In RC patients treated with chemoradiotherapy, KRAS mutation was not associated with lesser response to chemoradiotherapy or worse OS. While both p-AKT and p-ERK expression were correlated with better response to chemoradiotherapy, only high p-ERK expression was associated with better OS.
Constitutive activation of KRAS
via mutation is a common finding in CRC, and one that has been assumed to be of importance to tumorigenesis due to its frequent and early occurrence in the adenoma-carcinoma sequence. Recent studies in mCRC have demonstrated that responses to epidermal growth factor receptor (EGFR) inhibitors (cetuximab or panitumumab) occur only in patients with the wild-type (WT) KRAS
gene. These drugs are now used in the treatment of mCRC, with response rates ranging from 8% with single agent cetuximab[36
] to 23% in combination with irinotecan,[37
] and disease control rates (a combination of complete and partial responses and stable disease) of 39 and 56% respectively.
Preclinical studies have suggested that RAS mutation would lead to radioresistance, and conversely that targeting RAS or downstream effectors of activated KRAS
would identify potential tumor-specific radiosensitizers [38
]. This in part formed the rationale for targeting of EGFR in RC, an intervention that has to date been disappointing in clinical trials. Studies of RAS in human RC tumor samples have generally been small and have produced mixed conclusions [2
]. These studies have reported no change in downstaging[41
] or disease free status by KRAS
]. Other studies found that KRAS
WT was associated with improved survival,[2
] or trended to tumor downstaging[7
] or response,[42
] while another study found associations between KRAS
mutation with earlier stage and better survival [43
]. Our study set out to confirm that KRAS
mutational status was associated with radiosensitivity using more modern sequencing technology in a larger number of patients than had been studied previously. Our findings did not confirm KRAS
mutation as a prognostic factor, nor as a predictive factor for resistance to chemoradiotherapy, suggesting that KRAS
may not be a worthwhile target to pursue as a radiosensitizer in RC.
Our findings confirm a relatively low rate of BRAF mutation as is seen in mCRC, with numbers too small to draw conclusions about the effect of BRAF on chemoradiotherapy response. However, we note that all three patients with confirmed BRAF mutations were long-term survivors.
This study also represented an opportunity to assess the effects of KRAS
mutation on activation of its downstream targets ERK and AKT. Mutation or activation of KRAS
results in constitutive activation of several downstream effectors, and to date it has not been clear which effectors are most important in mediating the effects of mutated KRAS
in cancer cells [1
]. We found that AKT activation was significantly associated with KRAS
activation, while ERK activation was not. One potential weakness of this finding, however, is that anti-phosphoprotein antibodies were used in paraffin-embedded tissues of various ages with the possibility that there was variability in the stability of the phosphorylated proteins. Another small study of predominantly stage II and III CRC tumors evaluated the KRAS/BRAF
/ERK pathway. p-ERK was correlated with KRAS
codon 12 mutations (p = 0.016) but not with codon 13 mutations [14
mutations (V599) were infrequent (8.7%) and not associated with p-ERK status [14
]. We did not examine the association between mutation location and p-ERK in our study. Preoperative sample limitations prevented all phosphorylation and mutational analyses to be conducted on these untreated samples.
Among our most interesting findings was that higher p-ERK levels were associated with better survival and response to chemoradiotherapy, and that higher p-AKT expression was also associated with better response to chemoradiotherapy. Certain tumors exhibiting activated AKT, including squamous cell cancer of the head and neck[44
] and cervical cancer,[34
] have been shown to respond poorly to chemoradiotherapy. The positive association between ERK activation and response is less surprising, as ERK activation results in increased cell cycling,[45
] and presence of higher fractions of cycling cells has previously been associated with better outcome from chemoradiotherapy for RC [46
]. The positive association between ERK activation and survival is the first described in RC to our knowledge. In hepatocellular carcinoma, higher pERK staining intensity was associated with longer time to progression [47
It is more difficult to understand why AKT activation would be beneficial in terms of response. Data from other tumor types suggests that activation of the AKT cell survival pathway confers resistance to radiation [34
]. For example, in a PTEN-deficient glioma model, a case in which AKT is constitutively active, PTEN gene transfer resulted in significant radiosensitization [48
]. In a recent publication, it was demonstrated that PI3K mutations, which would be expected to result in constitutive AKT activation, were associated with a higher rate of local recurrence of RC (27.8 vs. 9.4%) [6