Osteolysis is a common complication of total joint replacement, and loss of support by the surrounding osseous architecture can lead to aseptic loosening of the prosthesis. The discovery of cytokines and the relationship between cellular signals and bone turnover has focused attention on biological processes that contribute to osteolysis6,12. However, the biological responses to plastics and metals are still poorly understood, and, to our knowledge, an immunological reaction directed against the implanted materials has not been seriously examined as a possible cause of prosthetic loosening.
Mechanical failure leads to the generation of small particles of biomaterials such as polymethylmethacrylate, ultra-high molecular weight polyethylene, and metal. Certain implanted polymers may cause immune reactions in a number of recipients10. It has been suggested that polymers may provide an adjuvant-like activity to native macromolecules, which adhere to hydrophobic surfaces and subsequently become immunogenic8. Furthermore, ultra-high molecular weight polyethylene-bound proteins may represent antigens to preexisting antibodies in a number of patients who have a joint replacement, as autologous antibodies to connective-tissue antigens are common in patients who have osteoarthritis or rheumatoid arthritis4,5,7,17.
The purpose of the current study was to investigate antibody reactions to proteins that adhere to ultra-high molecular weight polyethylene prosthetic components and to determine whether this polymer may provide a potential surface for the formation of immune complexes.
Materials and Methods
Patients who were scheduled to have a revision arthroplasty because of aseptic loosening, with the diagnosis based on the presence of osteolysis on roentgenograms, were selected for this study. All procedures for the study were approved by the Human Investigation Committee, and informed consent was obtained from all patients. Components of the failed prostheses and plain blood samples were obtained at the time of the revision operation.
Extraction of Proteins
Each failed ultra-high molecular weight polyethylene component was cleared of attached tissue by vigorous scrubbing with a nylon-bristle brush, and cement was removed mechanically. The component then was washed with use of end-over-end rotation in thirty milliliters of phosphate-buffered saline solution containing protease inhibitors, at 4 degrees Celsius, for thirty-four weeks with daily changes of the buffer. The entire component then was immersed in fifteen milliliters of 0.1-molar glycine-hydrogen chloride buffer (pH 2.8), vortexed for three minutes, and subjected to gentle agitation for an additional minute. The component was removed from the glycine-hydrogen chloride solution, which was adjusted to pH 7.5 by the addition of three-molar Tris buffer. It then was immersed in fifteen milliliters of four-molar guanidine hydrochloride solution, vortexed for three minutes, and subjected to gentle agitation for an additional minute. After removal of the component, this solution was adjusted to pH 7.5 by the addition of three-molar Tris buffer. After rapid dialysis to phosphate-buffered saline solution (pH 7.2), both solutions were concentrated with use of membrane ultrafiltration, and the protein content was determined with use of spectroscopic analysis at 280 nanometers and was adjusted to two milligrams per milliliter.
Analysis of Proteins with the Dot Blot Technique
The presence of connective-tissue antigens and immunoglobulin in the protein extracts was examined with use of the dot blot technique. Twenty microliters of saline solution containing five micrograms of protein extract was pipetted onto nitrocellulose paper strips and allowed to dry. The strips were blocked with 5 percent milk-phosphate-buffered saline solution at room temperature for 2.5 hours and were placed in five milliliters of saline solution containing a 1:500 dilution of polyclonal mouse antiserum to human type-I collagen, type-II collagen, or aggrecan proteoglycans. These antibodies to connective-tissue antigens were generated in DBA/1 mice and were evaluated for specificity with use of the Western blot technique as described previously21,22. The antibodies were demonstrated to be antigen specific and reactive with epitopes shared among several species (human, bovine, and porcine).
Control strips were placed in dilutions of normal mouse serum or in saline solution alone. All strips were incubated at 37 degrees Celsius for one hour; they then were washed three times with 0.5 percent Tween-80-phosphate-buffered saline solution for five minutes at five milliliters per track and once with phosphate-buffered saline solution for five minutes at five milliliters per track.
The strips were probed by placing goat anti-mouse antibody solution (39.0 milliliters of phosphate-buffered saline solution, 1.0 gram of milk, 1.0 milliliter of normal goat serum, and 0.08 milliliter of goat anti-mouse IgG conjugated to alkaline phosphatase [Fisher Scientific, Orangeburg, New York]) at room temperature for two hours. The control strips that had been placed in saline solution were reacted with adsorbed/eluted goat anti-human immunoglobulin (all isotypes) conjugated to alkaline phosphatase (Fisher Scientific) to assay for human immunoglobulins in the extracts. The strips were washed three times with 0.5 percent Tween-80-phosphate-buffered saline solution for five minutes at five milliliters per track and once with phosphate-buffered saline solution for five minutes at five milliliters per track. They then were placed in the aqueous substrate solution (0.2 milligram of nitro-blue tetrazolium per milliliter; 100-millimolar Tris-hydrogen chloride, pH 9.7; 0.5-millimolar MgCl2; 0.001-millimolar ZnCl2; 0.2 milligram of 5-bromo-4-chloro-3-indolyl phosphate [U.S. Biochemicals, Cleveland, Ohio] per milliliter, and 1.2 percent volume per volume dimethyl sulfoxide) for 2.5 minutes to develop. The development was stopped with water for twenty minutes at ten milliliters per track, and the strip was air-dried on parafilm.
The dot blots were scanned and analyzed with use of a Hewlett-Packard scanner (Mountain View, California) and with the Sigma-Scan image-analysis software package (Jandel Scientific, San Rafael, California).
Proteins recovered from the polyethylene components were analyzed with use of sodium dodecyl sulfate polyacrylamide gel electrophoresis9. The gel, containing the separated proteins or protein molecular weight standards, was fixed overnight in 250 milliliters of deionized water, fifty milliliters of 0.1-molar acetic acid, and 200 milliliters of methanol and then was silver-stained11. The number and molecular weight of the protein bands were determined with use of a gel image-analysis scanning apparatus (Alpha Innotech, San Leandro, California).
Western Blot Technique
After completion of the electrophoretic separation, the electrophoresis gel was electrotransferred onto nitrocellulose paper with use of methods described previously16. The strips were blocked with 5 percent milk-phosphate-buffered saline solution at 4 degrees Celsius for 2.5 hours and were placed overnight in autologous sera diluted to a concentration of 1:100 in 5 percent milk-phosphate-buffered saline solution or in serum-free 5 percent milk-phosphate-buffered saline solution at 4 degrees Celsius. The strips were washed, probed with secondary antibody, and developed as described previously16. They then were air-dried on parafilm and photographed. The number and size of the reactive bands that were present on strips that had been exposed to autologous serum but were absent on those that had been exposed to secondary antibody alone were determined with use of image analysis.
Recovery of Proteins from Failed Polyethylene Components
Polyethylene components from forty-nine failed total joint prostheses were recovered during revision operations. Thirty acetabular components from total hip replacements and nineteen tibial components from total knee replacements were obtained (Table I). The prostheses were recovered after they had been in place for a minimum of nine months, with the exception of two acetabular components (Implants 1 and 2, Table I) that had been removed during the primary operative procedure because of technical reasons. No proteins were recovered from these devices, and thus they were used as controls for the extraction procedure. A mean (and standard error) of 415 ± 59 micrograms (range, 4.70 to 2361 micrograms) of protein was recovered from the entire surface of the components. Forty-two extracts, which contained more than thirty-five micrograms of protein, were analyzed further with use of immunoblotting and electrophoresis.
Analysis of Proteins with Use of Immunoblotting
Forty-two protein extracts were analyzed with use of the immunoblot technique, which involved use of mouse antibodies specific for type-I and type-II collagen, aggrecan proteoglycans, and immunoglobulins (all isotypes). The immunoblots were classified as no anitibody-binding (-); clear, positive antibody-binding within the blot area (+); or saturation antibody-binding within the blot area (++) (Table I).
Type-I collagen was detected in thirty-four of the extracts, and aggrecan proteoglycans were detected in eight. Type-II collagen was absent in all extracts. Immunoglobulin was detected in thirty-three extracts. The immunoglobulin was concentrated in the glycine-hydrogen chloride extraction solution, whereas the connective-tissue antigens were readily detected in both extraction solutions. The level of protein that was recovered from the components corresponded with the ability to detect type-I collagen; collagen was found in all explant solutions containing more than eighty-five micrograms of protein. In most instances, this relationship was also applicable to the ability to detect immunoglobulins, but this class of proteins was absent from two samples containing more than 100 micrograms of protein.
Analysis of Protein with Use of Electrophoresis
Forty-two guanidine hydrochloride extracts were analyzed for protein content with use of polyacrylamide gel electrophoresis and gel-staining. Positive bands were detected in thirty-four of the extracts, which reflects the limits of sensitivity of this analytical technique. In the thirty-four extract samples with a positive result on polyacrylamide gel electrophoresis, a mean (and standard deviation) of 5.6 ± 3.06 bands were detected, with a minimum of one protein band and a maximum of twelve protein bands. Twenty-one different proteins were identified in the extracts, and they ranged in molecular weight from thirteen to 231 kilodaltons (Fig. 1). The most frequently occuring proteins had molecular weights of 103 and sixty-seven kilodaltons (twenty-two and twenty-one samples, respectively).
Antibody Responses to Polyethylene-Bound Proteins
The Western blot technique was used to assess the presence of autologous antibodies against polyethylene-bound proteins. Forty-two extracts were electrotransferred to nitrocellulose, and the resulting strips were reacted with autologous sera and were probed with goat anti-human immunoglobulin. Because many of the extracts were positive for immunoglobulin on dot-blotting, control strips reacted with goat anti-human immunoglobulin (secondary antibody) alone also were included in the assay. Bands that were identified on both the control strips and the autologous antibody strips were disregarded in this assessment.
In one illustrative case (Fig. 3), the protein bands that were detected with use of silver-staining (lane 1) resulted in three immunoreactive bands after electrotransfer and Western blotting with use of autologous serum (lane 2). Serum from this patient also generated a band at 103 kilodaltons with Western blotting against purified bovine type-I collagen (lane 3), which was coincidental with the band resulting from specific mouse anti-type-I collagen antibodies reacted against purified bovine type-I collagen (lane 4). This demonstrates the presence of an autologous antibody to denatured type-I collagen that reacts with the collagen bound to the ultra-high molecular weight polyethylene component.
Serum antibodies that were reactive with specific autologous polyethylene-bound proteins were detected in twenty-six of the forty-two patients. Fifteen different proteins generated specific positive reactions on the strips; the molecular weight of these proteins ranged from thirteen to 231 kilodaltons (Fig. 2). The most frequently immunoreactive proteins had molecular weights of 103 and 135 kilodaltons (twelve and ten samples, respectively).
Antibodies Specific for Type-I Collagen
Sera also were evaluated for reactivity with type-I collagen with use of the Western blot technique. All patients who were found to have an immunoreactive extract protein at 103 kilodaltons on assay with use of autologous sera also were found to have sera antibodies that were reactive against bovine type-I collagen on Western blots. Analysis of the sera from these patients revealed a single or duplex band that appeared to correspond with the 103-kilodalton molecular weight marker. This band position was identical to that of the strips that were developed with specific mouse anti-type-I collagen antiserum and were probed with goat anti-mouse antibodies. The control reactions of sera reacted against strips without collagen and the type-I collagen reacted against secondary antibody (goat anti-human IgG conjugated to alkaline phosphatase) alone were negative in this assay.
Proteins absorbed to biomaterials have been implicated in the inflammatory response to polymers, possibly involving the activation of macrophages19. Our data demonstrate that a variety of proteins are strongly bound to the polyethylene components of prostheses; more than twenty different protein bands were detected in extracts analyzed with polyacrylamide gel electrophoresis. Type-I collagen was readily detected in most extracts with use of the immunoblot technique, and a band at 103 kilodaltons (presumably type-I collagen) was observed in many of the extracts. Our findings also suggest that most patients express antibodies that are reactive with the proteins bound to polyethylene and that type-I collagen is a major antigenic target in these patients. Several other groups of investigators have found antibodies to type-I collagen (particularly denatured collagen) in patients who have osteoarthritis4,5,17, although the prevalence and titer of these antibodies have been invariably lower than those of the antibodies to type-II collagen in patients who have rheumatoid arthritis5,20 and the immune response to connective-tissue antigens has not been implicated in the pathogenesis of degenerative joint disease. Jasin7 reported that cartilage samples from patients who had osteoarthritis contained sequestered immune complexes with antibodies to type-I collagen, a finding suggesting that a response to connective-tissue antigens is not uncommon in patients who have a joint replacement. Antibody-binding to denatured collagen was common in a study of a large orthopaedic population1, although operative procedures do not appear to increase the titers of anti-collagen antibodies to any major degree3.
The titers of antibodies to type-I collagen increase with age13, and positive titers have been observed in two (4 percent) of fifty normal blood donors14. However, the prevalence of antibodies to collagen in patients who have a revision operation is high, and it is possible that the implantation of a biomaterial, followed by the deposition of collagen, may contribute to increased levels of antibodies. This hypothesis has been proposed to account for the development of antibodies to an unusual collagen epitope in recipients of silicone implants15.
We hypothesize that immunoglobulin complexed with polyethylene may fix complement and that the complement cascade may in turn attract inflammatory cells to the polyethylene surface. Immunoglobulin bound to biomaterials has previously been implicated in inflammatory responses, as IgG-coated polymer implants were shown to activate human neutrophils in vitro and to attract murine phagocytes when they were implanted in vivo18. Since small ultra-high molecular weight polyethylene particles inevitably result from motion and wear of an orthopaedic prosthesis, the presence of polyethylene-bound proteins may be important when the particulate debris is engulfed by phagocytes in the tissue adjoining the site of the prosthesis. The phagocytosis of these particles may result in activated cells that secrete both proinflammatory cytokines and proteolytic enzymes and also provide activation signals to lymphocytes. It was demonstrated previously that mononuclear cells respond in vitro to polymethylmethacrylate in culture both by the secretion of the cytokines interleukin-1β and interleukin-2 and by cellular proliferation23. Therefore, particles of polyethylene could provide an adjuvant effect for the development of a response to polyethylene-bound proteins. Chronic, immunologically mediated inflammation may result from the reaction to prosthetic debris and associated immune complexes, thus damaging the tissue in the region of the prosthesis. It also has been suggested that the localized tissue response to implanted material may have an adverse effect on the chemical integrity of the polymer2. Our data thus support the hypothesis that an immunological response to polyethylene-bound proteins may contribute to aseptic loosening.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was the Veterans Administration.
Investigation performed at the Department of Orthopaedic Surgery, Wayne State University School of Medicine, Hutzel Hospital, and the Veterans Administration Medical Center, Detroit
1. Adelmann, B., and Schoning, B.: Binding of native and denatured collagen to immunoglobulins and cold insoluble globulin in serum of patients undergoing orthopedic surgery. Klin. Wochenschr.
, 58: 625-629, 1980.
2. Ali, S. A.; Doherty, P. J.; and Williams, D. F.: Molecular biointeractions of biomedical polymers with extracellular exudate and inflammatory cells and their effects on the biocompatibility, in vivo. Biomaterials
, 15: 779-785, 1994.
3. Bujia, J.; Alsalameh, S.; Naumann, A.; Wilmes, E.; Sittinger, M.; and Burmester, G. R.: Humoral immune response against minor collagens type IX and XI in patients with cartilage graft resorption after reconstructive surgery. Ann. Rheumat. Dis.
, 53: 229-234, 1994.
4. Charriere, G.; Hartmann, D. J.; Vignon, E.; Ronziere, M. C.; Herbage, D.; and Ville, G.: Antibodies to types I, II, IX, and XI collagen in the serum of patients with rheumatic diseases. Arthrit. and Rheumat.
, 31: 325-332, 1988.
5. Choi, E. K.; Gatenby, P. A.; McGill, N. W.; Bateman, J. F.; Cole, W. G.; and York, J. R.: Autoantibodies to type II collagen: occurrence in rheumatoid arthritis, other arthritides, autoimmune connective tissue diseases, and chronic inflammatory syndromes. Ann. Rheumat. Dis.
, 47: 313-322, 1988.
6. Horowitz, S. M., and Purdon, M. A.: Mechanisms of cellular recruitment in aseptic loosening of prosthetic joint implants. Calcif. Tissue Internat.
, 57: 301-305, 1995.
7. Jasin, H. E.: Autoantibody specificities of immune complexes sequestered in articular cartilage of patients with rheumatoid arthritis and osteoarthritis. Arthrit. and Rheumat.
, 28: 241-248, 1985.
8. Kossovsky, N.; Heggers, J. P.; and Robson, M. C.: Bioreactivity of silicone. CRC Crit. Rev. Biocompat.
, 3: 53-85, 1987.
9. Laemmli, U. K.: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature
, 227: 680-685, 1970.
10. Lalor, P. A.; Revell, P. A.; Gray, A. B.; Wright, S.; Railton, G. T.; and Freeman, M. A. R.: Sensitivity to titanium. A cause of implant failure?. J. Bone and Joint Surg.
, 73-B(1): 25-28, 1991.
11. Merril, C. R.; Goldman, D.; and Van Keuren, M. L.: Gel protein stains: silver stain. Meth. Enzymol.
, 104: 441-447, 1990.
12. Mundy, G. R.: Cytokines and growth factors in the regulation of bone remodeling. J. Bone and Min. Res.
, 8 (Supplement 2): 505-S510, 1993.
13. Nicoloff, G.; Valtsova, D.; and Baydanoff, S.: Age-related changes in the levels of IgG, IgM and IgA antibodies to type I collagen in the sera of normal human subjects. Cor et Vasa
, 34: 365-372, 1992.
14. Rosenberg, A. M.; Hunt, D.W.; and Petty, R. E.: Antibodies to native type I collagen in childhood rheumatic diseases. J. Rheumatol.
, 11: 421-424, 1984.
15. Rowley, M. J.; Cook, A. D.; Teuber, S. S.; and Gershwin, M. E.: Antibodies to collagen: comparative epitope mapping in women with silicone breast implants, systemic lupus erythematosus and rheumatoid arthritis. J. Autoimmun.
, 7: 775-789, 1994.
16. Schaefer, C. J.; Whalen, J. D.; Knapp, T.; and Wooley, P. H.: The influence of silicone implantation on type II collagen-induced arthritis in mice. Arthrit. and Rheumat.
, 40: 1064-1072, 1997.
17. Stuart, J. M.; Huffstutter, E. H.; Townes, A. S.; and Kang, A. H.: Incidence and specificity of antibodies to types I, II, III, IV, and V collagen in rheumatoid arthritis and other rheumatic diseases as measured by 125I-radioimmunoassay. Arthrit. and Rheumat.
, 26: 832-840, 1983.
18. Tang, L.; Lucas, A. H.; and Eaton, J. W.: Inflammatory responses to implanted polymeric biomaterials: role of surface-adsorbed immunoglobulin G. J. Lab. and Clin. Med.
, 122: 292-300, 1993.
19. Tang, L.; Ugarova, T. P.; Plow, E. F.; and Eaton, J. W.: Molecular determinants of acute inflammatory responses to biomaterials. J. Clin. Invest.
, 97: 1329-1334, 1996.
20. Wooley, P. H.; Luthra, H. S.; O'Duffy, J. D.; Bunch, T. W.; Moore, S. B.; and Stuart, J. M.: Anti-type II collagen antibodies in rheumatoid arthritis. The influence of HLA phenotype. Tissue Antigens
, 23: 263-269, 1984.
21. Wooley, P. H.: Collagen-induced arthritis in the mouse. Meth. Enzymol.
, 162: 361-373, 1988.
22. Wooley, P. H.; Siegner, S. W.; Whalen, J. D.; Karvonen, R. L.; and Fernandez-Madrid, F.: Dependence of proteoglycan induced arthritis in BALB/c mice on the development of autoantibodies to high density proteoglycans. Ann. Rheumat. Dis.
, 51: 983-991, 1992.
23. Wooley, P. H.; Nasser, S.; and Fitzgerald, R. H., Jr.: The immune response to implant materials in humans. Clin. Orthop.
, 326: 63-70, 1996.