HER2-Positive Breast Cancer
A link between Akt/mTOR pathway activation and HER2 overexpression was first established in primary breast tumors, implicating the pathway in progression of HER2-positive breast cancer. Trastuzumab(Drug information on trastuzumab) resistance is caused by hyperactivation of the PI3K pathway, resulting from either PIK3CA mutation, loss of PTEN, or activation of parallel pathways.[35,36] Treatment with mTOR inhibitors is an effective strategy for overcoming preclinical trastuzumab resistance secondary to PTEN loss. Unlike trastuzumab, lapatinib (Tykerb) sensitivity is independent of PI3K pathway activation. A study aimed to identify pathway alterations that play a role in lapatinib resistance. Similar to prior studies that have assessed mechanisms of lapatinib sensitivity, the PI3K signaling pathway was not found to be hyperactivated in the lapatinib-resistant cells. However, the resistant cells demonstrated activation of mTORC1 by an Akt signaling–independent mechanism. This was evidenced by downstream activation of mTOR effectors; moreover, these lapatinib-resistant cells were sensitive to mTOR inhibition using a rapalog and a dual PI3K/mTOR inhibitor.
Metastatic setting. There is a biological rationale for the use of PI3K/mTOR pathway inhibitors, in both trastuzumab- and lapatinib-refractory disease, based on preclinical data. A dose-escalation study of everolimus in combination with paclitaxel(Drug information on paclitaxel) and trastuzumab in trastuzumab-refractory HER2-positive metastatic breast cancer has been reported. This study demonstrated clinical activity with an ORR of 44%, and control of disease for 6 months or more in 74% of the patients. Based on this study, paclitaxel and trastuzumab with either everolimus at a dose of 10 mg per day or placebo is being tested in a phase III randomized study (Table). An additional phase I study combined everolimus with weekly trastuzumab and vinorelbine in pretreated HER2-positive metastatic breast cancer. Everolimus doses of 5 mg per day and 30 mg per week were established as safe for further development, with grade 3 or 4 neutropenia as dose-limiting. Significant anti-tumor activity was noted, with an ORR of 19.1%, disease control rate of 83%, and median PFS of 30.7 weeks. This regimen is now being tested for the treatment of progressive or new brain metastases in patients with HER2-positive breast cancer. Everolimus crosses the blood-brain barrier, and patients who progress after radiation are eligible for treatment on this study (Table).
Early clinical studies have also combined dual PI3K/mTOR inhibitors with trastuzumab. A phase I/Ib dose-escalation study of BEZ235, a dual PI3K/mTOR inhibitor, with trastuzumab aimed to enrich for patients with PI3K pathway alterations by limiting the study to patients with mutations in PIK3CA or PTEN or loss of PTEN by IHC in tumor samples. The maximum tolerated dose (MTD) for BEZ235 is estimated to be 600 mg per day in combination with trastuzumab, and this dose is being carried forward to the dose-expansion cohort. The combination also appears to be clinically active, with a CBR of 27%. In this study, there does not appear to be an association between PI3K pathway alteration and response. Given the short half-life of BEZ235, a twice-daily schedule is under investigation in a phase Ib/II study in combination with trastuzumab (Table).
Studies with other combinations, including temsirolimus and neratinib (HKI272) (phase I/II) or everolimus and lapatinib (phase II) are ongoing (Table).
Adjuvant/neoadjuvant setting. A randomized, phase II study in HER2-positive early-stage breast cancer is assessing whether adding everolimus to trastuzumab in the neoadjuvant setting improves clinical tumor response rate (Table).
Triple-Negative Breast Cancer Preclinical studies
There is strong evidence that the PI3K pathway is commonly altered in TNBC. Copy number, gene expression, proteomic, and sequencing studies in TNBC have shown the following pathway alterations: (a) a loss of PTEN, which is a negative regulator of the PI3K pathway, in up to a third of TNBCs; (b) PIK3CA and PTEN mutations; and (c) activation of Akt with PTEN loss. A recent study compared TNBC to HER2-positive or ER-positive breast cancer and found a relatively low (9%) PIK3CA mutation frequency. However, using a pathway-based approach, this study highlighted that despite the low frequency of PIK3CA mutations, inferred PI3K activity was the highest in basal-like breast cancers. This increased pathway activity may be due to alternative mechanisms, such as PTEN loss; moreover, this finding provides strong support for using PI3K/mTOR inhibitors in this subtype. The basal-like subtype of TNBC has been shown to be sensitive to mTORC1 inhibitors, in both in vitro and in vivo studies, resulting in cell growth inhibition.[46,47]
TNBC behaves as a clinically heterogeneous disease, and this heterogeneity was well-characterized using gene expression profiling of primary TNBC tumors. Clustering analysis of TNBC gene expression data identified six subtypes, and tumors within the mesenchymal and mesenchymal stem cell-like subtypes were sensitive to dual PI3K/mTOR inhibitors. The frequent alterations in the PI3K pathway in TNBC and preclinical data with PI3K/mTOR inhibitors provide a strong rationale for further investigation of PI3K/mTOR inhibitors in TNBC.
Metastatic setting. There are very limited clinical data on TNBC treated with either rapalogs or PI3K/mTOR inhibitors. A phase II trial with everolimus and carboplatin(Drug information on carboplatin) in patients with metastatic TNBC found a CBR of 38% (8 of 21 assessable patients) in a pretreated population. A phase I study in metastatic TNBC, using a combination of temsirolimus, cisplatin(Drug information on cisplatin), and erlotinib (Tarceva), is ongoing (Table).
Adjuvant/neoadjuvant setting. A randomized phase II neoadjuvant study in TNBC compared paclitaxel with or without everolimus followed by FEC (T-FEC). Everolimus was administered at a dose of 30 mg orally weekly for a total 12 weeks, concurrently with paclitaxel. There was no significant difference with the addition of everolimus in the pathologic complete response rate (pCR; 26% vs 30% for everolimus arm) or 12-week response rate by ultrasound (30% vs 48% for everolimus arm). The clinical activity of everolimus in combination with cisplatin and paclitaxel is being evaluated in a randomized phase II neoadjuvant study in TNBC disease (Table).
PI3K Pathway Inhibitors and Drug Combinations in Clinical Trials
The PI3K pathway is very tightly regulated, and there is crosstalk between it and several other pathways. In complex pathways with negative regulatory feedback loops and other escape pathways, such as the PI3K pathway, blockade of one component of the pathway is unlikely to accomplish complete inhibition of the pathway. It has been speculated that this is one of the reasons for poor clinical responses with allosteric mTORC1 inhibitors or rapalogs alone that are effective against mTORC1 but not mTORC2. This has propelled the development of novel ATP-competitive inhibitors of mTOR kinase activity, which block both mTORC1 and mTORC2, as well as the development of dual PI3K/mTOR inhibitors. A number of dual mTORC1/2 inhibitors have been identified, including INK128 (Intellikine), CC223 (Celgene), OSI-027 (OSI Pharmaceuticals), AZD8055 (AstraZeneca), AZD2014 (AstraZeneca), and Palomid 529 (Paloma Pharmaceuticals) (Figure).
Dual PI3K/mTOR inhibitors (Figure) have also been developed in the hope of overcoming the loss of feedback inhibition or PI3K activation observed with rapalogs. The mTORC1 pathway is one of the prominent negative feedback regulators of the PI3K pathway; inhibition of mTORC1 can release this feedback inhibition and activate the PI3K pathway. BEZ235, a dual PI3K/mTOR inhibitor, showed higher anti-proliferative activity than rapamycins in a preclinical study with multiple cancer cell lines, and it had antitumor activity in a trastuzumab-resistant breast cancer cell line. In breast cancer, dual PI3K/mTOR inhibitors are being combined with everolimus, endocrine therapies (exemestane and letrozole(Drug information on letrozole)), chemotherapy (paclitaxel), and anti-HER2 therapy (trastuzumab) (Table).
Combinations of MEK and PI3K pathways inhibitors are being explored in attempts to block escape pathways, which may become prominent when the PI3K pathway is inhibited. There is also a rationale for combining insulin-like growth factor-1 receptor (IGF1R) signaling inhibitors with mTOR inhibitors, since IGF1R is hyperactivated as a result of the negative feedback loop release that occurs with mTORC1 inhibition.
Predictors of Response or Resistance
The identification of markers that predict for response to PI3K/mTOR inhibitors can significantly aid in selection of patients who are most likely to benefit from this targeted therapy. Preclinical data indicate that cell lines with PIK3CA or PTEN mutations, as well as elevated basal and post-treatment phospho-Akt levels, are likely to be sensitive to rapamycin. Based on this, there is a need to assess phospho-Akt as a potential biomarker for rapamycin response. Several clinical studies have evaluated PIK3CA and PTEN mutation status in tumors and have not conclusively shown that PI3K pathway mutations sensitize tumors to PI3K/mTOR inhibitors. PIK3CA mutations were analyzed in tumors including breast, cervical, endometrial, and ovarian, from a single-institution phase I program. Of a total of 140 patients, PIK3CA mutations were identified in 25 (18%) of these patients. A higher response rate was seen in patients with PIK3CA mutations, but six of the seven patients who experienced tumor response had received a combination of a rapalog and a cytotoxic drug (liposomal doxorubicin(Drug information on doxorubicin)). It is difficult to draw any conclusions from this study regarding an association of PI3K/mTOR inhibitor response with PIK3CA mutation status, since the study was small and only included patients being treated with rapalogs.
The majority of the clinical trials in the metastatic setting evaluate the primary tumors for PI3K pathway mutations; studies have shown that there may be discordance between the PIK3CA mutation status of the primary tumor and that of the metastatic site, likely due to tumor evolution. It is possible that an association between therapy response to PI3K/mTOR inhibitors and mutation status has been missed because the mutation analysis does not capture the current state of the tumor. In addition, most clinical trials have assessed only a small number of mutations in the PI3K pathway. Analysis of mutations has mostly been limited to PIK3CA or PTEN genes. Given the highly complex nature of the PI3K pathway, we are likely not capturing the entire spectrum of mutations in the this pathway.
Toxicity and Management
The most frequent adverse events seen with mTOR inhibitors include stomatitis, rash, asthenia, hyperlipidemia, thrombocytopenia, fatigue, anorexia, hyperglycemia, elevated transaminases, pruritus, and anemia. We will discuss screening and management of common metabolic side effects, including hyperlipidemia and hyperglycemia, as well as noninfectious pneumonitis.
The incidence rates of metabolic effects with either mTORC1 or dual mTORC1/mTORC2 inhibitors have a wide range; hyperglycemia is seen in 22% to 50%, hypertriglyceridemia in 27% to 71%, and hypercholesterolemia in 24% to 76%. High-grade adverse events are rare for all metabolic effects. The PI3K-Akt-mTOR (PAM) Task Force of the National Cancer Institute recently convened to establish guidelines for managing hyperlipidemia and hyperglycemia in cancer patients. They also discussed insights into the mechanisms associated with this phenomenon, which is predominantly related to promotion of insulin resistance by PAM pathway inhibition. Overall, these metabolic effects may be more prominent in patients with insulin resistance or who are at higher risk because of family history. Typically, with mTOR inhibitors, elevations in total cholesterol, lactate dehydrogenase (LDL), and triglyceride levels are primarily seen. The Task Force recommends obtaining a complete fasting lipid panel at baseline and then at least at every cycle for phase II and later-phase studies; for phase I studies, more intense monitoring, once per week for the first two cycles, is recommended. Thresholds for intervention with drug therapy vary depending on a patient’s estimated life expectancy. In general, the goals should be to keep fasting triglycerides < 300 mg/dL and LDL < 190 mg/dL, for patients without any cardiac risk factors. If the patient’s life expectancy is estimated to be less than 1 year, then drug therapy to lower triglycerides is needed for levels > 500 mg/dL, primarily to prevent complications of hypertriglyceridemia, such as pancreatitis. For hyperglycemia screening for patients on PAM pathway inhibitors, a random glucose test is recommended at every visit for non-diabetics, and at least once per day of the first cycle for high-risk patients. If hyperglycemia is sustained or of high grade even in asymptomatic patients, treatment with metformin(Drug information on metformin) as first-line therapy is recommended.
Incorporation of these guidelines is important for all ongoing clinical studies but may have an even higher impact in clinical trials in which therapies are being used to treat early-stage patients, with curative intent. The long-term consequences of uncontrolled hyperglycemia and hyperlipidemia may be more severe in these patients, and therefore they are likely to benefit the most from intense screening and treatment.
A known class effect of mTOR inhibitors is noninfectious pneumonitis, and it is of unclear etiology. Clinical data on pneumonitis were recently summarized from an advanced renal cell cancer study of 416 patients, in which 274 patients were randomized to receive everolimus. Clinical presentation of mTOR inhibitor–associated pneumonitis was typically with a cough or dyspnea or both. The pattern on radiologic imaging varied from ground-glass infiltrates to more diffuse infiltrates. Pneumonitis occurred in patients from 3.4 to 36.7 weeks (median of 15.4 weeks) following the start of therapy. For monitoring for pneumonitis, routine chest X-rays and CT scans were obtained every 8 weeks, and pulmonary function tests (PFTs) were performed at baseline. Among 13.5% of patients suspected to have clinical pneumonitis, 3.3% of cases were grade 1 (mostly asymptomatic), 6.6% were grade 2 (not interfering with daily living), and 3.6% were grade 3 (interfering with daily living or oxygen indicated), but there was no grade 4 (life-threatening) pneumonitis. Pneumonitis appeared to be increased in patients with baseline radiographic abnormalities. Of the patients who had grade 3 toxicity, 2 out of 10 patients eventually died of pulmonary complications related to infection and progression of disease. However, others had a good outcome when managed using guidelines as outlined in the study. Based on these data, the authors made recommendations for the treatment of pneumonitis during everolimus therapy; these included the use of steroids, dose adjustment, or treatment discontinuation. Everolimus should be held for grade 2 and 3 toxicity, and may be re-initiated at a lower dose after steroid treatment and resolution of clinical symptoms. If grade 4 toxicity occurs, everolimus should be discontinued permanently.
Among five breast cancer studies with everolimus, the adverse event of noninfectious pneumonitis was analyzed. In the majority of the subjects, pneumonitis was diagnosed and monitored by CT imaging and PFTs. The highest pneumonitis grade was 3, and its incidence ranged between 2.2% and 6%; in all cases, it resolved with appropriate management. A recent meta-analysis of 2233 patients assessed pulmonary toxicity with everolimus and temsirolimus in published trials of patients that included 989 breast cancer patients. The incidence of all pulmonary toxicity was 10.4% among patients taking mTOR inhibitors, and 2.4% had high-grade toxicity.
Clinical Practice Guidelines
National Comprehensive Cancer Network (NCCN) guidelines include the option of incorporating everolimus into endocrine therapy for patients with advanced HR-positive breast cancer who have progressed on at least one prior nonsteroidal AI in the metastatic setting. Currently, we have data on improvement in PFS but not overall survival with this regimen.
Thus, while the addition of everolimus to exemestane(Drug information on exemestane) appears to prolong PFS, the combination is substantially more toxic than exemestane alone. Without evidence of clear improvement in overall survival, the clinician and patient must balance the potential for improvement in anti-tumor activity with the increased risk of toxicity and corresponding reduction in overall quality of life. The decision-making is further complicated by the relatively high monthly cost of everolimus, which is estimated to be $7,000 to 8,000 per month for the drug only, not including the additional expense of everolimus toxicity monitoring or treatment.
In the investigation of breast cancer, the development of mTOR pathway inhibitors has evolved from single-agent rapalogs to dual mTORC1/mTORC2 inhibitors and dual PI3K/mTOR inhibitors. In advanced HR-positive breast cancer, there is now a potential role for the combination of everolimus with a steroidal AI in patients previously exposed to a nonsteroidal AI. It is reasonable to consider the use of this treatment strategy for patients with endocrine therapy–resistant breast cancer, as supported by the NCCN guidelines. The improvement in PFS however, is associated with increases in toxicity and expense, without a known impact on overall survival. There are no data currently supporting the combination of mTOR inhibitors with endocrine therapy in first-line treatment of metastatic breast cancer or in the adjuvant setting. For our patients with metastatic breast cancer, quality of life is an important consideration when choosing therapies. The adverse effects of PI3K/mTOR pathway inhibitors do need to be taken into account when considering these agents, especially in patients with heavily pretreated metastatic disease.
In patients with HER2-positive disease, the early clinical data on combinations of PI3K/mTOR inhibitors with anti-HER2 therapies are encouraging; however, the results from larger studies are not available yet. Unfortunately, no robust biomarkers to predict response or resistance to PI3K/mTOR pathway inhibitors have yet been identified to facilitate patient selection. Various combinations of dual PI3K/mTOR inhibitors and other pathway inhibitors, such as MEK or IGF1R, are being studied in clinical trials to either overcome loss of feedback inhibition or PI3K activation, as observed with rapalogs, or to block escape pathways, which may become prominent when the PI3K pathway is inhibited. The efficacy and safety of these combinations still need to be determined.
Financial Disclosure: Dr. Carlson has received grant support for research from AstraZeneca and Sanofi-Aventis. Dr. Vinayak has no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.