The RAS proteins participate in the RAS-RAF-MEK-ERK-MAPKinase pathway, which mediates cellular responses to growth signals.1 There are 3 RAF genes, each encoding cytoplasmic serine/threonine kinases that are regulated by binding to RAS.1, 2 We have recently reported that BRAF is somatically mutated in a number of human cancers, including malignant melanoma, colorectal carcinoma and ovarian borderline (low malignant potential) tumors.3 Mutations in BRAF occur in 2 regions of the BRAF kinase domain, the G loop (which mediates binding of ATP) and the activation segment (which protects the substrate binding site). Mutated forms of BRAF that have been studied so far have elevated kinase activity and can transform NIH3T3 cells.3 More recently, we have also documented that there are many similarities in the phenotypic patterns associated with BRAF and KRAS mutations in colorectal neoplasia,4 suggesting that BRAF mutations may be biologically equivalent to KRAS mutations in colorectal cancer development. In support of this notion, the BRAF and KRAS mutations exhibit a trend toward mutual exclusion in human tumors.3, 4, 5 Interestingly Rajagopalan et al. have observed a higher incidence of BRAF mutation in colorectal cancer with microsatellite instability (MSI) compared to the microsatellite stable (MSS) counterpart,5 although similar relationships were not documented in our series.
Gastric cancer are well known, sharing many phenotypic and molecular genetic changes with colorectal cancer. In particular, the intestinal type of gastric cancer is believed to have gone through the process of chronic gastritis, intestinal metaplasia, dysplasia followed by malignant transformation, the latter steps being analogous to the adenoma-carcinoma sequence in colorectal cancer. Somatic mutation of the RAS genes, in particular KRAS, is common in colorectal cancer, being found in more than one-third of cases.6, 7 Unexpectedly, recent molecular studies in gastric cancers have mostly found a low incidence of KRAS mutation, in the range of 7–20%.8, 9, 10 Similar to colorectal cancer, gastric cancers have been reported to show high-level microsatellite instability due to mismatch repair defects in 10–20% cases.11, 12, 13 We therefore assessed the presence of BRAF and KRAS mutations in 94 gastric cancers and their relationship to mismatch repair status. Frozen tumors and normal gastric mucosa were collected from gastrectomy specimens in Queen Mary Hospital, The University of Hong Kong. Fifty-four were from male and 40 from female patients. The patients' age ranged from 35–88 years (mean age 68 years). Seventy were located in the antrum, 15 were located in the body, 8 were located in the cardia and 1 diffusely involving the whole stomach. Seventy were of the intestinal type and 15 were of the diffuse type, whereas 9 were mixed/indeterminate type by Lauren's classification. There were 14 stage I, 20 stage II, 45 stage III and 15 stage IV cancers. Three were well differentiated, 42 moderately differentiated and 49 poorly differentiated. Frozen sections were cut for histology evaluation before DNA extraction. In the tumor blocks, contaminating nonneoplastic tissue was trimmed away as far as possible. DNA was extracted by standard protocols using proteinase K digestion, phenol-chloroform extraction and ethanol precipitation. Our study was approved by the Ethics Committee of the University of Hong Kong.
The mismatch repair status for most of these cases has been previously reported.11, 14 The MSI panel markers include both dinucleotide repeats (D2S123, D5S346, D17S250, Tp53, D18S58, D18S57) and polyadenine tracts (BAT25, BAT26, BAT40, BAT-RII). At least 5 loci, of which at least 2 were polyadenine tracts, were analysed in each case. Cases were referred to as MSI-H if there were more than 40% unstable loci, MSI-L if less than 40% unstable loci and MSS if no unstable loci. Twenty-one cases were classified as MSI-H, 10 as MSI-L and 63 as MSS. Immunohistochemical staining for hMSH2 and hMLH1 protein expression and assessment for hMLH1 promoter methylation using methylation specific-PCR and bisulphite genomic sequencing were similar to previously described.15 Of the 21 MSI-H cases, 19 cases lost hMLH1 protein, and these all showed hMLH1 promoter hypermethylation.
The complete coding sequences of exons 11 (G loop region) and 15 (activation segment) of BRAF and exon 2 of KRAS were amplified using intronic primers and directly sequenced using the DYEnamic ET Terminator Cycle Sequencing Kits (Amersham Pharmacia, Freiberg, Germany) and analysed by the Applied Biosystems 377 automated sequencer. These cover most of the previously known mutation hotspots of the 2 genes. The primers for BRAF sequencing were similar to those previously described,3 and primers for KRAS were as follows: forward: CTG AAA ATG ACT GAA TAT AAA CTT GT; reverse: ATA TGC ATA TTA AAA CAA GAT TTA CC. Direct sequencing was performed on both strands in all cases for BRAF. For KRAS, the reverse primer was used for sequencing. Cases with mutations were further confirmed by sequencing using the forward primer.
Of 94 gastric adenocarcinomas studied, 8 (8.5%) were found to have KRAS mutations (Table I). This frequency is consistent with other recent studies of gastric cancer in which a relatively low incidence of KRAS mutation (7–20%) has been reported.8, 9, 10 The relationship of KRAS mutation with other clinicopathologic and molecular parameters are summarised in Table II. All 8 gastric tumors with KRAS mutation arose from the gastric antrum (p = 0.002). Seven of 8 KRAS-positive cases showed the high-level microsatellite instability phenotype (p = 0.000076), and these were all associated with hMLH1 promoter methylation and loss of hMLH1 protein. None of the 94 gastric adenocarcinomas was found to have a BRAF mutation.
|Case||Nucleotide change||Amino acid change||MSI status||hMLH1 protein||hMLH1 promoter||Tumor site||Tumor type||IM|
|Percentage with KRAS mutation||p-value, Fisher exact test|
|Male||7.4 (54)a||p = 0.26|
|Well||33.3 (3)||p = 0.28|
|Poor||8.2 (49)||vs poor|
|Cardia||0 (29)||p = 0.002|
|Body||0 (20)||Antral vs|
|Whole stomach||0 (1)|
|Tumor type (Lauren)|
|Intestinal||10 (70)||p = 0.26|
|Diffuse||0 (15)||Intestinal vs|
|Tumor stage (UICC)|
|Stage I||21 (14)||p = 0.086|
|Stage II||10 (20)||Stage 1 & 2|
|Stage III||4.4 (45)||vs 3 & 4|
|Stage IV||6.7 (15)|
|Intestinal metaplasia at tumor edge|
|Nil/Minimal||4.3 (47)||p = 0.09|
|Helicobacter pylori infection|
|Negative||9.3 (43)||p = 0.277|
|MSI-H||33.3 (21)||p = 0.000076|
To our knowledge, this is the first systematic study of KRAS and BRAF mutation in gastric adenocarcinomas. Overall, we found only 8.5% of cases showing KRAS mutation and none of our cases showed mutation in BRAF. This is in contrast to colorectal adenocarcinomas, in which a high incidence of KRAS mutation and a modest incidence of BRAF mutation are found.4
KRAS mutations in gastric cancer appear to occur preferentially in mismatch repair-deficient tumours. Seven of 21 (33%) mismatch repair-deficient gastric cancers carried KRAS mutations compared to 1 of 73 (1.4%) mismatch repair-proficient (MSS and MSI-L were combined for this analysis) gastric cancers. There has been little data regarding the relationship of mismatch repair deficiency with KRAS mutation in gastric cancer. However, a similarly high incidence of KRAS mutations has been observed in mismatch repair-deficient endometrial carcinomas compared to mismatch repair-proficient endometrial cancers.16 By contrast, in colorectal cancer, KRAS mutations were found in 43% mismatch repair-deficient and 59% mismatch repair-proficient cases.5
The absence of BRAF mutations in gastric cancer may simply reflect the apparently overall lower frequency of ERK-MAPKinase pathway activation in this cancer type compared to colorectal cancer, since KRAS mutations were also found at a lower frequency. Alternatively, this pathway is important, yet there are as yet unknown components other than KRAS and BRAF that are the preferential targets of mutation in the stomach. Given the unique environmental factors associated with gastric carcinogenesis, such as infection by Helicobacter pylori, association with Epstein Barr virus and specific carcinogens such as nitrosamines, it is not surprising to find major differences in molecular targets even for the intestinal-type gastric cancer compared to colorectal cancer. Further investigations are needed to clarify and improve our understanding of the organ-specific pathway of gastric carcinogenesis.