2004 Landes Bioscience. Not for distribution.

Transcription

1 [Cancer Biology & Therapy 3:1, 73-78, January 2004]; 2004 Landes Bioscience Research Paper Evaluation of Microsatellite Instability, hmlh1 Expression and hmlh1 Promoter Hypermethylation in Defining the MSI Phenotype of Colorectal Cancer Christian N. Arnold 1, Ajay Goel 1, Carolyn Compton 2 Victoria Marcus 2 Donna Niedzwiecki 3 Jeannette M. Dowell Linda Wasserman 1 Toru Inoue 1 Robert J. Mayer 4 Monica M Bertagnolli 5 C. Richard Boland 1,6, * 1 Department of Medicine and Cancer Center; University of California San Diego; La Jolla, California USA 2 Department of Pathology; McGill University; Montreal, Quebec, Canada 3 CALGB Statistical Center; Duke University; Durham, North Carolina USA 4 Department of Medicine; Dana Farber Cancer Institute; Boston, Massachusetts USA 5 Brigham and Women s Hospital; Boston, Massachusetts USA 6 San Diego Veteran Affairs Medical Center; La Jolla, California USA These authors contributed equally to this study *Correspondence to: C. Richard Boland; Baylor University Medical Center; 2 Hoblitzelle, 3500 Gaston Avenue; Dallas, Texas USA; Tel.: ; Fax: ; RickBo@BaylorHealth.Edu Received 07/29/03; Accepted 09/05/03 Previously published online as a Cancer Biology & Therapy E-publication: KEY WORDS hmlh1, microsatellite instability, promoter methylation, colorectal cancer, immunohistochemistry ACKNOWLEDGEMENTS The research for CALGB study 9865 was supported, in part, by grants from the National Cancer Institute (CA31946) to the Cancer and Leukemia Group B (Richard L. Schilsky, Chairman). This work was also sponsored by a grant from the NIH (RO1-CA 72851) to C.R.B and in part by an American Cancer Society-IRG award to A.G. C.N.A. was funded by a grant of the Dr. Mildred- Scheel-Stiftung, Germany. The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute. For further information regarding who participated in this study, please see table on p. 78. ABSTRACT Introduction: About 15% of all colorectal cancers (CRCs) demonstrate high levels of microsatellite instability (MSI-H) and are currently best identified by molecular analysis of microsatellite markers. Most sporadic CRCs with MSI-H are known to be associated with the methylation of the hmlh1 promoter. Promoter methylation coincided with lack of hmlh1 expression. We aimed to investigate the association between MSI status, hmlh1 protein expression and methylation status of the hmlh1 promoter, and to determine the usefulness of each method in defining the MSI phenotype in sporadic CRCs. Materials and Methods: CRCs from 173 patients from the Cancer and Leukemia Group B (CALGB) were assessed for their MSI status. An additional cohort of 18 MSI-H tumors from the University of California San Diego (UCSD) was included in the analysis of the MSI-H subgroup. MSI testing was performed by PCR using five standard MSI markers. hmlh1 promoter analysis was investigated by methylation specific PCR (MSP), and expression of the MMR genes hmlh1 and hmsh2 was examined by immunohistochemistry (IHC). Results: Of the 173 CALGB tumors, 111 (64%) were MSS, 35 (20%) were MSI-L and 27 (16%) MSI-H, respectively. Data on hmlh1 protein expression, hmsh2 protein expression and hmlh1 methylation are available on 128, 173 and 81 of these tumors, respectively. Presence of hmlh1 and hmsh2 protein expression was significantly associated with MSI status. Four of 45 (8.9%) MSI-H tumors and 0 of 146 (0%) MSS/MSI-L tumors did not express hmsh2 (p = ). hmlh1 protein expression was present in 107 of 108 (99%) MSS and MSI-L tumors versus 11 of 20 (55%) MSI-H tumors (p < ). Of 61 MSS and MSI-L cancers studied for methylation, 11 (18%) were methylated at the hmlh1 promoter whereas 14 of 20 (70%) MSI-H cancers were methylated (p = ). In 27 MSI-H tumors studied for hmlh1 protein expression and methylation, 93% of tumors with loss of expression (93%) were also methylated while 42% (5/12) with positive immunostaining for hmlh1 were methylated at the hmlh1 promoter (p = 0.009). Conclusions: Promoter methylation and hmlh1 expression are significantly associated with the MSI-H phenotype in CRC. Promoter methylation analysis provides a useful means to screen for MSI-H tumors. Our data further suggests that hmlh1 promoter methylation analysis alone cannot replace MSI testing, as a significant number of MSI-H tumors could be potentially overseen by such an approach. We suggest that phenotypic evaluation of CRC is performed most reliably with MSI testing, although expression analysis and investigation of the promoter methylation status may complement the screening process. INTRODUCTION CRC pathogenesis proceeds through two well defined pathways of genomic instability, termed as the suppressor and mutator pathways. These pathways are characterized by successive accumulation of genetic events in neoplastic cells as a function of time. In most human neoplasias, the first mechanism is observed at the chromosomal level and leads to gene alterations by chromosomal gains and losses, accompanied by mutations at specific tumor suppressor genes and oncogenes. 1 The second mechanism was first described in 1993 by observing size variations of short repetitive DNA sequences in tumor DNA, termed microsatellites. 2-4 It was soon linked to hereditary non-polyposis colon cancer (HNPCC) and termed microsatellite instability (MSI) phenotype. 5 MSI is the consequence of failure of the DNA mismatch repair (MMR) system to faithfully correct errors occurring during DNA replication. MSI in HNPCC is mainly caused by germline mutations in two major MMR genes hmlh1 and hmsh2, and to a lesser extent by mutations in hmsh6, hpms2 and hmlh3. 6 Chromosomal abnormalities are infrequent in MSI-H cancers, and the tumor cells are most commonly diploid. Of all sporadic CRCs, 10 15% Cancer Biology & Therapy 73

2 possess the MSI-H phenotype. It has been demonstrated that the MMR defect in most of these tumors is caused by de novo methylation of the hmlh1 promoter. 7 Only MSI-H sporadic colon tumors were shown to demonstrate extensive hmlh1 promoter methylation, while a minor fraction of MSI-L and MSS cancers appear to be methylated at the hmlh1 promoter. 8 Promoter methylation causes transcriptional silencing of the hmlh1 gene resulting in loss of protein expression. 9 Tumors progressing through the mutator or the suppressor pathway differ in respect to their molecular and clinico-pathological features. It has been reported that MSS/MSI-L tumors, in contrast to MSI-H tumors, occur most commonly in the left colon and rectum and harbor k-ras, APC, and p53 mutations more frequently, and exhibit silencing of the DNA repair gene O-6-methylguanine DNA methyltransferase (O 6 -MGMT) through promoter methylation. 10,11 MSI-H tumors are predominantly located in the proximal colon and are often characterized histologically by poor differentiation, lymphocytic infiltration and extracellular mucin production. Both MSI-H and MSS/MSI-L tumors appear to arise from hyperplastic polyps evolving to serrated adenomas. 12,13,14 Despite their histology, HNPCC related cancers and sporadic MSI-H cancers tend to have a better outcome in terms of disease free and overall survival. 15 In contrast, data from in vitro cell models showed that MSI-H tumors have resistance against a variety of clinically used chemotherapeutic drugs These features necessitate to accurately diagnose and distinguish cancers among the mutator and the suppressor pathways for future evaluations and individualized treatment strategies. In the recent years, the gold standard in diagnosing MSI-H cancers has been MSI testing using standard PCR-methods. Several studies have evaluated various MSI markers with different degrees of sensitivity and specificity in their ability to diagnose MSI. 19,20 In 1998, five Bethesda MSI markers were recommended by a National Cancer Institute workshop on MSI, which are considered to be able to detect MSI-H cancers with high reliability. 21 hmlh1 expression determined by immunohistochemistry (IHC) has been compared to MSI testing in sporadic CRC and HNPCC, and a correlation was found in detecting the loss of hmlh1 expression and MSI-H. 22,23,24 However, these strategies could not exclude the possibility of missing some cases of MSI-H cancers when IHC was used as a solitary technique. Recently, much emphasis has also been placed on studying hmlh1 methylation in CRCs. Methylation specific PCR (MSP) appears to be the method of choice to reliably study the methylation status of different gene promoters. 7,25 Several studies proved that methylation of the hmlh1 promoter strongly correlates with the loss of its protein expression. Despite the use of all these methods to distinguish MSI-H cancers from other sporadic CRCs, there is no consensus opinion as to which strategy would be most specific and efficient for defining the MSI-H phenotype. We therefore intended to investigate the correlation among conventional MSI testing, protein expression of the two major MMR genes hmlh1 and hmsh2 and the promoter methylation status of hmlh1 in sporadic CRC to define the MSI phenotype. In an ideal scenario, MSI-H cancers would lack hmlh1 expression and demonstrate hmlh1 promoter methylation, which would facilitate screening methods by simply using standard and cost effective MSP techniques for categorizing CRC according to their pathway of origin. MATERIALS AND METHODS Tissue Selection and DNA Extraction. A total of 173 sporadic CRC as well as corresponding normal mucosa were collected from CALGB. For the further correlation of MSI-H tumors with hmlh1 methylation status and hmlh1 and hmsh2 protein expression, 18 formerly characterized MSI-H cancers from UCSD were included into the analyses. 26 Five micrometer thick serial sections from the paraffin-embedded matched normal and neoplastic primary tissues were stained with H&E, and representative normal and tumor regions were identified by microscopic examination. The reference slide was compared to an unstained slide and tumor areas and adjacent normal tissues were microdissected separately. Genomic DNA was isolated from the paraffin-embedded microdomains using GeneReleaser (Bioventure laboratories, CA). Subsequently, the DNA samples were incubated overnight at 55 C in a lysis buffer containing proteinase K and used as a template DNA for PCR analysis after heat inactivation of proteinase K for 15 minutes. Microsatellite Analysis. Five MSI markers were utilized, consisting of three dinucleotide repeats (D2S123, D5S346, D17S250) and two mononucleotide repeats (BAT25 and BAT26). In brief, genomic DNA was amplified and labeled with the radioisotope P 32 by PCR cycles with a denaturation step at 94 C for 1 min, an annealing step of C for sec, and an elongation step of 72 0 C for 1 min in a total of 5 µl of reaction mixture (0.5 µm of each primer, 1.5 mm MgCl 2, 0.2 mm of each deoxynucleotide triphosphate, 0.5U of Taq-DNA polymerase, and 0.5 µl of [γ-p]datp 10 µci/ml). The radioisotope-labeled MSI sequences were visualized by denaturing gel electrophoresis and autoradiography. Each PCR was repeated twice to ensure the reproducibility of results in cases where the band shifts were not clearly informative in the first attempt. Tumors with a shift in at least two of the five markers were classified as being MSI-H. MSI-L was defined as a shift in only one of the five markers. Tumors not showing allelic shifts were termed MSS. The scoring was undertaken independently by two authors (C.N.A. and A.G.) as described previously. 26 hmlh1 MSP. Methylation status of the hmlh1 promoter was determined by MSP. 27 The DNA methylation pattern within the CpG islands of the promoter was determined by chemical modification of unmethylated but not the methylated cytosines to uracil and subsequent PCR using specific primers for either methylated or the modified unmethylated DNA. DNA (1µg) in a volume of 50 µl was denatured by NaOH for 10 min at 37 C. Thirty microliters of 10 mm hydroquinone (Sigma) and 520 µl of 3 M sodium bisulfite (Sigma) at ph 5.5 were added, and samples were incubated at 50 C for h. The modified DNA was purified using the Wizard DNA purification resin (Promega, Madison, WI) according to the manufacturer s specifications. DNA was further treated with NaOH for 10 min at room temperature and precipitated with ethanol at -80 C. The specific primers for the methylated and unmethylated MSP were as follows: hmlh1-m 5 -ACGTAGACGTTTTATTAGGGTCGC-3 (forward), 5 - CCTCATCGTAACTACCCGCG-3 (backward); hmlh1-u 5 -TTTTG- ATGTAGATGTTTTATTAGGGTTGT-3 (forward), 5 -ACCACCTCAT- CATAACTACCCACA 3 (backward). The PCR mixture contained 1x PCR buffer ((NH 4 ) 2 SO mm, Tris ph mm, MgCl mm, β- mercaptoethanol 10 mm), 0.2 mm of each dntp, primers (500 ng each), 0.2 U Taq-polymerase (Gibco) and 50 ng bisulfite-modified or unmodified DNA. PCR reactions were hot-started at 95 C for 5 min. Amplification was performed for 35 cycles (30 sec at 95 C, 30 sec at 64 C, 30 sec at 72 C) followed by a final extension for 10 min at 72 C. The PCR product was electrophoresed on a 2.5 % agarose gel, stained with 0.5 µg/ml ethidium bromide, and visualized under UV-illumination. Immunohistochemical Analysis. Immunohistochemical staining was performed using mouse monoclonal antibodies against hmlh1 (clone G168-15, 0.5mg/ml, PharMingen, San Diego, CA) and hmsh2 (clone FE11, 0.1mg/ml, Oncogene Research Products, Boston, MA). Sections (5 µm) of formalin-fixed, paraffin-embedded tissue were picked up on charged slides and dried overnight at 37 C. Immunohistochemical staining was manually performed using the avidin-biotin complex (ABC) method. 74 Cancer Biology & Therapy 2004; Vol. 3 Issue 1

3 Table 1 DISTRIBUTION OF THE MSI PHENOTYPE MSS MSI-L MSI-H tumors (n=173) (n=111) (n=35) (n=27) 64% 20% 16% Briefly, sections were deparaffinized, and endogenous peroxidase blocked using 0.3% H 2 O 2 in methanol. Heat-induced antigen retrieval was performed in 10mM critrate buffer (ph 6.0) in a microwave oven for 15 min. Sections were incubated at 37 C for 1 hour with primary antibody and avidin-biotin complex (Histostain-Plus Bulk Kits, Zymed Laboratories Inc., San Francisco, CA). Sections were developed using the chromogen 3-amino- 9-ethylcarbazole, counterstained with hematoxylin, and mounted in aqueous medium. Negative control sections were treated in an identical manner with the exception of omission of the primary antibody. Normal colonic gland epithelium served as the internal positive control for both antibodies. The immunohistochemical staining for hmlh1 and hmsh2 was scored by two independent observers (CC and VM). The normal staining pattern for both antibodies is nuclear. Cancers with complete loss of the nuclear staining compared to normal tissue were scored as negative. In cases of disagreement, the slides were reviewed, and a consensus view was achieved. Staining for hmlh1 expression analysis was attempted in all tumors but was successful in 85 of 111 MSS tumors, 23 of 35 MSI-L tumors and 20 of 27 MSI-H tumors. MSH2 immunostaining was attempted and successful in all tumors. Cases that had inadequate internal control tissue or inadequate tumor in the section or in which the staining was technically inadequate were eliminated. Statistical Analysis. Fisher s exact test and the chi-square test were used as appropriate to test the associations between hmlh1 and hmsh2 protein expression and MSI status and to test associations within the MSI-H subgroup. The analyses involving MSI status were performed on 173 CALGB tumor samples. Analyses within the MSI-H subgroup included data from an additional 18 UCSD tumor samples. Sample sizes in each analysis varied according to the numbers of samples analysed for each of MSI status, hmlh1 and hmsh2 protein expression and hmlh1 methylation. All statistical analyses were performed at the CALGB statistical center. RESULTS MSI Status, Correlation with hmlh1 Promoter Methylation and MMR-Gene Expression. Of the 173 tumors from CALGB, 111 (64%) were MSS, 35 (20%) were MSI-L and 27 (16%) were MSI-H (Table 1). In order to rule out further undetected MSI, we performed MSI analysis with additional mononucleotide (TGFβRII, IGFIIR, BAX, hmsh3, hmsh6), di-nucleotide (D3S1029, D17S261, D18S64, D18S69 and D18S474) and tetra-nucleotide markers (MYCL1). Even by adding these markers, the overall MSI pattern did not change. We further investigated the methylation status of the hmlh1 promoter. We were able to successfully perform the MSP assay in 20 MSI-H and 27 MSI-L cancers (Fig. 1). Thirty-five MSS tumors were non-randomly selected for MSP analysis of which 34 were successful. As expected, most of the MSS (28/34, 82%) and MSI-L (22/27, 81%) cancers were unmethylated at the hmlh1 promoter, while 6 of 20 (30%) MSI-H cancers were unmethylated (p<0.0001) (Fig. 2). Subsequently, we performed expression analysis of the MMR proteins hmlh1 and hmsh2 in the CALGB tumor cohort (Table 2, Fig. 3). Of 108 MSS and MSI-L cancers investigated, 107 (99%) expressed hmlh1. All but one tumor were unmethylated. Of the MSI-H cancers, 11/20 (55%) expressed hmlh1 Figure 1. MSP of the hmlh1 promoter region in sporadic CRCs. Bisulfite treated DNA was amplified with hmlh1 methylated and unmethylated specific primers. The PCR product indicated by U represents an unmethylated allele, while the PCR product indicated by M represents a methylated allele. H2O represents a blank control, T1 is completely methylated, T2 is unmethylated and T3 carries both a methylated and an unmethylated allele. The lower bands represent nonspecific PCR products. (p<0.0001). Of the methylated MSI-H cases, 6/14 (43%) expressed hmlh1 protein while 5/6 (83%) of the unmethylated MSI-H were positive for hmlh1 staining. Sensitivity of the immunostaining in detecting MSI- H tumors was 45% and specificity was 99%, respectively. We further correlated hmlh1 expression with the hmlh1 promoter methylation status of the MSI-H tumors. Eighteen UCSD MSI-H tumors were included in this analysis. 26 Of the MSI-H cases, 4 lacked expression of the hmsh2 protein. These 4 tumors were not included in this analysis, because they were considered to belong to the HNPCC spectrum. Of 45 total MSI-H tumors, 14 were missing data on one or both of hmlh1 methylation and protein expression and 4 were negative on hmsh2. The 27 remaining MSI-H tumors were included in this analysis. Of 15 MSI-H cancers with loss of hmlh1 expression, 14 (93%) were also methylated while 42% MSI-H cancers (5/12) with positive immunostaining for hmlh1 were methylated at the hmlh1 promoter (p = 0.009) (Table 3). We thus were able to demonstrate the close relationship of hmlh1 protein expression with the hmlh1 promoter methylation status. Ability of MSI Markers to Determine the MSI Phenotype. Previous studies have reported a high accuracy of specific MSI markers in detecting the MSI-H phenotype, especially for the highly monomorphic mononucleotide marker BAT We compared the sensitivity and specificity of each individual marker of the Bethesda panel for its ability to define the MSI-H (n = 45) phenotype based on all 5 markers. All 45 MSI-H tumors (from CALGB and UCSD) were included in this analysis. We found that no individual marker on its own was able to reliably detect the MSI-H phenotype. Sensitivity of the individual MSI markers in detecting MSI-H was Table 2 hmlh1 EXPRESSION IN ALL TUMORS INVESTIGATED MSS MSI-L MSI-H tumors (n = 128) (n = 84) (n = 24) (n = 20) 99% 100% 55% Figure 2. Methylation status according to the degree of MSI. The hmlh1 promoter methylation was investigated by MSP in each tumor group. Cancer Biology & Therapy 75

4 Table 3 CORRELATION OF MMR GENE EXPRESSION AND METHYLATION STATUS IN MSI-H CANCERS (INCLUDING CALGB AND UCSD CASES) Figure 3. hmlh1 expression. hmlh1 protein expression was investigated in all tumor groups by IHC. highest for the dinucleotide markers D17S250 and D5S346 with 76% and 73%, respectively. Specificity was best for BAT26 and BAT25 with 99% and 95%, respectively (Table 4). DISCUSSION Molecular classification of colorectal cancer based on its molecular features has important implications regarding prognosis and might influence future treatment strategies. It has been shown that MSI-H sporadic colon cancers have in vitro resistance to commonly used chemotherapeutic drugs compared to cancers with the MSS and MSI-L phenotype. 17 It remains to be determined if this is also relevant in vivo. Up to now, only one retrospective study showed that 5-FU based chemotherapy might influence overall and disease free survival negatively in patients with stage II and III CRCs. 15 The MSI-H phenotype arises through the defects of the human MMR system, notably through transcriptional silencing of the hmlh1 gene by aberrant methylation of its promoter. 9 Promoter methylation has been demonstrated to inhibit hmlh1 protein expression. 25 So far, MSI-H cancers have been diagnosed by MSI testing using a panel of consensus MSI markers. 21 By definition, tumors with band shifts in 30 40% or more of the tested MSI markers are scored as MSI-H. However, no study has yet systematically compared the correlation of MSI-H sporadic colon cancers diagnosed by conventional MSI testing with the hmlh1 promoter methylation status and hmlh1 protein expression, respectively. It seems feasible that searching for cancers with a silenced hmlh1 gene by promoter methylation could all be performed by simply employing the MSP technique instead of PCR testing with 5 standard MSI markers. With this strategy, screening for these cancers could be facilitated. In our study we assessed the benefits of MSI testing, hmlh1 promoter methylation analysis and hmlh1 protein expression for the determination of sporadic CRC with the MSI-H phenotype. 23,29,30 First, we demonstrated the expected frequency of the MSI phenotype in sporadic colorectal cancer with 16% of all cancers being MSI-H. Interestingly, only 45% of these cancers showed absence of hmlh1 expression. Earlier studies have estimated sensitivity between 75% and 100% for negative immunostaining and the MSI-H phenotype. However, in our study only 1 of 108 MSS/MSI-L cancers did not express hmlh1 protein suggesting a high specificity of MMR gene expression in non-msi-h cancers. This is in accordance with others, all showing a specificity of 100% of immunohistochemistry in correctly diagnosing the MSS phenotype in cancers with positive hmlh1 and hmsh2 expression. 22,23,29-31 Despite Tumors Methylated Unmethylated Total Lack of hmlh1 14 (93%) 1 (7%) 15 expression hmlh1 expression 5 (42%) 7 (58%) 12 Total 19 (70%) 8 (30%) 27 this, an important issue still remains as to how the difference in sensitivity of immunohistochemistry in MSI-H cancers demonstrated in our and previous studies can be best explained. One issue is the definition of the MSI phenotype itself. It has been convincingly shown that the MSI-H phenotype is correlated in most cases with aberrant methylation of the hmlh1 promoter which influences the transcriptional activity of the gene. We therefore investigated the promoter methylation status of our entire tumor cohort with the MSI status. We found that most MSS/MSI-L cancers had an unmethylated and, presumably, transcriptionally active hmlh1 promoter. In 99% of these cases, the tumors were shown to express hmlh1 and hmsh2 proteins by immunohistochemistry. However, 18% of the MSS and 19% of the MSI-L cancers showed methylation. Surprisingly, only 70% of the MSI-H cancers demonstrated promoter hypermethylation. Correlating the immunohistochemistry findings and the promoter methylation status in MSI-H cancers, we found a strong association between negative immunohistochemistry and promoter methylation. However, the association between hmlh1 expression and promoter methylation status was not absolute, as 7% of the MSI-H tumors lacking hmlh1 expression were unmethylated and 42% of the hmlh1 expressing neoplasms were methylated. Although it would have been interesting to obtain data on DNA sequencing on those MSI-H cases that retained hmlh1 or hmsh2 expression in order to investigate if those tumors harbor somatic missense mutations that inactivate protein function. Unfortunately the amount of tumor DNA available did not allow us to perform such analysis for the current study. All MSI-H tumors in this study were defined using Bethesda criteria. Our findings raise two issues: 1) whether this definition correctly categorizes all MSI tumors; and 2) whether all tumors with the MSI-H phenotype develop through the same molecular mechanisms. It has been uniformly assumed that hmlh1 promoter methylation is responsible for transcriptional silencing and is related to the MSI-H phenotype. In our study, only 70% of the MSI-H cancers are hypermethylated and 45% of the MSI-H cancers expressed hmlh1. Some of these cases can be explained by defects of other MMR genes such as hmsh2 (4 cases in this study). Presumably, these were not sporadic cancers but rather belonged to Table 4 SENSITIVITY AND SPECIFICITY OF INDIVIDUAL MSI MARKER Sensitivity Specificity D2S123 69% 85% D5S346 76% 79% D17S250 73% 85% BAT25 67% 95% BAT26 71% 99% 76 Cancer Biology & Therapy 2004; Vol. 3 Issue 1

5 the familial tumor syndrome HNPCC. Our findings are in agreement with a recent study showing that only 47% of MSI positive sporadic colorectal tumors with absent hmlh1 expression were also hypermethylated at the hmlh1 promoter. 24 It is not known what role other MMR or DNA repair genes play in the development of the MSI-H phenotype. So far, it has not been demonstrated that any of the known MMR genes plays a major role in causing the MSI-H phenotype in sporadic colorectal cancer. Only the O 6 -MGMT gene, when methylated and transcriptionally silenced, has been clearly linked to cancers with the MSI-L, but not the MSI-H, phenotype. 10 The question of how the MSI phenotype should be truly defined remains unanswered. We demonstrated that most cancers which did not show immunoreactivity for hmlh1 or hmsh2 were also methylated for hmlh1. Conversely, most of the hmlh1 and hmsh2 expressing MSI-H cancers were unmethylated. MSI-H was diagnosed using the reference panel of 5 MSI standard markers and no significant difference was evident between the instability pattern of hmlh1 expressing and non-expressing cancers with respect to the markers tested. All MSI markers used in the study had a high specificity but a rather low sensitivity, suggesting that no marker on its own was able to reliably detect all MSI-H cancers. However, when performing MSI analysis comparing MSI-H and hmlh1 hypermethylated or MSI-H and unmethylated cancers, all markers but the dinucleotide repeat D2S123 had a high detection rate up to 90% for this selection of cancers. MSI-H cancers might originate from different pathways, one being caused by silencing of the hmlh1 gene and others by yet unrecognized causes. The issue which should be further addressed in future studies is the extent and amount of promoter methylation within the hmlh1 promoter causing transcriptional silencing of the gene. Earlier studies reported a significant concordance between methylation of a 5 promoter region (also called region A) with the MSI-H phenotype. 7 Others reported a further downstream area of the promoter (region C) being more closely related to MSI-H cancers. 32 However, their study was performed in various colorectal tumor cell lines and may not be directly extrapolated to the in vivo setting. 33 One recent study examined the methylation status of the whole hmlh1 promoter and concluded that MSI-H cancers usually possess a full methylation pattern of the whole promoter rather than a partial one. 34 We performed methylation analysis of the promoter region A in agreement with most other studies. From the findings of this study it seems feasible to continue MSI testing with standard MSI analysis. Promoter methylation analysis and staining for MMR gene expression are not yet capable by themselves to distinguish MSI-H from MSS/MSI-L cancers. However, they are able to complement the screening. References 1. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990; 61: Aaltonen LA, Peltomaki P, Leach FS, Sistonen P, Pylkkanen L, Mecklin JP, et al. Clues to the pathogenesis of familial colorectal cancer. Science 1993; 260: Ionov Y, Peinado MA, Malkhosyan S, Shibata D, Perucho M. Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature 1993; 363: Thibodeau SN, Bren G, Schaid D. Microsatellite instability in cancer of the proximal colon. Science 1993; 260: Marra G, Boland CR. DNA repair and colorectal cancer. Gastroenterol Clin North Am 1996; 25: Boland CR. Molecular genetics of hereditary nonpolyposis colorectal cancer. Ann N Y Acad Sci 2000; 910: Herman JG, Umar A, Polyak K, Graff JR, Ahuja N, Issa JP, et al. Incidence and functional consequences of hmlh1 promoter hypermethylation in colorectal carcinoma. Proc Natl Acad Sci USA 1998; 95: Laiho P, Launonen V, Lahermo P, Esteller M, Guo M, Herman JG, et al. Low-level microsatellite instability in most colorectal carcinomas. Cancer Res 2002; 62: Kane MF, Loda M, Gaida GM, Lipman J, Mishra R, Goldman H, 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 1997; 57: Whitehall VL, Walsh MD, Young J, Leggett BA, Jass JR. Methylation of O-6-methylguanine DNA methyltransferase characterizes a subset of colorectal cancer with low-level DNA microsatellite instability. Cancer Res 2001; 61: Jass JR, Biden KG, Cummings MC, Simms LA, Walsh M, Schoch E, et al. Characterisation of a subtype of colorectal cancer combining features of the suppressor and mild mutator pathways. J Clin Pathol 1999; 52: Chao A, Gilliland F, Willman C, Joste N, Chen IM, Stone N, et al. Patient and tumor characteristics of colon cancers with microsatellite instability: A population-based study. Cancer Epidemiol Biomarkers Prev 2000; 9: Hawkins NJ, Ward RL. Sporadic colorectal cancers with microsatellite instability and their possible origin in hyperplastic polyps and serrated adenomas. J Natl Cancer Inst 2001; 93: Jass JR, Iino H, Ruszkiewicz A, Painter D, Solomon MJ, Koorey DJ, et al. Neoplastic progression occurs through mutator pathways in hyperplastic polyposis of the colorectum. Gut 2000; 47: Ribic CM, Sargent DJ, Moore MJ, Thibodeau SN, French AJ, Goldberg RM, et al. Tumor microsatellite-instability status as a predictor of benefit from fluorouracil-based adjuvant chemotherapy for colon cancer. N Engl J Med 2003; 349: Aebi S, Kurdi-Haidar B, Gordon R, Cenni B, Zheng H, Fink D, et al. Loss of DNA mismatch repair in acquired resistance to cisplatin. Cancer Res 1996; 56: Carethers JM, Chauhan DP, Fink D, Nebel S, Bresalier RS, Howell SB, et al. Mismatch repair proficiency and in vitro response to 5-fluorouracil. Gastroenterology 1999; 117: Arnold CN, Goel A, Boland CR. Role of hmlh1 hypermethylation in drug resistance to 5-fluorouracil in colorectal cancer cell lines. Int J Cancer 2003; 106: Bocker T, Diermann J, Friedl W, Gebert J, Holinski-Feder E, Karner-Hanusch J, et al. Microsatellite instability analysis: A multicenter study for reliability and quality control. Cancer Res 1997; 57: Dietmaier W, Wallinger S, Bocker T, Kullmann F, Fishel R, Ruschoff J. Diagnostic microsatellite instability: Definition and correlation with mismatch repair protein expression. Cancer Res 1997; 57: Boland CR, Thibodeau SN, Hamilton SR, Sidransky D, Eshleman JR, Burt RW, 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 1998; 58: Chaves P, Cruz C, Lage P, Claro I, Cravo M, Leitao CN, et al. Immunohistochemical detection of mismatch repair gene proteins as a useful tool for the identification of colorectal carcinoma with the mutator phenotype. J Pathol 2000; 191: Lindor NM, Burgart LJ, Leontovich O, Goldberg RM, Cunningham JM, Sargent DJ, et al. Immunohistochemistry versus microsatellite instability testing in phenotyping colorectal tumors. J Clin Oncol 2002; 20: Salahshor S, Koelble K, Rubio C, Lindblom A. Microsatellite instability and hmlh1 and hmsh2 expression analysis in familial and sporadic colorectal cancer. Lab Invest 2001; 81: Menigatti M, Di Gregorio C, Borghi F, Sala E, Scarselli A, Pedroni M, et al. Methylation pattern of different regions of the MLH1 promoter and silencing of gene expression in hereditary and sporadic colorectal cancer. Genes Chromosomes Cancer 2001; 31: Goel A, Arnold CN, Niedzwiecki D, Chang DK, Ricciardiello L, Carethers JM, et al. Characterization of sporadic colon cancer by patterns of genomic instability. Cancer Res 2003; 63: Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB. Methylation-specific PCR: A novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci USA 1996; 93: Zhou XP, Hoang JM, Li YJ, Seruca R, Carneiro F, Sobrinho-Simoes M, et al. Determination of the replication error phenotype in human tumors without the requirement for matching normal DNA by analysis of mononucleotide repeat microsatellites. Genes Chromosomes Cancer 1998; 21: Cawkwell L, Gray S, Murgatroyd H, Sutherland F, Haine L, Longfellow M, et al. Choice of management strategy for colorectal cancer based on a diagnostic immunohistochemical test for defective mismatch repair. Gut 1999; 45: Terdiman JP, Gum Jr JR, Conrad PG, Miller GA, Weinberg V, Crawley SC, et al. Efficient detection of hereditary nonpolyposis colorectal cancer gene carriers by screening for tumor microsatellite instability before germline genetic testing. Gastroenterology 2001; 120: Marcus VA, Madlensky L, Gryfe R, Kim H, So K, Millar A et al. Immunohistochemistry for hmlh1 and hmsh2: A practical test for DNA mismatch repair-deficient tumors. Am J Surg Pathol 1999; 23: Deng G, Chen A, Hong J, Chae HS, Kim YS. Methylation of CpG in a small region of the hmlh1 promoter invariably correlates with the absence of gene expression. Cancer Res 1999; 59: Goel A, Arnold CN, Boland CR. Multistep progression of colorectal cancer in the setting of microsatellite instability: New details and novel insights. Gastroenterology 2001; 121: Miyakura Y, Sugano K, Konishi F, Ichikawa A, Maekawa M, Shitoh K, et al. Extensive methylation of hmlh1 promoter region predominates in proximal colon cancer with microsatellite instability. Gastroenterology 2001; 121: Cancer Biology & Therapy 77

6 THE FOLLOWING INSTITUTIONS PARTICIPATED IN THE STUDY Institution Name Location PI Grant # CALGB Statistical Office Durham, NC Stephen George, Ph.D supported by CA33601 Dana Farber Canc78er Institute Boston, MA George P. Canellos, MD supported by CA32291 Dartmouth Medical School-Norris Cotton Lebanon, NH Marc Ernstoff MD supported by CA04326 Cncr Ctr Massachusetts General Hospital Boston, MA Michael L. Grossbard, M.D. supported by CA12449 Mount Sinai School of Medicine New York, NY Lewis Silverman, MD supported by CA04457 Rhode Island Hospital Providence, RI William Sikov, M.D. supported by CA08025 Roswell Park Cancer Institute Buffalo, NY Ellis Levine, M.D. supported by CA02599 Southeast Cancer Control Consortium Inc. Goldsboro, NC James N. Atkins, M.D. supported by CA45808 SUNY Upstate Medical University Syracuse, NY Stephen L. Graziano, M.D. supported by CA21060 The Ohio State University Columbus, OH Clara D. Bloomfield, MD supported by CA77658 University of California at San Diego San Diego, CA Stephen Seagren, MD supported by CA11789 University of California at San Francisco San Francisco, CA Alan Venook, MD supported by CA60138 University of Chicago Medical Center Chicago, IL Gini Fleming, M.D. supported by CA41287 University of Illinois at Chicago Chicago, IL David Gustin, M.D. supported by CA74811 University of Iowa Iowa City, IA Gerald Clamon, MD supported by CA47642 University of Maryland Cancer Center Baltimore, MD David Van Echo, M.D. supported by CA31983 University of Massachusetts Medical Center Worcester, MA Mary Ellen Taplin, M.D. supported by CA37135 University of Minnesota Minneapolis, MN Bruce A Peterson, M.D. supported by CA16450 University of Missouri/Ellis Fischel Cancer Columbia, MO Michael C Perry, M.D. supported by CA12046 Center University of North Carolina at Chapel Hill Chapel Hill, NC Thomas C. Shea, M.D. supported by CA47559 University of Tennessee Memphis Memphis, TN Harvey B. Niell, M.D. supported by CA47555 Wake Forest University School of Medicine Winston-Salem, NC David D. Hurd, MD supported by CA03927 Walter Reed Army Medical Center Washington, DC John C. Byrd, M.D. supported by CA Cancer Biology & Therapy 2004; Vol. 3 Issue 1