Comparative genomic analysis of PIK3R1-mutated and wild-type breast cancers

PURPOSE

The PIK3R1 gene encodes the regulatory subunit-p85a-of the PI3K signaling complex. Prior studies have found that pathogenic somatic alterations in PIK3R1 are enriched in human breast cancers but the genomic landscape of breast cancer patients harboring PIK3R1 mutations has not been extensively characterized.

METHODS

We retrospectively analyzed 6,009 patient records that underwent next-generation sequencing (NGS) using the Tempus xT solid tumor assay. All patients had breast cancer with known HER2 (+/-) and hormone receptor (HR; +/-) status and were classified according to the presence of PIK3R1 mutations including short variants and copy number alterations.

RESULTS

The frequency of PIK3R1 mutations varied according to subtype: 6% in triple negative (TNBC, 89/1,475), 2% in HER2-/HR+ (80/3,893) and 2.3% in HER2+ (15/641) (p < 0.001). Co-mutations in PTEN, TP53 and NF1 were significantly enriched, co-mutations in PIK3CA were significantly less prevalent, and tumor mutational burden was significantly higher in PIK3R1-mutated HER2- samples relative to PIK3R1 wild-type. At the transcriptional-level, PIK3R1 RNA expression in HER2- disease was significantly higher in PIK3R1-mutated (excluding copy number loss) samples, regardless of subtype.

CONCLUSION

This is the largest investigation of the PIK3R1 mutational landscape in breast cancer patients (n = 6,009). PIK3R1 mutations were more common in triple-negative breast cancer (~ 6%) than in HER2 + or HER2-/HR + disease (approximately 2%). While alterations in the PI3K/AKT pathway are often actionable in HER2-/HR + breast cancer, our study suggests that PIK3R1 could be an important target in TNBC as well.

Binimetinib activity in PIK3R1-mutant patient-derived xenografts (PDX) implanted into immunodeficient mice

e13062

Background: The PIK3R1 gene is genetically altered in ̃3% of breast cancers in the Western and ̃17% in the Eastern worlds. We recently reported that breast cancer cells lacking protein expression of PIK3R1 had elevated levels of activated MEK, sensitizing them to the MEK inhibitor trametinib. To better understand whether MEK inhibition is a therapeutic option for breast cancer patients with mutated PIK3R1, we tested the newer generation MEK inhibitor, binimetinib, in a PDX mouse model. Methods: Two PDx models (table) were tested in athymic nude-foxn1nu (immune-compromised) mice. Both patients had had metastatic, grade 3 breast cancer and expressed moderate RNA levels of PIK3R1. Results: Body weights of treatment arms were unchanged throughout the experiment. The TNBC PDx study was halted early at 13 days because tumors reached the predefined maximum size of 1500 mm2. The binimetinib treated mice experienced slowed tumor growth (522.8 +/- 172.8%) compared to the vehicle control (720.7 +/- 229.1%), (p=0.0693). In the HR+ PDx model, the combination of binimetinib plus tamoxifen resulted in a statistically significant decrease (371.5 +/- 108.1%) in tumor volume compared to tamoxifen alone (525.0 +/- 154.1%; p = 0.0302). Conclusions: Our previous and current results suggest that mutation in PIK3R1 sensitizes cells to MEK inhibitors. Binimetinib inhibited growth in both models. This inhibition was of borderline significance in the TNBC model and was statistically significant in the ER+, HER2- model. To the best of our knowledge, these results provide the first evidence of binimetinib efficacy in PIK3R1-mutant, HR+, HER2- PDX immunodeficient mice. Plans for a phase II trial in patients with PIK3R1-mutant, HR+, HER2- breast cancer are underway.[Table: see text]

Abstract PS18-30: Loss of NF1 leads to rho GTPase activation and sensitivity to multiple agents in breast cancer

Background: The Ras signaling pathway is a key oncogenic growth signaling pathway. The neurofibromatosis gene (NF1) is a tumor suppressor gene and negative regulator of Ras. Approximately 20% of breast cancers experience loss of active NF1 protein, leading to unabated Ras activity. Germline inactivation of NF1 leads to neurofibromatosis, which can be treated with MEK inhibitors. The functional consequences of NF1 loss have not been thoroughly explored in breast cancer. Therefore, we created a cellular model for the loss of NF1 in breast cells in order to identify novel therapies that would target NF1 null breast cells.

Methods: We used CRISPR CAS9 to knock out (KO) NF1 through targeted disruption of both NF1 alleles in the human non-tumorigenic breast cell line MCF-10A. Immunoblotting confirmed loss of the protein and this cell line was used to identify changes in cellular signaling that resulted from the loss of NF1. We then used a rational approach to select drugs that may target the cellular changes in NF1 null cells. Finally, the candidate drugs were tested in a tumor xenograft model in mice.

Results: Loss of NF1 endowed cells with a more transformed phenotype, including EGF growth independence and anchorage-independent growth. We observed increased activation of MAPK and activated Rho GTPase. The MEK inhibitors trametinib and PD0325901 (IC50=0.25 nM and 0.16 nM, respectively) inhibited growth of NF1-null, but not parent cells. We explored inhibitors of other proteins in the MAPK signaling pathway, including the Raf inhibitors sorafenib and vemurafenib and the dual MAP3K1/MAP2K4 inhibitor LY2228820. However, these compounds did not selectively inhibit growth of NF1-null cells, suggesting that Raf and MAP3K1/MAP2K4 are not vulnerable targets in NF1-null cells.

Vincristine and zoledronic acid can indirectly affect Rho GTPase function. NF1-null cells were more sensitive to both as single agents (IC50=1.1 nM and 6.9 μM, respectively) than parental cells. We also tested multiple combinations of these drugs and observed benefits from several combinations that exceeded single agent use. Similar sensitivity was not observed to docetaxel or ixabepilone.

We then tested the drugs as single agents and in combination against mouse tumor xenograft models with the human breast cancer cell lines Hs578T (NF1 null) or MCF-7 (wild type NF1). Treatment with single agent vincristine or zoledronic acid for 19 days resulted in statistically-significant decreases of 37.3% and 38.9%, respectively (p&lt;0.05) in Hs578T tumor xenografts. The combination treatment resulted in an enhanced reduction in tumor size of 48.8% with a higher level of significance (p&lt;0.01). Treatment of mice with MCF-7 bearing tumors with single agent vincristine resulted in a statistically significant decrease of 35.2% (p&lt;0.05); however, this effect was not observed using the combination treatment, as the change in growth compared to saline treatment was not statistically significant.

Conclusions: Our results suggest that loss of NF1 results in a transformed phenotype. Our results provide the first evidence that vincristine in combination with zoledronic acid may be efficacious in treating breast cancers. Additionally, we provide evidence that other drugs and drug combinations may be more effective at treating breast cancers with NF1 loss than MEK inhibitors, warranting further exploration.

Citation Format: Melody A. Cobleigh, Matthew Najor, Satnam Brar, Sanja Turturro, Liam Portt, Timothy Yung, Abde M Abukhdeir. Loss of NF1 leads to rho GTPase activation and sensitivity to multiple agents in breast cancer [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS18-30.

FDA Approval Summary: Pertuzumab, Trastuzumab, and Hyaluronidase-zzxf Injection for Subcutaneous Use in Patients with HER2-positive Breast Cancer

On June 29, 2020, the FDA approved pertuzumab, trastuzumab, and hyaluronidase-zzxf subcutaneous injection (Phesgo) for the treatment of patients with HER2-positive early-stage and metastatic breast cancer. Patients should be selected for therapy based on an FDA-approved companion diagnostic test. Approval was primarily based on the FeDeriCa trial, a randomized, open-label, multicenter comparability study of pertuzumab, trastuzumab, and hyaluronidase-zzxf subcutaneous injection compared with intravenous pertuzumab and intravenous trastuzumab administered in the neoadjuvant and adjuvant settings with chemotherapy for the treatment of patients with early breast cancer. The pharmacokinetic endpoints were, first, to demonstrate that the exposure of subcutaneous pertuzumab was not inferior to that of intravenous pertuzumab, and then to demonstrate that the exposure of subcutaneous trastuzumab was not inferior to that of intravenous trastuzumab. The primary endpoints were met with the observed lower limit of the two-sided 90% confidence intervals above the prespecified noninferiority margins. The most common adverse reactions were alopecia, nausea, diarrhea, anemia, and asthenia. The totality of the evidence demonstrated comparability of the subcutaneous product to intravenous, allowing for extrapolation and approval of all breast cancer indications for which intravenous trastuzumab and pertuzumab are approved.

Loss of STAT6 leads to anchorage-independent growth and trastuzumab resistance in HER2+ breast cancer cells

Approximately 20% of breast cancers are HER2-positive. Trastuzumab has improved patient outcomes significantly for these cancers. However, acquired resistance remains a major hurdle in the clinical management of these patients. Therefore, identifying molecular changes that cause trastuzumab resistance is worthwhile. STAT6 is a transcription factor that regulates a variety of genes involved in cell cycle regulation, growth inhibition, and apoptosis. STAT6 expression is lost in approximately 3% of breast cancers, but little work has been done in the context of trastuzumab resistance in breast cancer. In isogenic cell line pairs, we observed that trastuzumab-resistant cells expressed significantly lower levels of STAT6 compared to trastuzumab-sensitive cells. Therefore, in order to study the consequences of STAT6 loss in HER2+ breast cancer, we knocked out both alleles of the STAT6 gene using somatic cell gene targeting. Interestingly, loss of STAT6 resulted in anchorage-independent growth and changes in several genes involved in epithelial to mesenchymal transition. This study suggests that STAT6 may play a role in the pathophysiology of HER2+ human breast cancer.

STAT6 expression and trastuzumab resistance in HER2+ breast cancer

e13006

Background: Approximately 20% of breast cancers are HER2-positive. Trastuzumab has improved patient outcomes significantly for these cancers. However, acquired resistance occurs in almost all patients with metastatic breast cancer. Therefore, identifying molecular changes that are associated with trastuzumab resistance is worthwhile. STAT6 is a transcription factor that regulates a variety of genes involved in cell cycle regulation, growth inhibition, and apoptosis. STAT6 expression is lost in approximately 3% of breast cancers, but little work has been done in the context of trastuzumab resistance in breast cancer. Methods: In isogenic cell line pairs, we observed that trastuzumab-resistant cells expressed significantly lower levels of STAT6 compared to trastuzumab-sensitive cells. Therefore, in order to study the consequences of STAT6 loss in HER2+ breast cancer, we knocked out both alleles of the STAT6 gene using somatic cell gene targeting. Results: We observed that loss of STAT6 resulted in resistance to trastuzumab treatment in HER2-over-expressing cells. Additionally, loss of STAT6 resulted in anchorage-independent growth and changed expression of several genes involved in epithelial to mesenchymal transition. Functional studies revealed that STAT6 loss caused a non-tumorigenic human breast cell line to form tumors in mice. Conclusions: Loss of STAT6 in breast cells results in enhanced growth properties and resistance to trastuzumab. This study suggests that STAT6 may play a role in the pathophysiology of HER2+ human breast cancer. Disclaimer: The work herein was completed while AMA was a faculty member at Rush University. AMA is currently an employee with the U.S. Food and Drug Administration. The views and data in this publication do not reflect the opinions of the U.S. Government or The U.S. Food and Drug Administration.

Expression and function of voltage gated proton channels (Hv1) in MDA-MB-231 cells

Expression of the voltage gated proton channel (H_v1) as identified by immunocytochemistry has been reported previously in breast cancer tissue. Increased expression of H_V1 was correlated with poor prognosis and decreased overall and disease-free survival but the mechanism of its involvement in the disease is unknown. Here we present electrophysiological recordings of H_V1 channel activity, confirming its presence and function in the plasma membrane of a breast cancer cell line, MDA-MB-231. With western blotting we identify significant levels of H_V1 expression in 3 out of 8 “triple negative” breast cancer cell lines (estrogen, progesterone, and HER2 receptor expression negative). We examine the function of H_V1 in breast cancer using MDA-MB-231 cells as a model by suppressing the expression of H_V1 using shRNA (knock-down; KD) and by eliminating H_V1 using CRISPR/Cas9 gene editing (knock-out; KO). Surprisingly, these two approaches produced incongruous effects. Knock-down of H_V1 using shRNA resulted in slower cell migration in a scratch assay and a significant reduction in H₂O₂ release. In contrast, H_V1 Knock-out cells did not show reduced migration or H₂O₂ release. H_V1 KO but not KD cells showed an increased glycolytic rate accompanied by an increase in p-AKT (phospho-AKT, Ser473) activity. The expression of CD171/LCAM-1, an adhesion molecule and prognostic indicator for breast cancer, was reduced in H_V1 KO cells. When we compared MDA-MB-231 xenograft growth rates in immunocompromised mice, tumors from H_V1 KO cells grew less than WT in mass, with lower staining for the Ki-67 marker for cell proliferation rate. Therefore, deletion of H_V1 expression in MDA-MB-231 cells limits tumor growth rate. The limited growth thus appears to be independent of oxidant production by NADPH oxidase molecules and to be mediated by cell adhesion molecules. Although H_V1 KO and KD affect certain cellular mechanisms differently, both implicate H_V1-mediated pathways for control of tumor growth in the MDA-MB-231 cell line.

Somatic loss of PIK3R1 may sensitize breast cancer to inhibitors of the MAPK pathway

PURPOSE

The PI3K pathway, which includes the PI3K catalytic subunits p110α (PIK3CA) and the PI3K regulatory subunit p85α (PIK3R1), is the most frequently altered pathway in cancer. We encountered a breast cancer patient whose tumor contained a somatic alteration in PIK3R1. Some commercial sequencing platforms suggest that somatic mutations in PIK3R1 may sensitize cancers to drugs that inhibit the mammalian target of rapamycin (mTOR). However, a review of the preclinical and clinical literature did not find evidence substantiating that hypothesis. The purpose of this study was to knock out PIK3R1 in order to determine the optimal therapeutic approach for breast cancers lacking p85α.

METHODS

We created an isogenic cellular system by knocking out both alleles of the PIK3R1 gene in the non-tumorigenic human breast cell line MCF-10A. Knockout cells were compared with wild-type cells by measuring growth, cellular signaling, and response to drugs.

RESULTS

We observed hyperphosphorylation of MEK in these knockouts, which sensitized PIK3R1-null cells to a MEK inhibitor, trametinib. However, they were not sensitized to the mTOR inhibitor, everolimus.

CONCLUSIONS

Our findings suggest that breast cancers with loss of p85α may not respond to mTOR inhibition, but may be sensitive to MEK inhibition.

KRAS mutation and epithelial-macrophage interplay in pancreatic neoplastic transformation

Pancreatic ductal adenocarcinoma (PDA) is characterized by epithelial mutations in KRAS and prominent tumor-associated inflammation, including macrophage infiltration. But knowledge of early interactions between neoplastic epithelium and macrophages in PDA carcinogenesis is limited. Using a pancreatic organoid model, we found that the expression of mutant KRAS in organoids increased (i) ductal to acinar gene expression ratios, (ii) epithelial cells proliferation and (iii) colony formation capacity in vitro, and endowed pancreatic cells with the ability to generate neoplastic tumors in vivo. KRAS mutations induced a protumorigenic phenotype in macrophages. Altered macrophages decreased epithelial pigment epithelial derived factor (PEDF) expression and induced a cancerous phenotype. We validated our findings using annotated patient samples from The Cancer Genome Atlas (TCGA) and in our human PDA specimens. Epithelium-macrophage cross-talk occurs early in pancreatic carcinogenesis where KRAS directly induces cancer-related phenotypes in epithelium, and also promotes a protumorigenic phenotype in macrophages, in turn augmenting neoplastic growth.

Abstract P2-09-30: A gene expression signature that predicts for trastuzumab response in HER2+ breast cancer

Background: Approximately 15-20% of breast cancers overexpress HER2. These patients are eligible for trastuzumab in combination with chemotherapy. However, some patients are extreme responders to single agent trastuzumab and we wanted to identify differences in cancer gene expression that could predict response to single agent trastuzumab.

Methods: We performed paired-end RNAseq on an isogenic cellular model of trastuzumab sensitivity and resistance. We reasoned that the isogenic nature of the cellular clones used in this study would enrich for differentially-expressed genes (DEGs) that were associated with response to single-agent trastuzumab. DEGs where chosen based upon either i) large fold changes in resistant vs. sensitive clones, ii) high frequency in human HER2+ breast cancers, or iii) were found to be enriched with other DEGs in signaling pathways selected by Ingenuity Pathway Analysis (IPA). DEGs were further scrutinized based upon associations with overall survival (OS) in HER2+ human breast cancers. The resulting genes were validated using qPCR and in several independent sample sets containing gene expression profiles of human breast cancers.

Results: Discovery: RNAseq yielded 3,241 statistically-significant DEGs. We used two independent filtering pipelines to obtain 175 DEGs. Ingenuity Pathway Analysis found signaling pathways associated with eukaryotic initiation factor, lysine specific demethylase 5B, and estrogen receptor alpha to be enriched in DEGs associated with trastuzumab resistance. Of these DEGs, six genes correlated with a statistically significant change in OS in the training dataset, and were validated by qPCR in the cell lines used for the analysis. We further determined that the six-gene signature was a negative predictor of overall survival in HER2+ breast cancer patients whose cancers carried at least one DEG. Validation: Using independent cohorts from TCGA and the website KMplot.com, we validated the predictive power of the six-gene signature. Of the 47 HER2+ patients from TCGA, eight patients carried two more DEGs, while 39 carried ≤ 1 DEG. Although the numbers are small, of the 8 patients followed for four or more years, only one patient was alive as compared with 7 out of 39 patients without the signature. Similarly, Kaplan Meier analysis of gene expression data from KMplot.com revealed that only 1 out of 23 patients (4.3%) who carried high mean expression of the six-gene signature were free of distant metastases after 87 months, compared to 4 out of 43 patients (9.3%) from the cohort carrying low mean expression of the six-gene signature. In both validation cohorts, the six DEG signature was not predictive in HER2-negative breast cancers.

Discussion: Patients whose tumors lack this gene expression signature are more likely to experience a favorable response to trastuzumab therapy. This signature requires validation in a clinical cohort treated with trastuzumab monotherapy.

Citation Format: Abukhdeir AM, Najor MS, Turturro SB, Armstrong AR, McDonald A, Fogg L, Cobleigh MA. A gene expression signature that predicts for trastuzumab response in HER2+ breast cancer [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P2-09-30.

Overexpression of lipid metabolism genes and PBX1 in the contralateral breasts of women with estrogen receptor-negative breast cancer

Risk biomarkers for estrogen receptor (ER)-negative breast cancer have clear value for breast cancer prevention. We previously reported a set of lipid metabolism (LiMe) genes with high expression in the contralateral unaffected breasts (CUBs) of ER-negative cancer cases. We now further examine LiMe gene expression in both tumor and CUB, and investigate the role of Pre-B-cell leukemia homeobox-1 (PBX1) as a candidate common transcription factor for LiMe gene expression. mRNA was extracted from laser-capture microdissected epithelium from tumor and CUB of 84 subjects (28 ER-positive cases, 28 ER-negative cases, 28 healthy controls). Gene expression was quantitated by qRT-PCR. Logistic regression models were generated to predict ER status of the contralateral cancer. Protein expression of HMGCS2 and PBX1 was measured using immunohistochemistry. The effect of PBX1 on LiMe gene expression was examined by overexpressing PBX1 in MCF10A cells with or without ER, and by suppressing PBX1 in MDA-MB-453 cells. The expression of DHRS2, HMGCS2, UGT2B7, UGT2B11, ALOX15B, HPGD, UGT2B28 and GLYATL1 was significantly higher in ER-negative versus ER-positive CUBs, and predicted ER status of the tumor in test and validation sets. In contrast, LiMe gene expression was significantly lower in ER-negative than ER-positive tumors. PBX1 overexpression in MCF10A cells up-regulated most LiMe genes, but not in MCF10A cells overexpressing ER. Suppressing PBX1 in MDA-MB-453 cells resulted in decrease of LiMe gene expression. Four binding sites of PBX1 and cofactor were identified in three lipid metabolism genes using ChIP-qPCR. These data suggest a novel role for PBX1 in the regulation of lipid metabolism genes in benign breast, which may contribute to ER-negative tumorigenesis.

Abstract P6-11-17: Mutations in PIK3R1 activate multiple pathways in breast cancer

It has been estimated that by the end of this year, 270,000 American women will be newly diagnosed with breast cancer, while 40,000 women that already have breast cancer will succumb to the disease. Considerable attention has been given to the PI3K signaling cascade following the discovery that PIK3CA is the most frequently mutated oncogene in breast cancer.However, few studies have explored the function of PIK3R1, which is the regulatory domain of the PI3K complex, despite being mutated in ~3% of all breast cancers.

Several studies have demonstrated that expression of PIK3R1 is downregulated in human cancers. Decreased expression of the PIK3R1 protein leads to tumor formation, suggesting its role as a tumor suppressor gene and a potential prognostic marker in breast cancer. However, PIK3R1 is a gene with little pre-clinical evidence to recommend experimental therapies. Despite this lack of evidence, commercial services that perform molecular analyses of tumors suggest the use of an mTOR inhibitor for patients whose breast cancers carry mutant PIK3R1.

In order to determine if mTOR inhibitors were indeed effective in mutant PIK3R1 tumors, we created and characterized a model for mutant PIK3R1 in the non-tumorigenic, human breast epithelial cell line, MCF-10A. Surprisingly, we observed that mTOR inhibitors were ineffective in these cells. However, in searching for other classes of small molecule inhibitors that were effective, we observed that mutations in PIK3R1 sensitized cells to MAPK inhibitors.

Herein, we present the first evidence for the use of targeted therapies in breast cancers carrying mutant PIK3R1. We provide evidence against the use of mTOR inhibitors and provide a rationale for the use of MAPK inhibitors.

Citation Format: Abukhdeir AM, Turturro SB, Najor MS, Brar SS, Cobleigh MA. Mutations in PIK3R1 activate multiple pathways in breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-11-17.

Mutations in PIK3CA sensitize breast cancer cells to physiologic levels of aspirin

A review of the literature finds that women diagnosed with breast cancer, who were on an aspirin regimen, experienced a decreased risk of distant metastases and death. Several recent studies have reported an improvement in overall survival in colorectal cancer patients who harbored mutations in the oncogene PIK3CA and received a daily aspirin regimen. Breast cancer patients on a daily aspirin regimen experienced decreased risk of distant metastases and death. PIK3CA is the most frequently mutated oncogene in breast cancer, occurring in up to 45 % of all breast cancers. In order to determine if mutations in PIK3CA sensitized breast cancers to aspirin treatment, we employed the use of isogenic cellular clones of the non-tumorigenic, breast epithelial cell line MCF-10A that harbored mutations in either PIK3CA or KRAS or both. We report that mutations in both PIK3CA and KRAS are required for the greatest aspirin sensitivity in breast cancer, and that the GSK3β protein was hyperphosphorylated in aspirin-treated double knockin cells, but not in other clones/treatments. A more modest effect was observed with single mutant PIK3CA, but not KRAS alone. These observations were further confirmed in a panel of breast cancer cell lines. Our findings provide the first evidence that mutations in PIK3CA sensitize breast cancer cells to aspirin.

Abstract P5-05-07: Elucidating molecular resistance to trastuzumab using next generation sequencing in isogenic cell models

A minority of all breast cancers will express increased levels of the ERBB2 protein. They are eligible for trastuzumab-based therapy. Some will respond, but all will progress. Thus, the problem of resistance to trastuzumab has generated an urgent need to determine the underlying mechanisms of that resistance.

Cancer is a genetic disease and the mechanism of trastuzumab resistance is likely also genetic in nature. However, the significant genomic heterogeneity between and within patient tumors greatly complicates the identification of a genetic mechanism of resistance for trastuzumab. In order to overcome some of these challenges, we looked to an isogenic model of trastuzumab resistance. We acquired the trastuzumab-sensitive breast cancer cell line, BT474 and two clones of this cell line that were conditioned to exhibit trastuzumab resistance.

To investigate a possible genetic mechanism of trastuzumab-resistance, we performed whole exome sequencing using Ampliseq chemistry on the Ion Torrent platform from Life Technologies and paired-end RNA-sequencing on the Illumina HiSeq platform. Next-generation sequencing data was bioinformatically analyzed using tools that allowed us to filter relevant variants based on statistical and functional significance. Variants of interest were those that that arose during drug treatment, which were identified as those in each of the resistant clones, which were novel compared to the parent clone. Proteins from whole cell lysates were resolved in two dimensions using 3-10 nonlinear strips for isoelectric focusing followed by resolution via 4-20% SDS-PAGE. Proteins were visualized via Gelcode blue and cored with a biopsy punch, trypsinized, and submitted for protein ID on an LTQ XL mass spectrometer. We performed functional validation of genetic alterations through the use of somatic cell gene targeting of an ERBB2-expressing clone of the MCF-10A cell lines, a non-tumorigenic model of breast cancer, which is sensitive to trastuzumab.

Exome sequencing initially yielded more than 10,000 unique DNA variants across the three clones, which after bioinformatic analysis resulted in ∼1000 variants of interest. Two-dimensional gel electrophoresis revealed 25-30 differentially expressed proteins per sample. Correlation between the sequencing and proteomics data provided us with a candidate gene list of less than 100 genes and several cellular pathways related to growth signaling and immunity. Genetic alterations were tested for their ability to cause trastuzumab resistance in MCF-10A clones.

We describe herein a detailed molecular analysis for a model of trastuzumab resistance. Validated genetic alterations will be investigated in a unique collection of archival specimens, which we hope will open the path towards the development of novel agents to augment the effects of trastuzumab.

Citation Format: Abde M Abukhdeir, Matthew Najor, Sanja Turturro, Melissa R Pergande, Jeffrey A Borgia, Hanif G Khalak, Melody Cobleigh. Elucidating molecular resistance to trastuzumab using next generation sequencing in isogenic cell models [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P5-05-07.

Single copies of mutant KRAS and mutant PIK3CA cooperate in immortalized human epithelial cells to induce tumor formation

The selective pressures leading to cancers with mutations in both KRAS and PIK3CA are unclear. Here, we show that somatic cell knockin of both KRAS G12V and oncogenic PIK3CA mutations in human breast epithelial cells results in cooperative activation of the phosphoinositide 3-kinase (PI3K) and mitogen-activated protein kinase (MAPK) pathways in vitro, and leads to tumor formation in immunocompromised mice. Xenografts from double-knockin cells retain single copies of mutant KRAS and PIK3CA, suggesting that tumor formation does not require increased copy number of either oncogene, and these results were also observed in human colorectal cancer specimens. Mechanistically, the cooperativity between mutant KRAS and PIK3CA is mediated in part by Ras/p110α binding, as inactivating point mutations within the Ras-binding domain of PIK3CA significantly abates pathway signaling. In addition, Pdk1 activation of the downstream effector p90RSK is also increased by the combined presence of mutant KRAS and PIK3CA. These results provide new insights into mutant KRAS function and its role in carcinogenesis.

The growth response to androgen receptor signaling in ERα-negative human breast cells is dependent on p21 and mediated by MAPK activation

Introduction

Although a high frequency of androgen receptor (AR) expression in human breast cancers has been described, exploiting this knowledge for therapy has been challenging. This is in part because androgens can either inhibit or stimulate cell proliferation in pre-clinical models of breast cancer. In addition, many breast cancers co-express other steroid hormone receptors that can affect AR signaling, further obfuscating the effects of androgens on breast cancer cells.

Methods

To create better-defined models of AR signaling in human breast epithelial cells, we took estrogen receptor (ER)-α-negative and progesterone receptor (PR)-negative human breast epithelial cell lines, both cancerous and non-cancerous, and engineered them to express AR, thus allowing the unambiguous study of AR signaling. We cloned a full-length cDNA of human AR, and expressed this transgene in MCF-10A non-tumorigenic human breast epithelial cells and MDA-MB-231 human breast-cancer cells. We characterized the responses to AR ligand binding using various assays, and used isogenic MCF-10A p21 knock-out cell lines expressing AR to demonstrate the requirement for p21 in mediating the proliferative responses to AR signaling in human breast epithelial cells.

Results

We found that hyperactivation of the mitogen-activated protein kinase (MAPK) pathway from both AR and epidermal growth factor receptor (EGFR) signaling resulted in a growth-inhibitory response, whereas MAPK signaling from either AR or EGFR activation resulted in cellular proliferation. Additionally, p21 gene knock-out studies confirmed that AR signaling/activation of the MAPK pathway is dependent on p21.

Conclusions

These studies present a new model for the analysis of AR signaling in human breast epithelial cells lacking ERα/PR expression, providing an experimental system without the potential confounding effects of ERα/PR crosstalk. Using this system, we provide a mechanistic explanation for previous observations ascribing a dual role for AR signaling in human breast cancer cells. As previous reports have shown that approximately 40% of breast cancers can lack p21 expression, our data also identify potential new caveats for exploiting AR as a target for breast cancer therapy.

PIK3CA mutations and EGFR overexpression predict for lithium sensitivity in human breast epithelial cells

A high frequency of somatic mutations has been found in breast cancers within the gene encoding the catalytic p110α subunit of PI3K, PIK3CA. Using isogenic human breast epithelial cells, we have previously demonstrated that oncogenic PIK3CA "hotspot" mutations predict for response to the toxic effects of lithium. However, other somatic genetic alterations occur within this pathway in breast cancers, and it is possible that these changes may also predict for lithium sensitivity. We overexpressed the epidermal growth factor receptor (EGFR) into the non-tumorigenic human breast epithelial cell line MCF-10A, and compared these cells to isogenic cell lines previously created via somatic cell gene targeting to model Pten loss, PIK3CA mutations, and the invariant AKT1 mutation, E17K. EGFR overexpressing clones were capable of cellular proliferation in the absence of EGF and were sensitive to lithium similar to the results previously seen with cells harboring PIK3CA mutations. In contrast, AKT1 E17K cells and PTEN -/- cells displayed resistance or partial sensitivity to lithium, respectively. Western blot analysis demonstrated that lithium sensitivity correlated with significant decreases in both PI3K and MAPK signaling that were observed only in EGFR overexpressing and mutant PIK3CA cell lines. These studies demonstrate that EGFR overexpression and PIK3CA mutations are predictors of response to lithium, whereas Pten loss and AKT1 E17K mutations do not predict for lithium sensitivity. Our findings may have important implications for the use of these genetic lesions in breast cancer patients as predictive markers of response to emerging PI3K pathway inhibitors.

Knock in of the AKT1 E17K mutation in human breast epithelial cells does not recapitulate oncogenic PIK3CA mutations

An oncogenic mutation (G49A:E17K) in the AKT1 gene has been described recently in human breast, colon, and ovarian cancers. The low frequency of this mutation and perhaps other selective pressures have prevented the isolation of human cancer cell lines that harbor this mutation thereby limiting functional analysis. Here, we create a physiologic in vitro model to study the effects of this mutation by using somatic cell gene targeting using the nontumorigenic human breast epithelial cell line, MCF10A. Surprisingly, knock in of E17K into the AKT1 gene had minimal phenotypic consequences and importantly, did not recapitulate the biochemical and growth characteristics seen with somatic cell knock in of PIK3CA hotspot mutations. These results suggest that mutations in critical genes within the PI3-kinase (PI3K) pathway are not functionally equivalent, and that other cooperative genetic events may be necessary to achieve oncogenic PI3K pathway activation in cancers that contain the AKT1 E17K mutation.

Shared TP53 gene mutation in morphologically and phenotypically distinct concurrent primary small cell neuroendocrine carcinoma and adenocarcinoma of the prostate

BACKGROUND

Small cell carcinoma of the prostate is an uncommon neoplasm, the origin of which has been controversial. To address this, we performed transcriptome profiling and TP53 sequencing of concurrent small cell and prostatic adenocarcinoma to determine the relationship between these entities.

METHODS

We identified an unusual case of primary prostate cancer that contained adjacent acinar adenocarcinoma (Gleason score 4 + 3 = 7) and small cell carcinoma. We performed laser capture microdissection to isolate tumor components and performed gene expression and TP53 gene sequence analysis on each component, with results validated by immunohistochemistry for PSA, PSAP, PSMA, androgen receptor, NKX 3.1 and neuroendocrine markers.

RESULTS

Transcriptome profiling of the carcinoma components identified 99 genes with a greater than 10-fold differential expression between prostatic adenocarcinoma and small cell carcinoma, many of which have not been previously reported in prostate cancer. The small cell carcinoma component demonstrated upregulation of proliferative and neuroendocrine markers and tyrosine kinase receptors, and downregulation of cell adhesion molecules, supporting the aggressive nature of this form of carcinoma. Sequencing of the TP53 gene suggested a common clonal origin for both components.

CONCLUSIONS

This is the first report of a primary small cell carcinoma of the prostate subjected to extensive molecular analysis and the first to show a clonal relation between two morphologically distinct prostate cancer types. The evidence of progression to small cell carcinoma may yield important insights into the pathogenesis of this entity and provide a novel spectrum of molecular markers to further dissect cellular pathways important in tumor progression.

P21 gene knock down does not identify genetic effectors seen with gene knock out

RNA interference (RNAi) has become a popular tool for analyzing gene function in cancer research. The feasibility of using RNAi in cellular and animal models as an alternative to conventional gene knock out approaches has been demonstrated. Although these studies show that RNAi can recapitulate phenotypes seen in knock out animals and their derived cell lines, a systematic study rigorously comparing downstream effector genes between RNAi and gene knock out has not been performed. Here we present data contrasting the phenotypic and genotypic changes that occur with either stable knock down via RNAi of the cyclin dependent kinase inhibitor p21 versus its somatic cell knock out counterpart in the human mammary epithelial cell line MCF-10A. Our results demonstrate that p21 knock down clones display a growth proliferative response upon exposure to Transforming Growth Factor-Beta Type 1 (TGFbeta) similar to p21 knock out clones. However, gene expression profiles were significantly different in p21 knock down cells versus p21 knock out clones. Importantly p21 knock down clones did not display increased gene expression of interleukin-1alpha (IL-1alpha), a critical effector of this growth response previously validated in p21 knock out cells. We conclude that gene knock out can yield additional vital information that may be missed with gene knock down strategies.

Knock-in of mutant K-ras in nontumorigenic human epithelial cells as a new model for studying K-ras mediated transformation

The oncogenic function of mutant ras in mammalian cells has been extensively investigated using multiple human and animal models. These systems include overexpression of exogenous mutant ras transgenes, conditionally expressed knock-in mouse models, and somatic cell knockout of mutant and wild-type ras genes in human cancer cell lines. However, phenotypic discrepancies between knock-in mice and transgenic mutant ras overexpression prompted us to evaluate the consequences of targeted knock-in of an oncogenic K-ras mutation in the nontumorigenic human breast epithelial cell line MCF-10A and hTERT-immortalized human mammary epithelial cells. Our results show several significant differences between mutant K-ras knock-in cells versus their transgene counterparts, including limited phosphorylation of the downstream molecules extracellular signal-regulated kinase and AKT, minor proliferative capacity in the absence of an exogenous growth factor, and the inability to form colonies in semisolid medium. Analysis of 16 cancer cell lines carrying mutant K-ras genes indicated that 50% of cancer cells harbor nonoverexpressed heterozygous K-ras mutations similar to the expression seen in our knock-in cell lines. Thus, this system serves as a new model for elucidating the oncogenic contribution of mutant K-ras as expressed in a large fraction of human cancer cells.

Physiologic estrogen receptor alpha signaling in non-tumorigenic human mammary epithelial cells

Currently, a number of breast cancer cell lines exist that serve as models for both estrogen receptor alpha (ERalpha) positive and ERalpha negative disease. Models are also available for pre-neoplastic breast epithelial cells that do not express ERalpha; however, there are no ideal systems for studying pre-neoplastic cells that are ERalpha positive. This has been largely due to the inability to establish an estrogen growth stimulated, non-tumorigenic breast epithelial cell line, as most human breast epithelial cells engineered to overexpress ERalpha have been found to be growth inhibited by estrogens. We have developed independently derived clones from the non-cancerous MCF-10A human breast cell line that express ERalpha and are growth stimulated by 17-beta-estradiol (E2) in the absence of epidermal growth factor (EGF), a cytokine normally required for MCF-10A cell proliferation. This effect is blocked by the selective estrogen receptor modulator (SERM), Tamoxifen and the selective estrogen receptor downregulator, ICI 182,780 (Faslodex, Fulvestrant). Exposure of these cells to EGF and E2 results in a growth inhibitory phenotype similar to previous reports. These data present a reconciling explanation for the previously described paradoxical effects of ERalpha overexpression, and provide a model for examining the carcinogenic effects of estrogens in non-tumorigenic human breast epithelial cells.

Identification of Li+ binding sites and the effect of Li+ treatment on phospholipid composition in human neuroblastoma cells: a 7Li and 31P NMR study

Li(+) binding in subcellular fractions of human neuroblastoma SH-SY 5 Y cells was investigated using (7)Li NMR spin-lattice (T(1)) and spin-spin (T(2)) relaxation measurements, as the T(1)/T(2) ratio is a sensitive parameter of Li(+) binding. The majority of Li(+) binding occurred in the plasma membrane, microsomes, and nuclear membrane fractions as demonstrated by the Li(+) binding constants and the values of the T(1)/T(2) ratios, which were drastically larger than those observed in the cytosol, nuclei, and mitochondria. We also investigated by (31)P NMR spectroscopy the effects of chronic Li(+) treatment for 4--6 weeks on the phospholipid composition of the plasma membrane and the cell homogenate and found that the levels of phosphatidylinositol and phosphatidylserine were significantly increased and decreased, respectively, in both fractions. From these observations, we propose that Li(+) binding occurs predominantly to membrane domains, and that chronic Li(+) treatment alters the phospholipid composition at these membrane sites. These findings support those from clinical studies that have indicated that Li(+) treatment of bipolar patients results in irregularities in Li(+) binding and phospholipid metabolism. Implications of our observations on putative mechanisms of Li(+) action, including the cell membrane abnormality, the inositol depletion and the G-protein hypotheses, are discussed.

Competition between lithium and magnesium ions for the G-protein transducin in the guanosine 5'-diphosphate bound conformation

Li(+) is the most effective drug used to treat bipolar disorder; however, its exact mechanism of action has yet to be elucidated. One hypothesis is that Li(+) competes with Mg2+ for the Mg2+ binding sites on guanine-nucleotide binding proteins (G-proteins). Using 7Li T1 relaxation measurements and fluorescence spectroscopy with the Mg2+ fluorophore furaptra, we detected Li(+)/Mg(2+) competition in three preparations: the purified G-protein transducin (Gt), stripped rod outer segment membranes (SROS), and SROS with purified Gt reattached (ROS-T). When purified ROS-T, SROS or transducin were titrated with Li+ in the presence of fixed amounts of Mg(2+), the apparent Li(+) binding constant decreased due to Li(+)/Mg(2+) competition. Whereas for SROS the competition mechanism was monophasic, for G(t), the competition was biphasic, suggesting that in G(t), Li(+)/Mg(2+) competition occurred with different affinities for Mg(2+) in two types of Mg(2+) binding sites. Moreover, as [Li(+)] increased, the fluorescence excitation spectra of both ROS-T and G(t) were blue shifted, indicating an increase in free [Mg(2+)] compatible with Li(+) displacement of Mg(2+) from two low affinity Mg(2+) binding sites of G(t). G(t) release from ROS-T membrane was also inhibited by Li(+) addition. In summary, we found evidence of Li(+)/Mg(2+) competition in G(t)-containing preparations.

Biochemical and psychiatric predictors of Li(+) response and toxicity in Li(+)-treated bipolar patients

BACKGROUND

It has not been determined whether biochemical or psychological variables predict clinical response and toxicity to Li(+) treatment.

METHODS

From 30 Li(+)-treated bipolar patients, we measured biochemical variables in red blood cells (RBCs) that encompassed the cell membrane abnormality and the Li(+)/Mg(2+) competition mechanism. Psychiatric measures of depression, mania, and side effects of Li(+) toxicity were correlated with these biochemical variables. Physician classification of Li(+) response and toxicity for each patient were used for determining whether significant differences in biochemical variables and psychiatric measures existed between full and partial responders, and as well as toxic and non-toxic Li(+)-treated bipolar patients.

RESULTS

Serum [Li(+)] ([Li(+)]e), the ratio of intracellular RBC to serum Li(+), [Li(+)]i/[Li(+)]e, and phosphatidylcholine shared moderate proportions of variance (10-15%) with several of the psychiatric measures. Physician assessment of full response was predicted by higher levels of [Li(+)]e and lower scores on the Hamilton Slowing subscale (95.6% accuracy), whereas higher lithium-binding constants and higher Hamilton total scores perfectly predicted physician classification of partial response. Higher scores on Hamilton Slowing subscale and General Side Effects (GSE) scale were strongly predictive of physician classified Li(+) toxicity (80% accuracy), whereas lower levels of [Li(+)]e and lower scores on the Hamilton Symptom Severity subscale perfectly predicted physician rated non-toxicity in these patients.

CONCLUSIONS

We found distinct [Li(+)]e levels that predict response and/or toxicity. Specifically, when [Li(+)]e was in the range of 0.93-1.42 mM, full response without toxicity was predicted; higher values predicted toxicity; lower values predicted partial response.

Effects of Li+ transport and Li+ immobilization on Li+/Mg2+ competition in cells: implications for bipolar disorder

Li(+)/Mg(2+) competition has been implicated in the therapeutic action of Li(+) treatment in bipolar illness. We hypothesized that this competition depended on cell-specific properties. To test this hypothesis, we determined the degree of Li(+) transport, immobilization, and Li(+)/Mg(2+) competition in lymphoblastomas, neuroblastomas, and erythrocytes. During a 50 mM/L Li(+)-loading incubation, Li(+) accumulation at 30 min (mmoles Li(+)/L cells) was the greatest in lymphoblastomas (11.1+/-0.3), followed by neuroblastomas (9.3+/-0.5), and then erythrocytes (4.0+/-0.5). Li(+) binding affinities to the plasma membrane in all three cell types were of the same order of magnitude; however, Li(+) immobilization in intact cells was greatest in neuroblastomas and least in erythrocytes. When cells were loaded for 30 min in a 50 mM/L Li(+)-containing medium, the percentage increase in free intracellular [Mg(2+)] in neuroblastoma and lymphoblastoma cells ( approximately 55 and approximately 52%, respectively) was similar, but erythrocytes did not exhibit any substantial increase ( approximately 6%). With the intracellular [Li(+)] at 15 mM/L, the free intracellular [Mg(2+)] increased by the greatest amount in neuroblastomas ( approximately 158%), followed by lymphoblastomas ( approximately 75%), and then erythrocytes ( approximately 50%). We conclude that Li(+) immobilization and transport are related to free intracellular [Mg(2+)] and to the extent of Li(+)/Mg(2+) competition in a cell-specific manner.

Testing competing path models linking the biochemical variables in red blood cells from Li+-treated bipolar patients

OBJECTIVES

Red blood cells (RBCs) from Li+-treated bipolar patients have shown abnormalities in intracellular Li+ concentration ([Li+]i), Na+/Li+ exchange rates, and membrane phospholipid levels. Based on Li+-loaded RBC studies, we hypothesized that Li+-treated bipolar patients also have varied intracellular free Mg2+ concentrations ([Mg2+]f) as compared with normotensive patients. We addressed how these experimentally determined values are intercorrelated. Assuming that Li+ treatment alters these biochemical parameters, we provide hypothetical pathways based upon structural equation modeling statistics.

METHODS

In RBCs from 30 Li+-treated bipolar patients, we determined [Li+]i, serum [Li+] ([Li+]e), Na+/Li+ exchange parameters, membrane phospholipid levels, [Mg2+]f, and Li+ membrane binding affinities. Comprehensive statistical analyses assessed correlations among the biochemical data. We used path analysis statistics to propose potential pathways in which the data were correlated.

RESULTS

We found significant correlations within the three Na+/Li+ exchange parameters and percentage composition of the membrane phospholipids. Additional correlations existed between [Mg2+]f and Vstd, Km, or phospholipid composition, between [Li+]i and percentage of phosphatidylcholine, and between percentage of phosphatidylserine and Km. Based on these findings, we hypothesized and statistically determined the most probable pathway through which these parameters were intercorrelated.

CONCLUSIONS

Significant correlations existed between the biochemical parameters that describe the cell membrane abnormality and the Li+/Mg2+ competition hypotheses. Using path analysis statistics, we identified a biochemical pathway by which Li+ may assert its cellular effects. This study serves as an illustrative example how path analysis is a valuable tool in determining the direction of a certain biochemical pathway.

Evaluation of [Co(gly)3]- as a 35Cl- NMR shift reagent for cellular studies

We studied the efficacy of the tris-glycinatocobaltate(II) complex ([Co(gly)(3)](-)) as a shift reagent (SR) for chloride by (35)Cl NMR spectroscopy and compared to that of Co(2+)((aq)). Due to the relatively low thermodynamic stability of [Co(gly)(3)](-), a 1:3 Co(II)/gly stoichiometric solution at physiological pH is approximately a 2:1 mixture of [Co(gly)(2)(H(2)O)(2)] and [Co(gly)(H(2)O)(4)](+). This SR was found to be stable up to higher pH values than Co(2+)((aq)), better preventing Co(OH)(2) formation at alkaline pH. No significant differences in the (35)Cl(-) NMR chemical shift induced by Co(II)/gly or Co(2+)((aq)) were observed in the presence of physiological concentrations of either Ca(2+) or Mg(2+), or of either Na(+) or K(+). Although Co(2+)((aq)) was almost twice as effective as Co(II)/gly in shifting the (35)Cl(-) NMR resonance at the same high rho ([SR]/[Cl(-)]) value and low ionic strength, Co(2+)((aq)) showed a significant decrease (p < 0.05) in the (35)Cl(-) chemical shift at higher ionic strength. Line widths at half-height were significantly (p < 0.05) less for Co(II)/gly than for Co(2+)((aq)) at rho values in the range 0.066-0.40. Intracellular chloride was clearly detectable by (35)Cl NMR spectroscopy in human skin fibroblast cells suspended in medium containing 40 mM Co(II)/gly SR. We determined that, although Co(2+)((aq)) provides a larger shift than Co(II)/gly at the same rho value, there are significant advantages for using Co(II)/gly, such as pH stability, ionic strength independent chemical shifts, narrow (35)Cl(-) NMR resonances, and reduced cellular toxicity, as a SR in biological systems.

Effect of chronic Li+ treatment on free intracellular Mg2+ in human neuroblastoma SH-SY5Y cells

OBJECTIVES

Previous findings have demonstrated Li+/Mg2+ competition at therapeutic intracellular Li+ levels after acute Li+ treatment in human neuroblastoma SH-SY5Y cells. In the current study, we examined whether Li+/Mg2+ competition exists at therapeutically relevant extra- and intracellular [Li+] after chronic Li+ loading times.

METHODS

In human neuroblastoma cells, intracellular free Mg2+ was determined by fluorescence spectroscopy with the fluorophore furaptra. Intracellular Li+ and Mg2+ were measured by atomic absorption spectrophotometry.

RESULTS

After loading of the neuroblastoma cells with 1-2 mM extracellular Li+ for 24-72 h, the observed, increased intracellular free [Mg2+] levels were significantly higher (p < 0.03) than those in matched Li+ free cells, and intracellular [Li+] was found to be at therapeutic intracellular levels (0.7-1.5 mM).

CONCLUSIONS

The results demonstrate that Li+/Mg2+ competition exists after chronic treatment with Li+ at therapeutically relevant intracellular Li+ levels in neuroblastoma cells. We found differences between acute and chronic Li+ treatment effects on the extent of Li+/Mg2+ competition. Possible reasons for these differences are discussed.