Development of fibroblast-seeded ligament analogs for ACL reconstruction
Michael G. Dunn,',* Janice B. Liesch,' Moti L. Tiku,' and Joseph P. Zawadsky' 1 Orthopaedic Research Laboratory, Division of Orthopaedic Surgery, and 'Department of Medicine, University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School, New Brunswick, New Jersey 08903
We fabricated "ligament analogs" in vitro by seeding high- strength resorbable collagen fiber scaffolds with intraartic- ular (anterior cruciate ligament, ACL) or extraarticular (pa- tellar tendon, PT) rabbit fibroblasts. Fibroblasts attached, proliferated, and secreted new collagen on the ligament analogs in vitro. Fibroblast function depended on the tissue culture substrate (ligament analog vs. tissue culture plate) and the origin of the fibroblasts (ACL vs. PT). PT fibroblasts proliferated more rapidly than ACL fibroblasts when cul- tured on ligament analogs. Collagen synthesis by ACL and PT fibroblasts was approximately tenfold greater on liga-
ment analogs than on tissue culture plates. The composi- tion, structure, and geometry of the collagen fiber scaffolds may promote collagen synthesis within ligament analogs in uitro. Ligament analogs roughly approximate the structure and strength of native ligament tissue. Ongoing in vivo studies suggest that autogenous fibroblast-seeded ligament analogs remain viable after implantation into the knee joint. With further development, ligament analogs may be useful as implants for ACL reconstruction surgery. 0 1995 John Wiley & Sons, Inc.
Severe injury to the anterior cruciate ligament (ACL) can cause knee instability, meniscal damage, and osteoarthritis.' Because the ACL heals poorly fol- lowing primary repair, surgical reconstruction is rec- ommended to improve joint function in young active patients.lT2 Patellar tendon (PT) autografts and al- lografts are widely used but are not ideally suited for this purpose.14 Problems associated with PT au- tografts include lengthy rehabilitation and persistent patellar paina5 PT allografts carry the risk of disease transmission,6 and their procurement is labor inten- sive and c o ~ t l y . ~ Both autografts and allografts may become necrotic and weak4 shortly after implanta- tion, and the knee must be protected from high me- chanical loads while the graft gradually gains strength. Permanent synthetic ACL prostheses may perform satisfactorily in the short term, but tend to break down and fail in the long Currently, no prosthesis is approved by the United States Food and Drug Administration for primary ACL reconstruc- tion.
Resorbable scaffolds seeded with cells are potential alternatives to biological grafts or permanent prosthe-
*To whom correspondence should be addressed.
ses. This tissue engineering" strategy has been used by others to repair large defects in skin" and carti- lage.'' Our previous studies suggest that the ACL can be regenerated using a similar approach. We showed that acellular collagen scaffolds can induce neotendon and neoligament formation in rab- bit Achilles tendon'517 and ACL." In our ACL re- construction study, however, nearly half of the col- lagen scaffold implants did not induce ingrowth of functional neoligament tissue. Tissue ingrowth into ACL prostheses and grafts is inconsistent and diffi- cult to control.
We hypothesize that autogenous fibroblast seeding of the collagen scaffolds prior to implantation might im- prove neoligament formation in the reconstructed ACL. Therefore, our objective was to fabricate "Iiga- ment analogs" by seeding high-strength resorbable collagen fiber scaffolds with viable fibroblasts from intraarticular (ACL) or extraarticular (PT) tissues. Fi- broblast attachment, morphology, proliferation, and collagen synthesis varied as a function of the culture substrate and the origin of the fibroblast. Collagen synthesis was tenfold greater on ligament analogs than on tissue culture plates. Ligament analogs roughly approximate the structure and strength of native ligament tissue. With further development,
Journal of Biomedical Materials Research, Vol. 29, 136S1371 (1995) 0 1995 John Wiley & Sons, Inc. CCC 0021-9304/95/111363-09
these ligament analogs may be useful as implants for ACL reconstruction.
MATERIALS AND METHODS
Fibroblasts were harvested from rabbit anterior cru- ciate ligament (ACL) and patellar tendon (PT), grown in culture, and seeded onto tissue culture plates and collagen fiber scaffolds, creating ligament analogs (Fig. 1). Fibroblast attachment, morphology, prolifer- ation, and collagen synthesis were measured in vitro as a function of substrate and fibroblast origin.
Fabrication of fibrous collagen
Freeze-dried acid-insoluble bovine dermal collagen was ground in a Wiley mill and dispersed at a con- centration of 1% (wt/vol) in pH 3.0 HCl in a blender for 3 min at 10,000 rpm. The dispersion was degassed under vacuum for 4 h and stored in a 30 cc plastic syringe at 4C. Collagen fibers (60 pm average dry diameter) were produced by extrusion of the collagen dispersion into fiber formation buffer, using a syringe pump and polyethylene tubing with an inner diam- eter of 580 p.m. The buffer was composed of 135 mM NaC1, 30 mM TES [N-tris(hydroxymethyI)methyl-2-
INS OL U B L E CO LL AG E N FIBERS ARE EXTRUDED AND
LSSEMBLED INTO SCAFFOLDS.
FIBROBLASTS FROM RABBIT ~ ACL OR PT ARE EXPLANTED
AND GROWN IN CULTURE. 1 1 I
FIBROBLAST-SEEDED LIGAMENT ANALOG
Figure 1. Schematic representation of ligament analog fabrication. Collagen fiber scaffolds were seeded with rab- bit ACL or PT fibroblasts. Fibroblasts attached, proliferated, and secreted collagen on the ligament analogs in vitro.
DUNN El AL.
aminoethane sulfonic acid] and 30 mM sodium phos- phate dibasic, at pH 7.4 and 37C. In this buffer, the pH, temperature, and salt concentration approach physiological values, and the acid-swollen collagen fibrils and fiber fragments within the dispersion de-swell and aggregate within the extruded collagen fibers .
After 45 min in the buffer, the extruded fibers were transferred to an isopropanol bath for 4 h, then washed for 20 min in distilled water. Fibers were dried overnight under tension (their own weight) to improve collagen orientation along the longitudinal fiber axis. Fibers were crosslinked using dehydrother- ma1 treatment* to avoid cytotoxic byproducts associ- ated with chemical crosslinking agents. Fibers were placed in an oven at 110C under a vacuum of less than 1 millitorr for three days.
Collagen fiber scaffolds (Fig. 2) were prepared by aligning 200 crosslinked collagen fibers (length = 15 cm) in parallel, coating the fibers with a 1% (wthol) collagen dispersion, rinsing extensively in distilled water, and drying overnight.
Establishment of rabbit fibroblast cultures
Tissue samples for primary explants were removed from mature male New Zealand white rabbits using general anesthesia and sterile surgical procedures. NIH guidelines for the care and use of laboratory an- imals were observed,20 and all procedures were IACUC approved. Samples of rabbit ACL and patel- lar tendon were obtained, cut into 1-2 mm pieces, and placed in polystyrene tissue-culture plates. The culture media (referred to as media I) used in these procedures and for cell propagation was Dulbeccos
Figure 2. The collagen scaffolds consisted of 200 ex- truded, dehydrothermally crosslinked collagen fibers aligned in parallel. Dry fiber diameters were approximately 50-70 km (Bar = 100 km).
LIGAMENT ANALOGS FOR ACL RECONSTRUCTION 1365
modified Eagle's medium, 10% fetal calf serum, 2 mM glutamine, 100 @mL penicillin, 100 pg/mL strepto- mycin, and 0.25 pg/mL amphotericin B. Media I (1 mL) was carefully added to the plates incubated at 37C in a humidified atmosphere of 5% C02. The media volume was gradually increased over the ini- tial 2-5 days to 5 mL. After approximately 7 days, the explant pieces were discarded and the outgrown cells were removed with trypsin-EDTA for subculture. Cells were maintained in culture using standard pro- cedures and used at passage 4.
Ligament analogs: Collagen scaffolds seeded with fibroblasts
Fibroblasts were released from culture plates by trypsin-EDTA treatment (terminated by the addition of 1 mg/mL soybean trypsin inhibitor) followed by washing and resuspension in media I and 10 mM HEPES pH 7.4. Cells were resuspended to a final con- centration of lo6 cells/mL. Cell aliquots of 0.1 mL were added to 24-well sterile tissue-culture plates containing 0.1 mL of media I and 10 mM HEPES pH 7.4, with a 1 cm length of sterile collagen scaffold on the bottom of the plate. After 24 h, the seeded colla- gen scaffolds (referred to hereafter as ligament ana- logs) were removed and placed in 96-well tissue- culture treated polystyrene plates with 0.2 mL fresh media I and 10 mM HEPES pH 7.4.
Determination of initial attachment of fibroblasts
Initial fibroblast attachment to tissue culture plates and ligament analogs was measured using 51Cr- labeling. Labeled fibroblasts (lo5 per well) were al- lowed to attach to the substrate for l , 2, 4, or 24 h. After those time periods, the ligament analogs and tissue-culture plates were rinsed extensively so only adherent cells remained, and the total radioactivity (51Cr counts per minute) was measured using a liquid scintillation counter.21
Determination of fibroblast morphology
Fibroblast morphology was examined using phase contrast microscopy and ultraviolet light microscopy for fluorescently labeled fibroblasts. Fibroblast mem- branes were labeled using the fluorescent lipophilic dye PKH2-GL (Zynaxis Cell Science, Malvern, PA). Diluent and PKH2-GL dye (0.5 mL each, with final dye concentration of 5 pMolar) were added to each pellet containing lo6 cells. Fibroblasts were incubated for 10 min at room temperature. Dye incorporation
was terminated by the addition of 1 mL of media I. Fibroblasts were separated from excess dye by cen- trifugation through serum, washed 3 times with me- dia, and resuspended to a concentration of 1 x lo6 cells/mL in media I. The PKHZGL dye has an exci- tation wavelength of 488 nm and emits at 525 nm. This fluorescent dye is incorporated and retained in the plasma membrane of viable cells for more than 200 days.*l During mitosis, the fluorescent mem- brane dye is transferred equally to progeny by shar- ing the parent cell membrane. Therefore, the dye is a specific marker for viable cells and their progeny.
Determination of fibroblast proliferation rate and collagen synthesis
Fibroblast proliferation rate (at days 2, 3, 4, and 7) and collagen synthesis (at day 4) were determined by radioactive labeling studies. Ligament analogs were fabricated by initially seeding collagen scaffolds with lo5 fibroblasts. At the designated time periods, the ligament analogs were removed from the wells and placed in new wells with 0.18 mL fresh media I. For proliferation measurements, 3[H] thymidine (1 pCi/ well, 5 pCi/mL, 0.02 mL) was added to each well, and the cells were incubated for 4 h at 37 "C in 5% CO,. For collagen synthesis measurements, 3[H] proline (1 pCi/well) was added to each well, and cells were in- cubated for 24 h.
After the incubation period, ligament analogs were removed from the wells, washed in HBSS, and placed in 0.3 mL 20% cold TCA for 30 min at 4 "C. Samples were centrifuged at 19,000 rpm for 10 min, and the supernatant was discarded. Two more times the pel- let was rewashed with cold 10% TCA, centrifuged, and the supernatant discarded. For proliferation mea- surements, perchloric acid (2N, 0.4 mL) was added, samples were incubated at 60 "C for 30 min and al- lowed to cool. For collagen synthesis measurements, 0.4 mL of 0.3 N NaOH with 1% sodium lauryl sulfate was added, and the samples were incubated at room temperature for 30 min. Ready Safe (4.6 mL) was added, and radioactivity (counts per min) was mea- sured by liquid scintillation counting. Data were nor- malized to radioactive counts per minute per cell.
For the initial attachment studies, a total of 48 mea- surements were made: 4 groups (ACL and PT fibro- blasts, cultured on plates and on ligament analogs) X 12 measurements per group. For the proliferation studies, a total of 48 measurements were made: 4 groups (as above) x 12 measurements per group. For
1366 DUNN ET AL.
collagen synthesis studies, a total of 12 measure- ments were made: 4 groups x 3 measurements per group. Analysis of variance was performed using Statgraphics software to determine the effects of fi- broblast type (ACL vs. PT) and substrate (ligament analogs vs. culture plates) on fibroblast behavior. These effects were considered significant for p values less than 0.05.
We fabricated ligament analogs resembling native ligament tissue by seeding ACL and PT fibroblasts onto collagen fiber scaffolds in vitro. ACL and PT fi- broblasts attached, proliferated, and secreted new matrix on ligament analogs in vitro. Fibroblast func- tion was dependent on both the origin of the fibro- blasts (ACL vs. PT) and the substrate on which the fibroblasts were seeded (ligament analogs vs. tissue culture plates).
Initial attachment of fibroblasts
The number of fibroblasts attached through the first 24 h to ligament analogs and tissue culture plates is shown in Figure 3 (51Cr counts per minute). No significant differences existed as a function of time (comparing cell numbers at 1, 2, 4, and 24 h, data not shown), so data for these 4 time periods were pooled.
Initial attachment of fibroblasts was significantly
INITIAL FIBROBLAST ATTACHMENT
TISSUE CULTURE PLATE LIGAMENT ANALOG
Figure 3. Initial attachment (radiolabeled 51Cr counts) of ACL and PT fibroblasts to ligament analogs and tissue cul- ture plates in vitro (mean values with standard deviations). ACL and PT fibroblast attachment data were not signifi- cantly different. However, significantly fewer (p < 0.05) fibroblasts initially were attached to ligament analogs (*)
influenced by the substrate but not by the origin of the fibroblast. The total number of attached fibro- blasts (not normalized per area) was significantly greater on tissue culture plates than on ligament an- alogs for both ACL and PT fibroblasts.
Fibroblast morphology was dependent on both the tissue culture substrate and the origin of the fibro- blast. ACL fibroblasts were well spread on tissue cul- ture plates and were ovoid in shape [Fig. 4(A)]. No preferred orientation was observed. PT fibroblasts on tissue culture plates [Fig. 4(B)] appeared longer and thinner than ACL fibroblasts, with increased cell-cell connections compared to ACL fibroblasts. Again, no preferred orientation was noted.
On ligament analogs, ACL fibroblasts remained plump and were arranged in columns along the col- lagen fibers in some areas [Fig. 4(C)]. Many PT fibro- blasts, in contrast, became bipolar and aligned their long axi...