PI3K pathway activation results in low efficacy of both trastuzumab and lapatinib

A global lapatinib Expanded Access Program was started to offer preapproval drug in order to provide clinical benefit to patients with HER2-positive metastatic breast cancer who had progressive diseases on treatment with regimens including anthracyclines, taxanes, and trastuzumab. Trastuzumab had to be used in metastatic setting. Tumors with either 3+ immunohistochemical staining for HER2 protein or HER2 gene amplification by fluorescence in situ hybridization were defined as HER2 positive in our institution. Women previously treated with capecitabine were eligible. Patients were required to have evaluable disease according to the Response Evaluation Criteria in Solid Tumors (RECIST); an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1; a left ventricular ejection fraction (LVEF) within the institution's normal range; a life expectancy of at least 12 weeks; and adequate renal, hepatic, and hematologic function. patients with central nervous system (CNS) metastases were eligible if they were clinically stable for at least 3 months after discontinuation of radiation therapy. patients with preexisting heart disease or conditions that could affect gastrointestinal absorption were ineligible. All patients gave written informed consent on recruitment into the global lapatinib Expanded Access Program and provision of the primary tumor sample for this translational study(http://​ClinicalTrials.​gov number, NCT00338247).

In this one-arm study, all patients receive the combination regimen consisting of lapatinib at a dose of 1250 mg daily on a continuous basis and capecitabine at a dose of 2000 mg per square meter of body-surface area in two divided doses on days 1 through 14 of a 21-day cycle. Standard recommendations for capecitabine dosage modifications were followed in the management of adverse events. Lapatinib was withheld for up to 14 days for grade 2 or more nonhematologic toxicity or any grade 3 or 4 hematologic toxicity. Patients were assessed every 6 weeks for the first 24 weeks, and then every 12 weeks while they were still receiving the study treatment. Patients who had no progressive disease but whose study treatment was withdrawn were assessed every 12 weeks until the commencement of alternative anticancer treatment, disease progression, or death. Efficacy was determined according to the RECIST criteria. Adverse events were assessed according to the National Cancer Institute's Common Terminology Criteria for Adverse Events (CTCAE, version 3.0). The clinical benefit was defined as a complete response, partial response, or stable disease for at least 6 months. Progression-free survival was calculated as the interval between the date of signing informed consent and the date of disease progression, or death from any cause.

DNA was extracted from formaldehyde-fixed, paraffin-embedded tumor tissue. PCR were performed with 10 to 100 ng of genomic DNA as template following a standard protocol. PIK3CA gene PCR primers were E9F CAAAGCAATTTCTACACGAGATCC; E9R GTAAAAACATGCTGAGATCAGCCACAT; E20F TGGAATGCCAGAACTACAATCTTT; E20R GGTCTTTGCCTGCTGAGAGTT. The PCR products were sequenced using the ABI3130XL automated capillary sequencer by Shanghai Tianhao Biotechnology Company.

The antibody for PTEN IHC staining was a rabbit monoclonal anti-PTEN (D4.3) XPTM (Cell Signaling Beverly, MA, USA) diluted 1:200. The antibody was applied overnight at 4ºC. Then the tissues were incubated with the second antibody (1:500, Jackson lab) for 30 min. The color was developed with DAB solution about 1 min at room temperature and then stained in Harris hematoxylin solution for 3 min. Each set of slides included positive and negative control slides and normal cells in a tumor were used as an internal control.

PTEN immunoreactivity was examined by two independent observers who were blinded to the clinical data. A third pathologist was invited when the discordance was present between them. The staining was mainly visible in the cytoplasm of tumor cells. PTEN expression levels were semiquantified using immunoreactive scores (IRS) calculated by multiplying the percentage of PTEN-positive tumor cells (scored 0 to 4) with the PTEN staining intensity [1‐3]. The tumor was graded as PTEN-negative (IRS 0-3), weak positive (IRS 4-6), positive (IRS 7-9), and strong positive (IRS 10-12)[6].

The relationships between different variables were assessed by Chi-square tests and the trends were also examined by Chi-square tests when required. Differences in progression-free survival (PFS) and overall survival (OS) between groups were determined using the log-rank test. After a univariate analysis, the variables with significant correlation with PFS and OS, continuous variables and PI3K pathway status were put in a Cox proportional hazard regression model to determine which was an independent prognostic factor for PFS and OS, respectively. The statistical difference was considered significant if the P value was less than 0.05.

Sixty-seven Chinese patients were enrolled from Fudan University Cancer Center from Aug. 2008 to Mar. 2010. The median age was 49.0 years old (range 26-75). Fifty-seven patients had their tumor tissues available for detection of PI3K pathway activation (PIK3CA mutation and PTEN expression loss). Clinical and pathological data for the patients are showed in Table 1.

The overall incidence of PIK3CA mutations was 12.3% (7 in 57 samples). The majority of mutations occurred at two hotspots, H1047R (7%, 4 samples) at exon 20 encoding the kinase domain (Figure 1a), and E542K (1.8%, 1 sample) at exon 9 encoding the helical domain (Figure 1b). L540F and T1052A mutations are rare and each was found in one tumor sample (Figure 1c-d).

PTEN expression loss was found in 18 patients (31.6%, Figure 2a). Thirty-nine patients were positive for PTEN expression [6], in which 17 (29.8%), 14 (24.6%), and 8 (14%) specimens were weak positive, positive and strong positive respectively (Figure 2b-d). In this study, PTEN loss was not mutually exclusive with PIK3CA mutations, since 3 of the 4 patients with H1047R mutation were also found to have no PTEN expression (Table 2).

Compared with the wild type, PI3K pathway activation (PIK3CA mutation and/or PTEN expression loss) was identified in a significantly older patient population (P = 0.016). The median age of patients with the PI3K pathway activation was 53.6 ± 7.9 years, while the median age of those with no PI3K pathway activation was 47.0 ± 10.9 years. The PI3K pathway activation status was not associated with all other clinicopathological parameters, such as hormone receptor status, HER2 protein expression status and disease free interval after radical mastectomy (all P > 0.05).

On September 30, 2010, preliminary analysis was made on the basis of 50 disease progression events and 28 deaths. Median follow-up time was 15.3 ± 6.3 months. The median PFS of all 67 patients was 6.5 months (95%CI, 5.7-7.3 months), and the median PFS of the 57 patients who provided their tumor tissues for detection of PI3K pathway activation was also 6.5 months (95%CI, 5.4-7.6 months). The overall response rate (complete response plus partial response) was 22.4% for all 67 patients and 22.8% for the 57 patients whose tumor sample were available. The corresponding clinical benefit rates were 58.2% and 56.1%. The median overall survivals for both cohorts were 17.0 months.

An analysis of our data showed that PIK3CA mutation didn't correlate with overall response rate, clinical benefit rate or progression-free survival. For PTEN expression status, patients with wild type gene enjoyed a clinical benefit rate of 66.7%, which was statistically higher than 33.3% in those with no PTEN expression (P = 0.018). The overall response rate of 28.2% and median PFS of 8 months in the patients with PTEN expression were substantially higher, although the differences were not statistically significant (Table 3). When analyzing PIK3CA mutation together with PTEN expression loss since both can activate PI3K pathway, the clinical benefit was still observed for patients with no activation of PI3K pathway (68.6% versus 36.4%, P = 0.017; Figure 3). The overall response rate was also higher (31.4% versus 9.1%, P = 0.05; Figure 3). Both overall response and clinical benefit significantly correlated with PFS (Figure 4a-b), however, there was no significant association of PI3K pathway activation status with PFS or OS (Figure 4c).

A retrospective analysis was done to explore the relationship between PI3K pathway activation status and the efficacy of the other anti-HER2 drug, trastuzumab. We chose the progression-free survival of the first trastuzumab-containing regimen as an indicator for trastuzumab efficacy. The regimens were trastuzumab combined with one or two chemotherapy drugs, including docetaxel, vinorelbine, paclitaxel, gemcitabine, capecitabine and cisplatin. As previously reported, PI3K pathway activation shortened the median progression-free survival significantly (4.5 versus 9.0 months, P = 0.013; Figure 4d). PTEN expression status had a similar differentiating effect (6.0 versus 9.0 months, P = 0.024). However, the difference of PFS (4.5 versus 8.0 months) resulting from PIK3CA mutation wasn't significant.

Response and survival of breast cancer might be affected by many other factors, such as age, ECOG performance status, hormone receptor status, HER2 expression, metastatic sites, number of metastatic sites and previous treatments. A univariate analysis of our data showed that only number of metastatic sites had a negative impact on overall response rate of the lapatinib and capecitbine doublet, while none of the above-mentioned factors was statistically associated with clinical benefit rate. Concerning survival, only previous capecitabine treatment was statistically associated with a shorter PFS (6.5 vs. 11.0 months, P = 0.011, Table 4) and OS (14.0 months vs. not reached, P = 0.003).

We put all the above-mentioned variants of PI3K pathway status, age and clinical benefit status into Cox's proportional hazards model to investigate how much PI3K pathway activation affected patient outcome. Clinical benefit rate was the only independent factor for PFS in HER2-positive breast cancer.