PO-447 CDK5 as a prognostic and predictive factor in colorectal cancer patients

Introduction Cdk5 has been related to cancer progression and metastasis and to response to DNA damage in cancer. In colorectal cancer (CRC), our previous results showed that Cdk5 was overexpressed in CRC tumours as compared to normal tissues. CDK5 gene knockdown was associated with decreased migration and invasion of colon cancer cells but has no effect on cell proliferation; this effect was not not observed in cell lines with BRAF or KRAS mutations. In cell lines with acquired resistance to oxaliplatin (OXA) CDK5 knockdown overcame this resistance. The aim of this work was to study the involvement of Cdk5 on migration and invasion according to KRAS status and to investigate the prognostic and predictive value of Cdk5 in representative patients cohorts. Material and methods SW48 (KRAS WT) and the isogenic SW48 KI G12V (KRAS mutant) cell lines were used (Horizon Discovery Ltd). CDK5 gene was silenced with specific siRNA (Ambion, 5 nM) to study its role in migration and invasion using Boyden chamber assays (HTS transwell, Corning). Comparisons between different experimental conditions were carried out using the T test. For clinical studies, we used 342 tumour samples from CRC patients (Non-treated stage II-III frozen n=135; stage IV FFPE=207). Stage IV tumours corresponded to patients treated with schedules containing irinotecan (n=139) or OXA (n=68) without antitarget drugs. Cdk5 expression was analysed by qPCR (Taqman assay) or by immunohistochemistry. Log rank (LR) and cox regression (CR) tests were used to study differences in disease-free (DFS) overall survival (OS) or time to progression (TTP); Chi-square or Fisher’s tests were used to study differences in response rates (ORR). Results and discussions CDK5 gene silencing was associated with decreased cell migration only in SW48 G12V cells (p=0.035). High Cdk5 mRNA levels was a bad prognostic factor in stage II (DFS CR HR=4.15; p=0.022; OS CR HR=8.25, p=0.048) and III (DFS LR 24 months vs. not reached; p=0.048). In stage II, Cdk5 prognostic value was associated with KRAS mutant status (CR p=0.045). In stage IV CRC, Cdk5 positive IHC staining was associated with worse TTP (LR 8 vs. 18 months, p=0.043; CR HR=2.17) and ORR (53% vs. 87%, p=0.029) only in OXA-treated patients. Conclusion These data confirm previous results establishing Cdk5 as a tumor-promoting factor in CRC, especially in a KRAS-mutated environment. Cdk5 immunostaining predicts outcome of stage IV CRC patients treated with OXA. Extensive data will be presented during the EACR meeting.