ALLELIC LOSS ON CHROMOSOMES 8 AND 9 CORRELATES WITH CLINICAL OUTCOME IN LOCALLY ADVANCED CLEAR CELL CARCINOMA OF THE KIDNEY

Transcription

1 /02/ /0 THE JOURNAL OF UROLOGY Vol. 167, , March 2002 Copyright 2002 by AMERICAN UROLOGICAL ASSOCIATION, INC. Printed in U.S.A. ALLELIC LOSS ON CHROMOSOMES 8 AND 9 CORRELATES WITH CLINICAL OUTCOME IN LOCALLY ADVANCED CLEAR CELL CARCINOMA OF THE KIDNEY JOSEPH C. PRESTI, JR., MONICA WILHELM, VICTOR REUTER, PAUL RUSSO, ROBERT MOTZER AND FREDERIC WALDMAN From the Department of Urology, Stanford University School of Medicine, Stanford and Division of Molecular Cytometry, Department of Laboratory Medicine, University of California, San Francisco, California, and Departments of Urology, Medicine and Pathology, Memorial Sloan-Kettering Cancer Center, New York, New York ABSTRACT Purpose: We correlated allelic loss on chromosomes 8p, 9p and 14q with clinical outcome in locally advanced conventional renal cell carcinoma. Materials and Methods: We analyzed radical nephrectomy specimens from 72 stage P3N0 conventional renal cell carcinomas by microsatellite loss of heterozygosity (LOH) analyses directed at chromosomes 3p, 8p, 9p and 14q (2 primers per chromosome). All patients were treated with surgery only. LOH results were correlated with disease-free survival using the Kaplan-Meier method. Results: We detected LOH on chromosome 3p in 60 of 64 (94%), on chromosome 8p in 19 of 59 (32%), on 9p in 21 of 67 (31%) and on 14q in 18 of 70 (26%) informative cases. Of the 72 patients 24 (33%) had recurrence during followup. On univariate analysis patients with tumors demonstrating LOH on chromosomes 8p and 9q were at high risk for recurrence (p 0.01). The correlation of recurrence with LOH approached significance for the individual chromosomes 8p and 9p (p 0.10 and 0.14, respectively). No correlation was observed of LOH on chromosome 14q with recurrence (p 0.42). The correlation of tumor grade with risk of relapse also approached significance (p 0.12). On multivariate analysis LOH on chromosome 8p was a more powerful predictor of recurrence than tumor grade. When combinations of LOH on chromosomes were tested, LOH on chromosomes 8p and 9p was the most powerful predictor of recurrence (p 0.006). Conclusions: LOH on chromosomes 8p or 9p may provide prognostic significance in patients with locally advanced renal cell carcinoma. KEY WORDS: kidney; carcinoma, renal cell; chromosomes, human, pair 8; chromosomes, human, pair 9; loss of heterozygosity Renal cell carcinoma accounted for about 2.4% of all new cancer cases diagnosed as well as 2.2% of all cancer deaths in the United States. 1 Recently the classification of renal cell carcinoma has been based on underlying genetic alterations. 2, 3 In this classification scheme conventional clear cell renal cell carcinoma, which accounts for approximately 70% to 80% of all renal neoplasms, is characterized by the loss of genetic material from the short arm of chromosome 3 (3p). This alteration, which most likely results in loss of function of the von Hippel-Lindau gene, is considered to be one of the earliest genetic events in the development of conventional renal cell carcinoma. In patients with renal cell carcinoma treated with surgery only survival correlates best with tumor stage at nephrectomy. 4 Patients with locally advanced stage P3N0 disease without evidence of metastases have 5-year disease specific survival of approximately 50% to 60%. Unfortunately stratification of these patients after nephrectomy into those at high and low risk for relapse is difficult. The usefulness of identifying patients destined to have relapse is optimal for adjuvant therapy trials. Underlying genetic alterations are thought to be responsible for the development and progression of renal cell carcinoma. Inactivation of tumor suppressor genes has a large role in tumorigenesis and typically occurs when the mutation Accepted for publication September 21, Supported in part by a Veterans Affairs Merit Review Grant of 1 allele becomes unmasked by the gross chromosomal deletion of the remaining wild-type allele. The latter event may be detected by microsatellite loss of heterozygosity (LOH) analysis. Several reports of renal cell carcinoma describe LOH involving additional chromosomal arms, including chromosomes 5, 6, 9, 10, 11, 13, 17, 18 and Generally additional deletions are associated with higher grade and stage tumors, leading to the hypothesis that these events result in tumor progression. However, these studies included small patient sets, precluding proper statistical analysis. We tested whether LOH on chromosomes 8p, 9p and 14q is associated with disease-free survival in a group of patients with well characterized, locally advanced renal cell carcinoma. MATERIALS AND METHODS Patient population. This retrospective study used an archival collection of specimens from a well characterized tissue bank at Memorial Sloan-Kettering Cancer Center. The study population consisted of 72 patients treated with radical nephrectomy only for stage pt3an0m0 (tumor invading perinephric tissues but not beyond Gerota s fascia) clear cell renal cell carcinoma. All samples were characterized by a single pathologist (V. R.) DNA extraction. Tumor and normal DNA were extracted from microdissected, paraffin embedded blocks. Tumor, ex-

2 ALLELIC LOSS AND CLEAR CELL CANCER OF KIDNEY 1465 cluding necrotic and normal tissue, was identified on a hematoxylin and eosin stained slide and manually microdissected from an adjacent 5 m. methyl green stained section. Each 1 mm. 2 microdissected tissue was digested in 15 l. 0.4 mg./ml. proteinase k-digestion buffer in phosphate buffered saline for 3 days at 55C. At the end of incubation the PK was heat inactivated and the DNA concentration was determined by quantitative real-time polymerase chain reaction (PCR) with an actin control primer-probe pair. DNA (0.5 ng.) was then subjected to PCR detection of allelic loss. LOH analysis. Microsatellite LOH studies directed at chromosomal regions commonly lost in renal cell carcinoma were performed using 2 primers from each chromosome arm, including D3S2405 (3p25), D3S4012 (3p24.2), D8S261 (8p22), ggaa20c10 (8p21), D9S925 (9p22), D9S921 (9p22), gata65g10 (14q31) and gata136b01 (14q32). Table 1 lists the primer sequences. Forward primers were labeled with 6-FAM, HEX or TET phosphoramidites (Applied Biosystems, Foster, California). PCR was performed in a DNA Thermal Cycler (Perkin-Elmer Cetus, Norwalk, Connecticut). The 25 l. reactions containing 0.5 ng. DNA, 10 pmol. of each primer, 2.5 mm. MgCl 2 and 0.6 units Taq-Gold (Perkin-Elmer Cetus, Norwalk, Connecticut) were topped with mineral oil. PCR consisted of 12 minutes of initial denaturation at 95C, followed by 34 cycles at 95C for 45 seconds, 55C for 45 seconds and 72C for 1 minute with final extension of 7 minutes at 72C after the last cycle. PCR products were electrophoresed on 2% agarose gels to confirm successful amplification before being applied onto 6% acrylamide gels containing 7 M. urea. Allelic imbalance was detected using the Genescan 672 software program (Genescan USA, Inc., New Orleans, Louisiana). Peak areas of the alleles in informative cases were used for calculating the allele ratio in normal and corresponding tumor specimens. The allelic imbalance was determined as the allele ratio of the normal sample divided by the allele ratio of its corresponding tumor. Thresholds for allelic imbalance were determined based on the variability of the ratio measurements in a series of normal archival samples using the same reaction conditions. Input DNA 0.5 ng. and above yielded a coefficient of variation of 10%. LOH was scored when the allelic imbalance was less than 0.5 and retention of heterozygosity was scored when the allelic imbalance was greater than 0.7. For samples with an allelic imbalance of between 0.5 and 0.7 the reaction was repeated. After the second run retention was scored for an allelic imbalance of greater than 0.7 and LOH for an allelic imbalance of less than 0.7. LOH on each chromosome was correlated with recurrence using Kaplan-Meier disease-free survival analysis and the log rank test. For multivariate analyses Cox proportional hazard models were constructed. RESULTS We noted LOH on chromosome 3p in 60 of 64 (94%), on 8p in 19 of 59 (32%), on 9p in 21 of 67 (31%) and on 14q in 18 of 70 (26%) informative cases. Table 2 lists all LOH and clinical data on all 72 specimens. Figure 1 shows representative tracings demonstrating LOH on chromosomes 3p and 8p. Of the 72 patients 24 (33%) had recurrence during followup. Figure 2 shows representative Kaplan-Meier diseasefree survival curves. The most powerful predictor of recurrence was seen in tumors with LOH on chromosomes 8p and 9p (p 0.01, fig. 2, A). The correlation of recurrence with LOH on chromosomes 8p and 9p only approached significance (p 0.1 and 0.14, respectively). No correlation was observed of LOH on chromosome 14q with recurrence (p 0.42). The ability of tumor grade to predict recurrence also showed a trend toward predicting recurrence (p 0.13, fig. 2, B). Cox proportional hazard models were constructed (table 3). LOH on chromosome 8p was the single most powerful predictor of recurrence (p 0.03). When combinations of LOH on various chromosomal arms were assessed, patients with tumors that showed LOH on chromosomes 8p and 9p were at greatest risk for recurrence (p 0.006). DISCUSSION The prognosis of patients with conventional renal cell carcinoma treated with radical nephrectomy depends on tumor stage at nephrectomy. Patients with stage P3N0M0 disease at nephrectomy are particularly important from a clinical standpoint because approximately half are destined to have recurrence. Current pathological parameters, including tumor grade and size, fail to predict which of these patients with stage P3N0 cancer are at risk for failure. Previous studies of the prognostic usefulness of genetic alterations in renal cell carcinoma have suffered from small sample sizes or have not specifically addressed on multivariate analysis the possible unique contribution of genetic alterations over standard pathological tumor grade and stage. A small cytogenetic study of 50 patients that included all tumor types as well as all tumor grades and stages showed that patients with a tumor with 5 or fewer alterations had a better prognosis than those with greater than 5 alterations. 10 This study indicated that tumor grade was the most powerful predictor of relapse in this population. A larger study of 148 patients that included a mixture of cell types as well as all tumor grades and stages showed that allelic imbalance on chromosome 14q correlated with tumor grade and stage. 11 Others evaluated 105 clear cell carcinomas of the kidney and noted that LOH on chromosomes 8p, 9p and 14q correlated with tumor grade and stage. 12 However, they included tumors of all grades and stages, and while the correlation was significant, it was rather weak (r to 0.419). 12 No analysis of clinical outcome was performed. Other investigators have observed that allelic loss of chromosome 9p may be involved with the progression of papillary renal cell carcinoma. 13 In a small subset analysis of 21 cases LOH on 9p was an independent predictor of survival. TABLE 1. Primer sequences Primer Locus Primer Pair Sequence Label D3S2405 3p25 Forward TAC CTT CCT TCC CCA CTC TTT C, HEX, none reverse CAA ACC AGA AGT GGG AGA GAG D3S4012 3p24.2 Forward CAT CTT TCT TTT CCT GTT CCC, HEX, none reverse CAT TTA CTG ATA GTA ACT AGG CA D8S261 8p22 Forward TGC CAC TGT CTT GAA AAT CC, reverse TAT GGC CCA GCA ATG TGT AT ggaa20c10 8p21 Forward AGC AAA CTT CAT CCA CTT GG, reverse CTG TCC TCT AAT GTA ACT TAC C D9S921 9p22 Forward GGT AAG TCA GCT ATA ATG ATC, FAM, none reverse CTC TTT CAT GTT GGC TCC TGT D9S925 9p22 Forward TGT GAG CCA AGG CCT TAT AG, FAM, none reverse TTA GTC TGG GTT CTC CAA AG Gata65g10 14q24.3/q31 Forward CTA TCT CCC CCT TTT TCC CT, reverse TTT GGA GAG GTG AGT GTA TTA GG Gata136b01 14q32 Forward CCT GGG CGA CAG TAA TGG, reverse GGG AGA GGC CCT GTA TTA GT

3 1466 ALLELIC LOSS AND CLEAR CELL CANCER OF KIDNEY Tumor No. TABLE 2. LOH and clinical results Chromosome (No. microsatellites) 3p 8p 9p 14q Fuhrman Grade Disease-Free Survival (days) 1 LOH LOH LOH Retention Recurrence 2 LOH Retention Retention Retention 2 2,220 Recurrence 3 LOH Retention LOH LOH 2 4,095 Disease-free 4 Retention Retention Retention Retention Recurrence 5 LOH Noninformative LOH Retention 2 4,099 Disease-free 6 Retention Retention Retention LOH Recurrence 7 LOH Retention LOH Retention 2 1,479 Recurrence 8 LOH LOH LOH LOH 3 81 Recurrence 9 LOH Retention Retention Retention 1 3,834 Disease-free 10 LOH Retention Retention Retention 2 3,639 Disease-free 11 LOH LOH Noninformative LOH 2 2,301 Recurrence 12 LOH Retention Noninformative Retention Recurrence 13 LOH Retention Retention Retention 3 3,067 Disease-free 14 Noninformative Retention Retention LOH 2 2,890 Disease-free 15 LOH Noninformative LOH Retention 3 1,252 Recurrence 16 Retention Retention Retention Retention 2 3,221 Disease-free 17 LOH Noninformative LOH Retention 3 3,008 Disease-free 18 LOH Retention Retention Retention Recurrence 19 LOH Retention Retention Noninformative 2 2,574 Disease-free 20 LOH Retention Retention Retention Disease-free 21 LOH Retention Retention Retention 1 2,633 Disease-free 22 LOH LOH Retention Retention 2 2,264 Disease-free 23 LOH Retention LOH Retention Recurrence 24 LOH Noninformative LOH LOH 2 2,018 Disease-free 25 LOH Retention LOH Retention 2 1,521 Recurrence 26 LOH Retention Retention Retention 2 1,134 Recurrence 27 LOH LOH Retention LOH 3 1,935 Disease-free 28 Noninformative LOH Retention Retention 3 1,830 Disease-free 29 LOH Retention Noninformative Retention Recurrence 30 LOH Retention Retention Noninformative 2 1,904 Disease-free 31 LOH Noninformative Retention Retention 3 30 Disease-free 32 LOH Retention LOH Retention 2 1,842 Disease-free 33 LOH Noninformative Retention Retention 4 1,772 Disease-free 34 Noninformative Retention LOH Retention 2 1,721 Disease-free 35 Retention Retention Retention Retention 2 1,561 Disease-free 36 LOH Retention Retention LOH Recurrence 37 Noninformative Retention LOH LOH 2 1,565 Disease-free 38 LOH Noninformative Retention Retention 3 1,730 Disease-free 39 LOH Retention Retention LOH 4 1,603 Disease-free 40 LOH Noninformative Retention LOH 2 1,576 Disease-free 41 LOH LOH LOH Retention Recurrence 42 LOH Noninformative Retention LOH 2 1,479 Disease-free 43 LOH Noninformative Retention Retention 2 1,766 Disease-free 44 LOH LOH Retention Retention 2 60 Disease-free 45 LOH LOH Retention Retention Disease-free 46 LOH Retention Noninformative LOH Recurrence 47 LOH Retention Retention Retention 2 1,083 Recurrence 48 LOH Retention Retention LOH Recurrence 49 LOH Noninformative Retention Retention 2 1,235 Disease-free 50 LOH Noninformative Retention Retention 2 1,489 Recurrence 51 Noninformative Retention Retention Retention 2 1,338 Disease-free 52 Noninformative LOH Retention LOH Recurrence 53 LOH LOH Retention Retention 3 1,295 Recurrence 54 LOH Retention Retention Retention 2 1,129 Disease-free 55 LOH Retention Retention LOH 3 1,073 Disease-free 56 LOH Retention LOH Retention Disease-free 57 LOH LOH Retention Retention Disease-free 58 LOH LOH LOH Retention 2 62 Recurrence 59 LOH Retention Retention Retention 4 1,029 Disease-free 60 LOH LOH Noninformative Retention Disease-free 61 LOH LOH LOH Retention Disease-free 62 LOH LOH Retention Retention Disease-free 63 LOH Retention LOH LOH Disease-free 64 LOH LOH Retention Retention Disease-free 65 LOH Retention Retention Retention Disease-free 66 Noninformative Retention LOH Retention Disease-free 67 LOH Retention LOH Retention Disease-free 68 LOH Retention Retention Retention Disease-free 69 LOH LOH Retention Retention Disease-free 70 LOH Noninformative Retention Retention Disease-free 71 LOH LOH LOH LOH 4 20 Recurrence 72 Noninformative Retention Retention Retention Disease-free Status In an earlier series of 41 patients with stage P3N0 renal cell carcinoma we reported that the number of genetic events detected by chromosome based comparative genomic hybridization provided some insight in this subgroup since patients with a loss of genetic material from 3 or more chromosomal arms had a higher rate of recurrence. 14 With respect to specific chromosomal alterations we noted a tendency toward recurrence and loss of chromosome 9p (p 0.04). We wished to study better this patient population with stage P3N0 in the current series and, thus, we used paraffin embedded archival tissue to provide a larger sample size. No current patient tumors were used in our previous series. Our study shows that patients who had tumors with LOH on chromosomes 8p and 9p were at high risk for failure

4 ALLELIC LOSS AND CLEAR CELL CANCER OF KIDNEY 1467 FIG. 2. Kaplan-Meier disease-free survival curves. A, chromosome 8p or 9p retained (upper curve) with LOH on chromosomes 8p and 9p (lower curve) (log rank test p 0.01). B, low (upper curve) and high (lower curve) grade tumors (log rank test p 0.13). FIG. 1. Normal tissue shows heterozygosity (top) and tumor shows LOH for larger allele on chromosome with relative area reduction of peak at right satisfying criteria for LOH. A, chromosome 3p (D3S2405) with 68% reduction. B, chromosome 8p (ggaa20c10) with 76% reduction. (relative risk 5.21). This genetic information was superior to tumor grade for predicting recurrence. This information is important because if such patients are accurately identified, imaging during followup may be more intensive and as discussed such patients may be appropriate candidates for adjuvant clinical trials. The importance of accurately identifying patients who may harbor microscopic metastatic disease after radical nephrectomy and before the development of overt metastases has been emphasized by recent reports of the response to systemic therapy after radical nephrectomy in the face of metastatic disease. Two large randomized trials evaluating the need for nephrectomy in the setting of overt metastatic disease before the initiation of systemic immunotherapy showed a survival advantage in patients undergoing nephrectomy. 15, 16 This finding suggests that tumor burden is an important factor for predicting the response to systemic immunotherapy. If one were to extrapolate these results into the adjuvant setting, perhaps patients with stage P3N0M0 renal cell carcinoma who undergo radical nephrectomy and are found to have genetic changes predicting a high likelihood of microscopic metastatic disease may receive additional benefit from systemic immunotherapy since by definition their residual tumor burden is microscopic. Alternative applications may include the identification of patients at high risk as candidates for clinical trials involving other novel therapeutics. In the current study we reliably identified LOH on chromosome 3p in 94% of the samples. This analysis served as our positive internal control since it confirmed the high frequency of 3p LOH reported in other studies. Thus, we are confident in our ability to detect LOH on other chromosomes TABLE 3. Cox proportional hazard model using individual and combinations of chromosomal arms Variable Relative Risk (95% CI) p Value Individual: LOH chromosome 3p 1.56 ( ) 0.26 LOH chromosome 8p 2.00 ( ) 0.03 LOH chromosome 9p 1.22 ( ) 0.60 LOH chromosome 14q 1.59 ( ) 0.26 Grade 1.60 ( ) 0.27 Combinations: LOH chromosome 8p 9p 5.21 ( ) LOH chromosome 8p or 9p 1.22 ( ) 0.67 Grade 1.78 ( ) 0.17 in the current series. Another strength of the study is the review by a single pathologist with extensive experience in renal pathology. A limitation was the lack of absolute uniformity in patient followup. However, most patients were followed regularly with imaging at 4 to 6-month intervals. Our sample size and, thus, study power was also limited, resulting from the fact that not all tumors were informative at each chromosomal arm studied. The refinement of newer technology, including array based DNA comparative genomic hybridization, which enables a high throughput evaluation of all loci simultaneously, may help to correct this problem. CONCLUSIONS LOH on chromosomes 8p and/or 9p provides additional prognostic information in patients with locally advanced renal cell carcinoma. Identifying patients in whom tumors demonstrate these abnormalities is important because this subset may warrant adjuvant therapy. REFERENCES 1. Greenlee, R. T., Hill-Harmon, M. B., Murray, T. et al: Cancer statistics, CA Cancer J Clin, 51: 15, Kovacs, G., Akhtar, M., Beckwith, B. J. et al: The Heidelberg

5 1468 ALLELIC LOSS AND CLEAR CELL CANCER OF KIDNEY classification of renal tumors. J Pathol, 183: 131, Storkel, S., Eble, J. N., Adlakha, K. et al: Classification of renal cell carcinoma: Workgroup No. 1. Union Internationale contre le Cancer (UICC) and American Joint Committee on Cancer (AJCC). Cancer, 80: 987, Skinner, D. G., Colvin, R. B., Vermillion, C. D. et al: Diagnosis and management of renal cell carcinoma: a clinical and pathological study of 309 cases. Cancer, 28: 1165, Anglard, P., Tory, K., Brauch, H. et al: Molecular analysis of genetic changes in the origin and development of renal cell carcinoma. Cancer Res, 51: 1071, Presti, J. C., Jr., Rao, P. H., Chen, Q. et al: Histopathological, cytogenetic, and molecular characterization of renal cortical tumors. Cancer Res, 51: 1544, Morita, R., Ishikawa, J., Tsutsumi, M. et al: Allelotype of renal cell carcinoma. Cancer Res, 51: 820, Bergerheim, U., Nordenskjold, M. and Collins, P.: Deletion mapping in human renal cell carcinoma. Cancer Res, 49: 1390, Cairns, P., Tokino, K., Eby, Y. et al: Localization of tumor suppressor loci on chromosome 9 in primary human renal cell carcinomas. Cancer Res, 55: 224, Elfving, P., Mandahl, N., Lundgren, R. et al: Prognostic implications of cytogenetic findings in kidney cancer. Br J Urol, 80: 698, Beroud, C., Fournet, J. C., Jeanpierre, C. et al: Correlations of allelic imbalance of chromosome 14 with adverse prognostic parameters in 148 renal cell carcinomas. Genes Chrom Cancer, 17: 215, Schullerus, D., Herbers, J., Chudek, J. et al: Loss of heterozygosity at chromosomes 8p, 9p, and 14q is associated with stage and grade of non-papillary renal cell carcinomas. J Pathol, 183: 151, Schraml, P., Muller, D., Bednar, R. et al: Allelic loss at the D9S171 locus on chromosome 9p13 is associated with progression of papillary renal cell carcinoma. J Pathol, 190: 457, Moch, H., Presti, J. C., Jr., Sauter, G. et al: Genetic aberrations detected by comparative genomic hybridization are associated with clinical outcome in renal cell carcinoma. Cancer Res, 56: 27, Flanigan, R. C., Blumenstein, B. A., Salmon, S. et al: Cytoreduction nephrectomy in metastatic renal cancer: the results of Southwest Oncology Group Trial J Urol, part 2, 163: 154, abstract 685, Mickisch, G. H., Garin, A., Madej, G. M. et al: Tumor nephrectomy plus interferon is superior to interferon alone in metastatic renal cell carcinoma. European Organization for the Research and Treatment of Cancer Genitourinary Group. J Urol, part 2, 163: 176, abstract 778, 2000