Targeting the PI3K Pathway in Gynecologic Malignancies

Abstract

Summary:

The phosphoinositide 3-kinase (PI3K) signaling pathway is implicated in tumorigenesis given its regulation over cell growth, cellular trafficking and angiogenesis. In gynecologic malignancies, alterations in PI3K signaling are common. Therefore, developing modulators of the PI3K pathway and identifying molecular markers to predict response is of great interest for these cancer types.

Purpose of Review:

This review explores the PI3K pathway aberrations common in gynecologic malignancies, the relevant therapeutic targets that have been explored to-date particularly given their success in endometrial cancers, and predictive biomarkers of response to therapy.

Recent Findings:

Landmark trials have been noted involving this pathway particulary in endometrial cancers. One phase II trial of the potent orally bioavailable mTOR inhibitor, everolimus, in combination with letrozole demonstrated an unprecedented clinical benefit rate (CBR) of 40% and high objective response rate (RR) of 32% in hormone agnostic endometrial cancers. This was followed by GOG 3007 that compared everolimus and letrozole to hormonal therapy yielding similar response rates but double progression free survival rates.

Introduction

The phosphoinositide 3-kinase (PI3K) signaling pathway is implicated in tumorigenesis given its regulation over cell growth, cellular trafficking, and angiogenesis. This complex pathway has a number of upstream targets, including HER2 and insulin growth factor 1 receptor, as well as downstream modulators such as AKT and the mTORC1 complex. Targeting this complex pathway has led to successful advances in cancer therapy, with several FDA approved therapies now widely available in multiple solid tumor types.

In gynecologic malignancies, alterations in PI3K signaling are common. Therefore, developing modulators of the PI3K pathway and identifying their molecular markers to predict response is integral. This review explores the PI3K signaling pathway aberrations that are common in gynecologic malignancies, the relevant therapeutic targets that have been explored to-date, and predictive biomarkers of response to therapy.

The PI3K Pathway

PI3Ks are part of the lipid-kinases family, originally discovered in the 1980s¹. Based on their structure and substrate specificity, PI3Ks have been categorized into 3 distinct classes: I , II and III. Class I PI3Ks are divided into class IA and class IB based on the differences in their activating receptors. The class most implicated in cancer is the Class IA PI3Ks consisting of one regulatory and one catalytic subunit. This includes a p85 regulatory subunit and a p110 catalytic subunit, which has 3 isomers (α, β and γ) respectively encoded by three distinct genes, PIK3CA, PIK3CB, and PIK3CD. Of these, the most frequently mutated in cancer appears to be PIK3CA².

The catalytic subunit PIK3CA (p110 α) and regulatory subunit PIK3R1 (p85) compose the heterodimer coupled to and activated by receptor tyrosine kinases (RTKs).³ Ligand binding to receptor tyrosine kinases (EGFR, HER2, VEGFR, FGFR2, IGF1R, PDGFR), leads to tyrosine phosphorylation of the intracellular receptor domain and activation of PI3K signaling. PI3K phosphorylates phosphatidylinositol-2-phosphate (PIP2) to PIP3. PIP3 propagates intracellular signaling by binding to AKT and the phosphoinositide dependent kinase 1 (PDK1). Ligand-independent activation of PI3K signaling is seen with specific somatic mutations of receptor tyrosine kinases or other PI3K pathway members rendering the pathway constitutively active.⁴

The tumor suppressor phosphatase and ten- sin homolog (PTEN) counteracts this by dephosphorylating PIP3 to PIP2, thereby negatively regulating the PI3K pathway⁵. A similar negative regulation action is carried out by Inositol polyphosphate 4-phosphatase type 2 (INPP4B) that hydrolyzes PIP3 to PIP2. Consequently, mutation or blockade of the above pathway can ultimately lead to an independent cell growth and enhanced overall motility, due to the lack of negative regulators of the PI3K/AKT/mTOR cascade⁶.

Activated AKT (also known as protein kinase B) regulates myriad downstream mediators involved in cell survival and metabolism by activating mammalian target of rapamycin (mTORC1)⁷. This occurs either directly via phosphorylation of Ser2448 or indirectly by phosphorylating tuberous sclerosis complex 2 (TSC2). Tuberous sclerosis complex proteins (TSC1/TSC2) are activating molecules for Ras homolog enriched in brain (Rheb)-GTPase, promoting the conversion of Rheb-GTP into its inactive GDP form, which in turn has feedback inhibition of mTORC1.By phosphorylating TSC2, AKT inactivates TSC1/TSC2, leading to indirect increase of mTOR activity.⁵

mTOR is a serine/threonine protein kinase, first identified in the budding yeast Saccharomyces cerevisiae during a test for resistance to the immunosuppressant drug rapamycin⁸. It performs its activities through 2 distinct complexes: mTORC1-Raptor and mTORC2-Rictor, which are now known to regulate differentiation of helper T cells⁹. Downstream, mTORC1 complex regulates factors that transition cells from G1 to S phase of the cell cycle and angiogenesis.⁶ mTORC2 leads to direct activation of AKT ultimately leading to further cell growth.⁶

PI3K Activity in Endometrial Cancer:

More than 80% of endometrial cancers harbor alterations in the PI3K pathway, which is the highest rate among all solid tumor types¹⁰. For both endometrioid and serous endometrial cancers, The Cancer Genome Atlas performed a large scale, integrated genomic analysis identifying four molecular subtypes based on somatic mutations, copy number alterations, and microsatellite instability status.¹¹ These four designated groups: 1) POLE ultramutated; 2) hypermutated/microsatellite unstable; 3) copy number low/microsatellite stable, and 4) copy number high (serous-like). The first group, called POLE “ultramutated” tumors, includes the clinically most favorable tumors with endometrioid histology. POLE refers to DNA polymerase epsilon, which is involved in DNA replication and harbors numerous hotspot mutations in tumors in this category. There are few copy number abnormalities in tumors of the POLE subtype, but a very high number of mutations. Mutations in PTEN, PIK3R1, PIK3CA, and KRAS are common in this group. The second group, called MSI “hypermutated” tumors, includes endometrioid tumors with microsatellite instability, few copy number abnormalities, and mutation rates lower than the “ultramutated” POLE group but higher than the copy number low/MSS group. Mutations in KRAS and PTEN are frequent. The third group is the copy number low/MSS group which is associated with PTEN and PIK3CA mutations and beta catenin mutations. The fourth group is the copy number high (serous-like) tumors with extensive copy number abnormalities.¹¹ In addition to most of the serous tumors, approximately one quarter of patients with grade 3 endometrioid fall in this category.

Relevant PI3K Pathway Trials in Endometrial Cancer

Single Agent Trials:

Several targeted agents along the PI3K pathway have been under investigation. Temsirolimus is a water-soluble ester of rapamycin, that functions as an mTORC1 inhibitor. It has been evaluated in several phase II trials in both treatment naive and heavily pretreated patients. In recurrent or metastatic endometrial cancer, a single-arm trial saw favorable responses in both treatment naïve patients and those treated with one prior line of therapy. In the chemotherapy naïve cohort, 14% of 29 evaluable patients had a partial response and 69% achieved SD. Among patients treated with prior chemotherapy, only 4% had partial responses but 48% had SD with a median of 4 months’ duration.¹² The AGO also performed a phase II study of temsirolimus in both endometrial and ovarian cancer with primary endpoint of PFS at 4 and 6 months. Here, women with advanced and recurrent endometrial cancer were only included if they were no longer amenable to curative surgery and/or radiation following no prior chemotherapy (or only prior adjuvant chemotherapy or endocrine therapy). Of the 20 evaluable endometrial cancer patients, 6 had PFS longer than 7 months and one had a PFS of over 31 months.¹³

The GOG looked at single agent temsirolimus in heavily pretreated patients with advanced or recurrent endometrial cancer. Of the 54 evaluable patients, 44% received 2 prior lines and 25% 3 prior lines of chemotherapy. This yielded 9% partial responses and 24% progression free survival at 6 months.¹⁴ Based on these response rates, temsirolimus was incorporated into NCCN guidelines for treatment of recurrent endometrial cancer. Trials incorporating temsirolimus have shown manageable toxicities, including hypertriglyceridemia, hyperglycemia, electrolyte abnormalities, and rash.

This benefit has also been demonstrated with other mTOR inhibitors. Ridaforolimus is an mTOR inhibitor that can be administered orally or intravenously. It was examined in the second and third line setting in a randomized fashion to chemotherapy and hormonal therapy in patients with no prior endocrine therapy. Here, mTOR inhibition showed improvement in PFS compared to chemotherapy or hormonal therapy (HR 0.53, 95% CI 0.31 to 0.91).¹⁵

Novel inhibitors of PI3K have also been explored. One phase II study evaluated the use of the PI3K inhibitor pilaralisib (SAR245408; XL147) in patients with advanced or recurrent endometrial carcinoma with up to two prior lines of chemotherapy. The objective response rate was 6% and PFS>6months was 11.9% in line with the single agent mTOR inhibitor data noted from temsirolimus.¹⁶ The GINECO group explored the PI3K inhibitor, BKM120, in a multicenter, single arm study in who received no more than one prior chemotherapy regimen stratified by low and high grade histology. The rate of non-progression following a dose-adjustment was 60% with a median PFS of 4.5 months (CI 95% 2.8–6.1) and OS of 10.1 months (CI 95% not feasible). Despite this, accrual was stopped at the higher dose rate given the large proportion of grade ¾ toxicities including cutaneous rash (54%), depressive events (47%), and anxiety (40%). At the lower dose, grade 3/4 AEs included hypertension (17%), cutaneous rash (13%), and fatigue (8%).¹⁷

Given favorable responses with PI3K inhibitors, other upstream targets such as AKT inhibitors have been explored as single agent or in combination. A Phase II study of the AKT inhibitor MK-2206 was performed in patients with recurrent endometrial cancer including all histologies except carcinosarcoma. However, of 36 evaluable patients only one achieved a partial response with 40% of patients requiring a dose reduction for a dermatologic related adverse event. In the exploratory analysis, only serous histology was linked to a greater than 6mo PFS when compared to other histologies.¹⁸ However, a recent basket trial evaluating AZD5363, an ATP-competitive pan-AKT kinase inhibitor, found decent results in a select population of patients with AKT1 E17K mutations. Considering a median 5 lines of prior therapy, the median PFS in gynecologic tumors was 6.6 months opening the door for exploration of AKT in genomic sub-populations.¹⁹ Another catalytic AKT inhibitor, Ipataserib is proving to be relatively safe in a new phase Ib study alongside various lines of chemotherapy following strong preclinical data in the gynecologic space.²⁰ This trial noted a singular dose-limiting toxicity of dehydration when in combination with taxol. As we demonstrate increased tolerability and safety with certain AKT inhibitors, their activity will need to be further explored.

Combination Trials: Endocrine and Antiangiogenic therapy

A landmark phase II trial of the potent orally bioavailable mTOR inhibitor, everolimus, in combination with letrozole demonstrated an unprecedented clinical benefit rate (CBR) of 40% and high objective response rate (RR) of 32%. Patients were on for a median of 15 cycles with 20% (7/35) of patients being taken off treatment given prolonged complete response at the discretion of the treating clinician. No patients required discontinuation of treatment because of toxicity. Most importantly, the trial was hormone receptor agnostic, but serous histology was the biggest predictor of lack of response. This eventually led to a subsequent trial, GOG 3007 that compared everolimus and letrozole to the hormonal regimen of alternating medroxyprogesterone acetate and tamoxifen. Everolimus and letrozole yielded similar response rates compared to hormonal therapy alone (24 vs 23%), however, median PFS was 6.3 months for the everolimus and letrozole group as compared to 3.8 months for patients receiving hormonal therapy. This benefit in progression free survival was improved when compared to the hormonal regimen in patients who were chemotherapy naïve at recurrence (21.6 vs 6.6mo). At a median follow-up of 14 months, median overall survival could not yet be calculated for the everolimus and letrozole group but was 16.6 months in the hormonal therapy group.. Adverse event rates were similar for both groups overall with two exceptions. Twenty-four percent of patients in the everolimus and letrozole group experienced grade 3 or higher anemia and 14% experienced grade 3 or higher hyperglycemia.²¹

Another phase II study evaluating the combination of everolimus, letrozole with metformin in advanced/recurrent endometrial cancer resulted in a 50% clinical benefit rate and 28% objective response rate (ORR), with the highest number of responders being progesterone receptor positive²². Temsirolimus was also evaluated in a randomized fashion alone vs with the addition of alternating megestrol acetate and tamoxifen. Given a high rate of thromboembolic disease (5/22 patients) in the combination arm, the combination arm was stopped early. The addition of megestrol acetate and tamoxifen did not enhance activity of temsirolimus with a response rate of 14 vs 22% respectively. In the temsirolimus alone arm, response rates were similar in patients who had prior chemotherapy vs those with no prior therapy (24 vs 19%).²³ One anticipated, ongoing phase II study is evaluating the combination of everolimus, letrozole and the cyclin dependent kinase (CDK) 4/6 inhibitor ribociblib.

Combination therapy with antiangiogenics has also been explored. A phase II study evaluated mTOR inhibition with temsirolimus in a three-arm design: paclitaxel/carboplatin/bevacizumab, paclitaxel/carboplatin/ temsirolimus or ixabepilone/carboplatin/bevacizumab as primary therapy for measurable stage III, IV or recurrent endometrial cancer. In this study, the temsirolimus regimen showed worse progression-free survival (PFS) compared to chemotherapy with bevacizumab (HR 1.43, 95% CI 1.06 to 1.93). Further, the PFS duration in each of these arms was not significantly increased in any experimental arm compared with historical controls (p > 0.039).²⁴

However, when temsirolimus was looked at in combination with bevacizumab alone, benefit was present but at the expense of much greater toxicity. Twenty-four percent of patients experienced clinical responses (one complete and 11 partial responses) and 46.9% survived progression free for at least six months. Median progression-free survival (PFS) and overall survival (OS) were 5.6 and 16.9 months, respectively. Of 49 evaluable patients, 32 patients had one or more grade 3 or 4 non-hematologic, non-metabolic toxicities at least possibly related to protocol therapy. Eight patients had one or more grade 3 or 4 hematologic toxicities and 14 patients had grade 3 or 4 metabolic toxicities at least possibly related to protocol therapy.²⁵

Combination Trials: Novel and Targeted Therapy

Further, it appears that mTOR inhibitors have synergistic effects while used in combination with other novel therapies that may yield promising results in vivo. In patients with loss of function PTEN mutations, PI3K inhibition sensitized endometrial cancer cells lines to PARP inhibition^26,27 In vivo cultures have also shown tumor response when combining a PI3K inhibitor with the RAS/RAF inhibitor AZD6244 in endometrial cancers.²⁸ Unfortunately, the combination of PI3K and MEK inhibition has been too toxic in clinical trials^{29, 30}. Further development will need to balance activity with toxicity.

PI3K Activity in Ovarian Cancer:

In ovarian cancer, PIK3CA appears to be the primary PI3K gene prone to somatic mutations, leading to a 12% missense mutation rate in epithelial ovarian cancers while others have shown that mutations in PIK3R1 have been found in 3.8% of ovarian cancer.^31,32 This appears to be histologically driven with the primary gain of function PIK3CA mutations noted in 30–40% of clear cell ovarian cancers³³. Intrinsic mechanisms of PI3K dysregulation leading to proliferation and angiogenesis can also include: mutations or amplifications in AKT, loss of the intrinsic regulators like PTEN, or loss of inactivating mutations in other tumor suppressor molecules such as TSC or LKB1 or INPP4B^34,35.

The expression levels of AKT and PIK3CA have been associated with ovarian cancer survival as designated by measuring AKT and mTOR phosphorylation levels^35,36. Therefore, it’s not surprising that presence of PIK3CA mutations predict the response to therapeutic targets such as mTOR inhibitors³⁷. Much like in endometrial cancers, the PI3K pathway serves a key role in ovarian cancer development, progression and eventual chemoresistance. This is largely due to activation mutations in the catalytic or regulatory subunit genes or in modification of downstream targets³⁸.

Relevant PI3K Trials in Ovarian Cancer

Only a few published studies have evaluated agents targeting the PI3K pathway in ovarian cancer. Temsirolimus, an mTOR inhibitor, was investigated in a phase II trial as a single agent in women with persistent or recurrent epithelial ovarian cancer. Of the 54 patients included in the trial, 5 (9.4%) experienced partial response, and 24.1% had PFS of ≥6 months¹⁴. In the AGO trial, temsirolimus monotherapy was evaluated in both patients with platinum refractory/resistant ovarian cancer and advanced/recurrent endometrial cancer. Among the 22 patients with ovarian cancer, a partial response was found in one patient (4.5%) and stable disease in 31.8% of patients¹³.

Another trial evaluated the use of the mTOR inhibitor everolimus with bevacizumab in recurrent ovarian, peritoneal, and fallopian tube cancers³⁹. Of 55 patients, only one patient had a complete response, while six had a partial response and 35 had stable disease as their best response. This was a lower response rate than has been previously seen with single agent bevacizumab, which can be up to 20%^40,41.

The inability to produce responses when transitioning from the phase I to 2 setting in ovarian cancer could be indicative of the low benchmark for response rates in refractory/recurrent disease or perhaps may be telling of the need for combination therapy given the complexity of the PI3K pathway. Much like in endometrial cancer, the phase II study of AKT inhibitor MK-2206 should menial response of stable disease in 4 patients with platinum resistant high grade serous cancers of ovary, peritoneum.⁴² The absence of response in this selected population suggests either pathway resistance in the setting of monotherapy or possibly partial inhibition of the AKT pathway opening the door for combinations. Currently, the AKT inhibitor afuresertib plus paclitaxel is being explored in a phase II clinical study to assess efficacy and safety in patients with platinum-resistant ovarian cancer.

PI3K Pathway in Other Gynecologic Malignancies

In the gynecologic malignancies, lower genital tract cancers such as cervical and vulvar are the least studied in PI3K signaling. Frequent PIK3CA amplification and AKT hyperphosphorylation has been noted in cervical cancer⁴³. The most common PIK3CA mutation in cervical cancer involves the p110α catalytic subunit of PI3K, leading to constitutive pathway activation and enhancing tumorigenesis^44,45. More commonly in cervical adenocarcinomas, mutations in PTEN and RAS have also been seen^43,46. The insulin signaling pathway that has been linked to PI3K inactivation has also been noted to be inhibited by the E6 protein that is present in the human papillomavirus (HPV)⁴⁷. This leads in increased PI3K activity in HPV positive cancer. HPV E6 also breaks down the inhibitor TSC leading to downstream activation of mTOR⁴⁸. Further, HPV E6 also deactivates LKB1 which is the activator of TSC itself⁴⁹. When this activity is blocked by mTOR inhibitors in in-vitro models, the phosphorylation of eukaryotic translation initiation factor 4E-binding protein 1 and markedly reduced the expression level of HPV E7 protein⁵⁰.

In vulvar cancer, mutations in the PI3K pathway, apart from TP53, including HRAS, KRAS, PIK3CA, KMT2D, PTEN, and FBXW7, have been consistently reported across small series. A large genomic profiling study showed that HPV-associated vulvar squamous cell carcinomas (VSCC) harbor alterations in the PI3K/mTOR pathway such as PIK3CA, PTEN, STK11, FBXW7, and SOX2 whereas HPV- independent VSCC showed more frequent mutations in TP53, TERT, CDKN2A, and CCND1. Further, it showed that approximately 50% of the HPV-associated VSCC have a potentially targetable alteration in the PI3K/mTOR pathway.⁵¹

Relevant PI3K trials in Other Gynecologic Malignancies:

To-date, there are very few studies evaluating PI3K or mTOR inhibitors in cervical, vulvar or vaginal cancers. The NCI conducted a study of single agent temsirolimus (CCI-779) in women with recurrent, unresectable, locally-advanced or metastatic carcinoma of the cervix. Of 38 evaluable patients, 1 patient had a PR and 19 patients had stable disease (57.6%) with a median duration 6.5 months. The 6-month progression free survival rate was 28% and the median progression free survival was 3.52 months.⁵² These responses, while not insignificant, are low compared to the prior established therapies specially in the era of immunotherapy.

MD Anderson published their retrospective, phase I experience in cervical, vulvar and vaginal carcinomas. Their retrospective review in cervical cancer patients evaluated response and survival in patients with PIK3CA mutation and PTEN loss/mutations. In patients with squamous cell carcinoma of the cervix, the presence of PIK3CA mutations was associated with a significantly longer median overall survival than those who did not (9.4 vs 4.2mo; p = 0.019). For patients who had matched therapy along the PI3K pathway, they had statistically significantly longer median progression-free survival than non-matched therapy (6 vs 1.5mo, p=0.026). When evaluating histologic differences, patients with adenocarcinoma had fewer PIK3CA mutations (14 vs 58%, p=0.016) and had longer median progression-free survival (14.2 vs 7.2mo; p=0.001). Matched therapy along the PI3K pathway led to a much higher clinical benefit rate when compared to non-matched therapy (53 vs 11%, p=0.08).⁵³

A second MD Anderson review focused on phase 1 patients with vaginal or vulvar cancers focused on investigative therapies including those with PI3K mutations. Of 27 evaluable patients who underwent phase 1 therapy, 5 patient with vaginal and 5 patients with vulvar cancers were administered either a PI3K or mTOR inhibitor. The mutational rate of 33% for PIK3CA and 67% for PTEN mutations in vaginal cancers and none in vulvar cancers. In this population, when comparing vaginal and vulvar malignancies there was no statistically significant difference in median overall survival at 6.2 and 4.6 months respectively amongst the two cancers. Among the 27 patients who underwent therapy, one experienced a partial response and three experienced stable disease for at least 6 months. Patients with a body mass index ≥30 had a significantly longer median overall survival duration than did those that did not.⁵⁴

Conclusion:

The PI3K pathway has been implicated in the pathogenesis and progression of gynecologic malignancies. As a result, several targeted therapies are currently being evaluated for therapeutic intent. Promising results in recent early phase trials should prompt further evaluation.

References of importance or outstanding importance:

1. Outstanding Importance. Slomovitz BM, Jiang Y, Yates MS, et al. Phase II study of everolimus and letrozole in patients with recurrent endometrial carcinoma. J Clin Oncol. 2015;33(8):930–936.

2. Importance. GOG 3007, a randomized phase II (RP2) trial of everolimus and letrozole (EL) or hormonal therapy (medroxyprogesterone acetate/tamoxifen, PT) in women with advanced, persistent or recurrent endometrial carcinoma (EC): A GOG Foundation study. Slomovitz, B.M. et al.Gynecologic Oncology, Volume 149, 2.

Abbreviations:

PI3K

phosphoinositide 3-kinases

RTK

receptor tyrosine kinase

PIP2

phosphatidylinositol-2-phosphate

PIP3

phosphatidylinositol-3-phosphate

mTOR

mammalian target of rapamycin

PTEN

phosphatase and ten- sin homolog

INPP4B

inositol polyphosphate 4-phosphatase type 2

TSC1

tuberous sclerosis complex 1

TSC2

tuberous sclerosis complex 2

IKB1

serine/threonine kinase 11

ER

estrogen receptor

PR

progesterone receptor

PFS

progression-free survival

CBR

clinical benefit rate

RR

response rate

CDK

cyclin dependent kinase

ORR

objective response rate

GOG

Gynecologic Oncology Group

AGO

Arbeitsgemeinschaft Gynäkologische Onkologie

HPV

human papillomavirus

VSCC

vulvar squamous cell carcinoma