Initiating breast cancer by PIK3CA mutation

Abstract

PIK3CA mutations confer constitutive activation of PI3K, which initiates intracellular kinase signaling cascades that promote cell proliferation and survival. Recent studies by Meyer and colleagues, and Liu and colleagues demonstrate that expression of the H1047R exon 20 mutant of PIK3CA in luminal mammary epithelial cells induces tumorigenesis, implying that PIK3CA mutation is an early event in breast cancer. PIK3CA-H1047R-initiated tumors exhibit variable dependence on the oncogene and variable sensitivity to PI3K inhibition. Amplification of the oncogenes MYC and MET was observed in tumors that recurred following silencing of PIK3CA-H1047R, suggesting that these pathways represent mechanisms of escape from PI3K inhibition.

Phosphatidylinositol 3-kinase (PI3K) phosphorylates phosphatidylinositol 4,5-bisphosphate (PIP₂) to produce phosphatidylinositol 3,4,5-trisphosphate (PIP₃) at the cytoplasmic face of the plasma membrane. PIP₃ promotes membrane recruitment and activation of several proteins that drive cell proliferation and survival, including AKT, PDK1, and SGK. Class IA PI3Ks are heterodimers composed of a p110 catalytic subunit and a p85/p55 regulatory subunit, each of which has several isoforms. The PI3K pathway is the most frequently mutated pathway in breast cancer, and mutations occur in signaling nodes both upstream and downstream of PI3K [1].

Activating mutations in PIK3CA (which encodes the p110α catalytic subunit) occur in approximately 30% of breast cancers and are more frequent in estrogen receptor-positive (ER⁺) breast cancers [2, 3]. Eighty percent of PIK3CA mutations occur in two 'hot spots' within exons 9 and 20, which encode the helical and kinase domains, respectively. The E542K and E545K (exon 9) mutations may confer a gain-of-function by disrupting an inhibitory interaction between p110α and p85 [4]. The H1047R (exon 20) mutation may induce an allosteric change that mimics Ras-GTP binding, making this mutant independent of interaction with Ras-GTP [5]. Both mutants are constitutively active, transform cells in culture, and promote tumorigenicity in xenograft models. Cancer cell lines harboring PIK3CA mutations are highly sensitive to PI3K pathway inhibitors [6, 7], rendering this pathway a drug target of high interest for cancer therapy. PIK3CA mutations have been found at similar frequencies in breast ductal carcinoma in situ (DCIS) lesions, DCIS adjacent to invasive ductal carcinoma (IDC), and IDC [8], suggesting that these mutations are early events in breast tumorigenesis and therefore may promote transformation of normal breast epithelial cells.

A recent study by Meyer and colleagues [9] revealed that expression of the PIK3CA-H1047R mutant in mammary epithelial cells is sufficient to induce tumor formation in transgenic mice. PIK3CA-H1047R expression driven by Cre-mediated recombination induced by either the WAP promoter (which is active in alveolar progenitor and differentiated secretory luminal epithelial cells) or the MMTV promoter (which is active in differentiated luminal mammary epithelial cells) induced the formation of mammary tumors of varying histologic subtypes. Tumor cells expressed markers associated with both luminal and basal epithelial lineages, suggesting that tumors with basal characteristics can arise from luminal cells. The authors postulate that PIK3CA-H1047R may (a) transform multi-potent progenitor cells to allow both luminal and basal differentiation, (b) induce de-differentiation of luminal cells to multi-potent progenitors, which then give rise to both lineages, or (c) do both. Involuting mammary glands (which undergo ductal pruning following pregnancy and lactation) from PIK3CA-H1047R mice showed a reduction in the number of apoptotic cells and delayed involution in comparison with controls. PIK3CA-H1047R tumors also showed very low rates of apoptosis and higher levels of phosphorylated AKT than mammary tumors from another model (MMTV-NeuNT), suggesting that PIK3CA-H1047R prevents cell death by increased PI3K/AKT pathway activation.

In another study, Liu and colleagues [10] reported that PIK3CA-H1047R-induced mammary tumors exhibit variable dependence on this oncogene. Transgenic mice expressed PIK3CA-H1047R under the control of an MMTV-regulated, doxycycline-inducible system. Mice treated with doxycycline showed increased phospho-AKT levels in mammary epithelial cells and formed mammary tumors of varying histologic subtypes. Silencing of PIK3CA-H1047R by withdrawal of doxycycline decreased tumor phospho-AKT levels, decreased proliferation, increased apoptosis, and induced complete tumor regression in one third of the mice. Two thirds of tumors partially regressed and then resumed growth. Some recurrent tumors that maintained high levels of P-AKT and P-S6 were sensitive to the PI3K inhibitor GDC-0941, whereas tumors with low P-AKT and P-S6 were insensitive to this agent. This suggests that some PIK3CA-H1047R-induced tumors escape from dependence on PI3K. GDC-0941-resistant and PIK3CA-H1047R-independent tumors exhibited amplification of the oncogenes MYC, MDM2, and/or MET. The authors demonstrated tumor dependence on MYC (using short-hairpin RNA knockdown) and MET (using a kinase inhibitor) and showed that MYC overexpression circumvented dependence on PI3K.

These studies have important implications for the role of PI3K mutations in breast cancer. First, these works show that PIK3CA-H1047R induces mammary epithelial cell transformation in vivo and support the notion that PIK3CA mutation is an early event in breast cancer. Second, the paper by Liu and colleagues [10] affirms that PIK3CA-mutant tumors are dependent, in whole or in part, on this oncogene. Some tumors that recurred following silencing of PIK3CA-H1047R showed sensitivity to a PI3K inhibitor, indicating continued addiction to PI3K. Since PI3K pathway inhibitors preferentially inhibit the growth of cancer cells harboring PIK3CA mutations [6, 7], such mutations are being used as an inclusion criterion in ongoing clinical trials with these agents to enroll patients who are most likely to benefit. The authors' findings support this concept. Third, in agreement with the existence of PIK3CA mutations in all subtypes of breast cancer, PIK3CA-H1047R expression induced mouse mammary tumors expressing both luminal and basal markers. Issues that remain to be addressed are the mechanism underlying the variable histologic subtypes of PIK3CA-H1047R-induced mammary tumors and the effects of PIK3CA-H1047R on luminal/basal cell differentiation. Fourth, the majority of PIK3CA-H1047R-induced mammary tumors recurred following an initial regression after oncogene silencing. Such recurrence was driven by MYC and MET, suggesting that therapies targeting the PI3K pathway may be most effective when used in combination with agents that block such escape mechanisms.