Date Published: July 29, 2011
Publisher: Public Library of Science
Author(s): Filip Janku, J. Jack Lee, Apostolia M. Tsimberidou, David S. Hong, Aung Naing, Gerald S. Falchook, Siqing Fu, Rajyalakshmi Luthra, Ignacio Garrido-Laguna, Razelle Kurzrock, Mikhail V. Blagosklonny. http://doi.org/10.1371/journal.pone.0022769
Abstract: Oncogenic mutations of PIK3CA, RAS (KRAS, NRAS), and BRAF have been identified in various malignancies, and activate the PI3K/AKT/mTOR and RAS/RAF/MEK pathways, respectively. Both pathways are critical drivers of tumorigenesis.
Tumor tissues from 504 patients with diverse cancers referred to the Clinical Center for Targeted Therapy at MD Anderson Cancer Center starting in October 2008 were analyzed for PIK3CA, RAS (KRAS, NRAS), and BRAF mutations using polymerase chain reaction-based DNA sequencing.
PIK3CA mutations were found in 54 (11%) of 504 patients tested; KRAS in 69 (19%) of 367; NRAS in 19 (8%) of 225; and BRAF in 31 (9%) of 361 patients. PIK3CA mutations were most frequent in squamous cervical (5/14, 36%), uterine (7/28, 25%), breast (6/29, 21%), and colorectal cancers (18/105, 17%); KRAS in pancreatic (5/9, 56%), colorectal (49/97, 51%), and uterine cancers (3/20, 15%); NRAS in melanoma (12/40, 30%), and uterine cancer (2/11, 18%); BRAF in melanoma (23/52, 44%), and colorectal cancer (5/88, 6%). Regardless of histology, KRAS mutations were found in 38% of patients with PIK3CA mutations compared to 16% of patients with wild-type (wt)PIK3CA (p = 0.001). In total, RAS (KRAS, NRAS) or BRAF mutations were found in 47% of patients with PIK3CA mutations vs. 24% of patients wtPIK3CA (p = 0.001). PIK3CA mutations were found in 28% of patients with KRAS mutations compared to 10% with wtKRAS (p = 0.001) and in 20% of patients with RAS (KRAS, NRAS) or BRAF mutations compared to 8% with wtRAS (KRAS, NRAS) or wtBRAF (p = 0.001).
PIK3CA, RAS (KRAS, NRAS), and BRAF mutations are frequent in diverse tumors. In a wide variety of tumors, PIK3CA mutations coexist with RAS (KRAS, NRAS) and BRAF mutations.
Partial Text: Recently, major discoveries in the molecular biology of human cancers along with an increased understanding of oncogenic mutations and cell signaling pathways led to the successful application of new targeted therapies in several cancers., , ,  These include the use of KIT kinase inhibitors in KIT-mutant gastrointestinal stromal tumors (GIST), ABL kinase inhibitors in BCR-ABL-positive chronic myelogenous leukemia (CML), EGFR tyrosine kinase inhibitors in EGFR-mutant lung cancers, and BRAF inhibitors in BRAF-mutant melanomas., , ,  It appears plausible that the most common cancers have been difficult to treat, in part because they are heterogeneous, with each subset of patients having different molecular abnormalities. Identifying relevant molecular subtypes within heterogeneous cancers is crucial to future targeted therapeutic progress., 
Across tumor types, we demonstrated a higher prevalence of RAS (KRAS, NRAS) or BRAF mutations (47%) and KRAS mutations (38%) in patients with mutant PIK3CA compared to those with wtPIK3CA (mutant RAS or BRAF present in 24%, p = 0.001; mutant KRAS present in 16%, p = 0.001). Most previously published studies investigating simultaneous PIK3CA, and RAS or BRAF mutations concentrated on colorectal cancer., , , ,  Some studies suggested that PIK3CA mutations are associated with KRAS mutations,, ,  whereas others did not report that.,  A large retrospective study that analyzed 1,022 tumor DNA samples from patients with colorectal cancer treated with cetuximab in multiple European institutions revealed association between exon 9 PIK3CA and KRAS mutations (14.7% in KRAS mutant vs. 6.8% in wtKRAS; p = 0.0006), but not between exon 20 PIK3CA and KRAS mutations (3.8% in KRAS mutant vs. 2.3% in wtKRAS; p = 0.27). In agreement with this paper, when we examined all histologies, we also found a strong association between exon 9 PIK3CA and KRAS mutations (19% in KRAS mutant vs. 6% in wtKRAS; p = 0.001), however, the frequency of exon 20 PIK3CA mutations also showed a trend towards being more common in patients with KRAS mutations compared to wtKRAS (12% in KRAS mutant vs. 5% in wtKRAS), albeit not reaching statistical significance (p = 0.07). In a disease-specific subanalysis in colorectal and ovarian cancer we noticed a trend toward an association between exon 9 PIK3CA and KRAS mutations in colorectal cancer (20% in KRAS mutant vs. 6% in wtKRAS; p = 0.07) and a statistically significant association between exon 20 PIK3CA and KRAS mutations in ovarian cancer (75% in KRAS mutant vs. 5% in wtKRAS; p = 0.003). However, the numbers of patients are low in colorectal, and ovarian cancer subgroup analyses suggesting that additional confirmatory studies will be necessary.