| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Correspondence |
1Institute of Cancer Research, Surrey, United Kingdom
To the Editor-in-Chief:
In a recent issue of The American Journal of Pathology, Samowitz et al1 reported on the instability of the monosatellite markers BAT-26 and BAT-40 in colorectal adenomas and carcinomas. The study identified a second, significantly shorter allele of BAT-26 and suggested that a germline polymorphism exists for BAT-26 in 7.7% of African-Americans and 0.08% of Caucasians. This raises the possibility that misclassification of replication error status in tumors may occur, especially in African-Americans, if based solely on analysis of tumor material.
We wish to draw attention to the possibility that methodological errors may have led to the observation of this shorter allele. In our experience2 using the ABI 377 (PE Applied Biosystems, Warrington, UK), constitutional DNA analyzed for BAT-26 with a fluorescently labeled PCR primer occasionally gives rise to a trace suggesting the presence of a shorter allele. However, this is never reproducible. Moreover, in both of the traces obtained with BAT-26 primers on constitutional DNA shown by Samowitz et al in their Figure 2a, the fluorescence signal is less than 150 relative units. According to PE Applied Biosystems, under 150 units "the signal-to-noise ratio is too low to discriminate between sample peaks and background."3 Furthermore, on both traces there are additional signals at around 85 length units, indicating a high background. It is unclear from the paper whether all cases with shortened alleles were confirmed using constitutional DNA extracted from lymphoblastoid cells. This, therefore, leaves open the possibility that an apparently shorter germline BAT-26 allele may have arisen from contamination with tumor tissue harboring that allele.
Unequivocal evidence for the presence of a shorter germline allele would come from detection of a homozygote. However, we acknowledge that given a heterozygote frequency of 7.7% in 65 African-Americans, there is only limited power to detect this. In the absence of unambiguous evidence, we believe there are sufficient methodological issues to draw into question the conclusion that there exists a germline BAT-26 allele that is significantly shorter than normal.
References
2University of Utah Health Sciences Center
Salt Lake City, Utah;
3The Fred Hutchinson Cancer Research Center Seattle,
Washington
Author’s Reply:
Bradshaw et al are concerned that our novel identification of
BAT-26 polymorphisms (and, presumably, the increased frequency of these
polymorphisms in African-Americans) could have been the result of
contamination with tumor and/or an artifact due to low signal and high
background. With regard to the first concern, shortened BAT-26 alleles
were identified in DNA extracted from peripheral blood lymphocytes from
6 out of the 12 individuals with germline polymorphisms listed in Table
2 or our paper.1
This result rules out tumor contamination
as the source of these novel alleles. With regard to the second
concern, we feel that our published data are convincing (and both of
the novel allele peaks are actually higher than the 150 relative units
cited as necessary to reliably distinguish signal from background).
Nevertheless, we are happy to present additional data from the study
that resolve this point. Figure 1
shows
BAT-26 results from peripheral blood lymphocyte DNA from two
individuals listed in Table 2 of our paper.1
A novel
allele (designated B), approximately 10 basepairs smaller than
full-length BAT-26 (designated A), is seen in both samples (the
absolute size of alleles A and B is increased because of a 6-bp
addition to the reverse primer2
). There is very low
background, and the amplitude of this shortened allele is above 1000
relative units.
|
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
