Mechanical properties of collagen fibres: a comparison of reconstituted and rat tail tendon fibres

Y. Pedro Kate”, David L. Christiansen*, Rita A. Hahn*, Sheu-Jane Shieh*, Jack D. Goldstein + and Frederick H. Silver* Biomatarials Center and Departments of Pathology* and Surgery ?, University of Medicine and Dentistry of New Jersey, Piscataway, New Jersey 08854, USA (Received 26 January 1988; accepted 26 April 1988)

This study involves comparison of the mechanical properties of reconstituted collagen fibres with those of collagen fibres obtained from rat tail tendons. Reconstituted collagen fibres were cross-linked in the presence of glutaraldehyde vapour for 2 and 4 d or using a combination of severe dehydration and carbodiimide treatment. Ultimate tensile strengths for reconstituted fibres cross-linked with glutaral- dehyde ranged from 50 to 66 MPa while those cross-linked by severe dehydration and carbodiimide treatment had ultimate tensile strengths between 24 and 31 MPa. Rat tail tendon fibres had tensile strengths that ranged from 33 to 39 MPa. These results indicate that high-strength collagen fibres can be reconstituted in vitro and that these fibres may be useful in repair of dermal, dental, cardiovascular and orthopaedic defects.

Keywords: Collagen, tendon, ligament, biomaterial

Collagen fibres are the structural elements that give shape to mammalian tissues as well as acting as the scaffold over which cells migrate and deposit new connective tissue. Types I, II and III collagen fibres in the form of cross-linked networks prevent over-expansion of the aorta and heart, limit shear deformation of cartilage and biaxial tensile stretching of skin and transmit tensile loads in tendons’,‘.

Several approaches have been tested to replace the function of tendons and ligaments damaged as a result of athletic injury. Replacement of tendon and ligament has been achieved using biodegradable and non-biodegradable synthetic polymers, and biological tissues from the knee3-“.

Non-porous polymeric implants have a limited fatigue life and ultimately fail when used to replace the function of a ligament such as the anterior cruciate”. Implanted tows of polymeric and carbon fibres result in the controlled formation of a ‘neo-tendon’ through the deposition of aligned collagen”. Biological structures such as glutaraldehyde- fixed bovine tendon and autogenous tendons and ligaments elongate after surgical implantation and eventually do not support loads in the knee”,“.

This paper presents initial experimental results on the development of a collagen fibre of small diameter and high strength, suitable for construction of orthopaedic implants. A comparison is made between the structure and mechanical properties of reconstituted collagen fibre and those of fibre

Correspondence to Dr F.H. Silver.

teased from rat tail tendons (RlT). Results of preliminary studies suggest that reconstituted collagen fibres with diameters of 25 pm have wet strengths of 30 to 60 MPa. In comparison, rat tail tendon fibres obtained by microdissection are 50 to 100 j.rm in diameter and have wet tensile strengths of about 40 MPa.

MATERIALS AND METHODS

Reconstituted collagen fibres

Insoluble collagen Type I from fresh, uncured corium was obtained from Devro Inc. (Somerville, NJ, USA). The corium was limed, fragmented, swollen in acid, precipitated, washed with distilled water and isopropanol, lyophilized and stored at -3O”C’*, 13. Collagen-derived peptides were characterized by SDS-PAGE and amino acid analysis as typical of Type I collagen without non-collagenous protein contamination14.

A 1% (w/v) dispersion of Type I collagen in dilute HCI pH 2.0 was prepared by adding 1.2 g of lyophilized collagen to 120 ml of HCI solution in a blender (Osterizer) and mixing at a speed of 10 000 rev min-’ for 4 min. The mixture was allowed to settle for IO min and then remixed at 10 000 rev min-’ for 4 min. The resulting dispersion was placed under a vacuum of 0.0 1 m torr at r.t. to remove any trapped air bubbles. Dispersion is then stored in disposable 30 cc syringes at 4°C.

0 1989 Butterworth Et Co (Publishers) Ltd. 014%9612/89/010038-05$03.00

38 Biomaterials 1989, Vol 10 January

Mechanical properties of collagen fibres: Y.P. Kato et al.

Collagen fibres were produced by extruding the collagen dispersion through polyethylene tubing with an inner diameter of 0.28 mm into a 37°C bath of aqueous fibre formation buffer composed of 135 mM NaCI, 30 mM TES (N-tris(hydroxylmethyl)methyl-2-aminoethane sulphonic acid) and 30 ITIM sodium phosphate dibasic. The final bath pH was adjusted to 7.5 by adding 5.0 N NaOH drop-wise. Fibres were allowed to remain in the buffer for 45 min. and then placed in 500 ml of isopropyl alcohol for at least 4 h. The fibres were immersed in distilled water for 15-20 min and air dried under tension.

Collagen fibres were cross-linked using glutaraldehyde or by a combination of severe dehydration and treatment with cyanamide as described previously’4. Glutaraldehyde cross-linking was accomplished by placing air-dried collagen fibres in a sealed dessicator containing 10 ml of a 25% (w/v) aqueous glutaraldehyde solution in a petri dish. The fibres were placed on a shelf in the dessicator and were cross- linked in glutaraldehyde vapour for 2-4 d at room tempera- ture. Collagen fibres were also cross-linked by placing in an oven at 110°C and at vacuum of 50-100 m torr for 3 d. Subsequent to dehydrothermal cross-linking (DHT)3, collagen fibres were placed on a shelf in a sealed dessicator containing a petri dish with 20 g of cyanamide in 5 ml of distilled water. Collagen fibres were cross-linked for one day in contact with cyanamide vapour (Cl ).

Collagen fibres 3-4 cm in length were isolated from the tails of Sprague Dawley rats weighing 200-300 g by dissection under a microscope. The skin was stripped from the tail and tendons were removed using a wire stripper placed on the free end of the tail (end opposite to dissected stump). Tendons were clamped using wire strippers placed on the free end of the tail and pulled free from the rest of the tail. Each tendon bundle was split in half along its axis repeatedly until the fibre diameter was about 50 pm. RlT fibres were air dried overnight by hanging the free ends over glass coverslips.

Mechanical testing

Stress-strain curves for reconstituted collagen fibres and RlT fibres were determined in tension using an lnstron Tester Model 1122. Fibres were tested dry and in phosphate-buffered saline (PBS) pH 7.5 using a gauge length of 2 cm and strain rates of IO,50 and 100% per min. Load extension curves obtained from the chart recorder as well as original cross-sectional areas determined by light microscopy were used to calculate stress-strain behaviour, ultimate tensile strength (UTS) and tensile modulus. For area calculations it was assumed that the fibres were circular. It was also assumed that the modulus was equivalent to the tangent to the stress-strain curve in the linear region.

Air-dried fibres were mounted on a paper frame by gluing the ends of the fibres to a vertical line drawn on the frame using an epoxy adhesive. The gauge length was determined by the vertical dimension of the window in the frame. Fibre diameters were determined dry and in PBS at three different locations within the window by comparison with the grid lines in a calibrated eyepiece of a microscope. Diameters of wet fibres were measured after immersion in PBS for at least 15 min.

Frames containing specimen fibres were then placed between the upper and lower pneumatic grips of the lnstron at a pressure of 40 psi. The sides of the frame were then cut leaving the fibres intact. Fibres were sprayed with PBS

before testing. Between 8 and 13 fibres were tested for each strain rate and treatment studied.

Moduli were determined bytaking a best-fit tangent to the load-extension curve. A two-tailed Students’ t-test (P = 0.0 1) was used for statistical evaluation of mechanical data.

RESULTS

Typical stress-strain curves are shown (Figure I) for reconstituted collagen fibres cross-linked dehydrothermally for 3 d and with cyanamide for 1 d (DHT3 + Cl), fibres cross-linked with glutaraldehyde for 2 and 4 d (Glut 2 and Glut 4) as well as for RlTfibres. An almost linear relationship between stress and strain is observed for wet and dry RlT and wet reconstituted collagen fibres, while the stress- strain behaviour of dry reconstituted collagen fibres is highly non-linear.

Tab/es 7-4 summarize the mechanical properties in tension of wet and dry RlTfibres and reconstituted collagen fibres. The ultimate mechanical properties were found to be independent of strain rate for strain rates between 10 and 100% per min.

The UTS of wet glutaraldehyde cross-linked recon- stituted fibres ranged from 50 to 66 MPa. Fibres cross- linked by a combination of dehydrothermal and cyanamide treatment (DHT3 + Cl) had wet tensile strengths of 24-31 MPa. UTS values for RlT fibres ranged from 33 to 39 MPa. Ultimate strains for reconstituted collagen fibres ranged from 14 to 18% while those for RlT were only 7 to 8%. Moduli for reconstituted collagen fibres ranged from 170-200 MPa (DHT3 + Cl) to 384-503 MPa (Glut 2 and Glut 4) and were below those observed for RlT (478-570 MPa). However, moduli values for RTT were not significantly different from those observed for glutaral- dehyde cross-linked collagen fibres.

DISCUSSION

Collagen fibres form structural networks throughout the body that limit tissue deformation and prevent mechanical failure’. The structural stability of collagen fibres is a consequence of the triple-helical structure of the molecule, high content of amino acid residues, organized packing of individual molecules into fibrils and the presence of cross- links within collagen fibrils that prevent intermolecular slippage’. Results of in vitro studies on collagen self- assembly indicate that fibrils formed by heating collagen solutions to 37°C at neutral pH under optimum conditions result in reconstituted collagen fibrils with a structure identical to that of collagen fibrils observed in tendons’ 5-21.

This study is directed at evaluating the mechanical properties of collagen fibres prepared under conditions that have been shown to be optimum for formation of collagen fibres that are characterized by the positive staining pattern of collagen in tissues. In addition, an attempt is made to determine if the strength of reconstituted collagen fibres is similar to that of fibres derived from tendon.

RlTfibres are used in this study, since it is possible to dissect fibres as small as 50 j.rrn in diameter from RTT without tearing the structure apart. It is well known that RTT is not a stress-bearing tendon and therefore the mechanical properties are somewhat different from those of other tendons such as the patellar. However, this difference is minimized during constant-strain-rate experiments’.

Biomaterials 1989, Vol 10 January 39

Mechanical properties of collagen fibres: Y. P. Kato et al.

200 -

175 -

150 -

3 IL

E

:: !?? 5

WET

0 5 10 15 20 25

Strain (%)

150

125

25

5 10 15

Strain (%)

80

60

-z 250

$ - 200

B L 150 tj

5 10 15

Strain (%)

5 10 15 20

Strain (%)

Figure 1 Typical tensile stress-strain cwves obtained at a strain rate of 5O%/min for collagen fibres cross-linked: (a) for 3 d using the dehydrothermal technique and for 1 d in cyanamide vapour (DHT3 + C 1); (b) for 2 d in glutaraldehyde vapour (Glut 2); and (c) for 4 d in glutaraldehyde vapour (Glut 4). For comparison, the cwve for rat tail tendon fibres is shown (d).

Results presented here suggest that reconstituted nation with cyanamide treatment. The higher tensile strength collagen fibres of small diameter have ultimate tensile of RlT in the dry state and the lower elongation at failure strengths that are approximately equal to those of RlTfibres compared with that observed for reconstituted collagen (30-40 MPa). This is achieved after extensive cross-linking fibres suggests that the degree of crystallinity is higher for with either glutaraldehyde or severe dehydration in combi- RlTfibres than for the reconstituted fibres. Crystallinity may

Table 1 Mechanical properties of wet and dry rat tail tendon fibres (2 cm gauge length)

Sample n Strain rate/min (%) UTS -t SD (MPa) US i- SD (%) E k SD (MPa)

m-r Pv 8 10 39.0* 11.10 8.10 * 1.700 570 + 84.9 10 50 34.9 t 14.40 7.10 * 1.200 552 + 172.3 10 100 32.6 + 12.00 6.70 + 3.10 478 + 129.9

RTT (D) 9 10 366 + 79.3 13.80 k 2.90 2690 k 42 1 9 50 366 k 87.5 15.60 + 3.70 2130 k 548 9 100 363 + 99.2 14.50 -t 3.30 2250 + 523

UTS, ultimate tensile stress; US, ultimate stram; E, modulus, tangent to stress-strain curve; RlT, rat tail tendon fibre; W, wet; D, dry; SD, standard deviation; n, number of samples.

Table 2 Mechanical properties of reconstituted collagen fibres cross-linked using DHT3 + C 1 vapour (2 cm gauge length)

Sample n Strain rate/min (%) UTS + SD (MPa) US f SD (%) El + SD (MPa) E2 k SD (MPa)

DHT3 + Cl (W) 15 10 27.4 k 5.60 17.50 + 4.40 179.5 + 54.7 12 50 23.9 f 3.80 14.70 f 2.10 198.1 k 56.7 10 100 31.3 f 4.70 17.70 + 2.20 170.1 + 32.9

DHT3 + Cl (D) 5 10 168.8 + 39.5 19.50 f 5.80 4320 k 1444 574 + 204 5 50 184.2 f 22.4 21 .O k 4.80 3600 + 785 569 _+ 345 5 100 177.6 k 32.1 22.2 k 3.90 3910 k 1205 544 + 271

UTS, ultimate tensile stress; US, ultimate strain; DHT3 + Cl, 3 d severe dehydrothermal treatment followed by 1 d exposure to cyanamidevapour; W, wet; D, dry; SD, standard deviation; n, number of samples; El, low strain modulus; E2, upper strain modulus.

40 Biomaterials 1989, Vol 10 January

Mechanical properties of collagen fibres: Y.P. Kato et al.

Table 3 Mechanical properties of reconstituted collagen fibres cross-linked for 2 d using glutaraldehyde vapour (2 cm gauge length)

Sample

Glut 2 (Wet)

Glut 2 (Dry)

n Strarn rate/min (%) UTS & SD (MPa) US + SD (%) El + SD (MPa)

10 10 66.2 ? 17.20 16.10 t 2.70 407 + 96.6

13 50 59.2 -t 17.90 13.80 * 2.90 503 i 127.7

12 100 50.0 f 17.40 14.90 f 3.50 412 i 83.4

5 10 180.7 it 27.6 15.30 r 2.10 3820 -t 376

5 50 152.0 + 43.6 14.80 t 1.800 4070 & 401

5 100 149.9 !I 45.0 12.80 t 3.20 3240 i 539

E2 t SD (MPa)

788 -f- 231 764 t 157.6

567 i 195.7

UTS, ultimate tensrle stress; US, ultimate strain; Glut 2. glutaraldehyde 2 d vapour; W. wet; D, dry; SD, standard devration; n, number of samples; El, low strain modulus; E2, upper strain modulus.

Table 4 Mechanical properties of reconstituted collagen fibres cross-linked for 4 d using glutaraldehyde vapour (2 cm gauge length)

Sample n Strain rate/min (%) UTS + SD (MPa) US + SD (%) El + SD (MPa) E2 f SD (MPa)

Glut 4 (Wet) 12 10 64.2 + 15.0 15.10 f 2.70 456.5 + 82.7

12 50 55.5 + 1 1.80 14.40 t 3.60 403 + 80.9

13 100 51.8 i 13.00 13.60 + 3.40 384 i- 57.5

Glut 4 (Dry) 5 10 139.4 * 47.0 14.60 f 1.700 3070 t 647 529 t 242

5 50 142.4 i 38.0 1 1.80 + 4.70 3550 -311 578+216

5 100 144.4 t 26.6 12.60 + 2.50 4860 + 1238 803 + 444

UTS. ultimatetensile stress; US, ultimate strain: Glut 4. glutaraldehyde 4 d vapour; SD, standard deviatron; n. number of samples; E 1, low strarn modulus; E2, upper strain modulus.

in part be introduced during drying before the sample is mounted for mechanical testing.

Reconstituted collagen fibres of high strength and small diameter are needed in the design of biomaterials that enhance the deposition of repair tissue in skin, tendon and the cardiovascular system. Although high-strength oriented and unoriented collagenous materials are reported in the literature22, no report exists of collagen fibres of small diameter that can be processed into woven and non-woven textile prostheses.

The effects of two different cross-linking procedures on mechanical properties of reconstituted collagen fibres were studied here. Glutaraldehyde reacts with free amino- containing groups on collagen and polymerizes to produce cross-links that can span the distance between two moleculesz3. The major drawback to glutaraldehyde cross- linking is the moderate inflammatory response that is generated by the residual glutaraldehyde that is released into tissues upon implantationz3, 24. Exposure to glutaraldehyde vapour may limit polymerization and resulting inflammation in vim, since the vapour pressure of the monomer is higher than that of polymers; however, biocompatibility studies are needed to evaluate this cross-linking procedure.

The alternative cross-linking process used in this study involves severe dehydration in combination with cyanamide treatment25S26. In theory, cyanamide acts as a catalyst in the formation of peptide bonds and is converted into urea after the reaction is completed. Previous implan- tation results on Type I collagen cross-linked using this method suggest that this material evokes a weak inflamma- tory response when tested subcutaneously or on excised wounds’4.25.

In summary, this study indicates that it is possible to prepare reconstituted collagen fibres of small diameter that have wet tensile strengths of up to 66 MPa while maintaining elongation at failure of about 15%. These fibres have applications in the repair of dermal, dental, cardiovascular and orthopaedic structures.

REFERENCES

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Srlver. F.H., Biological Materials, Structure, Properties and Modeling of Soft Tissues, NYU Press, New York, 1987, Ch. 1, 6. 7 Dunn, M.G. and Silver, F.H.. Viscoelastrc behavror of human connective tissues: relative contnbution of viscous and elastrc components, Connect. Tissue Res. 1983, 12, 59-70 Jenkins, D.H.R., Forster. I.W., McKibbin, 8. and Ralis, Z.A.. Induction of tendon and ligament formation by carbon implants, J. Bone Joint Surg. 1977, 59-B. 53-57 Hunter, J.M. and Salisbury. R.E.. Flexor-tendon reconstructron rn severely damaged hands,J. BoneJointSurg. 1971,53-A, 829-858 Alexander, H.. Weiss, A.8. and Parsons, J.R., Ligament and tendon

repair with an absorbable polymer-coated fiber stent, Bull. Hosp. Joint Dis. 1986, 46, 155-l 73 Bolton. C.W. and Bruchman, W.C., The GORE-TEX expanded poly- tetrafluoroethylene prosthetic ligament, C/m. Orthop. 1986, 196, 202-2 13

McMaster. W.C.. A hrstologrc xenograft assessment of canine antenor crucrate substitution wrth bovine.Clin. Orthop. 1985,196, 196-201 Shino, K.. Kimura.T.. Hirose, H., Inoue. M. and Ono, K.. Reconstructron of the anterior cruciate ligament by allogenrc tendon graft, J. Bone

Joint Surg. 1986. 68-B. 739-746 Amis,A.A.,Campbell, J.R., Kempson.S.A.and Mrller. J.H.,Companson of the structures of new tendons induced by implantatron, J. Bone Joint Surg. 1984, 66-B. 131-l 39 Goodship, A.E. and Cooke. P., Brocompatibility of tendon and Irgament prostheses, CRC Crit. Rev. Biocompat. 1986, 2, 303-334 Friedman, M.J., Sherman, O.H.. Fox, J.M., Del Pizzo, W., Snyder, S.J. and Ferkel, R.J., Autogenic anterior crucrate lrgament (ACL) anterior reconstruction of the knee, Clin. Orthop. 1985, 196, 9-l 4

Komanowsky. M.. Sinnamon, H.I., Elras. S., Heiland, W.K. and Aceto, N.C.. Production of commuted collagen for novel applrcatrons, Am. Leather Chemists Assoc. 1974, 69.41 O-41 1 Chernomorsky. A.. Effect of purification procedure on the bra- compatibility of collagen-based biomaterials. MS Theses, Rutgers University, 1987

Weadock, K.. Olson. R.M. and Silver, F.H.. Evaluatron of collagen crosslinking techniques, Biomater. Med. Devices Artif. Organs 1984. 11, 293-318 Gross, J.. Highberger. J.H. and Schmm. F.O., Collagen structures considered as states of aggregation of a kinetic unit, Proc. Nat/ Acad. Sci. USA 1954,40,679-688

Bensusan, H.6. and Scan”, A., Fiber formation from solutions of collagen. II The role of tyrosyl residues, J. Am. Chem. Sot. 1960,82. 4990-4998

Riomaterials 1989, Vol 10 January 4 1

Mechanical properties of collagen fibrest Y. P. Kato et al.

17 Fessler. J.H., Some properties of neutral salt soluble collagen, Biochem. J. 1960, 76,452-463

18 Cassel, J.M., Mandelkem, L. and Roberts, DE., The kinetics of the heat precipitation of collagen, J. Am. Leather Chemists Assoc. 1962,57, 556-575

19 Wood, G.C. and Keech, M.K., The formation of fibrils from collagen solutions, Biochem. J. 1960, 76, 588-598

20 Gelman, R.A., Poppke, D.C. and Piez, K.A., Collagen fibril formation in vitro, the role of the nonhelical terminal regions, J. Biol. Chem. 1979,254, 11741-l 1745

21 Silver, F.H., Type I collagen fibrillogenesis in vitro, J. Biol. Chem. 198 1, 256.4973-4977

22 Hughes, K.E., Fink, D.J., Hutson, T.8. and Veis, A., Oriented fibrillar

23

24

25

26

collagen and its application to biomedical devices, J. Am. Leather Chemists Assoc. 1984, 79, 146-l 58

Cheung, D.T. and Nimni, N.E., Mechanism of crosslinking by proteins by glutaraldehyde, Connect Tissue Res. 1982, 10, 201-216 Chvapil. M.. Speer, D., Mora, W. and Eskelson, C., Effect of tanning agent on tissue reaction to tissue implanted collagen sponge, J. Surg. Res. 1983, 35,402-409 Doillon, C.J. and Silver, F.H., Collagen-based wound dressing: effects of hyaluronic acid and fibronectin on wound healing, Biomaterials 1986, 7, 3-8 Silver, F.H.. Berg, R.A., Birk, D.A., Weadock, K. and Whyne, C., Biodegradable matrix and methods for producing same, US Patent 4, 703. 108. 1987

42 Biomaterials 1989, Vol 10 January