2324 Altered p53 Is Associated with Aggressive Behavior of Chondrosarcoma A Long Term Follow-Up Study Yumi Oshiro, M.D.1 Vijaya Chaturvedi, M.A.1 Dorothy Hayden, M.D.1 Tipu Nazeer, M.D.1 Mark Johnson, M.D.2 Dennis A. Johnston, Ph.D.3 Nelson G. Ordóñez, M.D.1 Alberto G. Ayala, M.D.1 Bogdan Czerniak, M.D.1 1 Department of Pathology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas. 2 Department of Surgical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas. 3 Department of Biomathematics, The University of Texas M. D. Anderson Cancer Center, Houston, Texas. BACKGROUND. p53 is a major tumor suppressor gene that has been implicated in the biology of a variety of human neoplasms, including some that affect the skeleton. Recent studies based on small numbers of cases have shown that overexpression or alteration of the p53 gene is frequently present in high grade, clinically aggressive chondrosarcomas of bone. In this study, the authors addressed the relation between overexpression and alteration of the p53 gene and the clinical aggressiveness of chondrosarcoma in a large series of patients for whom long term follow-up data were available. METHODS. The authors analyzed the expression and/or alteration of the p53 gene in 158 cases of chondrosarcoma of bone using immunohistochemistry, singlestrand conformation polymorphism, and direct sequencing. They then related the findings to various clinicopathologic parameters and long term follow-up data. RESULTS. The presence of overexpression and/or structural alterations of the p53 gene was documented in 38.1% of chondrosarcomas of bone. A statistically significant correlation was observed between overexpression or alteration of the p53 gene and both the histologic grade of the tumor and the presence of metastasis. The probability of local recurrence free, metastasis free, and overall survival was significantly higher for patients with no overexpression or alteration of p53 than for patients with p53 overexpression or alteration. CONCLUSIONS. Overexpression or alteration of the p53 gene is an important predictor of aggressive clinical behavior in chondrosarcoma of bone. Cancer 1998;83: 2324 –34. © 1998 American Cancer Society. KEYWORDS: chondrosarcoma, p53, immunohistochemistry, sequencing, clinical behavior. The authors thank Donna Sprabary for secretarial assistance. This work was supported in part by NIH grant CA66723-04 to Bogdan Czerniak. Dr. Oshiro’s current address: Yumi Oshiro, M.D., 2nd Dept. of Pathology, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, 812 Japan. Presented in part at the 1998 International Academy of Pathology meeting, Boston, Massachusetts, February 28 –March 6, 1998. Address for reprints: Bogdan Czerniak, M.D., Department of Pathology, The University of Texas M. D. Anderson Cancer Center, Box 085, 1515 Holcombe Blvd., Houston, TX 77030. Received January 14, 1998; revision received April 27, 1998; accepted April 27, 1998. © 1998 American Cancer Society C hondrosarcoma is the second most common malignant tumor of bone, accounting for approximately 25% of all primary sarcomas of bone.1 The incidence rate varies in different age groups, from 0.2 to 0.9 per 100,000 persons, and shows a gradual age-related increase that peaks among patients older than 50 years.1 The overall 5-year survival rate is about 70%, but the behavior of individual tumors varies widely, ranging from locally aggressive, nonmetastasizing tumors to high grade malignancies with a high propensity for metastasis. The behavior of these tumors can be predicted to some extent from their histologic grade.2– 6 Nonmetastasizing, locally aggressive lesions are typically of low histologic grade (Grade 1). Virtually all Grade 3 tumors are highly aggressive and exhibit a high propensity for distant metastasis. Grade 2 tumors represent an intermediate group. Some Grade 2 chondrosarcomas behave like locally aggressive lesions, whereas others may metastasize. In addition to histologic grade, other factors, Altered p53 in Chondrosarcoma of Bone/Oshiro et al. such as DNA ploidy, tumor location, and tumor resectability, also affect prognosis.7–13 Little is known about the molecular alterations involved in the pathogenesis of chondrosarcoma. On the chromosomal level, chondrosarcomas exhibit complex alterations with nonreciprocal translocations and deletions involving numerous chromosomes.14,15 It appears that alterations of chromosomes 1p, 4, 5, 9, and 20 are nonrandom and may play a role in the biology of these neoplasms.14,15 The presence of amplifications and deletions of genetic material on multiple chromosomes was recently confirmed by comparative genomic hybridization studies.16 Alterations of putative tumor suppressor genes (CDKN2 and MTS-2) located on chromosome 9 were identified in chondrosarcoma cell lines.17 Amplification of the cmyc protooncogene without any apparent correlation to histologic grade was found in another small series of chondrosarcoma cases.18 More recently, evidence of frequent loss of heterozygosity on 17p and the presence of multiple chromosomal alterations were correlated with high histologic grade19 –23 and aggressive clinical behavior21 in a few small series of chondrosarcoma of bone. Several recent studies suggest that, similar to many other human tumors, a major tumor suppressor gene located on 17p (the p53 gene) may play an important role in the pathogenesis of chondrosarcoma.9,19,21–23 In this study, we report alterations of the p53 gene in a series of 158 cases of chondrosarcoma of bone and relate the findings to clinicopathologic parameters, such as histologic grade, tumor extension, tumor location, and type of surgical procedure, as well as long term follow-up data. PATIENTS AND METHODS Patients One hundred fifty-eight patients with chondrosarcoma of bone were selected from the case files of the Department of Pathology at the University of Texas M. D. Anderson Cancer Center in Houston, Texas. All cases were reviewed to verify the diagnosis and tumor grade and to select a representative paraffin block for the study. In 124 of these cases, radiologic and long term follow-up data were available for analysis. In nine patients, secondary chondrosarcoma developed in association with multiple exostoses; in three patients, it was associated with Ollier’s disease. Three patients with chondrosarcoma had solitary osteochondroma in sites distinct from that of the chondrosarcoma. Cases of dedifferentiated, mesenchymal, and clear cell chondrosarcoma were excluded from the study. All lesions were unifocal chondrosarcomas involving one bone. Tumors were histologically classi- 2325 fied using a three-tier grading system. Grade 2 tumors with predominantly myxoid matrix were distinguished from tumors with predominantly hyaline cartilaginous matrix. None of the patients received chemotherapy or radiation therapy before surgery. Immunohistochemistry Tissue sections were cut 4 m thick, deparaffinized in xylene, and rehydrated in alcohol. They were then immersed in 3% hydrogen peroxide in methanol for 30 minutes to block endogenous peroxidase activity. After several washes in distilled water and phosphatebuffered saline, sections were incubated with normal horse serum to minimize background staining. This was followed by overnight incubation at 4°C with the DO-1 mouse monoclonal antibody to p53 (Oncogene Science, Uniondale, NY; 1:80 dilution). The immunoperoxidase staining was done using an ABC Elite kit (Vector Laboratories, Burlingame, CA) with 0.05% 3,3´-diaminobenzidine in Tris-HCl buffer containing 0.01% hydrogen peroxide (pH 7.6). Sections were counterstained with 0.01% toluidine blue and mounted in permount. Tumors that had 25% or more of positive tumor cell nuclei were defined as positive for overexpression of the p53 gene product. Single-Strand Conformation Polymorphism and Sequencing Single-strand conformation polymorphism (SSCP) and sequencing of exons 5 through 9 of the p53 gene were performed using template DNA extracted from formalin fixed, paraffin embedded tumor tissues. Tissue sections (50 m thick) were deparaffinized in xylene and alcohol, and genomic DNA was extracted by using the proteinase K–phenol chloroform procedure. SSCP analysis of the amplified gene fragments was performed using the following primers flanking coding sequences of exons 5 through 9 of the p53 gene: exon 5:5´-TTCCTCTTCCTGCAGTACTC-3´ 5´-ACCCTGGGCAACCAGCCCTGT-3´ exon 6:5´-ACAGGGCTGGTTGCCCATGGGT-3´ 5´-AGTTGCAAACCAGACCTAT-3´ exon 7:5´-GTGTTGTCTCCTAGGTTGGC-3´ 5´-GTCAGAGGCAAGCAGAGGCT-3´ exon 8:5´-TATCCTGAGTAGTGGTAATC-3´ 5´-AAGTGAATCTGAGGCATAAC-3´ exon 9:5´-GCAGTTATGCCTCAGATTCAC-3´ 5´-AAGACTTAGTACCTGAAGGGT-3´ Amplifications of the gene fragments were performed by polymerase chain reaction (PCR) using 150 ng of genomic DNA in a 10-L volume containing 1 M of each primer, 200 M of dNTPs, 2.5 Ci of [alpha-32P]dCTP, and 0.5 U of Ampli-taq DNA poly- 2326 CANCER December 1, 1998 / Volume 83 / Number 11 FIGURE 1. Distributions of chondrosarcoma are shown by age and anatomic site. merase (Perkin Elmer Cetus, Norwalk, CT). PCR was carried out using 33 cycles (1 minute at 94°C, 1 minute at 55°C, and 2 minutes at 72°C) on a programmable thermal cycler (Perkin Elmer Cetus). Samples were heat denatured at 95°C for 5 minutes and then loaded on a 6% polyacrylamide gel. A nondenatured sample from each case was loaded simultaneously. Electrophoresis was performed at room temperature with a constant power of 3 W for 16 –20 hours. Autoradiography was carried out for 18 –24 hours at – 80°C. PCR-amplified exons that showed band shifts on SSCP analysis suggestive of an altered p53 gene were sequenced. Direct sequencing of PCR-generated gene fragments was performed using the Sequenase PCR Product Sequencing kit (United States Biochemical Corporation, Cleveland, OH), according to the protocol supplied by the manufacturer. PCR was carried out as described above, but without the isotope. Briefly, 30 – 40 ng of the amplified PCR product was treated with 1 U of exonuclease 1 and 2 U of shrimp alkaline phosphatase at 37°C for 15 minutes to remove excess primers and dNTPs. Following heat inactivation of the added enzymes, the PCR product was annealed with 40 ng of the appropriate primer and subsequently labeled with [alpha-35S] dATP. Samples were heated to 80°C, loaded (3.5 L) onto a 6% polyacrylamide urea gel, and run at 1500 V, 55 W for 1.5– 4 hours. Autoradiographs were developed after 5– 8 days of ex- posure without intensifying screens at room temperature. Data Analysis The correlations between overexpression and alteration of the p53 gene and various clinicopathologic parameters were analyzed using chi-square statistics. The length of survival was tested by Kaplan–Meier analysis using both the Gehan–Wilcoxon and Peto log rank tests. To determine which of the clinical parameters would predict p53 status, a logistic multiparameter regression analysis was performed. RESULTS Clinicopathologic Data The patients’ ages at diagnosis ranged from 10 to 91 years (median age, 50 years). The highest incidence of chondrosarcoma occurred in the fifth and sixth decades of life. Seventy-eight of the patients were men, 71 were women, and 9 were of unknown gender. The skeletal and age-related distributions of the tumors are summarized in Figure 1. The most common site was the pelvis, followed by the femur and humerus. Eighty-three patients complained of pain related to the lesion. Five patients had pathologic fracture at presentation. Twenty-six tumors were treated by intralesional resection, 27 by marginal resection, 54 by wide resection, and 5 by radical resection. The exact Altered p53 in Chondrosarcoma of Bone/Oshiro et al. 2327 TABLE 1 Survival Rates and Development of Metastases in Chondrosarcoma by Histologic Grade Histologic grade Survival/metastases 1 2 2-myxoid 3 5-yr survival (%) 10-yr survival (%) Metastases (%) 91 91 5 66 66 30 68 61 42 60 60 50 evident that overall survival and metastasis free rates for patients with Grade 2–3 chondrosarcoma were lower than those for patients with Grade 1 chondrosarcoma. Survival rates for patients with Grade 1 chondrosarcoma after 5 and 10 years were similar and were within the range of 90% (Table 1). The survival rates for patients with higher grade chondrosarcomas (Grades 2–3) ranged from 60% to 68% after 5 and 10 years. The survival rates for patients who had chondrosarcoma involving trunk bones were significantly lower than the rates for those whose chondrosarcomas involved other anatomic sites (Fig. 2C). Alterations of p53 FIGURE 2. Kaplan–Meier plots of overall survival (A) and metastasis free survival (B) by histologic grade, as well as overall survival by tumor location (C), are shown for 158 cases of chondrosarcoma of bone. type of surgical procedure performed on 46 patients was unclear. Tumor size ranged from 2 to 42 cm (median size, 10 cm). Distribution by histologic grade was Grade 1 in 80 cases, Grade 2 in 37 cases, Grade 2-myxoid in 32 cases, and Grade 3 in 9 cases. The overall survival and metastasis free curves for patients with chondrosarcoma are shown by histologic grade in Figures 2A and 2B, respectively. It is Examples of p53 protein expression and alterations of the p53 gene documented by immunohistochemistry, SSCP analysis, and direct sequencing are shown in Figure 3. Complete immunohistochemical and SSCP data were available for 118 cases. The correlation between immunohistochemical and SSCP data is summarized in Table 2. The immunohistochemical analysis documented the presence of overexpression of the p53 gene product in 27 cases of chondrosarcoma (22.9%). In 17 of these cases the SSCP data suggested the presence of the altered p53 gene. Overall, the SSCP analysis suggested the presence of alterations of the p53 gene in 29.6% of the cases. In 11 cases, direct sequencing of the PCR-amplified fragments confirmed the presence of alterations in the p53 gene (Table 3). However, for technical reasons, we were able to confirm the presence of structural alterations of the p53 gene by sequencing in only 31.4% of the cases in which SSCP suggested the presence of an altered p53 gene. In the remaining cases, the quality and/or amount of DNA extracted from paraffin blocks was not adequate for successful sequencing of the PCR-amplified gene fragments. Theoretically, DNA that is used for SSCP could be subjected to sequence analysis. However, in paraffin embedded material, sometimes it is possible to generate readable SSCP radiograms, but the sequencing from the same sample may be unsuccessful, mainly because of insufficient 2328 CANCER December 1, 1998 / Volume 83 / Number 11 FIGURE 3. Alterations of p53 in chondrosarcoma. (A) Immunohistochemical overexpression of p53 protein. Intense nuclear staining is present in the tumor cells of Grade 1 (top), Grade 2-myxoid (middle), and Grade 3 (bottom) chondrosarcoma (original magnification ⫻200). (B) Single-strand conformation polymorphism analysis of exons 7 (top) and 8 (bottom) of p53. Arrows indicate the abnormal bands. N: nondenatured sample; D: denatured sample. (C) Sequence analysis of exon 7 (same case as in top panel of B) demonstrates the G deletion at codon 249. amplification and template contamination with degraded DNA fragments. Seventeen tumors (14.4%) exhibited immunohistochemical overexpression of p53 protein and band shifts suggestive of altered p53 gene by SSCP. In 73 cases (61.9%), no immunohistochemical overexpression or altered p53 gene by SSCP could be documented. Thus, the overall concordance rate between immunohistochemical and SSCP data was 76.3%. In 10 cases (8.5%), overexpression of the p53 protein with no evidence of an altered p53 gene by SSCP was present. In 18 cases (15.2%), no immunohistochemical overexpression of the p53 protein was documented, whereas SSCP showed band shifts sugges- Altered p53 in Chondrosarcoma of Bone/Oshiro et al. TABLE 2 Corelation Between Immunohistochemical Evidence of p53 Protein Expression and Alteration of the p53 Gene Documented by SSCP in 118 Cases of Chondrosarcoma Alteration of p53 by SSCP Immunohistochemistry Negative Positive Total Negative or positive in ⬍25% of cells Positive in ⬎25% of cells Total 73 (61.9)a 10 (8.5) 83 (70.4) 18 (15.2) 17 (14.4) 35 (29.6) 91 (77.1) 27 (22.9) 118 SSCP: single-strand conformation polymorphism. a No. of cases/% of cases within the group. tive of an altered gene. In 11 of these cases, the presence of structural alteration in the gene was confirmed by direct sequencing. Six of these cases showed alterations caused by frame shifts and point mutations resulting in amino acid substitutions that could be documented in four cases. A nucleotide insertion causing a stop signal sequence was found in one case. In one remaining case, a nucleotide substitution was located outside of the coding sequence. In summary, the absence of overexpression of the p53 protein and the presence of the normal p53 gene was suggested in 61.9% of cases. In 14.4% of cases, the presence of an overexpressed protein and altered gene was documented. In 15.2% of cases, the presence of an altered gene was documented by SSCP and sequencing with no evidence of immunohistochemical overexpression. The overexpression of an apparently normal p53 protein was present in 8.5% of chondrosarcomas. Finally, combined results from immunohistochemical analysis, SSCP, and direct sequencing suggested the overexpression and/or alteration of the p53 gene in 38.1% of chondrosarcomas. The correlation between the overexpression or alteration of the p53 gene and clinicopathologic features of chondrosarcoma of bone are summarized in Table 4. Strong, statistically significant correlations between the overexpression or alteration of the p53 gene and histologic grade (P ⫽ 0.003) and the presence of metastases (P ⫽ 0.034) were found. No statistically significant correlations were found between the overexpression or alteration of p53 and age, gender, presence or absence of extension into soft tissue, tumor size, or the presence or absence of local recurrence. However, the percentage of tumors with overexpression or alteration of the p53 gene was consistently slightly higher in association with the advanced age of patients, larger tumor size, and presence of local recurrence. Grade 2-myxoid and Grade 3 chondrosarcomas showed 2329 TABLE 3 Sequencing Analysis of Cases with Alteration of the p53 Gene Documented by SSCP Case no. Exon Codon Nucleotide change Amino acid alteration 2 9 10 10 10 41 51 79 101 104 110 124 128 128 7 6 7 7 8 6 6 8 8 7 7 6 6 6 240 212 246 249 274 212 220 276 279 Intron 241 Intron 190 192 T3A G ins. G ins. G del. T3C G ins. CG ins. T ins. T ins. a ins. C3G G3A T3G C3G Ser 3 Arg Frame shift Frame shift Frame shift Silent Frame shift Frame shift Frame shift Nonsense No change Ser 3 Cys Splice site mutation Silent Gln 3 Glu SSCP: single-strand conformation polymorphism; del: deletion; ins: insertion. overexpression or alteration of p53 in 60% and 100% of the cases, respectively. The survival rates of patients with overexpression or alteration of the p53 gene were significantly lower than the survival rates of patients who had no evidence of p53 overexpression or alteration (Fig. 4A). Similarly, the probability of longer local recurrence free and metastasis free survival was significantly greater for patients who had no evidence of p53 overexpression or alteration, compared with the same probability for patients whose tumors exhibited overexpression or alteration of the p53 gene (Fig. 4B and 4C). In patients with evidence of p53 overexpression or alteration, virtually all local recurrences and distant metastases occurred within 3 years. On the other hand, patients whose tumors showed no evidence of p53 overexpression or alteration had gradually declining relapse free rates, and their tumors continued to recur as long as 10 years after the primary diagnosis. Because the outcome of patients might be dependent on the type of surgical procedure, i.e., incomplete resection, positive margins, contamination by ruptured tumor, etc., all cases were divided into two groups: 1) cases treated with intralesional excision, marginal excision, or any other type of incomplete excision; and 2) cases treated with wide local or radical complete excision. There was a highly significant correlation between overall survival, local recurrence free and metastasis free survival, and p53 status for patients with chondrosarcoma who underwent complete excision. Completely excised lesions that exhibited overexpression or alteration of the p53 gene were associated with significantly lower overall survival, local recurrence 2330 CANCER December 1, 1998 / Volume 83 / Number 11 TABLE 4 Overexpression/Alteration of p53 Gene by Clinicopathologic Parameters in 118 Cases of Chondrosarcoma Clinicopathologic features No. of cases examined Positive cases (%) All patients Age ⬍50 yrs ⱖ50 yrs Gender Male Female Tumor extension Confined to bone Invasion of soft tissue Tumor size ⬍10 cm ⱖ10 cm Local recurrence No Yes Distant metastasis No Yes Histologic grade 1 2 2-myxoid 3 118 45 (38.1) 60 38 20 (33.3) 17 (50.0) 0.257 50 53 16 (32.0) 22 (41.5) 0.318 27 38 9 (33.3) 18 (47.4) 0.258 27 29 10 (37.0) 15 (48.4) 0.269 49 33 18 (36.7) 14 (42.4) 0.605 65 14 22 (33.8) 9 (64.3) 0.034 60 28 25 5 14 (23.3) 11 (39.3) 15 (60.0) 5 (100.0) 0.003 free survival, and metastasis free survival rates (Fig. 5). Conversely, for patients who underwent incomplete excision, there was no significant correlation between overexpression or alteration of the p53 gene and overall survival, local recurrence free survival, or metastasis free survival rates (data not shown). An analysis of the correlation between the p53 status restricted to the combined group of Grade 1 and 2 chondrosarcomas and treatment by complete excision showed that overexpression or alteration of p53 had a similar strong association with a high propensity for local recurrence and distant metastases as well as an association with lower overall survival rates (Fig. 6). A similar separate analysis of p53 status in Grade 1, 2, and 3 chondrosarcomas was not possible because of the insufficient number of tumors treated with complete excision or the inappropriate stratification of cases with different p53 status within these groups. A small proportion of Grade 1 tumors showed evidence of overexpression or alteration of p53, and all Grade 3 tumors exhibited evidence of p53 alteration or overexpression. The number of Grade 2 lesions treated with complete excision was too small for the analysis to be performed. It is noteworthy that the analysis of local recurrence free time according to the histologic grade of completely excised tumors disclosed an increased propensity for local recurrence of Grade 2 and 3 tumors as compared with Grade 1 lesions, but the differences Univariative analysis (P value) were not statistically significant (Fig. 7). Therefore, the propensity for local recurrence is unlikely to be strongly associated with histologic grade if the tumor is completely excised. This observation, together with a strong correlation of local recurrence rate, propensity for metastasis, and overall status of survival to the p53 altered status of low-to-intermediate-grade tumors, suggests that this correlation might be independent of histologic grade. To determine which of the clinical parameters would predict p53 status, a logistic multiparameter regression analysis was performed that included individual clinical parameters. As expected from the univariate analysis, histologic grade was the single best univariate parameter from the list (log likelihood [LL] ⫽ ⫺69.34; P ⬍0.001). An unexpected finding was that histologic grade and local advancement of the tumor (tumor with extension into soft tissue) was the best pair of clinical parameters (LL ⫽ ⫺37.23; P ⬍0.001), as opposed to tumor grade and distant metastasis (LL ⫽ ⫺43.28; P ⬍0.001). The three best clinical parameters associated with the overexpression or alteration of p53 were high histologic grade (Grades 2 and 3), the presence of distant metastasis, and tumor size ⬎10 cm (LL ⫽ ⫺29.04; P ⬍0.001). The four best were high histologic grade, the presence of distant metastases, large tumor size (⬎10 cm), and extension into soft tissue (LL ⫽ ⫺26.66; partial P ⫽ 0.017). The combina- Altered p53 in Chondrosarcoma of Bone/Oshiro et al. FIGURE 4. Kaplan–Meier plots of local recurrence free (A), metastasis free (B), and overall survival (C) are shown in relation to p53 status. Analysis is based on follow-up data of 118 cases of chondrosarcoma of bone, regardless of complete or incomplete excision of the primary tumor. p53⫹: presence of overexpression and/or alteration of the p53 gene; p53-: absence of overexpression and alteration of the p53 gene. tion of all five parameters (the four just cited and tumor site) was not significantly better than the best four (LL ⫽ ⫺25.63; partial P ⫽ 0.30). DISCUSSION In this study, we demonstrate the presence of overexpression and/or structural alterations of the p53 gene 2331 FIGURE 5. Kaplan–Meier plots of local recurrence free survival (A), metastasis free survival (B), and overall survival (C) are shown in relation to the p53 status of 59 patients treated by complete local or radical excision. The analysis of patients treated by intralesional, marginal, or in any other way incomplete excision showed no statistical significance of local recurrence free and metastasis free survival or of overall survival in relation to p53 status (data not shown). p53⫹: presence of overexpression and/or alteration of the p53 gene; p53-: absence of overexpression and alteration of the p53 gene. in 38.1% of chondrosarcomas of bone. The presence of an overexpressed, structurally altered gene product was documented in 14.4% of the cases. An apparently normal overexpressed p53 protein was found in 8.5% of the cases, and an altered gene without overexpres- 2332 CANCER December 1, 1998 / Volume 83 / Number 11 FIGURE 7. A Kaplan–Meier plot shows local recurrence free survival by histologic grade for 59 patients who underwent complete local or radical excision. Although patients with Grade 2 or 3 tumors showed decreased rates of freedom from local recurrence as compared with patients who had Grade 1 tumors, the difference was not statistically significant. FIGURE 6. Kaplan–Meier plots of local recurrence free survival (A), metastasis free survival (B), and overall survival (C) are shown according to the p53 status of 55 patients with Grade 1 or 2 chondrosarcoma who underwent complete local or radical excision. Analysis within other individual histologic grades of chondrosarcoma was not possible because of an insufficient number of cases and stratification of p53- versus p53⫹ groups. p53⫹: presence of overexpression and/or alteration of the p53 gene; p53-: absence of overexpression and alteration of the p53 gene. sion of its protein was present in 15.2% of the cases. It should be mentioned, however, that the tests and laboratory procedures used in this study were not capable of identifying all conceivable alterations of the p53 gene. Hence, for cases with immunohistochemical evidence of p53 overexpression but no evidence of p53 alterations by SSCP or direct sequencing, the presence of a structurally altered gene could not be completely ruled out.24 In such cases, the alterations might have been present outside of the screened sequences. Nonmutational stabilization of the wild p53 protein should also be considered in these cases. Moreover, some genes upstream or downstream from the p53 gene that are involved in cell-cycle regulation may be altered and indirectly cause p53 overexpression without evidence of structural gene alterations. The absence of overexpressed p53 protein in cases with a documented altered p53 gene can be explained by the presence of nonsense mutations, insertions, or deletions that lead to premature cessation of protein synthesis and/or generate a gene product that has a different antigenicity and cannot be identified by the antibody used in this study. In fact, our sequencing data invariably showed deletions and/or insertions that resulted in frame shifts in some cases without overexpression of the p53 protein. Previous reports based on limited numbers of cases have shown that the dominant type of point mutation within the p53 gene in chondrosarcoma results in amino acid substitution.19,23,25,26 On the other hand, a comparative analysis of p53 alterations has demonstrated that sarcomas in general have a higher rate of nonsense mutations that result in frame shift or larger deletions than epithelial malignancies have.26 Our results are in keeping with these observations and indicate that the p53 gene in chondrosarcoma of bone predominantly shows nonsense substitutions or insertions that result in frame shifts, as compared with mutations that cause amino acid change. Altered p53 in Chondrosarcoma of Bone/Oshiro et al. In our study, the concordance rate between gene alteration and protein overexpression was 76.3%. The reported concordance rates between immunohistochemical evidence of p53 overexpression and the presence of an altered p53 gene varied in different studies, from 55% up to 100%.27–32 The analysis of our data showed a strong correlation between overexpression or alteration of p53 and high histologic grade as well as the propensity for distant metastasis in chondrosarcoma of bone. In univariate analysis, no correlation was found between overexpression or alteration of p53 and age, gender, soft tissue extension, tumor size and location, or the presence of local recurrence. Analysis of our data has shown that overexpression or alteration of the p53 gene in chondrosarcoma of bone is strongly associated with a higher propensity for local recurrence and metastasis and overall shorter survival times, meaning that it signifies a clinically aggressive variant of the tumor. The association of the altered p53 status with a high propensity for local recurrence and distant metastasis as well as overall poor survival in a group of low and intermediate grade chondrosarcoma (Grades 1 and 2) suggests that the p53 status may represent a predictor of aggressive behavior of chondrosarcoma that is independent of grade. However, more detailed analysis of p53 status independent of histologic grade, i.e., with each histologic grade individually, was not possible in the cohort enrolled in this study. Multivariate analysis by logistic regression has shown that histologic grade is the best parameter to predict p53 status, because high grade (Grade 3) chondrosarcomas of bone often exhibit overexpression and/or alteration of the p53 gene. The three best univariate parameters associated with overexpression or alteration of the p53 gene were high histologic grade, presence of distant metastasis, and tumor size ⬎10 cm. In multivariate analysis, the association of p53 overexpression or alteration with larger tumor size (⬎10 cm) and higher stage (soft tissue extension) suggests that the alterations of this gene may represent a later phenomenon in the development of chondrosarcoma of bone. The published data on chondrosarcoma of bone are based on small series of cases without long term follow-up, and they suggest that p53 is predominantly altered in high grade lesions.9,19 –23,33 The correlation between immunohistochemical evidence of p53 overexpression and a lower 5-year survival rate was documented in a series of 29 cases of chondrosarcoma of bone.21 All dedifferentiated chondrosarcomas have been reported to exhibit immunohistochemical evidence of overexpression of the p53 protein.9,22 The presence of a mutated gene, as shown by SSCP and 2333 direct sequencing, was documented in several cases of Grade 3 and dedifferentiated chondrosarcomas based on the analysis of 2 small series consisting of 14 and 22 cases of chondrosarcomas of bone, respectively.19,23 Loss of heterozygosity on chromosome 17p in the region corresponding to the p53 locus was shown to be present in approximately 25% of chondrosarcomas of bone.23 This percentage is in the same range as the evidence of overexpression or alterations of p53 provided by our study and suggests that in chondrosarcoma of bone altered p53 frequently acts in a homozygotic state. In several reports, all immunohistochemically analyzed enchondromas were shown to be negative for p53 protein.9,20,21,33 Therefore, immunohistochemical evidence of p53 overexpression and/or the identification of an altered p53 gene in neoplastic cartilage lesions of bone signifies aggressiveness of the lesion. In summary, this study documents the frequent presence of p53 overexpression or alteration in chondrosarcoma of bone and establishes the usefulness of the gene as an important prognostic factor in the workup of patients with this cancer. REFERENCES 1. Dorfman HD, Czerniak B. Bone cancers. Cancer 1995;75: 203–10. 2. Evans HL, Ayala AG, Romsdahl MM. Prognostic factors in chondrosarcoma of bone: a clinicopathologic analysis with emphasis on histologic grading. Cancer 1977;40:818 –31. 3. Kreicbergs A, Slezak E, Soderberg G. The prognostic significance of different histomorphologic features in chondrosarcoma. Virchows Arch A Pathol Anat 1981;390:1–10. 4. Ishida T, Kikuchi F, Machinami R. Histological grading and morphometric analysis of cartilaginous tumours. Virchows Arch A Pathol Anat 1991;418:149 –55. 5. O’Neal LW, Ackerman LV. Chondrosarcoma of bone. Cancer 1951;551–77. 6. Unni KK, Dahlin DC. Grading of bone tumors. Semin Diagn Pathol 1984;1:165–72. 7. Alho A, Connor JF, Mankin HJ, Schiller AL, Campbell CJ. Assessment of malignancy of cartilage tumors using flow cytometry: a preliminary report. J Bone Joint Surg 1983;656A: 779 – 85. 8. Alho A, Skjedal S, Melvik JE, Pettersen EO, Larsen TE. The clinical importance of DNA synthesis and aneuploidy in bone and soft tissue tumors. Anticancer Res 1993;13:2383–7. 9. Coughlan B, Feliz A, Ishida T, Czerniak B, Dorfman HD. p53 expression and DNA ploidy of cartilage lesions. Hum Pathol 1995;26:620 – 4. 10. Helio H, Karaharju E, Bohling T, Kivioja A, Nordling S. Chondrosarcoma of bone: a clinical and DNA flow cytometric study. Eur J Surg Oncol 1995;21:408 –13. 11. Kreicbergs A, Boquist L, Borssen B, Larsson SE. Prognostic factors in chondrosarcoma: a comparative study of cellular DNA content and clinicopathologic features. Cancer 1982; 50:577– 83. 2334 CANCER December 1, 1998 / Volume 83 / Number 11 12. Mankin HJ, Connor JF, Schiller AL, Perlmutter N, Alho A, McGuire M. Grading of bone tumors by analysis of nuclear DNA content using flow cytometry. J Bone Joint Surg 1985; 67A:404 –13. 13. Pritchard DJ, Lunke RJ, Taylor WF, Dahlin DC, Medley BE. Chondrosarcoma: a clinicopathologic and statistical analysis. Cancer 1980;45:149 –57. 14. Mandahl N, Heim S, Arhedin K, Rydholm A, Willen H, Mitelman F. Chromosomal rearrangements in chondromatous tumors. Cancer 1990;65:242– 8. 15. Ozisik YY, Meloni AM, Peier A, Altungoz O, Spanier SS, Zalupski MM, et al. Cytogenetic findings in 19 malignant bone tumors. Cancer 1994;74:2268 –75. 16. Larramendy ML, Tarkkanen M, Valle J, Kivioja AH, Ervasti H, Karaharju E, et al. Gains, losses, and amplifications of DNA sequences evaluated by comparative genomic hybridization in chondrosarcomas. Am J Pathol 1997;150:685–91. 17. Jagasia AA, Block JA, Qureshi A, Diaz MO, Nobori T, Gitelis S, et al. Chromosome 9 related aberrations and deletions of the CDKNs and MTS2 putative tumor suppressor genes in human chondrosarcomas. Cancer Lett 1996;105:91–103. 18. Castresana JS, Barrios C, Gomez L, Kreicbergs A. Amplification of the c-myc protooncogene in human chondrosarcoma. Diagn Mol Pathol 1992;1:235– 8. 19. Dobashi Y, Sugimura H, Sato A, Hirabayashi T, Kanda H, Kitagawa T, et al. Possible association of p53 overexpression and mutation with high-grade chondrosarcoma. Diagn Mol Pathol 1993;2:257– 63. 20. Hasegawa T, Seki K, Yang P, Hirose T, Hizawa K, Wada T, et al. Differentiation and proliferative activity in benign and malignant cartilage tumors of bone. Hum Pathol 1995;26: 838 – 45. 21. Nawa G, Ueda T, Mori S, Yoshikawa H, Fukuda H, Ishiguro S, et al. Prognostic significance of Ki67 (MIB1) proliferation index and p53 overexpression in chondrosarcomas. Int J Cancer 1996;69:86 –91. 22. Simms WW, Ordonez NG, Johnston D, Ayala AG, Czerniak B. p53 expression in dedifferentiated chondrosarcoma. Cancer 1995;76:223–7. 23. Yamaguchi T, Toguchida J, Wadayama B, Kanoe H, Nakayama T, Ishizaki K, et al. Loss of heterozygosity and tumor suppressor gene mutations in chondrosarcomas. Anticancer Res 1996;16:2009 –15. 24. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 1994;54: 4855–78. 25. Schneider-Stock R, Radig K, Oda Y, Mellin W, Rys J, Niezabitowski A, et al. p53 gene mutations in soft-tissue sarcomas: correlations with p53 immunohistochemistry and DNA ploidy. J Cancer Res Clin Oncol 1997;123:211– 8. 26. Toguchida J, Yamaguchi T, Ritchie B, Beauchamp RL, Dayton SH, Herrera GE, et al. Mutation spectrum of the p53 gene in bone and soft tissue sarcomas. Cancer Res 1992;52: 6194 –9. 27. Allred DC, Clard GM, Elledge R. Association of p53 protein expression with tumor cell proliferation rate and clinical outcome in node-negative breast cancer. J Natl Cancer Inst 1993;85:200 – 6. 28. Calzolari A, Chiarelli I, Bianchi S, Messerini L, Gallo O, Porfirio B, et al. Immunohistochemical vs. molecular biology methods: complementary techniques for effective screening of p53 alterations in head and neck cancer. Am J Clin Pathol 1997;107:7–11. 29. Lang FF, Miller DC, Pisharody S, Koslow M, Newcomb EW. High frequency of p53 protein accumulation without p53 gene mutations in human juvenile pilocytic, low grade and anaplastic astrocytomas. Oncogene 1994;9:949 –54. 30. Leahy DT, Salman R, Mulcahy H, Sheahan K, O’Donoghue DP, Parfrey NA. Prognostic significance of p53 abnormalities in colorectal carcinoma detected by PCR-SSCP and immunohistochemical analysis. J Pathol 1996;180:364 –70. 31. Mitsudomi T, Oyama T, Nishida K, Ogami A, Osaki T, Nakanishi R, et al. p53 nuclear immunostaining and gene mutations in non-small-cell lung cancer and their effects on patient survival. Ann Oncol 1995;6:S9 –13. 32. Zheng W, Sung J, Cao P, Zhang ZF, Cai R, Godwin TA, et al. Early occurrence and prognostic significance of p53 alteration in primary carcinoma of the fallopian tube. Gynecol Oncol 1997;64:38 – 48. 33. Ueda Y, Dockhorn-Dworniczak B, Blasius S, Mellin W, Wuisman P, Bocker W, et al. Analysis of mutant p53 protein in osteosarcomas and other malignant and benign lesions of bone. J Cancer Res Clin Oncol 1993;119:172– 8.