|Home | About | Journals | Submit | Contact Us | Français|
Large-scale sequencing of cancer genomes holds great promise in identifying gene mutations that drive human cancer progression. A study was reported recently in which the coding region of 13,023 genes was systematically sequenced within a panel colorectal tumors . Six somatic heterozygous mutations of the MLL3 (Mixed-Lineage Leukemia 3) gene were identified in 35 colorectal tumors, making it the 4th highest mutation frequency in the disease . MLL3 had not been implicated previously in human cancer, however, the highly-related MLL1 is a known proto-oncogene in myeloid and lymphoid leukemia . MLL proteins are transcription factors that promote gene activation through histone methyltransferase activity performed by their conserved SET domain at the C-terminus . Importantly, two of the identified mutations from the prior study (R1906stop and 1290del40nt) resulted in truncation of the MLL3 protein eliminating its catalytic SET domain . Three mutations resided near the N-terminus within a stretch of PHD domains (C347G, D400N, L478W), which are known chromatin-binding modules . Since most of these mutations would be predicted to impair MLL3 function, these findings raised the tantalizing possibility that MLL3 mutations might play a role in the initiation and/or progression of the disease. It should be noted that considerable debate exists over the proper statistical methods used to distinguish passenger from driver status of mutations identified in such large-scale studies (see commentaries on ). This issue is particularly relevant to large genes (such as MLL3), which would be expected to contain a higher frequency of passenger mutations occurring at random in a given tumor. Hence, a reevaluation of the original mutation data using a more stringent filter for passenger mutation frequency deemed the identified MLL3 mutations as not achieving statistical significance . Based on these uncertainties regarding the true significance of MLL3 mutations in colorectal carcinoma, it is imperative that the frequency of mutations be evaluated in an independent cohort of disease.
In order to validate the mutation frequency of MLL3 in colon cancer, we sequenced all 59 exons of MLL3 (14.7 kb coding sequence) in an independent group of 57 colon tumor DNA samples (all processed from primary tumor specimens). Eighty-two PCR products spanning 867 kb of total tumor DNA were generated and subjected to direct sequencing in both orientations using Applied Biosystems 3730×l DNA Analyzer. Chromatogram analysis was performed using Consed software, which was verified by visual inspection. Nucleotide changes matching known SNPs were removed. All remaining nucleotide changes were then examined in matched normal control tissue DNA to examine somatic or germline status. All identified changes were verified by re-amplifying and sequencing the relevant PCR product.
Our analysis identified one somatic mutation of MLL3 among 57 colon tumors (1.8%), encoding a heterozygous missense mutation, S3449F. Serine 3449 lies outside of all conserved domains within the MLL3 protein, and the mutation is therefore of uncertain functional significance. Our study identified also several novel non-synonymous germline variants of MLL3, likely representing previously unidentified SNPs. The passenger mutation frequency in colon cancer has been estimated to be approximately 0.55 per Mb, although this number is likely to vary among distinct genomic regions and depend upon type and stage of the cancer . Therefore, our single identified mutation within 867 kb of sequenced tumor DNA approximately matches the number of passenger mutations to be expected by chance.
While our findings do not support a high frequency of MLL3 mutations in colon cancer, as was initially reported , our results are consistent with the subsequent reanalysis of the data using more stringent statistical evaluation of passenger mutation frequency . It should also be noted that the original screen was performed largely in cell lines and xenografts derived from metastases, whereas our analysis was performed exclusively in primary tumors. It is possible that passenger MLL3 mutations were more prevalent in the prior study due to prolonged cell incubations, whether in culture or following transplant into mice. Our findings are in concert with a prior sequencing study of 104 cases of colorectal carcinoma in which no mutations were identified in MLL3 . However, this study only examined 5 of the 59 exons of the MLL3 gene. MLL3 mutations might occur at higher rates in other human cancers, such as glioblastoma multiforme and pancreatic ductal adenocarcinoma . Further insights into a possible role of MLL3 in carcinogenesis would likely be gained from more expansive sequencing studies as well as through functional approaches.
We thank the Cooperative Human Tissue Network (CHTN) at the University of Pennsylvania, Yoshio Naomoto (Okayama University), and Daniel Chung (Massachusetts General Hospital) for providing human colon specimens. This work was supported by NIH grant DK58044 (G.A.B), “Genomes and Genetics at the BCM-HGSC” NHGRI grant # 2 U54 HG003273-05 (R.A.G), a pilot grant from the NIH/NIDDK Center for Molecular Studies in Digestive and Liver Diseases (P30 DK050306), and NIH/NIDDK DK056645 (CNJ, AKR).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.