Low Microsatellite Instability Is Associated With Poor Prognosis in Stage C Colon Cancer
http://www.100md.com
《临床肿瘤学》
the Cancer Research Program, Garvan Institute of Medical Research, St Vincent's Hospital, University of New South Wales
Disciplines of Medicine and Pathology, University of Sydney
Department of Molecular and Clinical Genetics, Royal Prince Alfred Hospital
Department of Anatomical Pathology, Concord Hospital
Department of Colorectal Surgery, Concord Hospital and University of Sydney, Sydney, Australia
ABSTRACT
PURPOSE: The significance of low microsatellite instability (MSI-L) in colorectal cancer is poorly understood. No clear biologic distinction has been found between MSI-L and microsatellite stable (MSS) colorectal cancer, and these two phenotypes are usually combined when analyzed against the well-defined high MSI (MSI-H) phenotype. Evidence is emerging that an O6-methylguanine DNA methyltransferase (MGMT) gene defect is associated with MSI-L. Therefore, to further define this phenotype, we undertook a detailed analysis of the prognostic significance of MSI-L and loss of MGMT expression in colon cancer.
PATIENTS AND METHODS: The study cohort was 183 patients with clinicopathologic stage C colon cancer who had not received adjuvant therapy. We analyzed MSI status, MGMT, and mismatch repair protein expression, as well as MGMT and p16 promoter hypermethylation.
RESULTS: We showed that MSI-L defines a group of patients with poorer survival (P = .026) than MSS patients, and that MSI-L was an independent prognostic indicator (P = .005) in stage C colon cancer. Loss of MGMT protein expression was associated with the MSI-L phenotype but was not a prognostic factor for overall survival in colon cancer. p16 methylation was significantly less frequent in MSI-L than in MSI-H and MSS tumors and was not associated with survival.
CONCLUSION: MSI-L characterizes a distinct subgroup of stage C colon cancer patients, including the MSI-L subset of proximal colon cancer, who have a poorer outcome. Neither the MGMT defect nor p16 methylation are likely to contribute to the worse prognosis of the MSI-L phenotype.
INTRODUCTION
The prognostic significance of microsatellite instability (MSI) has been studied intensely since the discovery of DNA mismatch repair defects in hereditary colorectal cancer.1,2 In common colorectal cancer, the high MSI (MSI-H) phenotype is present in approximately 10% to 15% of patients, predominantly as a result of the MLH1 gene silencing through promoter hypermethylation.3 This results in a somatic defect of DNA mismatch repair, which can be detected using the standardized microsatellite panel and is defined when at least two of the five standard markers show instability in tumor DNA.4,5 MSI-H is a marker of better prognosis in sporadic colorectal cancer,6-9 although some studies have demonstrated only weak or no survival advantage.10-12 MLH1 methylated cancers are a subset of the CpG island methylator phenotype (CIMP), which is characterized by methylation of several genes.13
The biologic defect causing the low MSI (MSI-L) phenotype, on the other hand, is not so well understood, and this phenotype is often combined with microsatellite stable (MSS) colorectal cancer in prognostic studies. At least one study has reported that there is no difference in survival between MSI-L and MSS in stage B2 and C colorectal cancer.7 MSI-L is found in approximately 20% of patients and was initially defined as the group of tumors that show instability in only one of the five standard microsatellite markers or in less than 30% to 40% of a larger number of markers.4,5 Recently, the definition of MSI-L has been questioned, as most tumors could be classified as MSI-L if a large enough number of markers is examined.14-16 In some studies, MSI typing has been carried out using only the mononucleotide markers BAT25 and BAT26, which reliably classify MSI-H. Both MSI-H and MSI-L can be determined using the standard panel of five markers supplemented with MYCL1,17 which is particularly sensitive for MSI-L.18
Although no clear biologic difference has yet been found between MSI-L and MSS colorectal cancer, it has been shown that K-ras mutations,19 O6-methylguanine DNA methyltransferase (MGMT) promoter methylation,17 and loss of MGMT gene expression20 are associated with the MSI-L phenotype. MGMT is a DNA repair enzyme that removes alkylating adducts from the O6-position of guanine. If this repair is defective, guanine is more easily mutated. In colorectal cancer, MGMT methylation is associated with G to A mutations in the K-ras and p53 genes.21,22 On the other hand, cancer cells deficient in MGMT are more sensitive to the cytotoxic effects of alkylating chemotherapy agents. This has been shown to be clinically relevant in the treatment of patients with glioma and large B-cell lymphoma.23,24 Methylation of the MGMT promoter is the most important mechanism by which this gene is silenced in cancer and should be taken into account in cancers where alkylating agents are commonly used. Furthermore, methylation of MGMT is associated with a poor outcome in non–small-cell lung cancer,25 but its prognostic significance in colon cancer is as yet unknown.
Histopathologic stage is the most important independent prognostic variable in colorectal cancer, but clinicopathologic stage C presents with a wide variation in survival.26,27 Currently, the recommended treatment for stage C patients is surgery followed by adjuvant chemotherapy, such as fluorouracil (5-FU). It is, however, unclear why all stage C patients do not benefit from adjuvant chemotherapy, and MSI/CIMP status is currently being investigated as a possible predictive marker for chemotherapeutic response. One study has suggested that the MSI-H phenotype is associated with a favorable response to adjuvant chemotherapy in stage C tumors.11 In contrast, two recent studies have found that 5-FU–based chemotherapy was of no benefit in this group when compared with the combined MSS and MSI-L group of cancers.9,28 This supports previous experimental data that show that mismatch repair deficiency confers a degree of resistance against 5-FU cytotoxicity.29
In this study, we set out to examine the significance of the reported association17 between MGMT methylation and the MSI-L phenotype. We hypothesized that these markers could be informative as prognostic indicators within stage C colon cancer. For MGMT, both promoter hypermethylation and protein expression were analyzed. Furthermore, as p16 is considered representative of the genes hypermethylated in the CIMP phenotype,12,13 we also examined whether there was any relationship between this marker and MSI-L.
PATIENTS AND METHODS
Patients and Specimens
Since 1971, data on all patients undergoing resection for colorectal cancer performed by members of the Concord Hospital Colorectal Surgical Unit have been entered into a prospective database and analyzed for clinicopathologic variables.26,27 The data set contains information on patient characteristics, presentation, investigations, surgical management, pathology, postoperative complications, adjuvant therapy, and follow-up. Since 1979, all data entry has been overseen by a single surgeon (P.H.C.). Between 1971 and 1999 inclusive, there were 260 patients with nonmetachronous stage C colon cancer who had not received adjuvant therapy and who did not have inflammatory bowel disease or familial adenomatous polyposis coli. A suitable specimen was unavailable in 74 patients and MSI was not assessable in three patients, leaving 183 patients for analysis. This group forms a subset of a cohort recently analyzed for adverse histopathologic factors30 and overlaps with a previous collaborative study of MSI-H in stage C colorectal cancer.31 Here we specifically wanted to concentrate on colon cancer, which forms a clinically more uniform group. Therefore, we excluded rectal and sigmoid tumors, which comprised 67% of the previous study,31 and included additional colon cancer cases. There was no significant difference between the 183 patients with an MSI result and the 77 patients for whom MSI could not be assessed in terms of patient age, sex, tumor site or size, direct spread, involvement of a free serosal surface, number of lymph node metastases, venous invasion, and histologic grade. Histopathology was based on a standard protocol, and tumors were staged according to the Australian Clinicopathological Staging System.32 This study had the approval of the Ethics Committee of the Central Sydney Area Health Service.
DNA Preparation
Archival hematoxylin and eosin sections of the primary tumor and lymph node metastases were first re-evaluated, and the most suitable blocks from each patient were chosen for subsequent analysis. Whole sections (4 μm) containing both primary tumor and adjacent normal mucosa were first cut for immunohistochemistry, and subsequent serial sections were manually microdissected for tumor DNA analysis. Sections of matched normal tissue were cut for control DNA specimens. If a matched lymph node metastasis contained a well-defined area of tumor involving at least one low-power field (x10 objective), sections were also cut for DNA analysis and nontumor portions removed. DNA was purified using the Puregene DNA Isolation Kit (Gentra, Minneapolis, MN).
Immunohistochemistry
The following antibodies were used: MLH1 (clone G168-15, BD Pharmingen 554072, San Diego, CA; diluted 1/1000), MSH2 (clone FE11, Oncogene NA27, San Diego, CA; diluted 1/1000), MSH6 (clone 44, BD Pharmingen 610919; diluted 1/1500), PMS2 (clone A16-4, BD Pharmingen 556415; diluted 1/200), MGMT (clone MT5.1, BD Pharmingen 557045; diluted 1/1200). Heat-induced antigen retrieval was required to unmask the antigenic sites. This was performed by heating the slides in a pressure cooker in preheated citrate buffer (10 nmol/L, pH6.0) for 5 minutes (MSH6 and MGMT) or for 10 minutes (MLH1, MSH2, and PMS2). In some cases, the tissue sections were microwave-treated in preheated EDTA buffer for 20 minutes (MSH6 and MGMT) or 40 minutes (MLH1, MSH2, and PMS2).
Immunostaining was carried out on a Sequenza rack with Coverplate (ThermoShandon, Pittsburgh, PA). The slides were treated with 1% goat serum and then incubated with primary antibody overnight at room temperature. On completion of a Tris buffer wash, the slides were incubated with goat antimouse immunoglobulin G-polymeric HRP linker conjugate (Poly-HRP IHC Amplification Reagent; Chemicon AP340P, Temecula, CA) for 30 minutes at room temperature, then washed in buffer. After reaction with 3,3'-diaminobenzidine/hydrogen peroxide (ICN Biomedicals, 980681, Aurora OH) for 5 minutes, the sections were counterstained in Mayer's hematoxylin. All slides were independently reviewed by two histopathologists (B.P.C.L. and C.C.). All 183 cases were analyzed for MLH1, MSH2 and MGMT expression, but staining was not conclusive in two cases for MLH1 and in five cases for MGMT. MSH6 and PMS2 were analyzed in all MSI-H cases that showed normal expression of MSH2 and MLH1.
MSI Analysis
We analyzed the six markers BAT25, BAT26, D2S123, D5S346, D17S250, and MYCL1, which have been used previously to classify both MSI-H and MSI-L.17 One primer from each primer pair had a 5' fluorescent tag. Polymerase chain reaction (PCR) amplification was performed in a total volume of 25 μL containing 25 to 100 ng of DNA, 250 μmol/L of dNTPs, 35 to 55 ng of primers, and 2U of Amplitaq Gold (Applied Biosystems, Foster City, CA). PCR was carried out at 94°C (30 seconds), 56°C (45 seconds), and 72°C (30 to 90 seconds) for 35 cycles (5 minutes initial denaturation and 7 minutes final elongation) with the GeneAmp PCR system 2400 (Applied Biosystems). Fragments were visualized using ABI 310 Genetic Analyzer (Applied Biosystems) capillary electrophoresis, and the microsatellite patterns of the tumors were compared with matching normal tissue. Specimens were scored as MSI-H if there was a significant shift in the microsatellite pattern for two or more of the markers and these included at least one of the mononucleotide repeats. If both of the mononucleotide markers were clearly normal and MSI was only seen in the dinucleotide repeats or the complex repeat MYCL1, the specimen was classified as MSI-L. If none of the six markers displayed MSI, the specimen was classified as MSS.
Methylation Specific PCR
Approximately 1 μg aliquots of the tumor and normal DNA specimens were bisulphite-treated as previously described.33 For MGMT, each specimen was amplified using primers specific for both the methylated and the unmethylated promoter sequences (predicted fragment sizes 81bp and 93bp, respectively).34 PCR conditions were 95°C (30 seconds), 59°C (30 seconds), and 72°C (30 seconds) for 40 cycles (methylated reaction). Unmodified DNA controls were used for PCR optimization to check for nonspecific amplification. For the unmethylated reaction, the annealing temperature was decreased from 64°C to 60°C at one-degree intervals (45 seconds, two cycles each) and then 59°C (45 seconds, 33 cycles). The p16 promoter was amplified using primers for the methylated and unmethylated sequences as previously described (predicted fragment sizes, 150 bp and 151 bp, respectively.33 PCR conditions were 95°C (30 seconds), 65°C methylated, 60°C unmethylated (30 seconds), 72°C (30 seconds) for 40 cycles. All PCR reactions were carried out using AmpliTaq Gold (Applied Biosystems) in a DNA engine DYAD (MJ Research, Waltham, MA) or the GeneAmp PCR system 2400 (Applied Biosystems). The cycling conditions included an initial denaturation step at 95°C for 12 minutes and final elongation at 72°C for 7 minutes. All methylation-specific PCR products were visualized on 6% polyacrylamide gels.
Statistical Analysis
The 2 test was used to evaluate the significance of differences in contingency tables. Patients were followed annually until death or February 2001. Overall survival was defined as time from the date of resection to the date of death from any cause, censored patients being those alive at the close of the study or lost to follow-up. Seven patients were lost to follow-up. In the 55 patients remaining alive at the close of the study, the minimum follow-up time was 15.9 months and the median was 108.0 months. Comparisons of survival time between strata of categoric variables were made by the Kaplan-Meier method and log-rank test. Regression analyses used Cox's method. The assumption of proportional hazards was checked by examining log cumulative hazard plots for parallelism. The level for statistical significance was set at .05.
RESULTS
Assessment of MSI and Protein Expression of Mismatch Repair Genes
MSI-H was observed in 42 (23.0%) and MSI-L in 51 (27.9%) of the 183 primary tumors studied. In a subgroup of 141 tumors proximal to the splenic flexure, the frequencies were 28.4% and 24.8% for MSI-H and MSI-L, respectively. In the overall series with both colon and rectal tumors, it was previously found that the frequency of MSI-H was 9%.31 Of the 41 MSI-H tumors where MLH1 was informative, 29 cases showed complete loss of MLH1 expression, and two cases had loss of MSH2. Ten MSI-H cases that did not show loss of expression for MLH1 and MSH2 were studied for MSH6 and PMS2. Eight of these showed loss (n = 5) or reduced expression (n = 3) of PMS2. No cases of MSH6 loss were identified. All MSI-L and MSS tumors showed positive expression for both MLH1 and MSH2.
Immunohistochemical Analysis of MGMT Expression
Complete loss of MGMT expression was demonstrated in 40 tumors, local loss of expression was demonstrated in four tumors, and reduced expression was demonstrated in a further five tumors. MGMT protein loss was more frequent in MSI-L specimens (42%) than in MSS (19%; P = .004). The MSI-H group was intermediate, with 30% of primary tumors showing loss or reduced expression of MGMT (Table 1).
Analysis of MGMT and p16 Promoter Hypermethylation
MGMT hypermethylation was detected in 53.0% (96 of 181) and p16 hypermethylation was detected in 34.9% (52 of 149) of the tumors. MGMT methylation was slightly more common in MSI-H (60%) and MSI-L (56%) than in MSS tumors (48%), but this was not statistically significant (Table 1). p16 methylation was lower in MSI-L (16%) than in MSS (36%; P = .029) or in MSI-H primary tumors (53%; P = .001). There was no association of p16 methylation with either MGMT methylation (P = .985) or loss of MGMT protein expression (P = .349). In the 100 patients where MGMT methylation was assessed in both primary cancer and lymph node tissue, there was concordance in 73 patients (73%). In 76 patients where p16 methylation was assessed, there was concordance in 54 patients (71%).
Both MGMT methylation and MGMT protein expression were successfully analyzed in 176 primary tumors. There were 43 cases where loss or reduced expression of MGMT corresponded with the presence of methylation. Five cases with MGMT loss did not show methylation indicating a second less common mechanism of MGMT gene down-regulation, possibly through sequestration of the MGMT enhancer binding protein.35 In a further 51 patients, MGMT methylation was detected but there was no loss or reduced protein expression.
MSI-L Is Associated With Shorter Survival
In bivariate Kaplan-Meier analysis, patients with the MSI-H and MSS phenotypes did not differ significantly in overall survival (P = .243), whereas MSI-L was clearly a marker of poorer prognosis than MSS (P = .026; Fig 1). In the subgroup of 141 patients with cancer proximal to the splenic flexure, MSI-L was also associated with shorter survival (P = .023). In multivariate Cox regression analysis, after exclusion of MSI-H tumors, MSI-L was significantly (P = .005) and independently associated with poorer prognosis compared with MSS tumors (Table 2), after adjustment for factors previously shown to be associated with the survival of stage C patients in this series27 and other potentially relevant variables. Some of these variables were not statistically significant in this study, probably due to smaller size of the cohort and the fact that rectal cancers were not analyzed. On the other hand, the result for tumor size (Table 2) may be further influenced by the exclusion of MSI-H patients from the multivariate analysis, as MSI-H is associated with larger tumor size.31 Although MSI-L is correlated with loss of MGMT protein expression, this was not associated with poorer survival in this cohort (P = .548; Fig 2). There was also no association between p16 hypermethylation and survival (P = .966).
DISCUSSION
This study has shown for the first time that MSI-L is associated with poorer survival in clinicopathologic stage C colon cancer, including the MSI-L subset of proximal colon cancer. This study is perhaps overdue, given the fact that the existence and definition of the mild mutator phenotype in colorectal cancer has been questioned. In many prognostic studies, the MSI-L phenotype has not been considered as a separate category, possibly due to the problem that a clear single marker characterizing this group of tumors is lacking. However, there is now emerging clear evidence that MSI-L tumors form a distinct molecular group when gene expression data from microarray experiments are analyzed using principal components analysis.20 This is consistent with some earlier molecular findings, for example, that K-ras mutations19 and MGMT gene promoter methylation17 are more prevalent in MSI-L tumors.
A recent large study demonstrated that CIMP may be a more relevant prognostic factor than MSI.12 CIMP is associated with adverse prognosis, but this is not observed in those patients that also have MSI-H tumor. No study has yet analyzed the relationship between MSI-L and CIMP, both of which define a patient group with poorer survival. It is possible that there is significant overlap between the two phenotypes, both of which are based on a similar definition using marker genes, although measuring two biologically different characteristics. In this study we analyzed p16 methylation as one of the genes that contribute to the CIMP phenotype. Interestingly, we observed that p16 methylation frequency was significantly lower in MSI-L tumors compared with both MSI-H and MSS tumors, which would argue against a large overlap of MSI-L with CIMP. In contrast, our study shows that MSI-L tumors have more frequent loss of MGMT protein expression, and this is associated with promoter hypermethylation. It has been hypothesized that inactivation of this gene can overload the mismatch repair system, resulting in the mild mutator phenotype,17 possibly mediated by an increase in methylG:T mismatches.36 However, in this patient cohort, neither the MGMT defect or p16 methylation showed any prognostic significance. Therefore, further studies are required to explore a possible functional link between loss of MGMT expression and the MSI-L phenotype as well as to pinpoint the gene defects responsible for the adverse prognosis associated with MSI-L cancers.
Although we have confirmed the association between MSI-L and the MGMT defect, we did not demonstrate an association between MGMT promoter methylation as such and the MSI-L phenotype. The frequency of MGMT methylation was the same in MSI-H and MSI-L tumors. This is probably a reflection of the higher frequency of MGMT methylation compared with detectable loss of protein expression. A previous study also reported that 26% of methylated cancers did not show immunohistochemical loss of MGMT expression.17 This raises the question of the biologic significance of MGMT methylation in primary tumors. It is possible that in some cancers, methylation is not biallelic, leaving one allele actively expressing the protein. Furthermore, the PCR method used here is so sensitive that it is able to detect small clusters of cells with MGMT promoter methylation among a majority of cells with normal MGMT promoter. Immunohistochemistry is a more robust method and may not detect small clusters of cells that have lost protein expression. Larger clusters of loss or reduced expression are more easily detected, as was demonstrated in our study, where only one of the eight specimens in this category did not show MGMT methylation. In contrast, we confirmed a previous finding that methylation of the p16 gene promoter was more frequent in MSI-H.12
Finally, this study may have implications for the treatment of stage C colon cancers that display the MSI-L phenotype. It is becoming well established that 5-FU–based adjuvant chemotherapy offers improved survival in non–MSI-H colorectal cancer as compared with patients who have MSI-H tumors.9,28 These studies analyzed MSI-L and MSS as one group, demonstrating improved overall survival with adjuvant chemotherapy. Against this background, it can be speculated that adjuvant treatment may be potentially valuable in MSI-L cancers, but our study once again raises the issue of the correct classification of this phenotype. There has been much debate on differentiating MSI-L from MSS but it is also very easy to misclassify MSI-L as MSI-H. This is particularly relevant in proximal colon cancer, where MSI-H is known to be more frequent than in distal (including rectal) cancer. We placed greater emphasis on the instability found in the mononucleotide markers BAT25 and BAT26 when determining MSI-H status. In only two of 41 MSI-H tumors classified this way, no reduced expression could be demonstrated for MLH1, MSH2, MSH6, or PMS2. No MSI-L tumors showed loss of MLH1 or MSH2 further supporting the view that this phenotype has a different biologic basis. To make better informed clinical decisions about treatment, the typing for at least the two BAT markers or immunohistochemistry for mismatch repair proteins therefore remains essential to reliably classify all MSI-H cases.
In conclusion, we have shown that MSI-L should not just be considered as an intermediate linear category between MSI-H and MSS but as a distinct phenotype that has adverse prognostic significance in clinicopathologic stage C colon cancer. We have also shown that MSI-L is inversely associated with p16 hypermethylation and have further confirmed the association with an MGMT defect. However, the MGMT defect is unlikely to contribute to the worse prognosis. It remains imperative to further determine the relationship between the MSI-L and other methylated genes in the CIMP phenotype and to find new genetic markers to reliably identify those patients who have a poorer prognosis.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank Ronald C. Newland for his expert advice on histopathology; Robert Sutherland and Bing Yu for valuable advice; and Jawad Saab, Han Qin, Huong Le, Andrew Kennedy, and Jacqui Burke for technical assistance.
NOTES
Supported by Strathfield Private Hospital Strathfield, Sydney, Australia.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Fishel R, Lescoe MK, Rao MRS, et al: The human mutator gene homolog MSH2 and its association with hereditary nonpolyposis colon cancer. Cell 75:1027-1038, 1993
Leach FS, Nicolaides NC, Papadopoulos N, et al: Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell 75:1215-1225, 1993
Kane MF, Loda M, Gaida GM, et al: Methylation of the hMLH1 promoter correlates with lack of expression of hMLH1 in sporadic colon tumors and mismatch repair-defective human tumor cell lines. Cancer Res 57:808-811, 1997
Dietmaier W, Wallinger S, Bocker T, et al: Diagnostic microsatellite instability: Definition and correlation with mismatch repair protein expression. Cancer Res 57:4749-4756, 1997
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
Thibodeau SN, Bren G, Schaid D: Microsatellite instability in cancer of the proximal colon. Science 260:816-819, 1993
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
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
Salahshor S, Kressner U, Fischer H, et al: Microsatellite instability in sporadic colorectal cancer is not an independent prognostic factor. Br J Cancer 81:190-193, 1999
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
Ward RL, Cheong K, Ku SL, et al: Adverse prognostic effect of methylation in colorectal cancer is reversed by microsatellite instability. J Clin Oncol 21:3729-3736, 2003
Toyota M, Ahuja N, Ohe-Toyota M, et al: CpG island methylator phenotype in colorectal cancer. Proc Natl Acad Sci U S A 96:8681-8686, 1999
Halford S, Sasieni P, Rowan A, et al: Low-level microsatellite instability occurs in most colorectal cancers and is a nonrandomly distributed quantitative trait. Cancer Res 62:53-57, 2002
Laiho P, Launonen V, Lahermo P, et al: Low-level microsatellite instability in most colorectal carcinomas. Cancer Res 62:1166-1170, 2002
Tomlinson I, Halford S, Aaltonen L, et al: Does MSI-low exist J Pathol 197:6-13, 2002
Whitehall VL, Walsh MD, Young J, et al: Methylation of O-6-methylguanine DNA methyltransferase characterizes a subset of colorectal cancer with low-level DNA microsatellite instability. Cancer Res 61:827-830, 2001
Iino H, Jass JR, Simms LA, et al: DNA microsatellite instability in hyperplastic polyps, serrated adenomas, and mixed polyps: A mild mutator pathway for colorectal cancer J Clin Pathol 52:5-9, 1999
Jass JR, Biden KG, Cummings MC, et al: Characterisation of a subtype of colorectal cancer combining features of the suppressor and mild mutator pathways. J Clin Pathol 52:455-460, 1999
Mori Y, Selaru FM, Sato F, et al: The impact of microsatellite instability on the molecular phenotype of colorectal tumors. Cancer Res 63:4577-4582, 2003
Esteller M, Toyota M, Sanchez-Cespedes M, et al: Inactivation of the DNA repair gene O6-methylguanine-DNA methyltransferase by promoter hypermethylation is associated with G to A mutations in K-ras in colorectal tumorigenesis. Cancer Res 60:2368-2371, 2000
Esteller M, Risques RA, Toyota M, et al: Promoter hypermethylation of the DNA repair gene O(6)-methylguanine-DNA methyltransferase is associated with the presence of G: C to A:T transition mutations in p53 in human colorectal tumorigenesis. Cancer Res 61:4689-4692, 2001
Esteller M, Garcia-Foncillas J, Andion E, et al: Inactivation of the DNA-repair gene MGMT and the clinical response of gliomas to alkylating agents. N Engl J Med 343:1350-1354, 2000
Esteller M, Gaidano G, Goodman SN, et al: Hypermethylation of the DNA repair gene O(6)-methylguanine DNA methyltransferase and survival of patients with diffuse large B-cell lymphoma. J Natl Cancer Inst 94:26-32, 2002
Brabender J, Usadel H, Metzger R, et al: Quantitative O(6)-methylguanine DNA methyltransferase methylation analysis in curatively resected non-small cell lung cancer: Associations with clinical outcome. Clin Cancer Res 9:223-227, 2003
Newland RC, Chapuis PH, Pheils MT, et al: The relationship of survival to staging and grading of colorectal carcinoma: A prospective study of 503 cases. Cancer 47:1424-1429, 1981
Newland RC, Dent OF, Lyttle MNB, et al: Pathologic determinants of survival associated with colorectal cancer with lymph node metastases. Cancer 73:2076-2082, 1994
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
Carethers JM, Chauhan DP, Fink D, et al: Mismatch repair proficiency and in vitro response to 5-fluorouracil. Gastroenterology 117:123-131, 1999
Chapuis PH, Dent OF, Bokey EL, et al: Adverse histopathological findings as a guide to patient management after curative resection of node-positive colonic cancer. Br J Surg 91:349-354, 2004
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
Fielding LP, Arsenault PA, Chapuis PH, et al: Clinicopathological staging for colorectal cancer: An International Documentation System (IDS) and an International Comprehensive Anatomical Terminology (ICAT). J Gastroenterol Hepatol 6:325-344, 1991
Herman JG, Graff JR, Myohanen S, et al: Methylation-specific PCR: A novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci U S A 93:9821-9826, 1996
Esteller M, Hamilton SR, Burger PC, et al: Inactivation of the DNA repair gene O6-methylguanine-DNA methyltransferase by promoter hypermethylation is a common event in primary human neoplasia. Cancer Res 59:793-797, 1999
Chen FY, Harris LC, Remack JS, et al: Cytoplasmic sequestration of an O6-methylguanine-DNA methyltransferase enhancer binding protein in DNA repair-deficient human cells. Proc Natl Acad Sci U S A 94:4348-4353, 1997
Jass JR, Whitehall VL, Young J, et al: Correspondence re: P. Laiho et al., Low-level microsatellite instability in most colorectal carcinomas, Cancer Res 62:1166-1170, 2002. Cancer Res 62:5988-5989, 2002(Maija R.J. Kohonen-Corish)
Disciplines of Medicine and Pathology, University of Sydney
Department of Molecular and Clinical Genetics, Royal Prince Alfred Hospital
Department of Anatomical Pathology, Concord Hospital
Department of Colorectal Surgery, Concord Hospital and University of Sydney, Sydney, Australia
ABSTRACT
PURPOSE: The significance of low microsatellite instability (MSI-L) in colorectal cancer is poorly understood. No clear biologic distinction has been found between MSI-L and microsatellite stable (MSS) colorectal cancer, and these two phenotypes are usually combined when analyzed against the well-defined high MSI (MSI-H) phenotype. Evidence is emerging that an O6-methylguanine DNA methyltransferase (MGMT) gene defect is associated with MSI-L. Therefore, to further define this phenotype, we undertook a detailed analysis of the prognostic significance of MSI-L and loss of MGMT expression in colon cancer.
PATIENTS AND METHODS: The study cohort was 183 patients with clinicopathologic stage C colon cancer who had not received adjuvant therapy. We analyzed MSI status, MGMT, and mismatch repair protein expression, as well as MGMT and p16 promoter hypermethylation.
RESULTS: We showed that MSI-L defines a group of patients with poorer survival (P = .026) than MSS patients, and that MSI-L was an independent prognostic indicator (P = .005) in stage C colon cancer. Loss of MGMT protein expression was associated with the MSI-L phenotype but was not a prognostic factor for overall survival in colon cancer. p16 methylation was significantly less frequent in MSI-L than in MSI-H and MSS tumors and was not associated with survival.
CONCLUSION: MSI-L characterizes a distinct subgroup of stage C colon cancer patients, including the MSI-L subset of proximal colon cancer, who have a poorer outcome. Neither the MGMT defect nor p16 methylation are likely to contribute to the worse prognosis of the MSI-L phenotype.
INTRODUCTION
The prognostic significance of microsatellite instability (MSI) has been studied intensely since the discovery of DNA mismatch repair defects in hereditary colorectal cancer.1,2 In common colorectal cancer, the high MSI (MSI-H) phenotype is present in approximately 10% to 15% of patients, predominantly as a result of the MLH1 gene silencing through promoter hypermethylation.3 This results in a somatic defect of DNA mismatch repair, which can be detected using the standardized microsatellite panel and is defined when at least two of the five standard markers show instability in tumor DNA.4,5 MSI-H is a marker of better prognosis in sporadic colorectal cancer,6-9 although some studies have demonstrated only weak or no survival advantage.10-12 MLH1 methylated cancers are a subset of the CpG island methylator phenotype (CIMP), which is characterized by methylation of several genes.13
The biologic defect causing the low MSI (MSI-L) phenotype, on the other hand, is not so well understood, and this phenotype is often combined with microsatellite stable (MSS) colorectal cancer in prognostic studies. At least one study has reported that there is no difference in survival between MSI-L and MSS in stage B2 and C colorectal cancer.7 MSI-L is found in approximately 20% of patients and was initially defined as the group of tumors that show instability in only one of the five standard microsatellite markers or in less than 30% to 40% of a larger number of markers.4,5 Recently, the definition of MSI-L has been questioned, as most tumors could be classified as MSI-L if a large enough number of markers is examined.14-16 In some studies, MSI typing has been carried out using only the mononucleotide markers BAT25 and BAT26, which reliably classify MSI-H. Both MSI-H and MSI-L can be determined using the standard panel of five markers supplemented with MYCL1,17 which is particularly sensitive for MSI-L.18
Although no clear biologic difference has yet been found between MSI-L and MSS colorectal cancer, it has been shown that K-ras mutations,19 O6-methylguanine DNA methyltransferase (MGMT) promoter methylation,17 and loss of MGMT gene expression20 are associated with the MSI-L phenotype. MGMT is a DNA repair enzyme that removes alkylating adducts from the O6-position of guanine. If this repair is defective, guanine is more easily mutated. In colorectal cancer, MGMT methylation is associated with G to A mutations in the K-ras and p53 genes.21,22 On the other hand, cancer cells deficient in MGMT are more sensitive to the cytotoxic effects of alkylating chemotherapy agents. This has been shown to be clinically relevant in the treatment of patients with glioma and large B-cell lymphoma.23,24 Methylation of the MGMT promoter is the most important mechanism by which this gene is silenced in cancer and should be taken into account in cancers where alkylating agents are commonly used. Furthermore, methylation of MGMT is associated with a poor outcome in non–small-cell lung cancer,25 but its prognostic significance in colon cancer is as yet unknown.
Histopathologic stage is the most important independent prognostic variable in colorectal cancer, but clinicopathologic stage C presents with a wide variation in survival.26,27 Currently, the recommended treatment for stage C patients is surgery followed by adjuvant chemotherapy, such as fluorouracil (5-FU). It is, however, unclear why all stage C patients do not benefit from adjuvant chemotherapy, and MSI/CIMP status is currently being investigated as a possible predictive marker for chemotherapeutic response. One study has suggested that the MSI-H phenotype is associated with a favorable response to adjuvant chemotherapy in stage C tumors.11 In contrast, two recent studies have found that 5-FU–based chemotherapy was of no benefit in this group when compared with the combined MSS and MSI-L group of cancers.9,28 This supports previous experimental data that show that mismatch repair deficiency confers a degree of resistance against 5-FU cytotoxicity.29
In this study, we set out to examine the significance of the reported association17 between MGMT methylation and the MSI-L phenotype. We hypothesized that these markers could be informative as prognostic indicators within stage C colon cancer. For MGMT, both promoter hypermethylation and protein expression were analyzed. Furthermore, as p16 is considered representative of the genes hypermethylated in the CIMP phenotype,12,13 we also examined whether there was any relationship between this marker and MSI-L.
PATIENTS AND METHODS
Patients and Specimens
Since 1971, data on all patients undergoing resection for colorectal cancer performed by members of the Concord Hospital Colorectal Surgical Unit have been entered into a prospective database and analyzed for clinicopathologic variables.26,27 The data set contains information on patient characteristics, presentation, investigations, surgical management, pathology, postoperative complications, adjuvant therapy, and follow-up. Since 1979, all data entry has been overseen by a single surgeon (P.H.C.). Between 1971 and 1999 inclusive, there were 260 patients with nonmetachronous stage C colon cancer who had not received adjuvant therapy and who did not have inflammatory bowel disease or familial adenomatous polyposis coli. A suitable specimen was unavailable in 74 patients and MSI was not assessable in three patients, leaving 183 patients for analysis. This group forms a subset of a cohort recently analyzed for adverse histopathologic factors30 and overlaps with a previous collaborative study of MSI-H in stage C colorectal cancer.31 Here we specifically wanted to concentrate on colon cancer, which forms a clinically more uniform group. Therefore, we excluded rectal and sigmoid tumors, which comprised 67% of the previous study,31 and included additional colon cancer cases. There was no significant difference between the 183 patients with an MSI result and the 77 patients for whom MSI could not be assessed in terms of patient age, sex, tumor site or size, direct spread, involvement of a free serosal surface, number of lymph node metastases, venous invasion, and histologic grade. Histopathology was based on a standard protocol, and tumors were staged according to the Australian Clinicopathological Staging System.32 This study had the approval of the Ethics Committee of the Central Sydney Area Health Service.
DNA Preparation
Archival hematoxylin and eosin sections of the primary tumor and lymph node metastases were first re-evaluated, and the most suitable blocks from each patient were chosen for subsequent analysis. Whole sections (4 μm) containing both primary tumor and adjacent normal mucosa were first cut for immunohistochemistry, and subsequent serial sections were manually microdissected for tumor DNA analysis. Sections of matched normal tissue were cut for control DNA specimens. If a matched lymph node metastasis contained a well-defined area of tumor involving at least one low-power field (x10 objective), sections were also cut for DNA analysis and nontumor portions removed. DNA was purified using the Puregene DNA Isolation Kit (Gentra, Minneapolis, MN).
Immunohistochemistry
The following antibodies were used: MLH1 (clone G168-15, BD Pharmingen 554072, San Diego, CA; diluted 1/1000), MSH2 (clone FE11, Oncogene NA27, San Diego, CA; diluted 1/1000), MSH6 (clone 44, BD Pharmingen 610919; diluted 1/1500), PMS2 (clone A16-4, BD Pharmingen 556415; diluted 1/200), MGMT (clone MT5.1, BD Pharmingen 557045; diluted 1/1200). Heat-induced antigen retrieval was required to unmask the antigenic sites. This was performed by heating the slides in a pressure cooker in preheated citrate buffer (10 nmol/L, pH6.0) for 5 minutes (MSH6 and MGMT) or for 10 minutes (MLH1, MSH2, and PMS2). In some cases, the tissue sections were microwave-treated in preheated EDTA buffer for 20 minutes (MSH6 and MGMT) or 40 minutes (MLH1, MSH2, and PMS2).
Immunostaining was carried out on a Sequenza rack with Coverplate (ThermoShandon, Pittsburgh, PA). The slides were treated with 1% goat serum and then incubated with primary antibody overnight at room temperature. On completion of a Tris buffer wash, the slides were incubated with goat antimouse immunoglobulin G-polymeric HRP linker conjugate (Poly-HRP IHC Amplification Reagent; Chemicon AP340P, Temecula, CA) for 30 minutes at room temperature, then washed in buffer. After reaction with 3,3'-diaminobenzidine/hydrogen peroxide (ICN Biomedicals, 980681, Aurora OH) for 5 minutes, the sections were counterstained in Mayer's hematoxylin. All slides were independently reviewed by two histopathologists (B.P.C.L. and C.C.). All 183 cases were analyzed for MLH1, MSH2 and MGMT expression, but staining was not conclusive in two cases for MLH1 and in five cases for MGMT. MSH6 and PMS2 were analyzed in all MSI-H cases that showed normal expression of MSH2 and MLH1.
MSI Analysis
We analyzed the six markers BAT25, BAT26, D2S123, D5S346, D17S250, and MYCL1, which have been used previously to classify both MSI-H and MSI-L.17 One primer from each primer pair had a 5' fluorescent tag. Polymerase chain reaction (PCR) amplification was performed in a total volume of 25 μL containing 25 to 100 ng of DNA, 250 μmol/L of dNTPs, 35 to 55 ng of primers, and 2U of Amplitaq Gold (Applied Biosystems, Foster City, CA). PCR was carried out at 94°C (30 seconds), 56°C (45 seconds), and 72°C (30 to 90 seconds) for 35 cycles (5 minutes initial denaturation and 7 minutes final elongation) with the GeneAmp PCR system 2400 (Applied Biosystems). Fragments were visualized using ABI 310 Genetic Analyzer (Applied Biosystems) capillary electrophoresis, and the microsatellite patterns of the tumors were compared with matching normal tissue. Specimens were scored as MSI-H if there was a significant shift in the microsatellite pattern for two or more of the markers and these included at least one of the mononucleotide repeats. If both of the mononucleotide markers were clearly normal and MSI was only seen in the dinucleotide repeats or the complex repeat MYCL1, the specimen was classified as MSI-L. If none of the six markers displayed MSI, the specimen was classified as MSS.
Methylation Specific PCR
Approximately 1 μg aliquots of the tumor and normal DNA specimens were bisulphite-treated as previously described.33 For MGMT, each specimen was amplified using primers specific for both the methylated and the unmethylated promoter sequences (predicted fragment sizes 81bp and 93bp, respectively).34 PCR conditions were 95°C (30 seconds), 59°C (30 seconds), and 72°C (30 seconds) for 40 cycles (methylated reaction). Unmodified DNA controls were used for PCR optimization to check for nonspecific amplification. For the unmethylated reaction, the annealing temperature was decreased from 64°C to 60°C at one-degree intervals (45 seconds, two cycles each) and then 59°C (45 seconds, 33 cycles). The p16 promoter was amplified using primers for the methylated and unmethylated sequences as previously described (predicted fragment sizes, 150 bp and 151 bp, respectively.33 PCR conditions were 95°C (30 seconds), 65°C methylated, 60°C unmethylated (30 seconds), 72°C (30 seconds) for 40 cycles. All PCR reactions were carried out using AmpliTaq Gold (Applied Biosystems) in a DNA engine DYAD (MJ Research, Waltham, MA) or the GeneAmp PCR system 2400 (Applied Biosystems). The cycling conditions included an initial denaturation step at 95°C for 12 minutes and final elongation at 72°C for 7 minutes. All methylation-specific PCR products were visualized on 6% polyacrylamide gels.
Statistical Analysis
The 2 test was used to evaluate the significance of differences in contingency tables. Patients were followed annually until death or February 2001. Overall survival was defined as time from the date of resection to the date of death from any cause, censored patients being those alive at the close of the study or lost to follow-up. Seven patients were lost to follow-up. In the 55 patients remaining alive at the close of the study, the minimum follow-up time was 15.9 months and the median was 108.0 months. Comparisons of survival time between strata of categoric variables were made by the Kaplan-Meier method and log-rank test. Regression analyses used Cox's method. The assumption of proportional hazards was checked by examining log cumulative hazard plots for parallelism. The level for statistical significance was set at .05.
RESULTS
Assessment of MSI and Protein Expression of Mismatch Repair Genes
MSI-H was observed in 42 (23.0%) and MSI-L in 51 (27.9%) of the 183 primary tumors studied. In a subgroup of 141 tumors proximal to the splenic flexure, the frequencies were 28.4% and 24.8% for MSI-H and MSI-L, respectively. In the overall series with both colon and rectal tumors, it was previously found that the frequency of MSI-H was 9%.31 Of the 41 MSI-H tumors where MLH1 was informative, 29 cases showed complete loss of MLH1 expression, and two cases had loss of MSH2. Ten MSI-H cases that did not show loss of expression for MLH1 and MSH2 were studied for MSH6 and PMS2. Eight of these showed loss (n = 5) or reduced expression (n = 3) of PMS2. No cases of MSH6 loss were identified. All MSI-L and MSS tumors showed positive expression for both MLH1 and MSH2.
Immunohistochemical Analysis of MGMT Expression
Complete loss of MGMT expression was demonstrated in 40 tumors, local loss of expression was demonstrated in four tumors, and reduced expression was demonstrated in a further five tumors. MGMT protein loss was more frequent in MSI-L specimens (42%) than in MSS (19%; P = .004). The MSI-H group was intermediate, with 30% of primary tumors showing loss or reduced expression of MGMT (Table 1).
Analysis of MGMT and p16 Promoter Hypermethylation
MGMT hypermethylation was detected in 53.0% (96 of 181) and p16 hypermethylation was detected in 34.9% (52 of 149) of the tumors. MGMT methylation was slightly more common in MSI-H (60%) and MSI-L (56%) than in MSS tumors (48%), but this was not statistically significant (Table 1). p16 methylation was lower in MSI-L (16%) than in MSS (36%; P = .029) or in MSI-H primary tumors (53%; P = .001). There was no association of p16 methylation with either MGMT methylation (P = .985) or loss of MGMT protein expression (P = .349). In the 100 patients where MGMT methylation was assessed in both primary cancer and lymph node tissue, there was concordance in 73 patients (73%). In 76 patients where p16 methylation was assessed, there was concordance in 54 patients (71%).
Both MGMT methylation and MGMT protein expression were successfully analyzed in 176 primary tumors. There were 43 cases where loss or reduced expression of MGMT corresponded with the presence of methylation. Five cases with MGMT loss did not show methylation indicating a second less common mechanism of MGMT gene down-regulation, possibly through sequestration of the MGMT enhancer binding protein.35 In a further 51 patients, MGMT methylation was detected but there was no loss or reduced protein expression.
MSI-L Is Associated With Shorter Survival
In bivariate Kaplan-Meier analysis, patients with the MSI-H and MSS phenotypes did not differ significantly in overall survival (P = .243), whereas MSI-L was clearly a marker of poorer prognosis than MSS (P = .026; Fig 1). In the subgroup of 141 patients with cancer proximal to the splenic flexure, MSI-L was also associated with shorter survival (P = .023). In multivariate Cox regression analysis, after exclusion of MSI-H tumors, MSI-L was significantly (P = .005) and independently associated with poorer prognosis compared with MSS tumors (Table 2), after adjustment for factors previously shown to be associated with the survival of stage C patients in this series27 and other potentially relevant variables. Some of these variables were not statistically significant in this study, probably due to smaller size of the cohort and the fact that rectal cancers were not analyzed. On the other hand, the result for tumor size (Table 2) may be further influenced by the exclusion of MSI-H patients from the multivariate analysis, as MSI-H is associated with larger tumor size.31 Although MSI-L is correlated with loss of MGMT protein expression, this was not associated with poorer survival in this cohort (P = .548; Fig 2). There was also no association between p16 hypermethylation and survival (P = .966).
DISCUSSION
This study has shown for the first time that MSI-L is associated with poorer survival in clinicopathologic stage C colon cancer, including the MSI-L subset of proximal colon cancer. This study is perhaps overdue, given the fact that the existence and definition of the mild mutator phenotype in colorectal cancer has been questioned. In many prognostic studies, the MSI-L phenotype has not been considered as a separate category, possibly due to the problem that a clear single marker characterizing this group of tumors is lacking. However, there is now emerging clear evidence that MSI-L tumors form a distinct molecular group when gene expression data from microarray experiments are analyzed using principal components analysis.20 This is consistent with some earlier molecular findings, for example, that K-ras mutations19 and MGMT gene promoter methylation17 are more prevalent in MSI-L tumors.
A recent large study demonstrated that CIMP may be a more relevant prognostic factor than MSI.12 CIMP is associated with adverse prognosis, but this is not observed in those patients that also have MSI-H tumor. No study has yet analyzed the relationship between MSI-L and CIMP, both of which define a patient group with poorer survival. It is possible that there is significant overlap between the two phenotypes, both of which are based on a similar definition using marker genes, although measuring two biologically different characteristics. In this study we analyzed p16 methylation as one of the genes that contribute to the CIMP phenotype. Interestingly, we observed that p16 methylation frequency was significantly lower in MSI-L tumors compared with both MSI-H and MSS tumors, which would argue against a large overlap of MSI-L with CIMP. In contrast, our study shows that MSI-L tumors have more frequent loss of MGMT protein expression, and this is associated with promoter hypermethylation. It has been hypothesized that inactivation of this gene can overload the mismatch repair system, resulting in the mild mutator phenotype,17 possibly mediated by an increase in methylG:T mismatches.36 However, in this patient cohort, neither the MGMT defect or p16 methylation showed any prognostic significance. Therefore, further studies are required to explore a possible functional link between loss of MGMT expression and the MSI-L phenotype as well as to pinpoint the gene defects responsible for the adverse prognosis associated with MSI-L cancers.
Although we have confirmed the association between MSI-L and the MGMT defect, we did not demonstrate an association between MGMT promoter methylation as such and the MSI-L phenotype. The frequency of MGMT methylation was the same in MSI-H and MSI-L tumors. This is probably a reflection of the higher frequency of MGMT methylation compared with detectable loss of protein expression. A previous study also reported that 26% of methylated cancers did not show immunohistochemical loss of MGMT expression.17 This raises the question of the biologic significance of MGMT methylation in primary tumors. It is possible that in some cancers, methylation is not biallelic, leaving one allele actively expressing the protein. Furthermore, the PCR method used here is so sensitive that it is able to detect small clusters of cells with MGMT promoter methylation among a majority of cells with normal MGMT promoter. Immunohistochemistry is a more robust method and may not detect small clusters of cells that have lost protein expression. Larger clusters of loss or reduced expression are more easily detected, as was demonstrated in our study, where only one of the eight specimens in this category did not show MGMT methylation. In contrast, we confirmed a previous finding that methylation of the p16 gene promoter was more frequent in MSI-H.12
Finally, this study may have implications for the treatment of stage C colon cancers that display the MSI-L phenotype. It is becoming well established that 5-FU–based adjuvant chemotherapy offers improved survival in non–MSI-H colorectal cancer as compared with patients who have MSI-H tumors.9,28 These studies analyzed MSI-L and MSS as one group, demonstrating improved overall survival with adjuvant chemotherapy. Against this background, it can be speculated that adjuvant treatment may be potentially valuable in MSI-L cancers, but our study once again raises the issue of the correct classification of this phenotype. There has been much debate on differentiating MSI-L from MSS but it is also very easy to misclassify MSI-L as MSI-H. This is particularly relevant in proximal colon cancer, where MSI-H is known to be more frequent than in distal (including rectal) cancer. We placed greater emphasis on the instability found in the mononucleotide markers BAT25 and BAT26 when determining MSI-H status. In only two of 41 MSI-H tumors classified this way, no reduced expression could be demonstrated for MLH1, MSH2, MSH6, or PMS2. No MSI-L tumors showed loss of MLH1 or MSH2 further supporting the view that this phenotype has a different biologic basis. To make better informed clinical decisions about treatment, the typing for at least the two BAT markers or immunohistochemistry for mismatch repair proteins therefore remains essential to reliably classify all MSI-H cases.
In conclusion, we have shown that MSI-L should not just be considered as an intermediate linear category between MSI-H and MSS but as a distinct phenotype that has adverse prognostic significance in clinicopathologic stage C colon cancer. We have also shown that MSI-L is inversely associated with p16 hypermethylation and have further confirmed the association with an MGMT defect. However, the MGMT defect is unlikely to contribute to the worse prognosis. It remains imperative to further determine the relationship between the MSI-L and other methylated genes in the CIMP phenotype and to find new genetic markers to reliably identify those patients who have a poorer prognosis.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank Ronald C. Newland for his expert advice on histopathology; Robert Sutherland and Bing Yu for valuable advice; and Jawad Saab, Han Qin, Huong Le, Andrew Kennedy, and Jacqui Burke for technical assistance.
NOTES
Supported by Strathfield Private Hospital Strathfield, Sydney, Australia.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Fishel R, Lescoe MK, Rao MRS, et al: The human mutator gene homolog MSH2 and its association with hereditary nonpolyposis colon cancer. Cell 75:1027-1038, 1993
Leach FS, Nicolaides NC, Papadopoulos N, et al: Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell 75:1215-1225, 1993
Kane MF, Loda M, Gaida GM, et al: Methylation of the hMLH1 promoter correlates with lack of expression of hMLH1 in sporadic colon tumors and mismatch repair-defective human tumor cell lines. Cancer Res 57:808-811, 1997
Dietmaier W, Wallinger S, Bocker T, et al: Diagnostic microsatellite instability: Definition and correlation with mismatch repair protein expression. Cancer Res 57:4749-4756, 1997
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
Thibodeau SN, Bren G, Schaid D: Microsatellite instability in cancer of the proximal colon. Science 260:816-819, 1993
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
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
Salahshor S, Kressner U, Fischer H, et al: Microsatellite instability in sporadic colorectal cancer is not an independent prognostic factor. Br J Cancer 81:190-193, 1999
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
Ward RL, Cheong K, Ku SL, et al: Adverse prognostic effect of methylation in colorectal cancer is reversed by microsatellite instability. J Clin Oncol 21:3729-3736, 2003
Toyota M, Ahuja N, Ohe-Toyota M, et al: CpG island methylator phenotype in colorectal cancer. Proc Natl Acad Sci U S A 96:8681-8686, 1999
Halford S, Sasieni P, Rowan A, et al: Low-level microsatellite instability occurs in most colorectal cancers and is a nonrandomly distributed quantitative trait. Cancer Res 62:53-57, 2002
Laiho P, Launonen V, Lahermo P, et al: Low-level microsatellite instability in most colorectal carcinomas. Cancer Res 62:1166-1170, 2002
Tomlinson I, Halford S, Aaltonen L, et al: Does MSI-low exist J Pathol 197:6-13, 2002
Whitehall VL, Walsh MD, Young J, et al: Methylation of O-6-methylguanine DNA methyltransferase characterizes a subset of colorectal cancer with low-level DNA microsatellite instability. Cancer Res 61:827-830, 2001
Iino H, Jass JR, Simms LA, et al: DNA microsatellite instability in hyperplastic polyps, serrated adenomas, and mixed polyps: A mild mutator pathway for colorectal cancer J Clin Pathol 52:5-9, 1999
Jass JR, Biden KG, Cummings MC, et al: Characterisation of a subtype of colorectal cancer combining features of the suppressor and mild mutator pathways. J Clin Pathol 52:455-460, 1999
Mori Y, Selaru FM, Sato F, et al: The impact of microsatellite instability on the molecular phenotype of colorectal tumors. Cancer Res 63:4577-4582, 2003
Esteller M, Toyota M, Sanchez-Cespedes M, et al: Inactivation of the DNA repair gene O6-methylguanine-DNA methyltransferase by promoter hypermethylation is associated with G to A mutations in K-ras in colorectal tumorigenesis. Cancer Res 60:2368-2371, 2000
Esteller M, Risques RA, Toyota M, et al: Promoter hypermethylation of the DNA repair gene O(6)-methylguanine-DNA methyltransferase is associated with the presence of G: C to A:T transition mutations in p53 in human colorectal tumorigenesis. Cancer Res 61:4689-4692, 2001
Esteller M, Garcia-Foncillas J, Andion E, et al: Inactivation of the DNA-repair gene MGMT and the clinical response of gliomas to alkylating agents. N Engl J Med 343:1350-1354, 2000
Esteller M, Gaidano G, Goodman SN, et al: Hypermethylation of the DNA repair gene O(6)-methylguanine DNA methyltransferase and survival of patients with diffuse large B-cell lymphoma. J Natl Cancer Inst 94:26-32, 2002
Brabender J, Usadel H, Metzger R, et al: Quantitative O(6)-methylguanine DNA methyltransferase methylation analysis in curatively resected non-small cell lung cancer: Associations with clinical outcome. Clin Cancer Res 9:223-227, 2003
Newland RC, Chapuis PH, Pheils MT, et al: The relationship of survival to staging and grading of colorectal carcinoma: A prospective study of 503 cases. Cancer 47:1424-1429, 1981
Newland RC, Dent OF, Lyttle MNB, et al: Pathologic determinants of survival associated with colorectal cancer with lymph node metastases. Cancer 73:2076-2082, 1994
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
Carethers JM, Chauhan DP, Fink D, et al: Mismatch repair proficiency and in vitro response to 5-fluorouracil. Gastroenterology 117:123-131, 1999
Chapuis PH, Dent OF, Bokey EL, et al: Adverse histopathological findings as a guide to patient management after curative resection of node-positive colonic cancer. Br J Surg 91:349-354, 2004
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
Fielding LP, Arsenault PA, Chapuis PH, et al: Clinicopathological staging for colorectal cancer: An International Documentation System (IDS) and an International Comprehensive Anatomical Terminology (ICAT). J Gastroenterol Hepatol 6:325-344, 1991
Herman JG, Graff JR, Myohanen S, et al: Methylation-specific PCR: A novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci U S A 93:9821-9826, 1996
Esteller M, Hamilton SR, Burger PC, et al: Inactivation of the DNA repair gene O6-methylguanine-DNA methyltransferase by promoter hypermethylation is a common event in primary human neoplasia. Cancer Res 59:793-797, 1999
Chen FY, Harris LC, Remack JS, et al: Cytoplasmic sequestration of an O6-methylguanine-DNA methyltransferase enhancer binding protein in DNA repair-deficient human cells. Proc Natl Acad Sci U S A 94:4348-4353, 1997
Jass JR, Whitehall VL, Young J, et al: Correspondence re: P. Laiho et al., Low-level microsatellite instability in most colorectal carcinomas, Cancer Res 62:1166-1170, 2002. Cancer Res 62:5988-5989, 2002(Maija R.J. Kohonen-Corish)