| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Letters to the Editor |
McGill University Department of Pathology Montreal, Quebec H3A 2B4 Canada
Conjoint Gastroenterology Laboratory Bancroft Centre Brisbane, Australia
University of Maryland Baltimore, MD
Department of Gastroenterological Okayama University Okayama, Japan
Kimmel Cancer Center Philadelphia, PA
Letter
The study by Laiho et al. (1) examines the distribution of MSI-L1 in 90 non-MSI-H colorectal cancers using a panel of 377 microsatellite markers. In this series, the percentage of markers showing instability ranges from 0% to only 3%, implying that the majority of markers in the large panel is insensitive to MSI-L. We question the value of employing a large panel of relatively stable markers to define MSI-L status. By a process of trial and error, markers have been identified that show relatively high frequencies of MSI in MSI-L cancers, and these have been introduced into reference panels that allow comparisons to be made between different cohorts (2) . In the series by Dietmaier et al., 3 of 12 MSI-L cancers showed instability at the dinucleotide marker APC and/or the compound marker MYCL (3) . Because these markers are still relatively insensitive (as compared with mononucleotide markers for MSI-H status), it is evident that a panel must be used to identify the subset of non-MSI-H cancers that shows more instability than would be expected by chance. When such a modest panel of relatively sensitive markers is used, approximately 8% of colorectal cancers are designated as MSI-L (4) .
The MSI-L subset does not show distinct clinical, morphological, or molecular differences from the remaining MSS cancers (i.e., not of the magnitude of those distinguishing MSI-H from non-MSI-H colorectal cancers). Nevertheless, certain molecular features are distributed on a nonrandom basis within the non-MSI-H subset of colorectal cancers. Two studies, including that of Laiho et al., provide evidence that MSI-L status is distributed nonrandomly (1 , 5) . The study by Halford et al. (5) distinguished within the non-MSI-H group a "super-MSI-L" group, a "super-stable" group, and an intermediate group with very mild background MSI. The distinction between the intermediate group and the super-stable group is obscured in the study of Laiho et al. by the use of very numerous, relatively stable markers.
Nonrandomly distributed features among non-MSI-H cancers include K-ras mutation (4 , 6 , 7) , DNA methylation (8) , and chromosomal alterations as shown by comparative genomic hybridization (9) . Furthermore, some of these features show patterns of cosegregation. Relatively high frequencies of K-ras mutation (90%) have been found in non-MSI-H cancers with the CpG island methylator phenotype (8) . G to A mutation in K-ras has been associated with methylation of the DNA repair gene MGMT (10) . The overall frequency of K-ras mutation in colorectal cancer is approximately 35%. A relatively high frequency of K-ras mutation in MSI-L versus MSS cancers has been demonstrated in three independent studies: 11 (50%) of 22 versus 24 (30%) of 79 (6) , 19 (50%) of 38 versus 13 (27%) of 48 (4) , and 12 (41%) of 29 versus 14 (18%) of 77, respectively (7) . One study failed to show a difference in the frequency of K-ras mutation in MSI-L versus MSS cancer, but this study also showed an unusually high frequency of K-ras mutation in MSI-H cancer (11) . This is probably explained by incorrect assignment of some MSI-L cancers within the MSI-H subset (through overdependence on the finding of instability limited to non-mononucleotide markers). The study by Gebert et al. (12) included only 79 colorectal cancers with no enrichment of MSI-L cases. An extremely low frequency of K-ras mutation (10%) was found in the super-stable subset as defined by Halford et al. (5) .
Laiho et al. (1) claim that the overall mutation rate in MSI-L cancers is equivalent to normal human fibroblasts and that, therefore, microsatellite mutation would accumulate merely on a stochastic basis. However, Laiho et al. provide no explanation for the considerably higher mutation rate demonstrated in selected markers for the MSI-L trait and, therefore, incorporated into the National Cancer Institute reference panel. It is reasonable to postulate that a DNA repair defect does underlie the small super-MSI-L subset and that this defect is associated with the methylation of DNA repair genes such as MGMT (13) . The mechanism in the case of MGMT silencing may be very simple. Inactivation of MGMT would result in methylG:T mismatches that are known to be recognized and repaired by the DNA mismatch repair system (14) . Excessive production of methylG:T mismatches may increase the probability of errors during DNA resynthesis that escape the mismatch repair system. An association between MGMT methylation and MSI-L status is, in fact, suggested in the study of Laiho et al. (1) , although the finding is difficult to interpret for the reasons stated above. Such a DNA repair defect is "subtle" only insofar as the production of genetic instability is secondary to the generation of excess mismatches as opposed to being the direct result of DNA mismatch repair deficiency.
It is interesting that the MSI-L phenotype was shown by Laiho et al. (1) to be established before the final clonal expansion, implying the presence of the phenotype within a precancerous lesion. This observation may be relevant to a previous demonstration of frequent MSI-L status in serrated polyps showing dysplasia (admixed polyps and serrated adenomas) and not within traditional adenomas (15) . This would bracket super-MSI-L and MSI-H cancers within a serrated pathway of pathogenesis associated with the methylator phenotype (16) .
Whereas the distinction between MSI-L and MSS cancers is clearly problematic and cannot be achieved in an absolute sense, the possibility that a "mild mutator" pathway may be driven by a DNA repair defect is of potential clinical importance because it may define a subgroup associated with altered response to chemotherapeutic agents. MSI-L status is interesting only insofar as MSI-L might serve as the by-product of a DNA repair defect. The data provided by Laiho et al. (1) do not discount this possibility. It should also be noted that the Dietmaier et al. study found several microsatellite markers that were completely ineffective at diagnosing any type of MSI. It is likely that Laiho et al. have now identified several more of these types of markers.
In conclusion, we reject the suggestion of Laiho et al. that the distinction between MSI-L and MSS status can no longer be justified, at least in the arena of research. After all, it was not so long ago that MSI-H and MSI-L cancers were not distinguished correctly and MSI-H and non-MSI-H cancers were argued to share a common molecular background (17 , 18) .
FOOTNOTES
1 The abbreviations used are: MSI, microsatallite instability; MSI-L, low-level MSI; MSI-H, high-level MSI; MSS, microsatellite stable. 
Received 3/23/02. Accepted 8/20/02.
REFERENCES
Biomedicum P. O. Box 63 (Haartmaninkatu 8) 00014 University of Helsinki, Finland
Jass et al. (1)
question the value of employing "a large panel of relatively stable markers" to define MSI-L1
in colorectal tumors, and propose that "a modest panel of relatively sensitive markers" should be used. In our study (2)
, any observed instability was taken into account to distinguish lesions displaying low-degree MSI, and if the "insensitive" markers showing no MSI were omitted from the study, the results would naturally remain exactly the same; 71 of 90 colorectal cancers showed some degree of instability. Therefore, it is difficult to imagine how the use of insensitive markers among other markers would prevent identification of MSI-L (see also below and Fig. 1
; the Bethesda dinucleotide markers (3)
were in no way superior to the random markers used).
|
. It is difficult to understand an argument that the only meaningful instability detected is one detected by markers mentioned in Ref. 3
.
We illustrate the issue by giving an example. Let’s follow the guidelines of Jass et al. (1)
: a modest panel of relatively sensitive markers should be used for the detection of MSI-L. If we had taken into account only the 11 dinucleotide markers that we used (2)
, which are listed in the Bethesda recommendations (3)
, MSI-L would have been detected in 9 (10%) of 90 samples, as indicated in our publication (Ref. 2
; Table 1
). For the purpose of this reply, we chose 11 of the most sensitive markers in our randomly picked set of 364 dinucleotide markers. These 11 markers, listed in Table 1
, are hereafter referred to as the Helsinki panel. MSI sensitivity of each of the Helsinki panel markers in the data set was equal or superior to the most sensitive Bethesda marker, D2S123 (four, five, or six MSI events detected; Fig. 1
). If we had used the Helsinki panel, the number of MSI-L tumors would have been 33 (37%; Table 1
). We do not wish to emphasize too much the "sensitivity" of these markers or the insensitivity of others; in another sample set, if we used the Helsinki markers, the proportion of MSI-L would probably be lower. The point and the concern here, which we would like all researchers studying the MSI-L phenotype to appreciate, is that these two sets of detected MSI-L tumors are almost entirely different. Only two lesions were shared between the groups (Table 1 ; three would be expected by chance alone). Meaningful comparisons between MSI-L and MSS are quite difficult if samples displaying comparable instability are scored differently.
|
|
FOOTNOTES
The abbreviations used are: MSI, microsatellite instability; MSI-L, low-level MSI; MSI-H, high-level MSI; MSS, microsatellite stable/stability.
Received 6/27/02. Accepted 8/20/02.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  M. R.J. Kohonen-Corish, J. J. Daniel, C. Chan, B. P.C. Lin, S. Y. Kwun, O. F. Dent, V. S. Dhillon, R. J.A. Trent, P. H. Chapuis, and E. L. Bokey Low Microsatellite Instability Is Associated With Poor Prognosis in Stage C Colon Cancer J. Clin. Oncol., April 1, 2005; 23(10): 2318 - 2324. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
