Determination of TP53 Mutation Is More Relevant Than Microsatellite Instability Status for the Prediction of Disease-Free Survival in Adjuva
http://www.100md.com
《临床肿瘤学》
the Department of Medical Genetics, University of Groningen
Comprehensive Cancer Center North Netherlands
the Departments of Pathology, Medical Oncology, and Surgery, University Hospital Groningen, the Netherlands
ABSTRACT
PURPOSE: Microsatellite instability (MSI), TP53 mutation, and KRAS mutation status have been reported as prognostic factors in colon cancer. Most studies, however, have included heterogeneous groups of patients with respect to cancer stage. We determined the prognostic relevance of high-frequency MSI (MSI-H), TP53 mutations, and KRAS mutations in a well-defined group of patients with stage III colon cancer (N = 391), randomly assigned for adjuvant treatment with fluorouracil-based chemotherapy.
METHODS: Three hundred ninety-one tumor specimens were available. MSI was determined in 273 specimens, and mutation analyses of TP53 and KRAS were performed in 220 and 205 specimens, respectively.
RESULTS: In a univariate analysis, MSI-H (44 of 273; 16%) was associated with a longer disease-free survival (DFS; P = .038), but in a multivariate model adjusting for nodal involvement, histology, invasion, and grade of tumor, the association of MSI status with DFS did no longer reach statistical significance, though the risk estimate for microsatellite stability versus MSI-H tumors did not change much. Mutant TP53, found in 116 (53%) of 220 tumors, was associated with a shorter DFS, both in univariate (P = .009) and multivariate analyses (P = .018), whereas KRAS mutations (58 of 205; 28%) did not show any prognostic significance.
CONCLUSION: Both mutant TP53 and MSI-H seem to be prognostic indicators for disease-free survival, but only TP53 retains statistical significance after adjusting for clinical heterogeneity. Thus, in adjuvantly treated patients with stage III colon cancer, presence or absence of a TP53 mutation should be considered as a better predictor for DFS than MSI status.
INTRODUCTION
Although the presently used TNM classification for colon cancer improves in the predictive value of its precursors, its value in choosing adjuvant treatment after surgery, which may be particularly important for patients with stage III colon cancer, is still limited. Therefore, many studies have been undertaken to explore molecular prognostic and predictive markers in this group of patients. Several well-known genetic alterations, associated with different forms of genomic instability,1-3 have been tested for their use in subclassification models using molecular criteria. A chromosomal instability (CIN) phenotype is found in approximately 85% of sporadic colon cancers and is characterized by aneuploidy, multiple chromosomal rearrangements, and accumulation of somatic mutations in oncogenes such as KRAS and tumor-suppressor-genes such as TP53 and adenomatous polyposis coli (APC).3,4 A high-frequency microsatellite instability (MSI-H or MIN) phenotype is associated with small insertions and deletions mainly in repetitive sequences (microsatellites).1,2 This instability is caused by defects in the mismatch repair system, which is involved in repairing errors arising during DNA replication. The MSI-H phenotype is observed in approximately 15% of the sporadic colon cancers.1,2,5
An inverse correlation with respect to clinicopathologic presentation has been found for TP53 and/or KRAS mutations, and for MSI-H conversely.6-9 MSI-H tumors are predominantly characterized by diploidy, more right-sided location, mucinous histology, and poor differentiation,2,10-13 whereas CIN tumors are usually characterized by left-sided location, aneuploidy, nonmucinous histology, predominantly moderate differentiation, and less tumor-infiltrating lymphocytes.13,14
The MSI-H tumor phenotype has been associated with a better prognosis than the microsatellite stability (MSS) tumor phenotype2,10,15-24; presence of a TP53 mutation in the tumor has often been associated with shorter survival of the patient compared with the presence of wild-type TP53.25-33
One of the problems in the interpretation of these data is that many studies are based on series of tumors not belonging to the same stage or include rectal tumors for which a different treatment strategy is used.34 Moreover, numbers of patients or tumors screened are often small. Another problem is that different methods of tissue isolation and mutation or MSI screening are used, which may hamper a direct comparison between studies. Our aim was to determine the prognostic value of MSI-H and of mutations in TP53 and KRAS in a well-defined homogeneous group of patients with stage III colon cancer, in combination with conventional clinicopathologic characteristics, such as the number of tumor-positive lymph nodes, and histologic criteria.
PATIENTS AND METHODS
Patients' Material
Formalin-fixed paraffin-embedded (FFPE) primary tumor tissue from 391 patients with stage III colon cancer included in a nation-wide randomized colon cancer (CKVO 90-11) trial (N = 500) was available for this study. In the remaining 109 patients no primary tumor specimens or insufficient material were available for DNA analyses. All patients had histologically negative (R0) resection margins and were treated adjuvantly with fluorouracil (FU) -based chemotherapy (FU/levamisole or FU/levamisole/leucovorin) as described earlier.35,36 The patients of a subset of 391 cases suitable for DNA extraction were equally distributed between both groups (FU/levamisole or FU/levamisole/leucovorin). Moreover, no differences in disease-free or overall survival were seen between both groups.
Clinical and Tumor Characteristics of the Patients
For all patients, clinical and reviewed tumor characteristics were derived from the clinical database maintained at the Comprehensive Cancer Center, North Netherlands.35 The mean patient age at the time of random assignment was 58.7 years (range, 26.2 to 75.8 years). Two hundred thirteen patients were men (54.5%); 182 tumors were right-sided (46.5%; ie, proximal of the splenic flexure). The tumors were defined as mucinous if more than 50% had a mucinous aspect. Furthermore, the tumors were histologically classified according to the WHO standard.37 With regard to histological grade, 17.5% of the tumors were well differentiated, 57.3% were moderately differentiated, and 25.2% were poorly differentiated. According to the TNM staging system, 13.6% of the tumors belonged to the T1 and T2 stages, and 86.4%, to stages T3 or T4. Furthermore, 74.2% of the resection specimens belonged to stage N1, while 101 (25.8%) were stage N2.38
The 5-year disease-free survival (DFS) was 55.5% for all participants of the CKVO 90-11 trial, which was equal to the 56.3% among the analyzed subgroup of 391 patients from whom FFPE primary tumor specimens were available. In the smallest subgroup of our study, the group of patients for whom we could determine both the TP53 mutation status and the MSI status, the 5-year DFS was 56.4% (n = 194). Based on survival patterns, there was no evidence for a selection bias for these subgroups (log-rank test P = .724).
DNA Extraction
After interpretation of hematoxylin and eosin staining from all FFPE tissues (acquired before chemotherapy had started), representative tumor tissue and corresponding normal tissue were selected by an experienced pathologist and were cut in 20-μm slices for DNA isolation. The QIAamp DNA mini kit (Qiagen, Westburg, Leusden, the Netherlands) was used for DNA extraction according to protocol, with some adjustments. These included skipping the xylene treatment, in the final elution step, twice putting the eluate (100 μL of preheated [70°C] elution buffer) on the column, and incubating it at 70°C for 5 to 10 minutes before spin down.
MSI Analysis
MSI analysis was performed on an ABI Prism 377 DNA sequencer after amplifying DNA from tumor and matching normal tissue, using fluorescent-labeled primers from the MSI multiplex kit (Promega, Madison, WI). The MSI multiplex kit includes primers for coamplification of nine microsatellite markers, including four mononucleotide repeat markers (BAT-25, BAT-26, BAT-40, and MONO27) and five tetranucleotide repeat markers (D3S2432, D7S1808, D7S3046, D7S3070, and D10S1426). Alternatively, the five consensus markers described by Boland et al39 were used, including two mononucleotide (BAT-25 and BAT-26) and three dinucleotide repeat markers (D2S123, D5S346, and D17S250). This was done as described previously40or by the use of the HNPCC MSI kit (Roche, Mannheim, Germany).
When using the Roche kit or the Boland consensus markers, those samples were included in which all five markers could be interpreted; four markers could be interpreted and all proved negative; or two (interpretable) markers proved positive. The Promega kit samples were included if five (or more) of nine markers were interpretable, all four mononucleotide markers were interpretable, or if three (interpretable) markers proved positive. These inclusion criteria were formulated because sometimes difficulties in interpretation or lack of results were met due to bad quality of DNA obtained from old FFPE specimens.
A tumor was classified as MSI-H when more than two of five or more than three of nine microsatellite markers showed instability. In all other cases, tumors were defined as MSS. All analyses were performed without previous knowledge of clinical outcome.
Mutation Analysis for KRAS and TP53
Mutation screening (of parts) of KRAS (exons 1 and 2) and TP53 (exons 4 to 8) was performed by denaturing gradient gel electrophoresis (DGGE) using the INGENYphorU system (Ingeny International, Goes, the Netherlands) and DNA sequencing using ABI Prism 377 DNA sequencer (Perkin Elmer Applied Biosystems, Foster City, CA) according to previously described protocols.41,42
Statistical Analysis
For comparison of the distribution of continuous variables, the analysis of variance test was used. The Fisher's exact test was used in case of 2 x 2 tables, otherwise the 2 test was used for comparison of categorical variables. DFS was used to assess the prognostic value of the various molecular markers, as it is the least biased indicator of treatment failure given the facts that disease-specific survival might be influenced by selective use of second-line treatments, and overall survival is biased due to deaths attributable to causes unrelated to colon cancer. DFS was defined as the period from the date of random assignment to the date of any documented relapse, to the date of death without documented relapse due to colon cancer, or to the date of last contact, whichever occurred first. Censoring occurred when a patient was alive without relapse at last contact or when the patient died of a cause other than colon cancer, not attributable to colon cancer. Time to event curves were estimated using the Kaplan-Meier method, and the survival distributions were compared with the log-rank test. Cox proportional hazard models were constructed to assess the simultaneous effects of various covariables on the survival experience in our cohort.43 Variables included in the model were tumor stage (invasion into the muscularis propria v beyond the muscularis propria), tumor grade, tumor cell type (mucinous v nonmucinous), and number of tumor-positive lymph nodes (< 4 v 4). Model fit was evaluated using graphical methods and the May-Hosmer goodness-of-fit test statistic.44 All reported P values are two-sided unless stated otherwise.
RESULTS
Molecular Parameters
MSI analysis. Based on the before-mentioned inclusion criteria, 273 of the 391 analyzed DNA specimens could be included. The Roche kit was applied to 37 tumors, of which 23 gave interpretable data, identifying 18 MSS and five MSI-H tumors. The Promega kit was applied to 223 tumors, of which 192 gave interpretable data, identifying 167 MSS and 25 MSI-H tumors. Furthermore, 67 tumors were analyzed by simplex PCR-analyses. Of these, 58 gave interpretable data, identifying 14 as MSI-H. In total, 44 (16%) of the 273 samples seemed to be instable (Table 1).
The difference in results between the methods used to determine the MSI status (Roche kit, 22%; Promega kit, 13%; simplex analysis, 24%) is not statistically significant (22 P = .097) and is most likely due to chance and not caused by the different repeat markers, since, if we determine the number of MSI-H cases by using only the shared mononucleotide repeats, which proved to be the most informative markers, no more than one tumor would change from MSS to MSI-H.
TP53 mutation analysis. Included in this analysis were 220 tumors screened for at least the seven amplicons covering the sequence coding for the conserved domains of TP53. A total of 123 mutations were identified in 116 (53%) of the 220 tumors. In Figure 1, the distribution over codons/exons/domains and frequency of TP53 mutations are presented. Multiple TP53 mutations were found in six tumors, of which five tumors had two, and one tumor had three TP53 mutations. Five frequently recurrent mutations occurred at codons 273, 248, 175, 282, and 245, with prevalences of 12, 12, six, seven, and seven times, respectively.
Three of the 220 tumors showed an obviously aberrant DGGE pattern, different from known polymorphisms, for which we were unable to produce a reliable sequence. The probability that one of these three would be a silent mutation is very low, since we observed only four silent mutations in a total of 120 mutations. We therefore decided to include the three in the group of causative mutations. In addition, we found 105 DGGE variants that we classified as polymorphisms, and four silent mutations: Lys120 (G360A), Thr125 (G375A), Cys135 (C405T), and Leu299 (C895T).
KRAS mutation analysis. In the 205 tumors that could be analyzed for both KRAS exons 1 and 2, 58 KRAS mutations were identified (28%). Except for one, in which we found a silent mutation in codon 60 next to a mutation in codon 61, none of the tumors had more than one KRAS mutation. Most mutations occurred in codon 12 (46 cases, 79%), followed by codon 61 (seven cases, 12%) and codon 13 (five cases, 9%). The Gly12Asp amino acid change was the most frequent mutation (n = 22), followed by Gly12Val (n = 12), Gly12Cys and Gln61His (both n = 6), Gly13Asp (n = 5), Gly12Ser (n = 4), Gly12Ala (n = 2) and Gln61Lys (n = 1). In addition to these well-known mutations, we identified four silent mutations (Leu6, Val8 [twice], and Gly60).
Combined molecular parameters MSI, TP53, and KRAS status. In nine (36%) of the 25 MSI-H tumors, a TP53 mutation was found, and 94 (56%) of 169 MSS tumors carried such a mutation (P = .067). KRAS mutations were found in four (16%) of 25 MSI-H tumors, compared with 42 of 126 MSS tumors (33%; P = .057).
For those tumors in which MSI as well as KRAS or TP53 mutation status could be determined, we found that in 10 (43%) of the 23 MSI-H tumors, either KRAS or TP53 were mutated, compared with 82 (79%) of 119 MSS tumors (P = .006). Both TP53 and KRAS mutations occurred in two (9%) of these 23 MSI-H tumors and in 19 (16%) of 119 MSS tumors.
Combined Molecular and Clinicopathologic Characteristics/Correlations
The results of the molecular analyses, in combination with the clinicopathologic characteristics, are summarized in Table 1. Compared with MSS tumors, tumors with an MSI-H phenotype were mostly right-sided (80%; P < .001) and more often showed a poor cellular differentiation. MSI-H tumors were more frequently of a mucinous histology, whereas the MSS tumors were predominantly nonmucinous. No differences were detected between patients with MSI-H and MSS tumor phenotype with respect to sex, number of tumor-positive lymph nodes (< 4 or 4), or tumor stage.
Tumors with mutant TP53 were more frequently of the nonmucinous cell type (71%), whereas tumors with wild-type TP53 had a more even distribution over nonmucinous and mucinous cell types. This association remained significant (P = .018) when cases with unspecified histological subtype were excluded. No significant associations were found with KRAS mutation status of the tumors (Table 1).
Survival in Relation to the Different Molecular Parameters
MSI. The 5-year DFS was 71.8% for patients with an MSI-H tumor and 49.7% for patients with an MSS tumor, patients with an MSS tumors had a worse DFS (P = .034; Fig 2). In our study, patients with a MSS tumor experienced a 1.85-fold increased risk of relapse (95% CI, 1.03 to 3.29) compared with patients with an MSI-H tumor. Table 2 presents the results of the multivariate analysis. Adjusted for tumor grade, histology, number of tumor-positive nodes, and depth of tumor invasion, the association of MSI status with DFS did, however, not reach statistical significance, although the risk estimate for MSS versus MSI-H tumors did hardly change (hazard ratio [HR] = 1.80; 95% CI, 0.97 to 3.32).
TP53. The 5-year DFS for patients with a tumor carrying a somatic TP53 mutation was 44.0%, compared with 62.7% for patients with a wild-type TP53 tumor. Patients having a tumor with a TP53 mutation had a significantly worse DFS than those with a wild-type TP53 tumor (P = .009; Fig 3). They experienced a 1.71-fold (95% CI, 1.14 to 2.56) increased risk of relapse. Adjusted for tumor grade, histology, number of tumor-positive nodes, and depth of tumor invasion, TP53 remained a significant independent predictor of DFS, with an HR of 1.71 (95% CI, 1.12 to 2.59) for patients who had a tumor with a TP53 mutation compared with patients with a wild-type TP53 tumor (Table 3).
KRAS. In a univariate analysis, KRAS status was not associated with DFS. Therefore, KRAS status was not included in the multivariate analyses.
Combined MSI and TP53 phenotype correlation with survival after adjusting for clinical heterogeneity. In a subgroup of patients, both MSI and TP53 status could be determined (n = 194). In a multivariate model comprising the degree of nodal involvement, tumor grade, histology, depth of invasion, MSI status, and TP53 mutation status, neither MSI nor TP53 were significantly associated with DFS (Table 4).
However, added to the model with nodal involvement, tumor grade, histology, and depth of invasion, TP53 status still reached borderline significance (likelihood-ratio test [21] 3.49; P = .062) while the MSI status explained only little variation in DFS (likelihood-ratio test [21] 1.41; P = .235). When adding MSI status to a model containing nodal involvement, tumor grade, histology, depth of tumor invasion, and TP53 status, the model-based log likelihood changed with 0.81 (likelihood-ratio test P = .368). Adding TP53 status to a model with the MSI status changed the log likelihood with 2.89 (likelihood-ratio test P = .089).
DISCUSSION
This study used data from 391 stage III colon cancer patients who participated in a prospective randomized trial studying two adjuvant regimens of FU-based chemotherapy. Genetic alterations in TP53 (exons 4 to 8) and KRAS (exons 1 and 2) were found in 53% and 28% of cases, respectively. MSI was exhibited by 16% of tumors. These frequencies are in agreement with other published data.1,2,5,32,45,46 Mutation analysis of TP53 was limited to exons 4 to 8 for most tumors, because an initial screening of the whole gene in 69 tumors revealed 39 mutations, with all but one located in exons 4 to 8. The International Agency for Research on Cancer TP53 mutation database also contains few mutations (2% overall or 5% colon carcinoma) outside exons 4 to 8.45
We first investigated whether associations existed between clinical parameters and the genetic alterations. The associations observed with MSI-H status, proximal location of the tumor, poor cellular differentiation, and mucinous tumor histology are in agreement with those found by other investigators.12,13,47,48 The observation that TP53 mutation was associated with an increased frequency of nonmucinous tumors is also in agreement with results from others.49,50
In univariate analysis, a better prognosis resulted for patients with MSI-H tumors than for patients with MSS tumors. Comparing our data with findings reported on similar groups of patients, our results are in agreement with results obtained by several others9,19,21,23,51-55 who included only stage III (and/or II) colon (and/or rectal) cancer patients, irrespective of adjuvant treatment. It should be noted, however, that if only the non–adjuvantly treated patients were selected, no association of MSI status and survival was found by Elsaleh et al,9,51,54 whereas Ribic et al55 recently reported that for patients who did not receive adjuvant chemotherapy, MSI-H status is a positive prognostic factor in univariate analysis. For adjuvantly treated patients, no difference was observed in DFS or overall survival of patients with MSI-H versus MSS tumors. This is in disagreement with our results. A partial explanation for the difference between the study of Ribic et al55 and our study may be that they included not only stage III but also stage II colon cancer patients. A subgroup analysis of the data of Ribic et al55 according to stage is missing. Carethers et al,56 conversely, found no prognostic significance for MSI status in combined adjuvantly treated and untreated, or selected adjuvantly treated or untreated subgroups of stage II/III colorectal cancer (CRC) patients.
Our study showed that having tumors with a TP53 mutation is associated with a shorter DFS (P = .018). This is in agreement with results for non–adjuvantly treated groups of stage III/II colon cancer patients studied by Pricolo et al27 and Tang et al,33 and from the stage III subgroup analysis of CRC patients in the study by Goh et al,31 but disagrees with results obtained by others51,57-59 who found no significant association between TP53 mutation and overall or disease-specific survival. Soong et al57 even reported a trend toward better survival. All disagreeing studies included both adjuvantly treated and non–adjuvantly treated patients. Only Elsaleh et al51 included merely stage III colon cancer. Others also included stage II colon tumors and rectal carcinomas that are clinically and genetically different cancers.34,60
Presence of a TP53 mutation is more common in CIN tumors than in MIN tumors.7-9 This suggests that tumors with a MIN phenotype are less aggressive than those with a CIN phenotype. TP53 mutations were found in both the MSS (presumed CIN) and the MSI-H groups. The difference between the mutation frequencies in these groups was not statistically significant, but certainly the estimated proportion of patients having a tumor with a TP53 mutation was higher in the MSS group than in the MSI-H group (56% v 36%).
We also did not find a definite negative association between MSI-H and KRAS as suggested by others.6,8 In our study, four (16%) of 25 MSI-H tumors had a KRAS mutation, compared with 42 (33%) of 126 of MSS tumors. Presence of a KRAS mutation in the tumor did not correlate with any clinicopathologic parameter, nor with DFS.
Our results do not exclude a coexistence of MIN and CIN developmental routes in the same tumor. This is in line with the findings of Goel et al,61 who found a significant degree of overlap between MIN and CIN characteristics, by performing loss of heterozygosity analyses on a number of chromosome arms (18q, 5q, 17p, 3p, and 2p), instead of TP53 analysis.
our univariate analyses, it could be concluded that an MSI-H phenotype acts as a positive prognostic factor and that TP53 mutation is an adverse prognostic factor for patients with stage III colon cancer. Adjusted for clinical heterogeneity, only a TP53 mutation retained its independent significant adverse prognostic effect. Although the risk estimate for MSI status in multivariate analysis did not change much compared to the univariate analysis, the MSI status did no longer reach statistical significance with DFS, which can be explained by the low occurrence of MSI-H tumors (approximately 15%) in sporadic CRCs.
In a multivariate model including both MSI and TP53 mutation status, the effect of TP53 appeared to be stronger than that of MSI. The power of that analysis, however, was too small to prove that effect at a 5% significance level. The likelihood-ratio test showed that TP53 mutation status was more powerful in predicting DFS than MSI status. Presence or absence of a TP53 mutation, and not MSI, therefore seems a better predictor for DFS in adjuvantly treated stage III colon cancer patients.
It has been suggested that more than 50% of the adjuvant therapies given to stage III patients may be unnecessary.62 This suggestion again underscores the need for predictive markers that will allow identification of patients who will or will not benefit from adjuvant FU treatment. Unfortunately, from our study, we cannot conclude that MSI and TP53 mutation status of the primary tumor can be used as such. Although (small) effects of both are seen, they cannot be used for the individual patient in a decision protocol concerning adjuvant chemotherapy.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank Jeff Bacher from the Promega Corporation for generously supplying the MSI multiplex system: prototype kits; Ton Tiebosch MD, PhD, for pathologic assistance; and Gerwin Jansen, Annet Kooistra, Ludolf Boven, Jan Osinga, and Edwin Verlind for technical assistance. We also thank the anonymous reviewers for their useful comments, improving the quality of this manuscript.
NOTES
Supported by a grant from Dutch Cancer Society, NKB number RUG 99-1962.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Ionov Y, Peinado MA, Malkhosyan S, et al: Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature 363:558-561, 1993
Thibodeau SN, Bren G, Schaid D: Microsatellite instability in cancer of the proximal colon. Science 260:816-819, 1993
Lengauer C, Kinzler KW, Vogelstein B: Genetic instability in colorectal cancers. Nature 386:623-627, 1997
Fearon ER, Vogelstein B: A genetic model for colorectal tumorigenesis. Cell 61:759-767, 1990
Aaltonen LA, Peltomaki P, Leach FS, et al: Clues to the pathogenesis of familial colorectal cancer. Science 260:812-816, 1993
Breivik J, Lothe RA, Meling GI, et al: Different genetic pathways to proximal and distal colorectal cancer influenced by sex-related factors. Int J Cancer 74:664-669, 1997
Salahshor S, Kressner U, Pahlman L, et al: Colorectal cancer with and without microsatellite instability involves different genes. Genes Chromosomes Cancer 26:247-252, 1999
Samowitz WS, Holden JA, Curtin, et al: Inverse relationship between microsatellite instability and K-ras and p53 gene alterations in colon cancer. Am J Pathol 158:1517-1524, 2001
Elsaleh H, Powell B, McCaul K, et al: p53 alteration and microsatellite instability have predictive value for survival benefit from chemotherapy in stage III colorectal carcinoma. Clin Cancer Res 7:1343-1349, 2001
Lothe RA, Peltomaki P, Meling GI, et al: Genomic instability in colorectal cancer: Relationship to clinicopathologicalvariables and family history. Cancer Res 53:5849-5852, 1993
Kim H, Jen J, Vogelstein B, et al: Clinical and pathological characteristics of sporadic colorectal carcinomas with DNA replication errors in microsatellite sequences. Am J Pathol 145:148-156, 1994
Jass JR, Do KA, Simms LA, et al: Morphology of sporadic colorectal cancer with DNA replication errors. Gut 42:673-679, 1998
Young J, Simms LA, Biden KG, et al: Features of colorectal cancers with high-level microsatellite instability occurring in familial and sporadic settings: Parallel pathways of tumorigenesis. Am J Pathol 159:2107-2116, 2001
Delattre O, Olschwang S, Law DJ, et al: Multiple genetic alterations in distal and proximal colorectal cancer. Lancet 2:353-356, 1989
Bubb VJ, Curtis LJ, Cunningham C, et al: Microsatellite instability and the role of hMSH2 in sporadic colorectal cancer. Oncogene 12:2641-2649, 1996
Lukish JR, Muro K, DeNobile J, et al: Prognostic significance of DNA replication errors in young patients with colorectal cancer. Ann Surg 227:51-56, 1998
Massa MJ, Iniesta P, Gonzalez-Quevedo R, et al: Differential prognosis of replication error phenotype and loss of heterozygosity in sporadic colorectal cancer. Eur J Cancer 35:1676-1682, 1999
Jernvall P, Makinen MJ, Karttunen TJ, et al: Microsatellite instability: Impact on cancer progression in proximal and distal colorectal cancers, Eur J Cancer 35:197-201, 1999
Halling KC, French AJ, McDonnell SK, et al: Microsatellite instability and 8p allelic imbalance in stage B2 and C colorectal cancers. J Natl Cancer Inst 91:1295-1303, 1999
Gryfe R, Kim H, Hsieh ET, et al: Tumor microsatellite instability and clinical outcome in young patients with colorectal cancer. N Engl J Med 342:69-77, 2000
Wright CM, Dent OF, Barker M, et al: Prognostic significance of extensive microsatellite instability in sporadic clinicopathological stage C colorectal cancer. Br J Surg 87:1197-1202, 2000
Choi SW, Lee KJ, Bae YA, et al: Genetic classification of colorectal cancer based on chromosomal loss and microsatellite instability predicts survival. Clin Cancer Res 8:2311-2322, 2002
Watanabe T, Wu TT, Catalano PJ, et al: Molecular predictors of survival after adjuvant chemotherapy for colon cancer. N Engl J Med 344:1196-1206, 2001
Lim SB, Jeong SY, Lee MR, et al: Prognostic significance of microsatellite instability in sporadic colorectal cancer. Int J Colorectal Dis 19:533-537, 2004
Goh HS, Yao J, Smith DR: p53 point mutation and survival in colorectal cancer patients. Cancer Res 55:5217-5221, 1995
Smith DR, Ji CY, Goh HS: Prognostic significance of p53 overexpression and mutation in colorectal adenocarcinomas. Br J Cancer 74:216-223, 1996
Pricolo VE, Finkelstein SD, Hansen K, et al: Mutated p53 gene is an independent adverse predictor of survival in colon carcinoma. Arch Surg 132:371-374, 1997
Borresen-Dale AL, Lothe RA, Meling GI, et al: TP53 and long-term prognosis in colorectal cancer: Mutations in the L3 zinc-binding domain predict poor survival. Clin Cancer Res 4:203-210, 1998
Goh HS, Elnatan J, Low CH, et al: p53 point mutation and survival in colorectal cancer patients: Effect of disease dissemination and tumor location. Int J Oncol 15:491-498, 1999
Kressner U, Inganas M, Byding S, et al: Prognostic value of p53 genetic changes in colorectal cancer. J Clin Oncol 17:593-599, 1999
Tortola S, Marcuello E, Gonzalez I, et al: p53 and K-ras gene mutations correlate with tumor aggressiveness but are not of routine prognostic value in colorectal cancer. J Clin Oncol 17:1375-1381, 1999
Samowitz WS, Curtin K, Ma KN, et al: Prognostic significance of p53 mutations in colon cancer at the population level, Int J Cancer 99:597-602, 2002
Tang R, Wang JY, Fan CW, et al: p53 is an independent pre-treatment markers for long-term survival in stage II and III colorectal cancers: An analysis of interaction between genetic markers and fluorouracil-based adjuvant therapy. Cancer Lett 210:101-109, 2004
Kapiteijn E, Marijnen CA, Nagtegaal ID, et al: Dutch Colorectal Cancer Group: Preoperative radiotherapy combined with total mesorectal excision for resectable rectal cancer. N Engl J Med 345:638-646, 2001
Bleeker WA, Mulder NH, Hermans J, et al: The addition of low-dose leucovorin to the combination of 5-fluorouracil-levamisole does not improve survival in the adjuvant treatment of Dukes' C colon cancer. Ann Oncol 11:547-552, 2000
Bleeker WA, Mulder NH, Hermans J, et al: Value and cost of follow-up after adjuvant treatment of patients with Dukes' C colonic cancer. Br J Surg 88:101-106, 2001
Hamilton SR and Aaltonen LA. Ed: WHO Classification of Tumours: Pathology and Genetics of Tumours of Digestive System. Geneva, Switzerland, WHO, 2000
Sobin LH, Ch Wittekind: UICC International Union Against Cancer TNM Classification of Malignant Tumors (ed 6). New York, NY, Wiley-Liss, 2002
Boland CR, Thibodeau SN, Hamilton SR, et al: A National Cancer Institute workshop on microsatellite instability for cancer detection and familial predisposition: Development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 58:5248-5257, 1998
Wu Y, Berends MJW, Mensink RGJ, et al: Association of hereditary nonpolyposis colorectal cancer-related tumors displaying low microsatellite instability with MSH6 germline mutations. Am J Hum Genet 65:1291-1298, 1999
Hayes VM, Westra JL, Verlind E, et al: New comprehensive denaturing-gradient-gel- electrophoresis assay for KRAS mutation detection applied to paraffin-embedded tumors. Genes Chromosomes Cancer 29:309-314, 2000
Hayes VM, Bleeker W, Verlind E, et al: Comprehensive TP53-denaturing gradient gel electrophoresis mutation detection assay also applicable to archival paraffin-embedded tissue. Diagn Mol Pathol 8:2-10, 1999
Grambsch PM, Thernau TM: Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 81:515-526, 1994
May S, Hosmer DW: A simplified method of calculating an overall goodness-of-fit test for the Cox proportional hazards model. Lifetime Data Anal 4:109-120, 1998
Olivier M, Eeles R, Hollstein M, et al: The IARC TP53 database: New online mutation analysis and recommendations to users. Hum Mutat 19:607-614, 2002
Andreyev HJ, Norman AR, Cunningham D, et al: Kirsten ras mutations in patients with colorectal cancer: The ‘RASCAL II’ study. Br J Cancer 85:692-696, 2001
Gafa R, Maestri I, Matteuzzi M, et al: Sporadic colorectal adenocarcinomas with high-frequency microsatellite instability. Cancer 89:2025-2037, 2000
Alexander J, Watanabe T, Wu TT, et al: Histopathological identification of colon cancer with microsatellite instability. Am J Pathol 158:527-535, 2001
Costa A, Marasca R, Valentinis B, et al: p53 gene point mutations in relation to p53 nuclear protein accumulation in colorectal cancers. J Pathol 176:45-53, 1995
Hanski C, Tiecke F, Hummel M, et al: Low frequency of p53 gene mutation and protein expression in mucinous colorectal carcinomas. Cancer Lett 103:163-170, 1996
Elsaleh H, Powell B, Soontrapornchai P, et al: p53 gene mutation, microsatellite instability and adjuvant chemotherapy: Impact on survival of 388 patients with Dukes' C colon carcinoma. Oncology 58:52-59, 2000
Samowitz WS, Curtin K, Ma KN, et al: Microsatellite instability in sporadic colon cancer is associated with an improved prognosis at the population level. Cancer Epidemiol Biomarkers Prev 10:917-923, 2001
Hemminki A, Mecklin JP, Jarvinen H, et al: Microsatellite instability is a favorable prognostic indicator in patients with colorectal cancer receiving chemotherapy. Gastroenterology 119:921-928, 2000
Elsaleh H, Joseph D, Grieu F, et al: Association of tumor site and sex with survival benefit from adjuvant chemotherapy in colorectal cancer. Lancet 355:1745-1750, 2000
Ribic CM, Sargent DJ, Moore MJ, et al: Tumor microsatellite-instability status as a predictor of benefit from fluorouracil-based adjuvant chemotherapy for colon cancer. N Engl J Med 349:247-257, 2003
Carethers JM, Smith EJ, Behling CA, et al: Use of 5-fluorouracil and survival in patients with microsatellite-unstable colorectal cancer. Gastroenterology 126:394-401, 2004
Dix BR, Robbins P, Soong R, et al: The common molecular genetic alterations in Dukes' B and C colorectal carcinomas are not short-term prognostic indicators of survival. Int J Cancer 59:747-751, 1994
Soong R, Grieu F, Robbins P, et al: p53 alterations are associated with improved prognosis in distal colonic carcinomas. Clin Cancer Res 3:1405-1411, 1997
Soong R, Powell B, Elsaleh H, et al: Prognostic significance of TP53 gene mutation in 995 cases of colorectal carcinoma: Influence of tumor site, stage, adjuvant chemotherapy and type of mutation. Eur J Cancer 36:2053-2060, 2000
Kapiteijn E, Liefers GJ, Los LC, et al: Mechanisms of oncogenesis in colon versus rectal cancer. J Pathol 195:171-178, 2001
Goel A, Arnold CN, Niedzwiecki D, et al: Characterization of sporadic colon cancer by patterns of genomic instability. Cancer Res 63:1608-1614, 2003
Johnston PG: Of what value genomics in colorectal cancer? Opportunities and challenges. J Clin Oncol 22:1538-1539, 2004(Jantine L. Westra, Michae)
Comprehensive Cancer Center North Netherlands
the Departments of Pathology, Medical Oncology, and Surgery, University Hospital Groningen, the Netherlands
ABSTRACT
PURPOSE: Microsatellite instability (MSI), TP53 mutation, and KRAS mutation status have been reported as prognostic factors in colon cancer. Most studies, however, have included heterogeneous groups of patients with respect to cancer stage. We determined the prognostic relevance of high-frequency MSI (MSI-H), TP53 mutations, and KRAS mutations in a well-defined group of patients with stage III colon cancer (N = 391), randomly assigned for adjuvant treatment with fluorouracil-based chemotherapy.
METHODS: Three hundred ninety-one tumor specimens were available. MSI was determined in 273 specimens, and mutation analyses of TP53 and KRAS were performed in 220 and 205 specimens, respectively.
RESULTS: In a univariate analysis, MSI-H (44 of 273; 16%) was associated with a longer disease-free survival (DFS; P = .038), but in a multivariate model adjusting for nodal involvement, histology, invasion, and grade of tumor, the association of MSI status with DFS did no longer reach statistical significance, though the risk estimate for microsatellite stability versus MSI-H tumors did not change much. Mutant TP53, found in 116 (53%) of 220 tumors, was associated with a shorter DFS, both in univariate (P = .009) and multivariate analyses (P = .018), whereas KRAS mutations (58 of 205; 28%) did not show any prognostic significance.
CONCLUSION: Both mutant TP53 and MSI-H seem to be prognostic indicators for disease-free survival, but only TP53 retains statistical significance after adjusting for clinical heterogeneity. Thus, in adjuvantly treated patients with stage III colon cancer, presence or absence of a TP53 mutation should be considered as a better predictor for DFS than MSI status.
INTRODUCTION
Although the presently used TNM classification for colon cancer improves in the predictive value of its precursors, its value in choosing adjuvant treatment after surgery, which may be particularly important for patients with stage III colon cancer, is still limited. Therefore, many studies have been undertaken to explore molecular prognostic and predictive markers in this group of patients. Several well-known genetic alterations, associated with different forms of genomic instability,1-3 have been tested for their use in subclassification models using molecular criteria. A chromosomal instability (CIN) phenotype is found in approximately 85% of sporadic colon cancers and is characterized by aneuploidy, multiple chromosomal rearrangements, and accumulation of somatic mutations in oncogenes such as KRAS and tumor-suppressor-genes such as TP53 and adenomatous polyposis coli (APC).3,4 A high-frequency microsatellite instability (MSI-H or MIN) phenotype is associated with small insertions and deletions mainly in repetitive sequences (microsatellites).1,2 This instability is caused by defects in the mismatch repair system, which is involved in repairing errors arising during DNA replication. The MSI-H phenotype is observed in approximately 15% of the sporadic colon cancers.1,2,5
An inverse correlation with respect to clinicopathologic presentation has been found for TP53 and/or KRAS mutations, and for MSI-H conversely.6-9 MSI-H tumors are predominantly characterized by diploidy, more right-sided location, mucinous histology, and poor differentiation,2,10-13 whereas CIN tumors are usually characterized by left-sided location, aneuploidy, nonmucinous histology, predominantly moderate differentiation, and less tumor-infiltrating lymphocytes.13,14
The MSI-H tumor phenotype has been associated with a better prognosis than the microsatellite stability (MSS) tumor phenotype2,10,15-24; presence of a TP53 mutation in the tumor has often been associated with shorter survival of the patient compared with the presence of wild-type TP53.25-33
One of the problems in the interpretation of these data is that many studies are based on series of tumors not belonging to the same stage or include rectal tumors for which a different treatment strategy is used.34 Moreover, numbers of patients or tumors screened are often small. Another problem is that different methods of tissue isolation and mutation or MSI screening are used, which may hamper a direct comparison between studies. Our aim was to determine the prognostic value of MSI-H and of mutations in TP53 and KRAS in a well-defined homogeneous group of patients with stage III colon cancer, in combination with conventional clinicopathologic characteristics, such as the number of tumor-positive lymph nodes, and histologic criteria.
PATIENTS AND METHODS
Patients' Material
Formalin-fixed paraffin-embedded (FFPE) primary tumor tissue from 391 patients with stage III colon cancer included in a nation-wide randomized colon cancer (CKVO 90-11) trial (N = 500) was available for this study. In the remaining 109 patients no primary tumor specimens or insufficient material were available for DNA analyses. All patients had histologically negative (R0) resection margins and were treated adjuvantly with fluorouracil (FU) -based chemotherapy (FU/levamisole or FU/levamisole/leucovorin) as described earlier.35,36 The patients of a subset of 391 cases suitable for DNA extraction were equally distributed between both groups (FU/levamisole or FU/levamisole/leucovorin). Moreover, no differences in disease-free or overall survival were seen between both groups.
Clinical and Tumor Characteristics of the Patients
For all patients, clinical and reviewed tumor characteristics were derived from the clinical database maintained at the Comprehensive Cancer Center, North Netherlands.35 The mean patient age at the time of random assignment was 58.7 years (range, 26.2 to 75.8 years). Two hundred thirteen patients were men (54.5%); 182 tumors were right-sided (46.5%; ie, proximal of the splenic flexure). The tumors were defined as mucinous if more than 50% had a mucinous aspect. Furthermore, the tumors were histologically classified according to the WHO standard.37 With regard to histological grade, 17.5% of the tumors were well differentiated, 57.3% were moderately differentiated, and 25.2% were poorly differentiated. According to the TNM staging system, 13.6% of the tumors belonged to the T1 and T2 stages, and 86.4%, to stages T3 or T4. Furthermore, 74.2% of the resection specimens belonged to stage N1, while 101 (25.8%) were stage N2.38
The 5-year disease-free survival (DFS) was 55.5% for all participants of the CKVO 90-11 trial, which was equal to the 56.3% among the analyzed subgroup of 391 patients from whom FFPE primary tumor specimens were available. In the smallest subgroup of our study, the group of patients for whom we could determine both the TP53 mutation status and the MSI status, the 5-year DFS was 56.4% (n = 194). Based on survival patterns, there was no evidence for a selection bias for these subgroups (log-rank test P = .724).
DNA Extraction
After interpretation of hematoxylin and eosin staining from all FFPE tissues (acquired before chemotherapy had started), representative tumor tissue and corresponding normal tissue were selected by an experienced pathologist and were cut in 20-μm slices for DNA isolation. The QIAamp DNA mini kit (Qiagen, Westburg, Leusden, the Netherlands) was used for DNA extraction according to protocol, with some adjustments. These included skipping the xylene treatment, in the final elution step, twice putting the eluate (100 μL of preheated [70°C] elution buffer) on the column, and incubating it at 70°C for 5 to 10 minutes before spin down.
MSI Analysis
MSI analysis was performed on an ABI Prism 377 DNA sequencer after amplifying DNA from tumor and matching normal tissue, using fluorescent-labeled primers from the MSI multiplex kit (Promega, Madison, WI). The MSI multiplex kit includes primers for coamplification of nine microsatellite markers, including four mononucleotide repeat markers (BAT-25, BAT-26, BAT-40, and MONO27) and five tetranucleotide repeat markers (D3S2432, D7S1808, D7S3046, D7S3070, and D10S1426). Alternatively, the five consensus markers described by Boland et al39 were used, including two mononucleotide (BAT-25 and BAT-26) and three dinucleotide repeat markers (D2S123, D5S346, and D17S250). This was done as described previously40or by the use of the HNPCC MSI kit (Roche, Mannheim, Germany).
When using the Roche kit or the Boland consensus markers, those samples were included in which all five markers could be interpreted; four markers could be interpreted and all proved negative; or two (interpretable) markers proved positive. The Promega kit samples were included if five (or more) of nine markers were interpretable, all four mononucleotide markers were interpretable, or if three (interpretable) markers proved positive. These inclusion criteria were formulated because sometimes difficulties in interpretation or lack of results were met due to bad quality of DNA obtained from old FFPE specimens.
A tumor was classified as MSI-H when more than two of five or more than three of nine microsatellite markers showed instability. In all other cases, tumors were defined as MSS. All analyses were performed without previous knowledge of clinical outcome.
Mutation Analysis for KRAS and TP53
Mutation screening (of parts) of KRAS (exons 1 and 2) and TP53 (exons 4 to 8) was performed by denaturing gradient gel electrophoresis (DGGE) using the INGENYphorU system (Ingeny International, Goes, the Netherlands) and DNA sequencing using ABI Prism 377 DNA sequencer (Perkin Elmer Applied Biosystems, Foster City, CA) according to previously described protocols.41,42
Statistical Analysis
For comparison of the distribution of continuous variables, the analysis of variance test was used. The Fisher's exact test was used in case of 2 x 2 tables, otherwise the 2 test was used for comparison of categorical variables. DFS was used to assess the prognostic value of the various molecular markers, as it is the least biased indicator of treatment failure given the facts that disease-specific survival might be influenced by selective use of second-line treatments, and overall survival is biased due to deaths attributable to causes unrelated to colon cancer. DFS was defined as the period from the date of random assignment to the date of any documented relapse, to the date of death without documented relapse due to colon cancer, or to the date of last contact, whichever occurred first. Censoring occurred when a patient was alive without relapse at last contact or when the patient died of a cause other than colon cancer, not attributable to colon cancer. Time to event curves were estimated using the Kaplan-Meier method, and the survival distributions were compared with the log-rank test. Cox proportional hazard models were constructed to assess the simultaneous effects of various covariables on the survival experience in our cohort.43 Variables included in the model were tumor stage (invasion into the muscularis propria v beyond the muscularis propria), tumor grade, tumor cell type (mucinous v nonmucinous), and number of tumor-positive lymph nodes (< 4 v 4). Model fit was evaluated using graphical methods and the May-Hosmer goodness-of-fit test statistic.44 All reported P values are two-sided unless stated otherwise.
RESULTS
Molecular Parameters
MSI analysis. Based on the before-mentioned inclusion criteria, 273 of the 391 analyzed DNA specimens could be included. The Roche kit was applied to 37 tumors, of which 23 gave interpretable data, identifying 18 MSS and five MSI-H tumors. The Promega kit was applied to 223 tumors, of which 192 gave interpretable data, identifying 167 MSS and 25 MSI-H tumors. Furthermore, 67 tumors were analyzed by simplex PCR-analyses. Of these, 58 gave interpretable data, identifying 14 as MSI-H. In total, 44 (16%) of the 273 samples seemed to be instable (Table 1).
The difference in results between the methods used to determine the MSI status (Roche kit, 22%; Promega kit, 13%; simplex analysis, 24%) is not statistically significant (22 P = .097) and is most likely due to chance and not caused by the different repeat markers, since, if we determine the number of MSI-H cases by using only the shared mononucleotide repeats, which proved to be the most informative markers, no more than one tumor would change from MSS to MSI-H.
TP53 mutation analysis. Included in this analysis were 220 tumors screened for at least the seven amplicons covering the sequence coding for the conserved domains of TP53. A total of 123 mutations were identified in 116 (53%) of the 220 tumors. In Figure 1, the distribution over codons/exons/domains and frequency of TP53 mutations are presented. Multiple TP53 mutations were found in six tumors, of which five tumors had two, and one tumor had three TP53 mutations. Five frequently recurrent mutations occurred at codons 273, 248, 175, 282, and 245, with prevalences of 12, 12, six, seven, and seven times, respectively.
Three of the 220 tumors showed an obviously aberrant DGGE pattern, different from known polymorphisms, for which we were unable to produce a reliable sequence. The probability that one of these three would be a silent mutation is very low, since we observed only four silent mutations in a total of 120 mutations. We therefore decided to include the three in the group of causative mutations. In addition, we found 105 DGGE variants that we classified as polymorphisms, and four silent mutations: Lys120 (G360A), Thr125 (G375A), Cys135 (C405T), and Leu299 (C895T).
KRAS mutation analysis. In the 205 tumors that could be analyzed for both KRAS exons 1 and 2, 58 KRAS mutations were identified (28%). Except for one, in which we found a silent mutation in codon 60 next to a mutation in codon 61, none of the tumors had more than one KRAS mutation. Most mutations occurred in codon 12 (46 cases, 79%), followed by codon 61 (seven cases, 12%) and codon 13 (five cases, 9%). The Gly12Asp amino acid change was the most frequent mutation (n = 22), followed by Gly12Val (n = 12), Gly12Cys and Gln61His (both n = 6), Gly13Asp (n = 5), Gly12Ser (n = 4), Gly12Ala (n = 2) and Gln61Lys (n = 1). In addition to these well-known mutations, we identified four silent mutations (Leu6, Val8 [twice], and Gly60).
Combined molecular parameters MSI, TP53, and KRAS status. In nine (36%) of the 25 MSI-H tumors, a TP53 mutation was found, and 94 (56%) of 169 MSS tumors carried such a mutation (P = .067). KRAS mutations were found in four (16%) of 25 MSI-H tumors, compared with 42 of 126 MSS tumors (33%; P = .057).
For those tumors in which MSI as well as KRAS or TP53 mutation status could be determined, we found that in 10 (43%) of the 23 MSI-H tumors, either KRAS or TP53 were mutated, compared with 82 (79%) of 119 MSS tumors (P = .006). Both TP53 and KRAS mutations occurred in two (9%) of these 23 MSI-H tumors and in 19 (16%) of 119 MSS tumors.
Combined Molecular and Clinicopathologic Characteristics/Correlations
The results of the molecular analyses, in combination with the clinicopathologic characteristics, are summarized in Table 1. Compared with MSS tumors, tumors with an MSI-H phenotype were mostly right-sided (80%; P < .001) and more often showed a poor cellular differentiation. MSI-H tumors were more frequently of a mucinous histology, whereas the MSS tumors were predominantly nonmucinous. No differences were detected between patients with MSI-H and MSS tumor phenotype with respect to sex, number of tumor-positive lymph nodes (< 4 or 4), or tumor stage.
Tumors with mutant TP53 were more frequently of the nonmucinous cell type (71%), whereas tumors with wild-type TP53 had a more even distribution over nonmucinous and mucinous cell types. This association remained significant (P = .018) when cases with unspecified histological subtype were excluded. No significant associations were found with KRAS mutation status of the tumors (Table 1).
Survival in Relation to the Different Molecular Parameters
MSI. The 5-year DFS was 71.8% for patients with an MSI-H tumor and 49.7% for patients with an MSS tumor, patients with an MSS tumors had a worse DFS (P = .034; Fig 2). In our study, patients with a MSS tumor experienced a 1.85-fold increased risk of relapse (95% CI, 1.03 to 3.29) compared with patients with an MSI-H tumor. Table 2 presents the results of the multivariate analysis. Adjusted for tumor grade, histology, number of tumor-positive nodes, and depth of tumor invasion, the association of MSI status with DFS did, however, not reach statistical significance, although the risk estimate for MSS versus MSI-H tumors did hardly change (hazard ratio [HR] = 1.80; 95% CI, 0.97 to 3.32).
TP53. The 5-year DFS for patients with a tumor carrying a somatic TP53 mutation was 44.0%, compared with 62.7% for patients with a wild-type TP53 tumor. Patients having a tumor with a TP53 mutation had a significantly worse DFS than those with a wild-type TP53 tumor (P = .009; Fig 3). They experienced a 1.71-fold (95% CI, 1.14 to 2.56) increased risk of relapse. Adjusted for tumor grade, histology, number of tumor-positive nodes, and depth of tumor invasion, TP53 remained a significant independent predictor of DFS, with an HR of 1.71 (95% CI, 1.12 to 2.59) for patients who had a tumor with a TP53 mutation compared with patients with a wild-type TP53 tumor (Table 3).
KRAS. In a univariate analysis, KRAS status was not associated with DFS. Therefore, KRAS status was not included in the multivariate analyses.
Combined MSI and TP53 phenotype correlation with survival after adjusting for clinical heterogeneity. In a subgroup of patients, both MSI and TP53 status could be determined (n = 194). In a multivariate model comprising the degree of nodal involvement, tumor grade, histology, depth of invasion, MSI status, and TP53 mutation status, neither MSI nor TP53 were significantly associated with DFS (Table 4).
However, added to the model with nodal involvement, tumor grade, histology, and depth of invasion, TP53 status still reached borderline significance (likelihood-ratio test [21] 3.49; P = .062) while the MSI status explained only little variation in DFS (likelihood-ratio test [21] 1.41; P = .235). When adding MSI status to a model containing nodal involvement, tumor grade, histology, depth of tumor invasion, and TP53 status, the model-based log likelihood changed with 0.81 (likelihood-ratio test P = .368). Adding TP53 status to a model with the MSI status changed the log likelihood with 2.89 (likelihood-ratio test P = .089).
DISCUSSION
This study used data from 391 stage III colon cancer patients who participated in a prospective randomized trial studying two adjuvant regimens of FU-based chemotherapy. Genetic alterations in TP53 (exons 4 to 8) and KRAS (exons 1 and 2) were found in 53% and 28% of cases, respectively. MSI was exhibited by 16% of tumors. These frequencies are in agreement with other published data.1,2,5,32,45,46 Mutation analysis of TP53 was limited to exons 4 to 8 for most tumors, because an initial screening of the whole gene in 69 tumors revealed 39 mutations, with all but one located in exons 4 to 8. The International Agency for Research on Cancer TP53 mutation database also contains few mutations (2% overall or 5% colon carcinoma) outside exons 4 to 8.45
We first investigated whether associations existed between clinical parameters and the genetic alterations. The associations observed with MSI-H status, proximal location of the tumor, poor cellular differentiation, and mucinous tumor histology are in agreement with those found by other investigators.12,13,47,48 The observation that TP53 mutation was associated with an increased frequency of nonmucinous tumors is also in agreement with results from others.49,50
In univariate analysis, a better prognosis resulted for patients with MSI-H tumors than for patients with MSS tumors. Comparing our data with findings reported on similar groups of patients, our results are in agreement with results obtained by several others9,19,21,23,51-55 who included only stage III (and/or II) colon (and/or rectal) cancer patients, irrespective of adjuvant treatment. It should be noted, however, that if only the non–adjuvantly treated patients were selected, no association of MSI status and survival was found by Elsaleh et al,9,51,54 whereas Ribic et al55 recently reported that for patients who did not receive adjuvant chemotherapy, MSI-H status is a positive prognostic factor in univariate analysis. For adjuvantly treated patients, no difference was observed in DFS or overall survival of patients with MSI-H versus MSS tumors. This is in disagreement with our results. A partial explanation for the difference between the study of Ribic et al55 and our study may be that they included not only stage III but also stage II colon cancer patients. A subgroup analysis of the data of Ribic et al55 according to stage is missing. Carethers et al,56 conversely, found no prognostic significance for MSI status in combined adjuvantly treated and untreated, or selected adjuvantly treated or untreated subgroups of stage II/III colorectal cancer (CRC) patients.
Our study showed that having tumors with a TP53 mutation is associated with a shorter DFS (P = .018). This is in agreement with results for non–adjuvantly treated groups of stage III/II colon cancer patients studied by Pricolo et al27 and Tang et al,33 and from the stage III subgroup analysis of CRC patients in the study by Goh et al,31 but disagrees with results obtained by others51,57-59 who found no significant association between TP53 mutation and overall or disease-specific survival. Soong et al57 even reported a trend toward better survival. All disagreeing studies included both adjuvantly treated and non–adjuvantly treated patients. Only Elsaleh et al51 included merely stage III colon cancer. Others also included stage II colon tumors and rectal carcinomas that are clinically and genetically different cancers.34,60
Presence of a TP53 mutation is more common in CIN tumors than in MIN tumors.7-9 This suggests that tumors with a MIN phenotype are less aggressive than those with a CIN phenotype. TP53 mutations were found in both the MSS (presumed CIN) and the MSI-H groups. The difference between the mutation frequencies in these groups was not statistically significant, but certainly the estimated proportion of patients having a tumor with a TP53 mutation was higher in the MSS group than in the MSI-H group (56% v 36%).
We also did not find a definite negative association between MSI-H and KRAS as suggested by others.6,8 In our study, four (16%) of 25 MSI-H tumors had a KRAS mutation, compared with 42 (33%) of 126 of MSS tumors. Presence of a KRAS mutation in the tumor did not correlate with any clinicopathologic parameter, nor with DFS.
Our results do not exclude a coexistence of MIN and CIN developmental routes in the same tumor. This is in line with the findings of Goel et al,61 who found a significant degree of overlap between MIN and CIN characteristics, by performing loss of heterozygosity analyses on a number of chromosome arms (18q, 5q, 17p, 3p, and 2p), instead of TP53 analysis.
our univariate analyses, it could be concluded that an MSI-H phenotype acts as a positive prognostic factor and that TP53 mutation is an adverse prognostic factor for patients with stage III colon cancer. Adjusted for clinical heterogeneity, only a TP53 mutation retained its independent significant adverse prognostic effect. Although the risk estimate for MSI status in multivariate analysis did not change much compared to the univariate analysis, the MSI status did no longer reach statistical significance with DFS, which can be explained by the low occurrence of MSI-H tumors (approximately 15%) in sporadic CRCs.
In a multivariate model including both MSI and TP53 mutation status, the effect of TP53 appeared to be stronger than that of MSI. The power of that analysis, however, was too small to prove that effect at a 5% significance level. The likelihood-ratio test showed that TP53 mutation status was more powerful in predicting DFS than MSI status. Presence or absence of a TP53 mutation, and not MSI, therefore seems a better predictor for DFS in adjuvantly treated stage III colon cancer patients.
It has been suggested that more than 50% of the adjuvant therapies given to stage III patients may be unnecessary.62 This suggestion again underscores the need for predictive markers that will allow identification of patients who will or will not benefit from adjuvant FU treatment. Unfortunately, from our study, we cannot conclude that MSI and TP53 mutation status of the primary tumor can be used as such. Although (small) effects of both are seen, they cannot be used for the individual patient in a decision protocol concerning adjuvant chemotherapy.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank Jeff Bacher from the Promega Corporation for generously supplying the MSI multiplex system: prototype kits; Ton Tiebosch MD, PhD, for pathologic assistance; and Gerwin Jansen, Annet Kooistra, Ludolf Boven, Jan Osinga, and Edwin Verlind for technical assistance. We also thank the anonymous reviewers for their useful comments, improving the quality of this manuscript.
NOTES
Supported by a grant from Dutch Cancer Society, NKB number RUG 99-1962.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Ionov Y, Peinado MA, Malkhosyan S, et al: Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature 363:558-561, 1993
Thibodeau SN, Bren G, Schaid D: Microsatellite instability in cancer of the proximal colon. Science 260:816-819, 1993
Lengauer C, Kinzler KW, Vogelstein B: Genetic instability in colorectal cancers. Nature 386:623-627, 1997
Fearon ER, Vogelstein B: A genetic model for colorectal tumorigenesis. Cell 61:759-767, 1990
Aaltonen LA, Peltomaki P, Leach FS, et al: Clues to the pathogenesis of familial colorectal cancer. Science 260:812-816, 1993
Breivik J, Lothe RA, Meling GI, et al: Different genetic pathways to proximal and distal colorectal cancer influenced by sex-related factors. Int J Cancer 74:664-669, 1997
Salahshor S, Kressner U, Pahlman L, et al: Colorectal cancer with and without microsatellite instability involves different genes. Genes Chromosomes Cancer 26:247-252, 1999
Samowitz WS, Holden JA, Curtin, et al: Inverse relationship between microsatellite instability and K-ras and p53 gene alterations in colon cancer. Am J Pathol 158:1517-1524, 2001
Elsaleh H, Powell B, McCaul K, et al: p53 alteration and microsatellite instability have predictive value for survival benefit from chemotherapy in stage III colorectal carcinoma. Clin Cancer Res 7:1343-1349, 2001
Lothe RA, Peltomaki P, Meling GI, et al: Genomic instability in colorectal cancer: Relationship to clinicopathologicalvariables and family history. Cancer Res 53:5849-5852, 1993
Kim H, Jen J, Vogelstein B, et al: Clinical and pathological characteristics of sporadic colorectal carcinomas with DNA replication errors in microsatellite sequences. Am J Pathol 145:148-156, 1994
Jass JR, Do KA, Simms LA, et al: Morphology of sporadic colorectal cancer with DNA replication errors. Gut 42:673-679, 1998
Young J, Simms LA, Biden KG, et al: Features of colorectal cancers with high-level microsatellite instability occurring in familial and sporadic settings: Parallel pathways of tumorigenesis. Am J Pathol 159:2107-2116, 2001
Delattre O, Olschwang S, Law DJ, et al: Multiple genetic alterations in distal and proximal colorectal cancer. Lancet 2:353-356, 1989
Bubb VJ, Curtis LJ, Cunningham C, et al: Microsatellite instability and the role of hMSH2 in sporadic colorectal cancer. Oncogene 12:2641-2649, 1996
Lukish JR, Muro K, DeNobile J, et al: Prognostic significance of DNA replication errors in young patients with colorectal cancer. Ann Surg 227:51-56, 1998
Massa MJ, Iniesta P, Gonzalez-Quevedo R, et al: Differential prognosis of replication error phenotype and loss of heterozygosity in sporadic colorectal cancer. Eur J Cancer 35:1676-1682, 1999
Jernvall P, Makinen MJ, Karttunen TJ, et al: Microsatellite instability: Impact on cancer progression in proximal and distal colorectal cancers, Eur J Cancer 35:197-201, 1999
Halling KC, French AJ, McDonnell SK, et al: Microsatellite instability and 8p allelic imbalance in stage B2 and C colorectal cancers. J Natl Cancer Inst 91:1295-1303, 1999
Gryfe R, Kim H, Hsieh ET, et al: Tumor microsatellite instability and clinical outcome in young patients with colorectal cancer. N Engl J Med 342:69-77, 2000
Wright CM, Dent OF, Barker M, et al: Prognostic significance of extensive microsatellite instability in sporadic clinicopathological stage C colorectal cancer. Br J Surg 87:1197-1202, 2000
Choi SW, Lee KJ, Bae YA, et al: Genetic classification of colorectal cancer based on chromosomal loss and microsatellite instability predicts survival. Clin Cancer Res 8:2311-2322, 2002
Watanabe T, Wu TT, Catalano PJ, et al: Molecular predictors of survival after adjuvant chemotherapy for colon cancer. N Engl J Med 344:1196-1206, 2001
Lim SB, Jeong SY, Lee MR, et al: Prognostic significance of microsatellite instability in sporadic colorectal cancer. Int J Colorectal Dis 19:533-537, 2004
Goh HS, Yao J, Smith DR: p53 point mutation and survival in colorectal cancer patients. Cancer Res 55:5217-5221, 1995
Smith DR, Ji CY, Goh HS: Prognostic significance of p53 overexpression and mutation in colorectal adenocarcinomas. Br J Cancer 74:216-223, 1996
Pricolo VE, Finkelstein SD, Hansen K, et al: Mutated p53 gene is an independent adverse predictor of survival in colon carcinoma. Arch Surg 132:371-374, 1997
Borresen-Dale AL, Lothe RA, Meling GI, et al: TP53 and long-term prognosis in colorectal cancer: Mutations in the L3 zinc-binding domain predict poor survival. Clin Cancer Res 4:203-210, 1998
Goh HS, Elnatan J, Low CH, et al: p53 point mutation and survival in colorectal cancer patients: Effect of disease dissemination and tumor location. Int J Oncol 15:491-498, 1999
Kressner U, Inganas M, Byding S, et al: Prognostic value of p53 genetic changes in colorectal cancer. J Clin Oncol 17:593-599, 1999
Tortola S, Marcuello E, Gonzalez I, et al: p53 and K-ras gene mutations correlate with tumor aggressiveness but are not of routine prognostic value in colorectal cancer. J Clin Oncol 17:1375-1381, 1999
Samowitz WS, Curtin K, Ma KN, et al: Prognostic significance of p53 mutations in colon cancer at the population level, Int J Cancer 99:597-602, 2002
Tang R, Wang JY, Fan CW, et al: p53 is an independent pre-treatment markers for long-term survival in stage II and III colorectal cancers: An analysis of interaction between genetic markers and fluorouracil-based adjuvant therapy. Cancer Lett 210:101-109, 2004
Kapiteijn E, Marijnen CA, Nagtegaal ID, et al: Dutch Colorectal Cancer Group: Preoperative radiotherapy combined with total mesorectal excision for resectable rectal cancer. N Engl J Med 345:638-646, 2001
Bleeker WA, Mulder NH, Hermans J, et al: The addition of low-dose leucovorin to the combination of 5-fluorouracil-levamisole does not improve survival in the adjuvant treatment of Dukes' C colon cancer. Ann Oncol 11:547-552, 2000
Bleeker WA, Mulder NH, Hermans J, et al: Value and cost of follow-up after adjuvant treatment of patients with Dukes' C colonic cancer. Br J Surg 88:101-106, 2001
Hamilton SR and Aaltonen LA. Ed: WHO Classification of Tumours: Pathology and Genetics of Tumours of Digestive System. Geneva, Switzerland, WHO, 2000
Sobin LH, Ch Wittekind: UICC International Union Against Cancer TNM Classification of Malignant Tumors (ed 6). New York, NY, Wiley-Liss, 2002
Boland CR, Thibodeau SN, Hamilton SR, et al: A National Cancer Institute workshop on microsatellite instability for cancer detection and familial predisposition: Development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 58:5248-5257, 1998
Wu Y, Berends MJW, Mensink RGJ, et al: Association of hereditary nonpolyposis colorectal cancer-related tumors displaying low microsatellite instability with MSH6 germline mutations. Am J Hum Genet 65:1291-1298, 1999
Hayes VM, Westra JL, Verlind E, et al: New comprehensive denaturing-gradient-gel- electrophoresis assay for KRAS mutation detection applied to paraffin-embedded tumors. Genes Chromosomes Cancer 29:309-314, 2000
Hayes VM, Bleeker W, Verlind E, et al: Comprehensive TP53-denaturing gradient gel electrophoresis mutation detection assay also applicable to archival paraffin-embedded tissue. Diagn Mol Pathol 8:2-10, 1999
Grambsch PM, Thernau TM: Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 81:515-526, 1994
May S, Hosmer DW: A simplified method of calculating an overall goodness-of-fit test for the Cox proportional hazards model. Lifetime Data Anal 4:109-120, 1998
Olivier M, Eeles R, Hollstein M, et al: The IARC TP53 database: New online mutation analysis and recommendations to users. Hum Mutat 19:607-614, 2002
Andreyev HJ, Norman AR, Cunningham D, et al: Kirsten ras mutations in patients with colorectal cancer: The ‘RASCAL II’ study. Br J Cancer 85:692-696, 2001
Gafa R, Maestri I, Matteuzzi M, et al: Sporadic colorectal adenocarcinomas with high-frequency microsatellite instability. Cancer 89:2025-2037, 2000
Alexander J, Watanabe T, Wu TT, et al: Histopathological identification of colon cancer with microsatellite instability. Am J Pathol 158:527-535, 2001
Costa A, Marasca R, Valentinis B, et al: p53 gene point mutations in relation to p53 nuclear protein accumulation in colorectal cancers. J Pathol 176:45-53, 1995
Hanski C, Tiecke F, Hummel M, et al: Low frequency of p53 gene mutation and protein expression in mucinous colorectal carcinomas. Cancer Lett 103:163-170, 1996
Elsaleh H, Powell B, Soontrapornchai P, et al: p53 gene mutation, microsatellite instability and adjuvant chemotherapy: Impact on survival of 388 patients with Dukes' C colon carcinoma. Oncology 58:52-59, 2000
Samowitz WS, Curtin K, Ma KN, et al: Microsatellite instability in sporadic colon cancer is associated with an improved prognosis at the population level. Cancer Epidemiol Biomarkers Prev 10:917-923, 2001
Hemminki A, Mecklin JP, Jarvinen H, et al: Microsatellite instability is a favorable prognostic indicator in patients with colorectal cancer receiving chemotherapy. Gastroenterology 119:921-928, 2000
Elsaleh H, Joseph D, Grieu F, et al: Association of tumor site and sex with survival benefit from adjuvant chemotherapy in colorectal cancer. Lancet 355:1745-1750, 2000
Ribic CM, Sargent DJ, Moore MJ, et al: Tumor microsatellite-instability status as a predictor of benefit from fluorouracil-based adjuvant chemotherapy for colon cancer. N Engl J Med 349:247-257, 2003
Carethers JM, Smith EJ, Behling CA, et al: Use of 5-fluorouracil and survival in patients with microsatellite-unstable colorectal cancer. Gastroenterology 126:394-401, 2004
Dix BR, Robbins P, Soong R, et al: The common molecular genetic alterations in Dukes' B and C colorectal carcinomas are not short-term prognostic indicators of survival. Int J Cancer 59:747-751, 1994
Soong R, Grieu F, Robbins P, et al: p53 alterations are associated with improved prognosis in distal colonic carcinomas. Clin Cancer Res 3:1405-1411, 1997
Soong R, Powell B, Elsaleh H, et al: Prognostic significance of TP53 gene mutation in 995 cases of colorectal carcinoma: Influence of tumor site, stage, adjuvant chemotherapy and type of mutation. Eur J Cancer 36:2053-2060, 2000
Kapiteijn E, Liefers GJ, Los LC, et al: Mechanisms of oncogenesis in colon versus rectal cancer. J Pathol 195:171-178, 2001
Goel A, Arnold CN, Niedzwiecki D, et al: Characterization of sporadic colon cancer by patterns of genomic instability. Cancer Res 63:1608-1614, 2003
Johnston PG: Of what value genomics in colorectal cancer? Opportunities and challenges. J Clin Oncol 22:1538-1539, 2004(Jantine L. Westra, Michae)