneration of cartilaginous tissue. Rib perichondrium was found to contain both types land II collagen, as established in our laboratories through differential salt precipitation of the collagen types. In order to evaluate the neocartilage tissues resulting from perichondrial transplants, it was neces­sary to estimate type II to type I ratios in both neocartilage and perichondrium tissues.

It was deemed preferable to analyze the CNBr digests of whole tissues (after removal of non-collagenous proteins) instead of pepsin extracts because it was thought that pepsin might preferentially extract one collagen type over another. Furthermore, neocartilage sampie amounts were so limited that pepsin extraction prior to CNBr digestion was impractical. However, in order to investigate the effect of pepsin on CNBr digestion, pooled sampies of rib perichondrium and neocartilage were treated with pepsin and the solubilized collagens were then digested with CNBr. The elec­trophoretic analyses of these digests were compared to those of whole tissue digests, and no significant differences in marker peptide ratios were observed. Hydroxyproline analysis of the supernatant and residue fractions of centrifuged CNBr digests of whole tissues revealed that greater than 95 percent of the neocartilage and perichondrial collagens was solubilized by CNBr under the conditions described in Methods.

The validity of our proposed procedure was reinforced by utilizing an internal standard in some sampies. Therefore, to individual neocartilage CNBr digests of known type I-type II composition, standards type I1type II of known proportions (75%/25% and 25%/75%) were added separately. Electrophoresis on each of the composite mixtures was performed and gel scans yielded values which corresponded to 100% ± 5 of the theroretical values calculated for each of the neocartilage standard composites.

Figure 4 shows representative gel scans of the CNBr digests of rib perichondrium and neocartilage (6 and 12 weeks post transplant). a2(1)-CB3,5 was chosen for the Type I marker peptide since, as can be seen in the scans, the other potential type I

344 D. Amiei, F. L. Harwood, M. F. Abel and W. H. Akeson

a b c

Fig.4. Linear transport scans of the SDS gels after electrophoresis of the CNBr peptide on 10% polyacrylamide.

A. Normal perichondrium. B. 6 week post-surgery neocartilage. C. 12 week post-surgery neocartilage.

marker, ut(I)-CB6, is considerably less distinct and therefore difficult to integrate accurately. The spreading of the u1 (I)-CB6 band proved to be a generalized phenome­non in neocartilage tissues. Applying equation (1), an estimate of the type II to type I ratio was obtained for these tissues. It can be seen in Table I that type 11 collagen is a minor component in rib perichondrium, but is the major component in the neocartilage tissues, especially in 12-week post transplant tissue. Since articular cartilage contains predominantly type 11 collagen (> 95%), there is an apparent trend toward the growth of hyaline-like cartilage collagen qs.opposed to fibrocartilage collagen which is predo­minantly type I (> 85%) in natbre. This is significant because a previous study of arthroplasty tissue has shown a tendency towards fibrocartilage growth (type I colla­gen) (Coutts et al., 1984 b).

Discussion

In a perichondrial transplant model using rabbits, it is essential that collagen typing estimation be performed on small sampies, since tissue availability is severely limited.

Table L Collagen Type Distribution.

Tissue

Rib Perichondrium 6 Week Neocartilage

12 Week Neocartilage

Type II: Type I Ratio'

0.4 ± .1 1.0 ± .2 1.4 ± .2

• Values expressed as me an ± standard error based on six specimens in each group.

Rib Perichondrial Graft in Articular Defect 345

Typically a dissection of neocartilage yields less than 10 mg (dry) tissue for analysis. A semiquantitative electrophoretic analysis of marker peptides resulting from CNBr di­gestion of such tissues provides a rapid and reliable method for estimating collagen type ratios.

This paper details the relations hip that exists between the ratio of type-specific CNBr peptide markers (as measured by the integrated peptide bands on SDS-I0% polyacrylamide) and the corresponding ratio of parent collagen types yielding those peptides. Previous studies have shown that SDS polyacrylamide gel electrophoresis is applicable to the estimation of components in a mixture of peptides (Fishbein, 1972) and, when specifically applied to the CNBr peptides of collagen under carefully con­trolled conditions, the absorbances of the Coomassie-stained peptide bands on poly­acryl amide gels are proportional to the concentration of peptides applied to those gels (Scott, 1976).

Relying on this principle, various techniques have been described for estimating relative amounts of collagen types in CNBr digests of small tissue sampies (Cole and Bean, 1979; and Weber et al., 1977). Problems with some of these techniques arise, however, because of the necessity of applying identical quantities of sampie digests to each gel for accurate comparisons. There is the further disadvantage of the need with some methodologies to quantitate and compare marker peptide band intensities from different gels. This is a potential problem, since staining and destaining inconsistencies are common. Also, the smaller molecular weight bands tend to fade rather quickly with time. The technique described in this study attempts to avoid some of these potential sources of error in quantitation of CNBr peptides by SBS-gel electrophoresis. This study demonstrates that in a rabbit model the ratio of the areas of certain type-specific marker peptide bands is proportional to the ratio of corresponding collagen types applied to that gel. Since peptide band intensities are compared only within a gel and not between gels, staining and destaining variations are of less concern. Also, the actual amounts of CNBr peptides applied to the gels are not critical to subsequent calculation of marker peptide ratios.

Certain limitations should, however, be considered when applying this technique to the analyses of tissues. The peptide al (1)-CB6, a potential marker for type I collagen, is known to contain an intermolecular cross-linking site (Kang, 1972), and its yield may thus depend in part on the susceptibility of these linkages to acid cleavage. Because variations in the amounts of al(I)-CB6 released could therefore occur, it should be used with caution as a type I marker. Indeed, in the gel scans of the CNBr digests of neocartilage, the al(I)-CB6 peak was poorly defined and was considered unsuitable for use as a quantitative marker of type I collagen. Conversely, a2(I)-CB3,5 was weil defined in the digests of all the tissues and proved a reliable type I marker, provided that type 11 to type I ratios fell into the range of 0.1 to 3.0. When this ratio surpasses 3.0 (> 75% type 11), the type I marker, a2(I)-CB3,5, becomes too small to integrate accurately and reproducibly.

The methodology could presumably be applied to the quantitation of collagen types comprising tissues of other species. However, it would be necessary to first standardize the relationships with collagen types purified from those species. Keeping these limita­tions under consideration, the method provides a rapid and simple technique for estimating collagen type ratios in tissues which contain a mixture of types land 11 collagens. The method requires very sm all amounts of tissue and, because of its speed and simplicity, is particularly useful in evaluating large numbers of small tissue sam­pies.

24 Collagen 5/4

346 D. Amiel, F. L. Harwood, M. F. Abel and W. H. Akeson

Aeknowledgements

Support from NIH grant AM 28467 and the Malcolm and Dorothy Coutts Institute for Joint Reconstruetion and Research is gratefully acknowledged. The authors also wish to

thank Dr. Y. K. Woo and Mr. Mike Furniss for their technical assistance.

References

Bailey, A. J., Bazin, S., Sims, T. J., LeLous, M., Nicoletis, C. and Delaunay, A.: Charaeteri­zation of the collagen of hypertrophie and normal sears. Biochim. Biophys. Acta 405: 412-421,1975.

Chung, E., Keele, E. M. and Miller, E. ].: Isolation and eharacterization of the cyanogen bromide peptides from the al (III) chain of human collagen. Biochemistry 13: 3459-3464, 1974.

Click, E. M. and Bornstein, P.: Isolation and characterization of the cyanogen bromide peptides from the al and a2 chains of human skin collagen. Biochemistry 9: 4699-4706, 1970.

Cole, William G. and Bean, Deborah, A.: Analysis of collagen cyanogen bromide peptides using electrophoresis in continuous concave gradient polyacrylamide gels. Anal. Biochem. 92: 183-188, 1979.

Coutts, R. D., Amiel, D., Woo, S. L-Y., Woo, Y-K. and Akeson, W. H.: Teehnieal aspeets of perichondrial grafting in the rabbit. Eur. Surg. Res. 16 (5): Oet., 1984a.

Coutts, R. D., Amiel, D., Harwood, F. L., and Bradley, G.: Charaeterization of arthroplasty tissue after 14 years post-cup arthroplasty: A morphologieal and bioehemieal assessment. J. Orth. Res.: in press, 1985-

Deshmukh, K. and Hemriek, S.: Metabolie ehanges in rabbit artieular cartilage due to inflammation. Arthr. Rheum. 19: 199-208, 1976.

Epstein, E. H.: [a(III)h Human skin collagen - Release by pepsin digestion and preponder­anee in fetallife. J. Biol. Chem. 249: 3225-3231, 1974.

Epstein, E. H., Seott, R. D., Miller, E. ]. and Piez, K. A.: Isolation and eharacterization of the peptides derived from soluble human and baboon skin collagen after cyanogen bromide cleavage. J. Biol. Chem. 246: 1718-1724, 1971.

Eyre, D. R. and Muir, H.: The distribution of different molecular species of collagen in fibrous, elastic and hyaline cartilages of pig. Biochem. J. 151: 595-602, 1975.

Fietzek, P. and Piez, K.: Isolation and characterization of the cyanogen bromide peptides from the a2 chain of rat skin collagen. Biochemistry 8: 2129-2133, 1969.

Fishbein, W. N.: Quantitative densitometry of 1-50 f!gm pro tein in acrylamide gel slabs with Coomassie Blue. Anal. Biochem. 46: 388-401, 1972.

Gay, S., Mueller, P. K., Lemmen, c., Remberger, K., Matzen, K. and Kuhn, K.: Immunohis­tological study on collagen in cartilage-bone metamorphosis and degenerative osteoarth­rosis. Klin. Wochenschr. 54: 969-976, 1976.

Gay, S., Velzanto, ]., Raekellio, J. and Penttinen, R.: Collagen types in early phases of wo und healing in children. Acta Chir. Scand. 144: 205-211, 1978.

Hanson, A. N. and Bentley, T. P.: Quantitation of type I to type III collagen ratios in small sampies of human tendon, blood vessels and artherosclerotic plaque. Anal. Biochem. 130: 32-40, 1983.

Kang, A. H.: Studies on the location of intermolecular cross-links in collagen. Isolation of a CNBr peptide containing ö-hydroxylysinonorleucine. Biochemistry 11: 1828-1835, 1972.

Lane, J. M., Suda, M., von der Mark, K. and Timpi, R.: Immunofluorescent localization of structural collagen types in endochondral fracture healing. Trans. Orthoped. Res. Soc. 7: 94-95, 1982.

Laurent, G. ]., Cockerill, P., McAnnity, R. T. and Hastings, T.: A simplified method for quantification of the relative amounts of type I and type III collagen in small tissue sampies. Anal. Biochem. 113: 301-312, 1981.

Rib Perichondrial Graft in Articular Defect 347

Miller, E. J.: Isolation and characterization of the cyanogen bromide peptides from the a1(II) chain of chick cartilage collagen. Biachemistry 10: 3030-3035, 1971.

Nimni, M. E. and Deshmukh, K.: Differences in collagen metabolism between normal and osteoarthritic human articular cartilage. Science 181: 751-752, 1973.

Reiser, Karen M. and Last, Terold, A.: Quantitation of specific collagen types from lungs of small mammals. Anal. Biachem. 104: 87-98, 1980.

Scott, P. G., Telser, A. G. and Veib, A.: Semiquantitative determination of cyanogen bromide peptides of collagen in SDS-polyacrylamide gels. Anal. Biachem. 70: 251-257, 1976.

Weber, L., Meigel, W. N. and Rauterberg, J.: SDS-polyacrylamide gel electrophoretic deter­mination of Type I and Type III collagen in sm all skin sampies. Arch. Dermata!. Res. 258: 251-257, 1977.

Dr. David Amiel, Orthopaedic Biochemistry, University of California at San Diego, M-030, La Jolla, CA 92093, USA.