Double trouble for cancer gene

See allHide authors and affiliations

Science  08 Nov 2019:
Vol. 366, Issue 6466, pp. 685-686
DOI: 10.1126/science.aaz4016

Cancer is predominantly a genetic disease. Numerous gain-of-function mutations and gene amplifications that promote cell growth and survival have been identified in human cancer genomes through the efforts of large-scale sequencing projects. One of the most frequently mutated oncogenes in all human cancers is PIK3CA [phosphatidylinositol 4,5-bisphosphate (PIP2) 3-kinase catalytic subunit α], which encodes the catalytic subunit (p110α) of phosphoinositide 3-kinase (PI3K) (1). Oncogenic mutations in PIK3CA hyperactivate downstream signaling and promote phenotypes associated with malignancy. On page 714 of this issue, Vasan et al. (2) find that double mutations (two different mutations in one allele) in PIK3CA occur with much higher frequency in cancer genomes, particularly breast cancers, than previously thought. Double mutations result in increased PI3K pathway activity and tumor growth and predict increased sensitivity of human breast cancer to PI3K inhibitors.

Small-molecule inhibitors that target the PI3K pathway for cancer therapy have been developed. In May 2019, the U.S. Food and Drug Administration approved alpelisib, a p110α inhibitor for the treatment of postmenopausal women with estrogen receptor–positive (ER+), PIK3CA-mutant advanced or metastatic breast cancer, in combination with the ER antagonist fulvestrant. Approval was prompted by a phase 3 clinical trial that showed doubling of progression-free survival in patients with ER+ and PIK3CA-mutant breast cancer treated with alpelisib and fulvestrant, compared with patients with wild-type PIK3CA (3). A recent study also identified an exceptional responder (a patient who responded to monotherapy in early clinical trials). This patient had double PIK3CA mutations, whereas the majority of recurrent PIK3CA mutations identified in sequencing projects were mostly heterozygous, single mutations (4).

The enhanced sensitivity to alpelisib was likely due to double mutation in PIK3CA, so Vasan et al. investigated the pattern and frequency of oncogenic mutations in breast and other cancers. They found that double PIK3CA mutations recur in different tumors with much higher frequencies than previously appreciated. Double mutations in PIK3CA occurred in 8 to 12% of breast cancer patients with primary as well as metastatic tumors, depending on the cohort analyzed. Double PIK3CA mutations were enriched in luminal-subtype ER+ breast tumors, which have a high frequency of single-hotspot PIK3CA mutations (∼40%). Additionally, double PIK3CA mutations occurred in uterine cancers (27%) and colorectal cancers (12%) and also recur to a lesser extent in numerous other solid tumors.

The three most frequent single-hotspot mutations in PIK3CA are His1047Arg in the kinase domain and Glu542Lys or Glu545Lys in the helical domain (5). Perhaps surprisingly, the double mutations comprise one of these major hotspot mutations and a second minor site (Glu453, Glu726, or Met1043). Double hotspot mutations—His1047Arg plus Glu542Lys or Glu545Lys—were not found, and if they do arise in the cancer cell of origin, presumably they are not selected for during clonal expansion of the primary tumor. The major-plus-minor double mutations occur in the same cancer cell and on the same allele, in cis, as opposed to on different alleles, in trans. Therefore, the resulting mutant p110α protein harbors both mutations as opposed to two separate species of p110α with a different single mutation.

PI3K activation is a complex mechanism that includes engagement of the regulatory p85 subunit of PI3K to phosphotyrosine-containing sequences in upstream receptor tyrosine kinases (RTKs), which in turn relieves the catalytic inhibition of the p110α subunit (1). Vasan et al. show that the double PIK3CA mutations induce PI3K hyperactivation by a combination of disruption of the p85-p110α interaction and enhanced binding of p110α to membranes, where its substrate (PIP2) is located. The net effect is increased production of the PI3K lipid product and second messenger phosphatidylinositol 3,4,5-triphosphate (PIP3), which in turn engages downstream effectors, such as the kinase AKT (see the figure).

The consequences of harboring double PIK3CA mutations on the same allele are far-reaching. Vasan et al. show that double PIK3CA mutations lead to significantly increased PI3K activity and downstream pathway activation when compared with those of single-hotspot mutations. This results in enhanced tumor growth in vivo. Therapeutically, breast cancer cells with double PIK3CA mutations show enhanced sensitivity to alpelisib in vitro and in vivo, compared with that of single-hotspot mutants. Moreover, a retrospective analysis of clinical responses to PI3K inhibitors in breast cancer trials showed that patients with tumors with multiple PIK3CA mutations experience a greater overall response to alpelisib as compared with patients with single-mutant tumors.

Although single and double mutations in PIK3CA are prevalent in some cancers, hyperactivation of the PI3K-AKT pathway is observed in more than 50% of human tumors (6). Multiple other genetic alterations in genes that either regulate or transduce PI3K signaling are also frequent. These include amplification or mutations of RTKs, such as members of the epidermal growth factor receptor (EGFR) family, and oncogenic activating mutations or amplification in the three AKT genes: AKT1, AKT2, and AKT3 (6). Signal termination in the PI3K pathway is achieved primarily through the action of lipid phosphatases, including the tumor suppressor proteins phosphatase and tensin homolog (PTEN), inositol polyphosphate 4-phosphatase type II B (INPP4B), PH domain and leucine-rich repeat protein phosphatase 1 (PHLPP1), and PHLPP2 (7). Genetic inactivation of these tumor suppressors in mice leads to enhanced PI3K-AKT signaling and occurs in many human cancers. Thus, genetic mutations in components of the PI3K pathway render it the most frequently mutated pathway in human cancer.

Growth factor signaling in normal tissues and cancer

In normal tissues, growth factors activate RTKs, leading to recruitment of PI3K, which converts PIP2 to PIP3. This leads to recruitment of downstream effectors, such as AKT, that stimulate cell growth, proliferation, and survival. In cancer, the double mutant PIK3CA oncogene encodes hyperactive p110α that is independent of RTK signaling, producing excess PIP3, which leads to hyperactivation of AKT and uncontrolled cell growth and survival.


However, single-hotspot mutations in PIK3CA are typically insufficient to promote malignancy, and additional “second hit” mutations in cancer-causing genes are required. Vasan et al. propose that the presence of double PIK3CA mutations in the same allele follows the “oncogene addiction” paradigm (8), in which tumors depend on a single gene for malignant transformation and are thus likely to die when the corresponding oncoprotein is therapeutically targeted, whereas single PIK3CA hotspot mutations coexist in the same cell and tumor with other PI3K pathway lesions, such as PTEN inactivation or AKT oncogenic mutations.

These findings are likely to renew interest in the clinical development of PI3K inhibitors. Dose-limiting toxicities and acquired resistance have been noted in patients treated with PI3K inhibitors (9), and therefore combination strategies with chemotherapy, immunotherapy, and other targeted agents will likely be most effective. Although alpelisib is potent and highly selective, it is not a p110α-mutant–specific inhibitor, and this may limit efficacy. PI3K inhibitors under clinical evaluation, such as GDC-0077, appear to be selective for mutant p110α (10) and therefore may be more effective in patients with double PIK3CA mutations. Could double mutations recur in other oncogenes? The approach of Vasan et al. could reveal a more complex mutational spectrum in other oncogenes than previously appreciated.

References and Notes

Acknowledgments: This work was supported by NIH grants R01-CA177910 and R01-CA200671 and the Ludwig Center at Harvard.
View Abstract

Navigate This Article