Abstract
Our goal in this study was to evaluate the effects of and the interaction between the hypothesized principal determinants of the incorporation of grafts: antigenicity and treatment of the graft. We implanted fresh and frozen cortical bone grafts that were matched for both major and non-major histocompatibility complex antigens (syngeneic grafts), matched for major but not for non-major histocompatibility complex antigens (minor mismatch), and mismatched for both major and non-major histocompatibility complex antigens (major mismatch). We used a rat model with an eight-millimeter segmental defect in the femur. The construct was stabilized with a plastic plate, threaded Kirschner wires, and cerclage wires. We evaluated the grafts at one, two, and four months after implantation.We measured the immune response; assessed the incorporation of the graft with use of histological examination, biomechanical testing, and quantitative isotopic kinetics; and statistically analyzed the effects of and the interactions among three independent variables: time, the degree of matching for major histocompatibility complex antigens, and the treatment of the graft (whether it was fresh or frozen). These three independent variables had profound effects on the pattern, rate, and quality of the incorporation of the graft. Two-way and three-way interactions among these variables were also noted. Serial changes in every dependent variable were observed with time.Systemic antibody specific for donor antigens was measurable only in the serum of animals that had a major mismatch, but freezing markedly attenuated the systemic antibody response. Revascularization was profoundly affected by histocompatibility-antigen matching; the syngeneic grafts were revascularized more quickly and to a greater degree than the grafts with either a minor or a major mismatch. Freezing significantly (p < 0.001) reduced the revascularization of the syngeneic grafts but had no discernible effect on the grafts with a minor mismatch.CLINICAL RELEVANCE: The findings of this investigation are clinically important because they help to explain the unpredictability of incorporation of cortical bone grafts. The graft that is most commonly implanted clinically, the frozen (or processed) mismatched allograft, had the least predictable process of incorporation. However, our findings suggest that the process of incorporation may be manipulated; for example, by the addition or removal of cells and, indirectly, of cytokines.
Despite the widespread use of cortical bone grafts for reconstruction after resection of a tumor, for traumatic bone loss, or in conjunction with a total joint procedure, the physiological and biological events that are crucial to the process of incorporation and the mechanisms that control these events are only superficially understood. In 25 to 35 per cent of patients who are managed with a massive allograft, some complication develops within the first three years after the operation27. Although a humoral immune response to a massive allograft has been demonstrated, the effect of the immune response on the function of an allograft is less clear29. Serious failures, other than those due to a technical cause or infection, have occurred, and minimum revascularization has been demonstrated15. To complicate matters further, the available imaging methods are of limited use for assessing host-graft union, revascularization of the graft, new-bone formation, and the strength of the graft1.
Animal models have demonstrated that at least two factors are important influences on the process of incorporation: the degree of histocompatibility matching between the graft and the recipient and the treatment of the graft (whether it is fresh or frozen). Allografts of bone mismatched for major histocompatibility complex antigens have functioned poorly compared with autogenous grafts2-4,7,14,30. A specific immunological response has been observed after implantation of a fresh mismatched allogenic cortical bone graft9,12,13,37, and the rate and pattern of revascularization of these grafts have been delayed and altered compared with those of autogenous grafts19. Freezing of the graft also interferes with the process of incorporation8,9. However, only limited conclusions can be drawn from these studies because investigators have separately studied the immune response, the biomechanical strength of the graft, or the histological incorporation. Grafts frequently have been implanted in heterotopic sites10, in orthotopic unstable sites22, or in orthotopic non-weight-bearing sites3, which has confounded the analyses of incorporation. Excellent studies have been performed on dogs, rats, and rabbits5,28,38-40, but the incorporation of a bone graft has been difficult to quantify. Often, only one time-point has been studied, and usually the numbers of animals have not been sufficient for the investigators to evaluate interactions among variables.
Our goal in this study was to evaluate the effects of and the interactions among the hypothesized principal determinants of the incorporation of grafts. Our purposes were to study a large number of animals, to standardize the model, to study sequential time-points, and to measure multiple functional parameters of incorporation.
*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 sources were grants from the National Institutes of Health.
†Read in part at the Annual Meeting of the Orthopaedic Research Society, Washington, D.C., February 20, 1992.
‡Departments of Orthopaedics (S. S., L. K., and V. M. G.) and Mechanical Aerospace Engineering (D. T. D.), Case Western Reserve University, 11100 Euclid Avenue, Cleveland, Ohio 44106. Please address requests for reprints to Dr. Stevenson.
§200 Carmen Avenue, Apartment 23-F, East Meadow, New York 11554.
Experimental Design
A fixed-factor model was used in this experiment41. The three independent variables were the time of the observation (one, two, or four months after the operation), the degree of histocompatibility matching between the recipient and the graft (matched for both major and non-major histocompatibility complex antigens, matched for only major histocompatibility complex antigens, and mismatched for both major and non-major histocompatibility complex antigens), and the treatment of the graft (whether it was fresh or frozen). All eighteen possible combinations of the variables were studied. Separate groups of six animals each were studied by histological and isotopic examination at each of the three time-points, and by biomechanical examination at two and four months. Thus, grafts were implanted in 288 rats in this study.
Immunogenetics
Inbred strains of rats were used. The major histocompatibility complex in rats is termed Rt.1, and it is analogous to the human leukocyte antigen. Loci B and D code for class-II histocompatibility antigens, and loci E and A code for class-I major histocompatibility antigens20,26,31. The inbred rat strains DA, DA1U, and L1U were used. DA rats differ from L1U rats with regard to both the major histocompatibility complex antigens and the non-major histocompatibility complex (minor) antigens. DA1U rats are congenic with L1U rats, but they express DA non-major histocompatibility complex antigens (Table I). Two hundred and eighty-eight male L1U rats, each weighing more than 270 grams, were the recipients, and male or female L1U rats, DA1U rats, and DA rats were the graft donors.
Operative Model
After the induction of anesthesia, the rat was aseptically prepared and draped, with the right hindlimb exposed in abduction and partial flexion. Through an anterolateral approach, the vastus lateralis and the biceps femoris were separated, exposing the femur. The soft tissue was freed circumferentially from the diaphysis. A small (twenty-three by five by five-millimeter) polyacetyl plate was fastened to the anterolateral aspect of the femur with two threaded Kirschner wires and a cerclage wire at both the proximal and the distal end. An eight-millimeter osteoperiosteal segment was cut from the mid-portion of the diaphysis, with use of a high-speed drill with a one-millimeter side-cutting burr (Zimmer USA, Warsaw, Indiana), under irrigation with saline solution. The donor segment was placed into the defect and was secured in place with another full cerclage wire. The graft was approximately 25 per cent of the total length of the femur and 50 per cent of the length of the diaphysis. The muscles were apposed, and the fascia and the skin were closed routinely. The animals were fully ambulatory by fewer than twelve hours after the operation.
Treatment of the Graft
Fresh grafts were kept at room temperature in tissue-culture medium (RPMI-1640; Gibco, Grand Island, New York) and were implanted within two hours. The periosteum was routinely elevated when the graft was obtained, and the intramedullary canal of each graft was reamed and was lavaged gently to remove bone marrow. Grafts to be frozen were stored at -20 degrees Celsius for at least two weeks but not for more than six months, and they were thawed in room-temperature physiological saline solution before they were implanted. Freezing to -20 degrees Celsius for this short period was considered to be analogous to freezing at -80 degrees Celsius for longer periods, as might be done clinically.
Immunological Analysis
Serum for immunological assays was collected from each animal before it was killed. Each assay was run in triplicate, and the mean for three experimental wells was calculated. The maximum release of isotope (51chromium) was determined by measuring the radioactivity of the supernatant after detergent lysis (with use of Triton X-100) of target cells. The spontaneous release of isotope was ascertained by measuring the radioactivity of the supernatant after target cells had been incubated with medium alone. Donor-strain lymph-node cells stimulated with concanavalin A and labeled with 51chromium were used as target cells. For all assays, the percentage of chromium (%R) was determined with the formula: %R = ([experimental release - spontaneous release]/[maximum release - spontaneous release]) x 100. Negative values were reported as 0 per cent release of chromium. Serum from each recipient was heat-inactivated at 56 degrees Celsius for thirty minutes. It was then tested, in dilutions ranging from 1:5 to 1:320, for the presence of cytotoxic antibody specific for donor antigens. Negative controls were target cells in medium alone and with autologous (donor) serum and rabbit complement. A release of chromium of more than 10 per cent was considered positive37.
Biomechanical Analysis
Each femur was removed en bloc, with the graft and the fixation device intact, and was stored frozen in saline solution until testing. After thawing, the central five-millimeter segment of the graft was excised and subjected to uniaxial compression testing under displacement control at a rate of eight millimeters per minute. The displacement and load histories were recorded for each test. The compressive strength was considered to be the maximum load achieved before collapse, as evidenced by a decrease in load. Biomechanical testing was performed only at two and four months after the operation. Segmental compression testing was used so that we could evaluate the material properties of the graft itself rather than of the host-graft interface.
Isotopic Analysis
The skeletons of weanling rats were labeled with isotopes by repeated administration of 3H-tetracycline and 45calcium for six weeks. After a one-month equilibration period, eight-millimeter grafts were obtained from both femora of each rat. These control grafts were weighed. Then, one was frozen and stored and the other was handled in the same manner as the other grafts, as already described. After each recipient was killed, the implanted graft was removed and analyzed simultaneously with a paired control graft. The dry weight and the total counts per minute of 45calcium and 3H-tetracycline were measured in the paired grafts. The absolute change in the dry weight, the percentage change in the dry weight, and the percentage loss of 45calcium and of 3H-tetracycline were calculated by comparing the values for the implanted graft with those for the control graft.
Histomorphometric Analysis
The fluorochromes DCAF, fluorescein (twenty milligrams per kilogram of body weight; Sigma Chemical, St. Louis, Missouri), tetracycline (twenty-five milligrams per kilogram of body weight; Lederle Laboratories, Pearl River, New York), and xylenol orange (ninety milligrams per kilogram of body weight; Sigma Chemical) were given intraperitoneally. For animals that were to be killed at one month, tetracycline was given at fourteen days and DCAF, at twenty-eight days. For animals that were to be killed at two months, tetracycline was given at twenty-eight days; DCAF, at forty-two days; and xylenol orange, at fifty-six days. For animals that were to be killed at four months, tetracycline was given at eighty-four days; DCAF, at ninety-eight days; and xylenol orange, at 112 days.
After the femur was removed, it was fixed in cold 40 per cent ethanol, dehydrated, cleared, and embedded in polymethylmethacrylate. Cross sections of the graft and longitudinal sections of the host-graft interface (200 to 300 micrometers) were cut with use of a water-cooled diamond saw. Sections were glued to opalescent two-millimeter-thick plastic blocks. The sections then were ground to a thickness of 100 micrometers, polished, and surface-stained with toluidine blue with the method of Schenk et al.33. This method of preparation differentiates old, well mineralized bone from new bone as well as living bone from dead bone (as determined by the presence or absence of osteocytes), and it demonstrates cement lines. Alternate sections were mounted, ground, and polished in the same manner but were not stained, to allow analysis of fluorochromes. Some specimens were block-stained with 1 per cent basic fuchsin and processed without being embedded.
Five or six cross sections from each graft were analyzed. With use of the Bioquant system (R and M Biometrics, Memphis, Tennessee), the total cross-sectional area of bone and the areas of old bone (original graft) and of new bone were measured. The percentages of the total cross-sectional area of bone that consisted of old or new bone were calculated. Additionally, the number of vessels that were rimmed with fluorochrome in the original graft were counted as an index of revascularization. This value was expressed as the total number of vessels per cross section and as the number of vessels per square millimeter of the original graft. For each measured parameter, the reported mean was calculated from all of the cross sections of a given graft. The union at the proximal and distal host-graft interfaces was evaluated with use of a semiquantitative grading scale (Table II).
Statistical Analysis
Eight dependent variables were measured: the total area of bone, the area of new bone, the area of old bone, the total number of vessels in the original graft, the absolute change in the dry weight of the graft, the compressive strength of the graft, the presence of cytotoxic antibody specific for donor antigens, and the score for the host-graft union. An additional six dependent variables were calculated: the number of vessels per square millimeter of the original graft, the percentage of the total area of bone that was old bone, the percentage of the total area of bone that was new bone, the percentage loss of 45calcium, the percentage loss of 3H-tetracycline, and the percentage change in the dry weight. The mean and the standard error for each dependent variable were calculated for each experimental group at each time-point, and the values were inspected for homogeneity of variance. We used the SPSS/PC+ statistical package (SPSS, Chicago, Illinois) and the PC Statistician package (Human Systems Dynamics, Northridge, California) for statistical analysis. The effects of all independent variables on each dependent variable were analyzed with one-way, two-way, and three-way analysis of variance. The F ratio, the degrees of freedom, and the probability of occurrence are given for each reported analysis of variance. When the F ratio was significant, the Student-Newman-Keuls test was used post hoc to identify significant differences among the groups. Additionally, the data were divided according to the eighteen groups defined by the combinations of the three independent variables. Certain dependent variables were evaluated with a one-way analysis of variance with each of these groups used as the independent variable. When a significant effect was found, the Scheffé test was applied to identify significant differences among the groups.
All three independent variables had significant effects on the pattern, rate, and quality of the incorporation of the graft. Not only were there direct effects of an independent variable but there were also two-way and three-way interactions among the independent variables.
Quantitative Evaluation
Main Effects of Time
We found that the incorporation of a cortical bone graft was a progressive phenomenon: serial changes in most of the dependent variables were observed with time (Table III and Fig. 1). When the main effects were evaluated, the percentage of the total area of bone that was old bone (original graft) significantly decreased (p < 0.001) with time. The total area of bone (p = 0.047), the area of new bone (p < 0.001), the percentage of the total bone area that was new bone (p < 0.001), the percentage loss of 3H-tetracycline and of 45calcium (p < 0.001 for both), the total number of vessels in the original graft (p = 0.02), the number of vessels per square millimeter of the original graft (p = 0.017), and the score for the host-graft union (p < 0.001) significantly increased with time.
Main Effects of the Degree of Histocompatibility Matching
Systemic donor-specific antibody was measurable in serum only after implantation of a graft with a major mismatch, but an interaction was demonstrated between the presence of the antibody and the treatment of the graft (whether it was fresh or frozen), as will be described. Systemic donor-specific antibody was not detectable in the serum after implantation of the fresh or the frozen syngeneic grafts or after implantation of the fresh or the frozen grafts with a minor mismatch. Revascularization was profoundly affected by histocompatibility-antigen matching: the syngeneic grafts were revascularized more quickly and to a greater extent than the grafts with either a major or a minor mismatch (Table IV).
Main Effects of the Treatment of the Graft
Freezing markedly attenuated the systemic antibody response to the grafts with a major mismatch. Compared with the fresh grafts, the frozen grafts had a smaller total area of bone (p = 0.06), had a smaller area of new bone (p = 0.025), had a smaller percentage of total bone area that was new bone (p = 0.09), and had fewer vessels in the original graft (total number of vessels, p = 0.033; number of vessels per square millimeter, p = 0.054) (Table V and Fig. 2). In contrast, frozen grafts, as a group, had an increased absolute change in the dry weight (p < 0.001) and a greater increase in the percentage change in dry weight (p < 0.001).
Interactive Effects
The treatment of the graft (whether it was fresh or frozen) and the degree of histocompatibility matching interacted to affect the systemic antibody response to the graft. Cytotoxic antibody specific for donor antigens was produced consistently in the animals that had received a fresh graft with a major mismatch: at one month six of seven animals tested positive (mean titer, 1:73), at two months all of fourteen animals tested positive (mean titer, 1:54), and at four months nine of ten animals tested positive (mean titer, 1:33). Cytotoxic donor-specific antibody was produced intermittently in the animals that had received a frozen graft with a major mismatch: at one month one of five animals tested positive (mean titer, 1:160); at two months one of eight animals tested positive (mean titer, 1:40); and at four months one of seven animals tested positive (mean titer, 1:40).
The treatment of the graft and the degree of histocompatibility matching also interacted to affect the revascularization of the graft. Freezing significantly (p < 0.001) reduced the revascularization of the syngeneic grafts, had no discernible effect on the grafts with a minor mismatch, and tended to increase (but not significantly) the revascularization of the grafts with a major mismatch. This finding was observed regardless of whether revascularization was measured by the total number of vessels or the number of vessels per square millimeter of the original graft (Table VI).
The treatment of the graft also interacted with time to affect the revascularization of the grafts. As a group, the fresh grafts were well revascularized by one month and remained so throughout the four months of the study. The revascularization of the frozen grafts, as a group, was significantly (p < 0.006) retarded, achieving parity with that of the fresh grafts only at four months.
A three-way interaction among time, the degree of histocompatibility matching, and the treatment of the graft affected the total number of vessels and the number of vessels per square millimeter of the original graft. When the data were analyzed with a three-way analysis of variance, with time, treatment of the graft, and degree of histocompatibility matching used as the independent variables, a significant interaction was not identified for either the total number of vessels in the original graft (F ratio = 0.810; degrees of freedom = 4, 98; and p = 0.523) or the number of vessels per square millimeter of the original graft (F ratio = 1.266; degrees of freedom = 4, 98; and p = 0.306). However, when the data were organized by groups with the same descriptors—for example, fresh syngeneic grafts at one month, fresh grafts with a minor mismatch at two months, or frozen grafts with a major mismatch at four months—and were analyzed with one-way analysis of variance with the eighteen groups used as an independent variable, a significant effect was found (F ratio = 27.411; degrees of freedom = 17, 224; and p < 0.001). The post hoc Scheffé procedure identified many differences among the groups that were significant at a level of p < 0.05. These differences illustrate the interactive effects of the degree of histocompatibility matching between the recipient and the donor, the treatment of the graft (whether it was fresh or frozen), and time on the revascularization of these non-vascularized cortical bone grafts (Fig. 3).
Qualitative Evaluation
The fresh syngeneic grafts were noted to be minimally resorbed. New woven bone was forming periosteally and endosteally by one month. Endosteal and periosteal new bone was a mixture of lamellar and woven bone at two months and consisted wholly of lamellar bone at four months. All new bone tightly adhered to and was integrated with the original graft. Many vascularized channels that were rimmed with fluorochromes were evident beginning at one month (Figs. 4-A and 4-B).
There was also little resorption of the frozen sygeneic grafts. New-bone formation around these grafts was delayed until two months, and bone always consisted of a mixture of woven and lamellar bone. There were a few vascular channels at two months, and there were many at four months (Figs. 5-A and 5-B).
The fresh grafts with a minor mismatch underwent slight periosteal and endosteal resorption. In the periosteum, there was some new woven bone at one month and abundant woven bone at two months. By four months, new lamellar bone had formed and was well integrated with the original graft. Few vascular channels were evident at one month; a moderate number, at two months; and many, at four months (Figs. 6-A, 6-B, and 6-C).
There was moderate periosteal and endosteal resorption of the frozen grafts with a minor mismatch starting at one month. Endosteal formation of woven bone was demonstrated at two and four months, but little periosteal formation of bone was seen until four months. All new bone was a mixture of woven and lamellar bone. There were fewer vascular channels in these grafts than in the fresh grafts with a minor mismatch until four months, at which time they had a similar number of vessels.
The fresh grafts with a major mismatch underwent marked periosteal and endosteal resorption, sometimes resulting in almost complete loss of the graft (Figs. 7-A and 7-B). In the periosteum, small foci of woven bone were noted at two months and woven bone was abundant at four months; however, new bone was never well integrated with the original graft. No vascular channels were present within the substance of the original graft at one and two months, and only a few were present at four months.
The frozen grafts with a major mismatch underwent some endosteal and periosteal resorption. New bone did not form until two months and consisted of mostly woven bone. By four months, lamellar new bone had formed in the periosteum and the endosteum. Very few vascular channels were present in the original graft at one and two months. Slightly more vascular channels were noted at four months than at the earlier time-points (Fig. 8).
The host-graft interfaces first filled with fibrous and chondroid tissue, which, over time, was variably replaced by woven and lamellar bone. All interfaces became more mature with time, but there were no demonstrated effects of histocompatibility matching or treatment of the graft.
We found that incorporation of a cortical bone graft is a progressive phenomenon: every functional parameter of incorporation that we measured changed with time and was continuing to change even after four months. No one functional parameter adequately defined successful incorporation; for example, larger grafts were not necessarily stronger. In addition, no one independent variable was an autonomous determinant of incorporation. In certain circumstances (such as with a fresh graft with a major mismatch), the role of histocompatibility matching may dominate; in other circumstances (such as with a frozen syngeneic graft), the treatment of the graft may strongly influence the outcome. Incorporation of a cortical bone graft is an intricate process, and the cortical graft itself becomes a very complex structure. Our findings suggest that, to comprehend fully the complexity of the process and of the structure, multiple parameters of incorporation must be measured at sequential time-points and these data must be integrated.
In the animal model that we used, the graft was placed in a stably fixed weight-bearing environment. Stable fixation is important when evaluating the incorporation of cortical bone grafts, as motion interferes with revascularization and affects new-bone formation32. We saw very little resorption of these cortical bone grafts compared with that reported in other experimental models16 and in clinical practice27, and this finding is probably due to the stable fixation and the weight-bearing environment. The use of only male rats of a similar body weight as the recipients also substantially reduced the biological variability36. We used a variety of methods to evaluate the incorporation of the graft and found that no one method adequately demonstrated the process of incorporation. Histological examination of the six serial cross sections of the non-decalcified specimens, together with evaluation with fluorochromes, allowed us to measure the areas of old and new bone and to quantitate revascularization. Our quantitative histological methods, however, were not as helpful for appraising the organization of bone (woven or lamellar bone) or the degree of mineralization of the matrix. These two parameters were best evaluated with qualitative histological assessment, biomechanical testing, isotopic prelabeling, and measurement of dry weight. Histological evaluation was most useful for the study of cell-matrix interactions between the graft and the host soft tissue, such as new-bone formation and revascularization, but was not predictive of changes in the mass or strength of the graft. Biomechanical testing measured the compressive strength of the composite original graft-new bone unit, but the findings did not necessarily correspond with the total bone area or the changes in mass17. With time, a cortical bone graft becomes a very complex structure. It is not homogeneous structurally or functionally; rather, it is an admixture of old bone (original graft), variably revascularized, and new host bone, variably mineralized.
The degree of histocompatibility matching between the graft and the recipient was an influential determinant of the incorporation of the graft. Syngeneic grafts were more successful than grafts with a minor mismatch (matched for major histocompatibility complex antigens only), which in turn were more successful than grafts with a major mismatch (mismatched for major and non-major histocompatibility complex antigens). Investigators have shown that bone cells display class-I and class-II histocompatibility antigens28,34,35, that there are humoral and cellular responses to bone allografts18,21,23,37, and that the magnitude of the disparity in histocompatibility antigens negatively corresponds with incorporation of the graft22,25,38. The influence of histocompatibility matching in our study, even in the absence of a measurable systemic immune response, demonstrates that the incorporation of the graft is a locally governed process. We believe that the sum of the interactions between the cells of the inflammatory-wound healing response in the host and the net biological activity of the graft (its cells, its capacity for osteoinduction, and its ability to support osteoconduction) determines the process of incorporation. Others have shown that cortical bone allografts are revascularized and mineralized in a manner similar to that in autogenous grafts in athymic rats but not in normal, immunologically intact rats24.
Freezing had two main effects on the process of incorporation of the graft: it muted the effects of histocompatibility mismatching and reduced the biological activity of the graft. Both effects are probably a consequence of the killing of cells in the graft. Because the grafts were not cryopreserved in any way, it is unlikely that any cells in the graft survived the freezing and thawing processes. These two effects of freezing resulted in a decrease in the incorporation of the syngeneic grafts, an increase in the incorporation of the grafts with a major mismatch, and had little effect on the incorporation of the grafts with a minor mismatch. The killing of cells and the disruption of the membrane also account for the loss of detectable systemic immunogenicity and the increased revascularization that was seen in the frozen grafts with a major mismatch. Membrane-associated histocompatibility antigens are probably disrupted by the rupture and death of cells. The net biological activity of a graft is the sum of its inherent biological activity (living cells and their products), its capacity to induce surrounding host tissues to produce relevant biological activity (osteoinduction mediated by bioactive factors within the matrix), and its ability to support the ingrowth of osteogenic host tissue (osteoconduction). New bone still formed around and within the frozen grafts, and this formation was probably mediated by matrix-derived factors, but the process was slow and the new bone was scanty compared with that in the fresh grafts with living surface cells. However, the slow rate of new-bone formation may favor more complete mineralization of the matrix. In the frozen grafts, as a group, the absolute change in the dry weight, the percentage change in the dry weight, and the mineral density were greater than in the fresh grafts. All of these parameters reflect the amount of matrix and the degree of mineralization of the matrix.
Although both the degree of histocompatibility matching between the graft and the recipient and the treatment of the graft (whether it was fresh or frozen) had significant independent effects on the incorporation of the graft, there were consistent and important interactions among these properties. For example, the fresh grafts as a group had more vessels than the frozen grafts as a group, but the fresh syngeneic grafts had substantially more vessels than the fresh grafts with a major mismatch. Time also interacted with the properties of the graft. For example, there were significantly fewer vessels in the fresh grafts with a minor mismatch than in the fresh syngeneic grafts at one month, but the number of vessels was equal at four months. Under certain circumstances, a single determinant dominated the control of the process of incorporation. For example, the degree of histocompatibility matching was the critical functional determinant of incorporation of the fresh grafts. However, among the frozen grafts, the effects of freezing were more dominant than the effects of histocompatibility matching. Under other circumstances, the functional outcome was modulated by the aggregate interaction of variables, with no one variable having a predominant influence. Some of these individual observations have been made by other investigators6,11,19,22, but as far as we know ours is the first large study involving a well controlled model allowing extensive statistical analysis of time, the treatment of the graft (whether it was fresh or frozen), the degree of histocompatibility matching, and the interactions among these variables. All 288 grafts were equivalent in size, were subject to equivalent strain, were placed in equivalent soft-tissue beds, and were stabilized by internal fixation. We measured fourteen different parameters of incorporation with use of three methodologies: histological examination, biomechanical testing, and quantitative isotopic kinetics.
We defined successful incorporation of a cortical bone graft as concurrent revascularization and substitution with host bone without a significant loss of strength. Successful incorporation results in a composite that can bear physiological loads and can repair and remodel itself in response to changes in load or to fatigue damage. Our data demonstrate that there are multiple patterns of incorporation and that the sum of the biological properties of the graft determine the direction and rate of incorporation. Grafts with different biological properties had different pathways toward incorporation, some leading toward success and others leading toward failure. The route to successful incorporation of fresh syngeneic grafts is short and direct. Frozen syngeneic grafts and fresh and frozen grafts with a minor mismatch also progress toward successful incorporation with time, but the process is slower. The ultimate outcome with regard to incorporation of frozen grafts with a major mismatch is harder to predict; these grafts seem destined for failure (a lack of revascularization and loss of strength), but the process is slow and variable. Fresh grafts with a major mismatch clearly seem destined for fairly rapid failure of incorporation.
These findings are clinically important because they help to explain the unpredictability of incorporation of cortical bone grafts. The grafts in our study were placed in healthy, well vascularized soft-tissue beds and were stably fixed. Even under these optimum conditions, the process of incorporation was demonstrated to be complex and to be controlled by multiple factors. All of the determinants that we studied—time, the degree of histocompatibility matching, and the treatment of the graft (whether it was fresh or frozen)—affected the rate and direction of the continuum of events leading to the incorporation of cortical bone grafts. The graft that is most commonly implanted clinically, the frozen (or processed) mismatched allograft, had the least predictable process of incorporation. Routine tissue-antigen matching and the use of fresh matched cortical bone grafts may improve the clinical outcome, but they are unlikely to be feasible in the near future, if at all. However, the findings of the present study suggest that the process of incorporation may be manipulated by, for example, the addition or removal of cells (and, indirectly, of cytokines). This observation raises the possibility of enhancing the incorporation of frozen mismatched grafts.
Berggren, A.; Weiland, A. J.; and |and |Östrup, L. T.: Bone scintigraphy in evaluating the viability of composite bone grafts revascularized by microvascular anastomoses, conventional autogenous bone grafts, and free non-vascularized periosteal grafts. J. Bone and Joint Surg.,64-A: 799-809, July 1982.64-A799
1982
Bonfiglio, M.: Repair of bone-transplant fractures. J. Bone and Joint Surg.,40-A: 446-456, April 1958.40-A446
1958
Bonfiglio, M., and |and |Jeter, W. S.: Immunological responses to bone. Clin. Orthop.,87: 19-27, 1972.8719
1972
[PubMed]
Bonfiglio, M.; Jeter, W. S.; and |and |Smith, C. L.: The immune concept: its relation to bone transplantation. Ann. New York Acad. Sci.,59: 417-432, 1955.59417
1955
[CrossRef]
Bos, G. D.; Goldberg, V. M.; Zika, J. M.; Heiple, K. G.; and |and |Powell, A. E.: Immune responses of rats to frozen bone allografts. J. Bone and Joint Surg.,65-A: 239-246, Feb. 1983.65-A239
1983
Burchardt, H.: The biology of bone graft repair. Clin. Orthop.,174: 28-42, 1983.17428
1983
[PubMed]
Burchardt, H., and |and |Enneking, W. F.: Transplantation of bone. Surg. Clin. North America,58: 403-427, 1978.58403
1978
Burwell, R. G.: Studies in the transplantation of bone. V. The capacity of fresh and treated homografts of bone to evoke transplantation immunity. J. Bone and Joint Surg.,45-B(2): 386-401, 1963.45-B(2)386
1963
Burwell, R. G.: The fate of bone grafts. In Recent Advances in Orthopaedics, pp. 115-207. Edited by A. G. Apley. London, J. and A. Churchill, 1969.
Burwell, R. G.; Gowland, G.; and |and |Dexter, F.: Studies in the transplantation of bone. VI. Further observations concerning the antigenicity of homologous cortical and cancellous bone. J. Bone and Joint Surg.,45-B(3): 597-608, 1963.45-B(3)597
1963
Dambe, L. T.; Saur, K.; and |and |Schweiberer, L.: Revaskularisation frischer homologer Knochentransplantate in der Diaphyse des Röhrenknochens beim Hund. Arch. Orthop. and Trauma Surg.,92: 35-45, 1978.9235
1978
[CrossRef]
Eitel, F.; Klapp, F.; Jacobson, W.; and |and |Schweiberer, L.: Bone regeneration in animals and in man. A contribution to understanding the relative value of animal experiments to human pathophysiology. Arch. Orthop. and Trauma Surg.,99: 59-64, 1981.9959
1981
[CrossRef]
Elves, M. W.: Studies of the behaviour of allogeneic cancellous bone grafts in inbred rats. Transplantation,19: 416-425, 1975.19416
1975
[PubMed][CrossRef]
Elves, M. W.; Gray, J. C.; and |and |Thorogood, P. V.: The cellular changes occurring with allografts of marrow-containing cortical bone. J. Anat.,122: 253-269, 1976.122253
1976
[PubMed]
Enneking, W. F., and |and |Mindell, E. R.: Observations on massive retrieved human allografts. J. Bone and Joint Surg.,73-A: 1123-1142, Sept. 1991.73-A1123
1991
Enneking, W. F.; Burchardt, H.; Puhl, J. J.; and |and |Piotrowski, G.: Physical and biological aspects of repair in dog cortical-bone transplants. J. Bone and Joint Surg.,57-A: 237-252, March 1975.57-A237
1975
Ferretti, J. L.; Capozza, R. F.; Mondelo, N.; and |and |Zanchetta, J. R.: Interrelationships between densitometric, geometric, and mechanical properties of rat femora: inferences concerning mechanical regulation of bone modeling. J. Bone and Min. Res.,8: 1389-1396, 1993.81389
1993
[CrossRef]
Friedlaender, G. E.; Strong, D. M.; and |and |Sell, K. W.: Studies on the antigenicity of bone. I. Freeze-dried and deep-frozen allograft in rabbits. J. Bone and Joint Surg.,58-A: 854-858, Sept. 1976.58-A854
1976
Goldberg, V. M., and |and |Lance, E. M.: Revascularization and accretion in transplantation. Quantitative studies of the role of the allograft barrier. J. Bone and Joint Surg.,54-A: 807-816, June 1972.54-A807
1972
Goldner-Sauve, A. J.; Fuks, A.; Guttmann, R. D.; and |and |Cramer, D. V.: A class II monoclonal antibody specific for the RT1.B, rather than the RT1.D, product of the rat major histocompatibility complex. Transplantation,35: 495-497, 1983.35495
1983
[PubMed][CrossRef]
Halloran, P. F.; Lee, E. H.; Ziv, I.; Langer, F.; and |and |Gross, A. E.: Orthotopic bone transplantation in mice. II. Studies of the alloantibody response. Transplantation,27: 420-426, 1979.27420
1979
[PubMed]
Heiple, K. G.; Chase, S. W.; and |and |Herndon, C. H.: A comparative study of the healing process following different types of bone transplantation. J. Bone and Joint Surg.,45-A: 1593-1616, Dec. 1963.45-A1593
1963
Horowitz, M. C., and |and |Friedlaender, G. E.: Induction of specific T-cell responsiveness to allogeneic bone. J. Bone and Joint Surg.,73-A: 1157-1168, Sept. 1991.73-A1157
1991
Kirkeby, O. J.; Nordsletten, L.; and |and |Skjeldal, S.: Healing of cortical bone grafts in athymic rats. Acta Orthop. Scandinavica,63: 318-322, 1992.63318
1992
[CrossRef]
Kliman, M.; Halloran, P. F.; Lee, E.; Esses, S.; Fortner, P.; and |and |Langer, F.: Orthotopic bone transplantation in mice. III. Methods of reducing the immune response and their effect on healing. Transplantation,31: 34-40, 1981.3134
1981
[PubMed][CrossRef]
Kunz, H. W.; Gill, T. J., III.; and |and |Misra, D. N.: The identification and mapping of a second class I locus in the major histocompatibility complex of the rat. J. Immunol.,128: 402-408, 1982.128402
1982
[PubMed]
Mankin, H. J.: Complications of allograft surgery. In Osteochondral Allografts: Biology, Banking, and Clinical Applications, pp. 259-274. Edited by G. E. Friedlaender, H. J. Mankin, and K. W. Sell. Boston, Little, Brown, 1983.
Muscolo, D. L.; Kawai, S.; and |and |Ray, R. D.: In vitro studies of transplantation antigens present on bone cells in the rat. J. Bone and Joint Surg.,59-B(3): 342-348, 1977.59-B(3)342
1977
Muscolo, D. L.; Caletti, E.; Schajowicz, F.; Araujo, E. S.; and |and |Makino, A.: Tissue-typing in human massive allografts of frozen bone. J. Bone and Joint Surg.,69-A: 583-595, April 1987.69-A583
1987
Ray, R. D.: Vascularization of bone grafts and implants. Clin. Orthop.,87: 43-48, 1972.8743
1972
[PubMed]
Reske, K., and |and |Weitzel, R.: Immunologically discrete conformation isomers of I-A locus-equivalent class II molecules detected in Lewis rats. European J. Immunol.,15: 1229-1239, 1985.151229
1985
[CrossRef]
Sauer, H. D., and |and |Schoettle, H.: The stability of osteosyntheses bridging defects. Arch. Orthop. and Trauma Surg.,95: 27-30, 1979.9527
1979
[CrossRef]
Schenk, R. K.; Olah, A. J.; and Herrmann, W.: Preparation of calcified tissues for light microscopy. In Methods of Calcified Tissue Preparation, pp. 1-56. Edited by G. R. Dickson. New York, Elsevier Science, 1984.
Skjodt, H.; Moller, T.; and |and |Freiesleben, S. F.: Human osteoblast-like cells expressing MHC class II determinants stimulate allogeneic and autologous peripheral blood mononuclear cells and function as antigen-presenting cells. Immunology,68: 416-420, 1989.68416
1989
[PubMed]
Skjodt, H.; Hughes, D. E.; Dobson, P. R.; and |and |Russell, R. G.: Constitutive and inducible expression of HLA class II determinants by human osteoblast-like cells in vitro. J. Clin. Invest.,85: 1421-1426, 1990.851421
1990
[PubMed][CrossRef]
Sontag, W.: Quantitative measurements of periosteal and cortical-endosteal bone formation and resorption in the midshaft of male rat femur. Bone,7: 63-70, 1986.763
1986
[PubMed][CrossRef]
Stevenson, S.: The immune response to osteochondral allografts in dogs. J. Bone and Joint Surg.,69-A: 573-582, April 1987.69-A573
1987
Stevenson, S.; Hohn, R. B.; and |and |Templeton, J. W.: Effects of tissue antigen matching on the healing of fresh cancellous bone allografts in dogs. Am. J. Vet. Res.,44: 202-207, 1983.44202
1983
Stevenson, S.; Li, X. Q.; and |and |Martin, B.: The fate of cancellous and cortical bone after transplantation of fresh and frozen tissue-antigen-matched and mismatched osteochondral allografts in dogs. J. Bone and Joint Surg.,73-A: 1143-1156, Sept. 1991.73-A1143
1991
Stevenson, S.; Dannucci, G. A.; Sharkey, N. A.; and |and |Pool, R. R.: The fate of articular cartilage after transplantation of fresh and cryopreserved tissue-antigen-matched and mismatched osteochondral allografts in dogs. J. Bone and Joint Surg.,71-A: 1297-1307, Oct. 1989.71-A1297
1989
Zolman, J. F.: Biostatistics: Experimental Design and Statistical Inference. New York, Oxford University Press, 1993.