The impact of PI3K inhibitors on breast cancer cell and its tumor microenvironment

PI3K and RTK Inhibitors

Several studies have shown that inhibition of the PI3K pathway induces the expression and phosphorylation of multiple RTKs, such as HER2, via the forkhead box O-regulated transcriptional process in breast cancer cells (Chandarlapaty et al., 2011; Garrett et al., 2011; O’Reilly et al., 2006; Tao et al., 2014). Especially in HER2-positive breast cancer, the upregulation of HER3 was detected in response to PI3K inhibition (Chakrabarty et al., 2012; Chandarlapaty et al., 2011; Garrett et al., 2011, 2013; Tao et al., 2014). HER3 is a kinase-defective member of the HER kinase family, and overexpressed HER2 can dimerize with HER3 to form HER2-HER3 heterodimers, which bind to and activate PI3K (Chandarlapaty et al., 2011). With sustained inhibition of mTOR, breast cancer cells can take advantage of upstream RTK feedback signaling to acquire resistance to drugs. Preclinical studies have shown that a combination of HER3-neutralizing antibody and a p110α-selective inhibitor markedly reduced the growth of breast cancer xenografts (Garrett et al., 2013). Another study showed the effectiveness of a combination of PI3K inhibitor and a dual EGFR and HER3 inhibitor to bypass this resistance mechanism (Tao et al., 2014). In addition, more evidence has indicated PIK3CA mutations mediating resistance to HER2 targeted agent in patients with HER2-positive solid cancer (Berns et al., 2007; Black et al., 2015; Rimawi et al., 2018). The wild-type breast cancer cell lines transfected with constitutively oncogenic PIK3CA mutations showed almost insensitive toward trastuzumab, suggesting the major role of PIK3CA in the development of resistance to trastuzumab (Berns et al., 2007). Importantly, it is demonstrated that PIK3CA mutations are associated with lower pathological complete response (pCR) rates to HER2 targeted therapy in primary HER2-positive breast cancer (Loibl et al., 2014, 2016; Majewski et al., 2015; Rimawi et al., 2018). In a pooled analysis of five prospective clinical trials evaluating lapatinib and trastuzumab, the pCR rate for PIK3CA mutation versus wild-type was 16.7% versus 39.1% (P < 0.001) (Loibl et al., 2016). Taken together, these studies suggest that simultaneous inhibition of PI3K and RTK is an effective therapeutic strategy to enhance drug efficacy.

A phase I trail was conducted to evaluate the safety of combination AKT inhibitor MK-2206 with trastuzumab, results indicating its safety and clinical activity (Hudis et al., 2013). The other two phase IB studies aimed to determine tolerability and maximum tolerated dose for buparlisib in combination with HER2 targeted drugs in HER2-positive, trastuzumab-resistant, advanced breast cancer (Guerin et al., 2017; Saura et al., 2014). In this patient population, the combination was well tolerated and preliminary evidence of antitumor activity were observed. Based on the recommended phase II dose of burparlisib as 100 mg/day in combination with 2 mg/kg weekly tratuzumab from phase IB, 50 patients with HER2+ locally advanced breast cancer resistant to trastuzumab-based treatment were treated with this therapeutic regimen (Pistilli et al., 2018). The overall response rate was only 10% and the primary endpoint was not met, demonstrating limited efficacy in this phase II study. Therefore, more evidence needs to be gained from ongoing clinical trials.

PI3K inhibitors and endocrine therapy

The crosstalk between PI3K activity and estrogen receptor (ER) signaling has been identified, as the PI3K pathway regulates the ER pathway both directly and indirectly (Miller, Balko & Arteaga, 2011). AKT phosphorylates ER at Ser167 to increase estrogen-induced ER transcriptional activity, and PI3K indirectly activates ER transcription by means of c-Jun complexes interacting with c-Foc to form AP-1 complexes (Bosch et al., 2015; Petz et al., 2002). In addition, emerging evidence has shown that estrogen signaling also has an impact on the PI3K pathway, as estrogen stimulation activates intracellular kinase pathways, including PI3K, IGF-1R, and EGFR (Miller et al., 2009; Song et al., 2006). Preclinical research indicated that suppression of PI3K signaling in an ER/PIK3CAmut model induced activation of ER-dependent transcription, including the ER promoter and genes with ER-binding sites in the coding sequence (Bosch et al., 2015). Interestingly, another study showed that reduced PTEN levels led to endocrine resistance in ER-positive breast cancer, which can be overcome by combination endocrine therapy with a PI3K inhibitor (Fu et al., 2014). Therefore, combined inhibition of PI3K and ER may be an effective strategy for breast cancer, and several clinical trials are undergoing.

A randomized double-blind phase II trial in ER-positive, HER2-negative metastatic breast cancer that had been resistant to treatment with aromatase inhibitor indicated that therapy efficiency did not benefit from the addition of the pan-class I PI3K inhibitor pictilisib (Genentech; GDC-0941) to fulvestrant (an ER antagonist) (Krop et al., 2016). The reason underlying the lack of significantly improved progression-free survival was because of toxicity limitations with pictilisib. Another phase II randomized preoperative window-of-opportunity study in postmenopausal women with newly diagnosed, operable ER-positive breast cancer showed that adding pictilisib to anastrozole (an aromatase inhibitor) significantly enhanced suppression of breast cancer cell proliferation compared with anastrozole alone (Ki 67 staining; 83.8% vs. 66.0%) (Schmid et al., 2016). Further, the sensitivity to pictilisib was independent of PIK3CA mutations and the combination treatment showed a remarkable anti-proliferation effect in luminal B primary breast cancer (Ki 67 staining; 37%). In addition, a randomized double-blind placebo-controlled phase III trial in postmenopausal, HR-positive, HER2-nagative, advanced breast cancer was performed to assess buparlisib plus fulvestrant (Baselga et al., 2017). The results of this BELLE-2 trail indicated that pan-PI3K inhibitor with endocrine therapy was clinically meaningful benefit in the total patient population (median progression-free survival: 6.9 months vs. 5.0 months), while serious adverse were reported as high as 23% in the buparlisib group compared with 16% in the placebo group. In consideration of PI3K inhibitor down-regulating the phosphorylation of AKT following mTOR inhibition, the phase III, randomized, placebo-controlled BELLE-3 trail was designed to test buparlisib combination with fulvestrant in patients who relapsed on or after endocrine therapy and mTOR inhibitors (Di Leo et al., 2018). However, the safety profile of the therapeutic regimen was not recommended to further development. Interestingly, preclinical studies have suggested that oncogenic PIK3CA driven the development of breast cancer (Liu et al., 2011; Utermark et al., 2012), and the PI3Kα-selective inhibitors shown more effective and more tolerable toxicity profile in clinical trails (Juric et al., 2018; Mayer et al., 2017). Phase III trials investigating α-selective PI3K inhibitors combined with ER antagonist in the endocrine resistant setting are currently ongoing, including SOLAR-1 and SANDPIPER.

PI3K and poly-ADP-ribose polymerase inhibitors

Two studies provided the first evidence that a PI3K inhibitor induces impairment of DNA homologous recombination in BRCA-proficient triple-negative breast cancer, with an increase in poly-ADP-ribosylation and downregulation of BRCA1/2 (Condorelli & Andre, 2017; Ibrahim et al., 2012; Juvekar et al., 2012; Rehman, Lord & Ashworth, 2012). The authors also showed that adding the poly-ADP-ribose polymerase (PARP) inhibitor olaparib to PI3K inhibitors markedly attenuated tumor activity in triple negative breast cancer with wild-type BRCA1. Pradip et al., conducted a study using BRCA-competent breast cancer-bearing mouse models and showed that GDC-0980 (a dual PI3K-mTOR inhibitor) enhanced the antitumor effects of the PARP inhibitor ABT888 plus carboplatin by inhibition of tumor cell proliferation and tumor-induced angiogenesis; however, combination with GDC-0941, a pan-PI3K inhibitor, failed to suppress tumor growth in MDA-MB231 breast cancer cells (De et al., 2014). These results indicate that co-administration of PI3K inhibitor and PARP inhibitor may a rational therapeutic strategy.

Based on these preclinical studies in animal models, a phase I clinical trial was performed on the combination of the PI3K inhibitor BKM120 and the PARP inhibitor olaparib to test side effects and the maximum tolerated dose (Matulonis et al., 2017). Clinical benefit of the combination treatment was observed in patients with germline BRCA mutation or germline wild-type BRCA, but the dose of BKM120 is limited by the co-administered drugs because of its toxicity. With further clinical study of PI3K inhibitors, other PI3K inhibitors might be more suitable for combination therapy. Randomized phase II studies will be conducted to further confirm the efficacy of the PI3K/PARP inhibitor combination.

Combination therapy in breast cancer: further evidence to explore

Preclinical and clinical studies have provided several potential combination PI3K inhibitors and other therapeutics, but more evidence is needed to clarify additional questions. Whether targeted kinase inhibitors are dependent on the PI3K molecular alteration should be examined. Interestingly, several studies revealed a key effect of oncogenic PIK3CA on mammary cell fate, which activates a multipotent genetic program at the early stage of tumor initiation and controls tumor heterogeneity (Koren et al., 2015; Van Keymeulen et al., 2015). Whether inhibition of PI3K has an impact on blocking PIK3CA-induced initiation of malignancy still needs to be verified by more research. Additional studies should also focus on identifying the patients who benefit from the combination treatment.

Impact of PI3K inhibitors on stromal fibroblasts

The interactions of stromal cells and epithelial cells play a pivotal role in normal mammary development, including in regulating epithelial cell polarity and determining mammary duct formation, and the dysfunction of this interaction results in mammary epithelial cell proliferation and even malignant transformation (Mao et al., 2013; Place, Jin Huh & Polyak, 2011). Fibroblasts, the abundant and critical cell type in the normal mammary stroma, serve as the basic cellular component of connective tissue and maintain the extracellular environment by secreting and remodeling the extracellular matrix (ECM) (Mao et al., 2013; Soysal, Tzankov & Muenst, 2015). Cancer-associated fibroblasts, with higher proliferative and other cellular activities, produce increasing levels of various growth factors, cytokines, and ECM-degrading proteases, and substantial data support a role for these cancer-associated fibroblasts in breast cancer progression (Aboussekhra, 2011; Orimo et al., 2005; Tyan et al., 2011). Orimo and colleagues showed that cancer-associated fibroblasts extracted from human breast cancer significantly promote the growth of breast cancer cells and increase angiogenesis by recruiting endothelial progenitor cells (Orimo et al., 2005).

Phosphatase and tensin homolog, as a negative regulator of PI3K, has been extensively demonstrated to function in the stromal fibroblasts to suppress mammary epithelial tumors (Trimboli et al., 2009; Wallace et al., 2011). Genetic ablation of PTEN in stromal fibroblasts of mouse mammary glands have been shown to remodel the ECM by increasing the level of the transcription factor Ets2, which directly targets the promoters of MMP9 and CCL3 (Trimboli et al., 2009). As a result, the PTEN stromal-deleted breast cancers exhibit high levels of innate immune cell infiltration and tumor vascular structures (Trimboli et al., 2009). Another recent study further indicated that fibroblast-specific PTEN inactivation results in failure of elongation of mammary ducts and promotes abnormal alveolar side-branching (Sizemore et al., 2017). Importantly, the PTEN-null mice showed an expansion of the mammary epithelial stem cell-enriched basal/myoepithelial population and well-defined activity of the stem cells tested in vitro (Sizemore et al., 2017). In a pancreatic cancer mouse model, the inhibition of PI3Kγ was demonstrated to reduce cancer-associated fibroblast-induced collagen production by regulation of tumor associated macrophages (Kaneda et al., 2016a).

Pleiotropic impact of PI3K inhibitors on tumor angiogenesis

Tumor vessel formation is critical for persistently supplying nutrients and oxygen for high proliferating cancer cells, and indeed, aberrant tumor vessel structure and function are hallmarks of cancer (Hanahan & Weinberg, 2011). Microvascular density is an independent prognostic indicator in invasive breast cancer (Arora et al., 2002; Nieto et al., 2007; Tsutsui, Kume & Era, 2003), and high microvascular density in breast cancer is associated with tumor invasiveness and metastasis (Ribatti et al., 2016). A variety of tumor microenvironment cells are involved in the formation of tumor-associated neovasculature. Endothelial cells respond to tumor overexpressed angiogenic signals and are activated to form the inner surface of tubular vessel (Potente, Gerhardt & Carmeliet, 2011), and the infiltrating immune cells are recruited to promote tumor angiogenesis by releasing angiogenic factors (Hanahan & Coussens, 2012). Remarkably, PI3K isoforms each function in these cell types, such as PI3Kα in endothelial cells (Soler et al., 2013), PI3Kγ/PI3Kδ in immune cells and PI3Kβ in platelets (Martin et al., 2010; Rivera et al., 2015). It is therefore likely that PI3K inhibition has pleiotropic effects on angiogenesis by different mechanisms.

Impact of PI3K inhibitors on tumor-infiltrating immune cells

Tumor-infiltrating immune cells constitute a major cell population of the breast cancer microenvironment and are reprogrammed by breast cancer cells to support tumor onset and progression by suppressing the antitumor immune response (Jezierska-Drutel, Rosenzweig & Neumann, 2013; Place, Jin Huh & Polyak, 2011; Soysal, Tzankov & Muenst, 2015). Immunosuppression regulatory T cells (Tregs) are involved in immune tolerance to turn into more aggressive and higher grade breast cancer (Bohling & Allison, 2008; Ohara et al., 2009; Plitas et al., 2016), while CD8+ T lymphocytes infiltrating in invasive breast cancer have been considered as a good predictive and prognostic marker (Al-Saleh et al., 2017; Mahmoud et al., 2012). Myeloid-derived suppressor cells play a crucial role in breast cancer malignant individualities and are recruited by tumor cells to produce and secrete pro-tumor cytokines (Chen, Zhang & Massague, 2011; Peng et al., 2016b; Shou et al., 2016). Further, the immunosuppressive effect of these cells is associated with weakening effectiveness of immunotherapy and even resistance to checkpoint blockade therapy in breast cancer (De Henau et al., 2016; Shou et al., 2016). Surprisingly, more and more evidence has suggested that PI3K isoforms in leukocytes have an effect on innate and adaptive immune responses, such as PI3Kδ in regulatory T cells and PI3Kγ in myeloid cells (Ahmad et al., 2017; Ali et al., 2014; De Henau et al., 2016). Given the significance of the immune reaction in breast cancer, we focus on the the immunomodulatory impact of PI3K inhibition on tumor regression and the promising benefit to enhance immunotherapy.