Key Points: PIK3CA and NRAS mutations are recurrent in BRAFV600E wild-type ECD patients. 57.5% (46/80) of ECD patients have a BRAFV600E mutation, and an additional 10.9% and 3.7% have PIK3CA and NRAS mutations, respectively.

Purpose: KRAS mutations are predictive of nonresponse to anti-EGFR therapies in metastatic colorectal cancer (mCRC). However, only 50% of nonmutated patients benefit from them. KRAS-mutated subclonal populations nondetectable by conventional methods have been suggested as the cause of early progression. Molecular analysis technology with high sensitivity and precision is required to test this hypothesis. Experimental Design: From two cohorts of patients with mCRC, 136 KRAS, NRAS, and BRAF wild-type tumors with sufficient tumor material to perform highly sensitive picodroplet digital PCR (dPCR) and 41 KRAS-mutated tumors were selected. All these patients were treated by anti-EGFR therapy. dPCR was used for KRAS or BRAF mutation screening and compared with qPCR. Progression-free survival (PFS) and overall survival (OS) were analyzed according to the KRAS-mutated allele fraction. Results: In addition to the confirmation of the 41 patients with KRAS-mutated tumors, dPCR also identified KRAS mutations in 22 samples considered as KRAS wild-type by qPCR. The fraction of KRAS-mutated allele quantified by dPCR was inversely correlated with anti-EGFR therapy response rate (P < 0.001). In a Cox model, the fraction of KRAS-mutated allele was associated with worse PFS and OS. Patients with less than 1% of mutant KRAS allele have similar PFS and OS than those with wild-type KRAS tumors. Conclusions: This study suggests that patients with mCRC with KRAS-mutated subclones (at least those with a KRAS-mutated subclones fraction lower or equal to 1%) had a benefit from anti-EGFR therapies. Clin Cancer Res; 21(5); 1087–97. ©2014 AACR.

BACKGROUND: Circulating tumor DNA (ctDNA) has emerged as a good candidate for tracking tumor dynamics in different cancer types, potentially avoiding repeated tumor biopsies. Many different genes can be mutated within a tumor, complicating procedures for tumor monitoring, even with highly sensitive next-generation sequencing (NGS) strategies. Droplet-based digital PCR (dPCR) is a highly sensitive and quantitative procedure, allowing detection of very low amounts of circulating tumor genetic material, but can be limited in the total number of target loci monitored.METHODS: We analyzed hypermethylation of 3 genes, by use of droplet-based dPCR in different stages of colorectal cancer (CRC), to identify universal markers for tumor follow-up.RESULTS: Hypermethylation of WIF1 (WNT inhibitory factor 1) and NPY (neuropeptide Y) genes was significantly higher in tumor tissue compared to normal tissue, independently of tumor stage. All tumor tissues appeared positive for one of the 2 markers. Methylated ctDNA (MetctDNA) was detected in 80% of metastatic CRC and 45% of localized CRC. For samples with detectable mutations in ctDNA, MetctDNA and mutant ctDNA (MutctDNA) fractions were correlated. During follow-up of different stage CRC patients, MetctDNA changes allowed monitoring of tumor evolution.CONCLUSIONS: These results indicate that MetctDNA could be used as a universal surrogate marker for tumor follow-up in CRC patients, and monitoring MetctDNA by droplet-based dPCR could avoid the need for monitoring mutations.

Clusterin is a secreted protein chaperone up-regulated in several pathologies, including cancer and neurodegenerative diseases. The present study shows that accumulation of aberrant proteins, caused by the proteasome inhibitor MG132 or the incorporation of the amino acid analogue AZC (L-azetidine-2-carboxylic acid), increased both clusterin protein and mRNA levels in the human glial cell line U-251 MG. Consistently, MG132 treatment was capable of stimulating a 1.3 kb clusterin gene promoter. Promoter deletion and mutation studies revealed a critical MG132-responsive region between −218 and −106 bp, which contains a particular heat-shock element, named CLE for ‘clusterin element’. Gel mobility-shift assays demonstrated that MG132 and AZC treatments induced the formation of a protein complex that bound to CLE. As shown by supershift and chromatin-immunoprecipitation experiments, CLE is bound by HSF1 (heat-shock factor 1) and HSF2 upon proteasome inhibition. Furthermore, co-immunoprecipitation assays indicated that these two transcription factors interact. Gel-filtration analyses revealed that the HSF1–HSF2 heterocomplexes bound to CLE after proteasome inhibition have the same apparent mass as HSF1 homotrimers after heat shock, suggesting that HSF1 and HSF2 could heterotrimerize. Therefore these studies indicate that the clusterin is a good candidate to be part of a cellular defence mechanism against neurodegenerative diseases associated with misfolded protein accumulation or decrease in proteasome activity.