Abstract 144: Genome-wide DNA methylation changes in AI-resistant breast cancer models during E2-treatment

Disruption of estrogen signaling with tamoxifen or aromatase inhibitors (AI) is the clinical standard for the treatment of estrogen receptor-positive breast cancer, yet resistance to antiestrogen therapies is inevitable. The long-term E2-deprived breast cancer models MCF-7:5C and MCF-7:2A mimic AI resistance and moreover are vulnerable to E2-induced apoptosis due to the reconfiguration of their survival signaling. E2-induced apoptosis which results from the accumulation of endoplasmic reticulum (ER), oxidative, and inflammatory stresses, is a new paradigm to subvert endocrine resistance, with sensors of the unfolded protein response (UPR) being important mediators of ER stress (Ariazi et al. 2011; Fan et al. 2013, 2015). Here we asked whether epigenetic modifications may play a role in this and investigated genome-wide changes of 5C and 2A DNA methylation patterns that occurred during E2 stimulation during a 72h time course with WS8 cells as a reference. Methylation patterns were established using the Illumina Infinium HumanMethylation450 BeadChip (Service XS, Leiden, NL), and raw intensities were normalized and filtered using the bioconductor minfi package. A linear model (bioconductor limma package) has been built to detect differentially methylated CpGs. F-statistics were calculated by empirical Bayes moderation of standard errors with "trend" and "robust" set “on”. We also performed whole-genome bisulfite sequencing (WGBS) in untreated cells using the Illumina HiSeq platform (GATC Biotech, Konstanz, DE) to obtain maximum baseline CpG coverage. Trimmed WGBS reads were aligned to the hg38 genome (Bismarck, PMC3102221), duplicates removed, and methylation levels extracted with MethylDackel (https://github.com/dpryan79/MethylDackel) with a 3x coverage threshold (Wreczycka et al. 2017). Raw methylation levels were smoothed and differential methylation called using the bsseq package (Hansen et al. 2012). At baseline, WGBS-derived profiles showed a 2-fold higher degree of hypomethylation (beta <0.2) in 5C compared to 2A and WS8. In time course E2-treatments we identified 134 dynamic CpGs across all three models (Padj < 0.05). Methylation status mainly decreased, with few specific changes in 5C and 2A. Among 120 affected CpG sites matching to GeneHancer regulatory elements 23 locate to ERE sites suggesting a direct control of inferred target genes (UCSC genome browser). Their matching to ER stress-, UPR- and apoptosis-related gene sets identified among others ATF6B, EIF2AK4, HSPB1, TNF, GANAB and PPIA. Currently, we are performing functional enrichment analyses and E2-responsive CpGs are being correlated with respective transcriptome signatures. Our preliminary data suggest that E2-triggered methylation changes in cell models vulnerable to E2-induced apoptosis may aid the identification of potential therapeutic targets towards overcoming endocrine resistance.

Citation Format: Reiner Hoppe, Siarhei Kandabarau, Ping Fan, Stefan Winter, Balkees Abderrahman, Heather Cunliffe, Werner Schroth, V. Craig Jordan, Hiltrud B. Brauch. Genome-wide DNA methylation changes in AI-resistant breast cancer models during E2-treatment [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 144.

Abstract 1689: Comparative whole genome analyses of breast cancer aromatase inhibitor resistance models

One third of breast cancer patients treated with Aromatase Inhibitors (AI) or Tamoxifen acquire resistance leading to disease recurrence and death. The clinically observed AI resistance is mimicked by in vitro models MCF-7:5C and MCF-7:2A that are susceptible to apoptosis upon E2-treatment. These models have been instrumental to the current understanding of underlying mechanisms such as the reconfiguration of survival signaling, alterations of stress-related pathways (Ariazi et al. 2011; Fan et al. 2013, 2015), and changes in miRNA expression profiles (Hoppe et al. 2016). How the model phenotypes relate to a specific genomic background is poorly understood. In order to identify somatic variants indicative of AI resistance and potentially useful as novel therapeutic targets we performed NGS analyses of the 5C and 2A clones compared to the non-resistant WS8 and MCF-7 (ATCC). Whole genome/exome libraries were sequenced on Illumina HiSeq/NovaSeq platforms, respectively. Previously we reported first results from a preliminary analysis of gains and losses of variants in both resistant derivatives (Hoppe at al. AACR 2018). Here we report an updated analysis of raw reads trimmed and aligned to GRCh38 reference using BWAkit v 0.7.15. Duplicates were removed by Picard MarkDuplicates. Base quality scores were recalibrated using GATK4 BQSR. WGS and WES alignments were merged and tested for somatic variants using GATK4 Mutect2 (Tumor-only mode) with population variation data from gnomAD as a germline source. Called and filtered variants were uploaded to IVA for functional and clinical annotation. Copy number segments were modeled using both read- and allele-counts from WGS alignments by Gaussian-kernel binary-segmentation algorithm applied in GATK4 ModelSegments tool. To interpret copy number events causative for LOH we recovered germline variants and annotated CNVs that led to loss of the corresponding wild-type allele. In line with the COSMIC database we confirmed 461 mutated loci of which 228 were lost in both resistant derivatives. Examples of 5C-specific coding mutations classified as pathogenic include splice site alterations (i.e. SEMA3A), frameshifts (i.e. ARID1A), and stop gains (i.e. CACNA1S, CDH2, NOTCH4). 2A-specific mutations include for example a stop gain affecting ADGRV1. Copy number variation analyses revealed that both resistance models share prominent genetic defects such as amplified regions at Chr6q25.1 and Chr21q22.2. A 5C-specific aberration is a large Chr18 deletion affecting most of the q-arm, and a 2A-specific aberration is the amplification of most parts of Chr6q23.3-q24.3. We will present the status of this currently ongoing comparative genome analysis of AI resistance models and discuss the putative effects of homozygous mutations at loci with copy number aberrations within the context of AI resistance.

Citation Format: Reiner Hoppe, Siarhei Kandabarau, Ping Fan, V. Craig Jordan, Heather Cunliffe, Hiltrud B. Brauch. Comparative whole genome analyses of breast cancer aromatase inhibitor resistance models [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1689.

Regulation of the interferon-gamma (IFN-γ) pathway by p63 and Δ133p53 isoform in different breast cancer subtypes

The TP53 family consists of three sets of transcription factor genes, TP53, TP63 and TP73, each of which expresses multiple RNA variants and protein isoforms. Of these, TP53 is mutated in 25-30% of breast cancers. How TP53 mutations affect the interaction of TP53 family members and their isoforms in breast cancer is unknown. To investigate this, 3 independent breast cancer cohorts were stratified into 4 groups based on oestrogen receptor (ER) and TP53 mutation status. Using bioinformatic methodologies, principal signalling pathways associated with the expression of TP53 family members were identified. Results show an enrichment of IFN-γ signalling associated with TP63 RNA in wild type TP53 (wtTP53), ER negative (ER-) tumours and with Δ133TP53 RNA in mutant TP53 (mTP53) ER positive (ER+) tumours. Moreover, tumours with low IFN-γ signalling were associated with significantly poorer patient outcome. The predicted changes in expression of a subset of RNAs involved in IFN-γ signalling were confirmed in vitro. Our data show that different members of the TP53 family can drive transcription of genes involved in IFN-γ signalling in different breast cancer subgroups.

Profiles of miRNAs matched to biology in aromatase inhibitor resistant breast cancer

Aromatase inhibitor (AI) resistance during breast cancer treatment is mimicked by MCF-7:5C (5C) and MCF-7:2A (2A) cell lines that grow spontaneously. Survival signaling is reconfigured but cells are vulnerable to estradiol (E₂)-inducible apoptosis. These model systems have alterations of stress related pathways including the accumulation of endoplasmic reticulum, oxidative, and inflammatory stress that occur prior to E₂-induced apoptosis. We investigated miRNA expression profiles of 5C and 2A to characterize their AI resistance phenotypes. Affymetrix GeneChip miRNA2.0 arrays identified 184 miRNAs differentially expressed in 2A and 5C compared to E₂-free wild-type MCF-7:WS8. In 5C, 34 miRNAs of the DLK1-DIO3 locus and miR-31 were overexpressed, whereas miR-222 was low. TCGA data revealed poor and favorable overall survival for low miR-31 and miR-222 levels, respectively (HR=3.0, 95% CI:1.9-4.8; HR=0.3, 95% CI:0.1-0.6). Targets of deregulated miRNAs were identified using CLIP-confirmed TargetScan predictions. KEGG enrichment analyses for 5C- and 2A-specific target gene sets revealed pathways associated with cell proliferation including insulin, mTOR, and ErbB signaling as well as immune response and metabolism. Key genes overrepresented in 5C- and 2A-specific pathway interaction networks including EGFR, IGF1R and PIK3R1 had lower protein levels in 5C compared to 2A and were found to be differentially modulated by respective miRNA sets. Distinct up-regulated miRNAs from the DLK1-DIO3 locus may cause these attenuative effects as they are predicted to interact with corresponding 3′ untranslated regions. These new miRNA profiles become an important regulatory database to explore E₂-induced apoptotic mechanisms of clinical relevance for the treatment of resistant breast cancer.

Abstract 1073: Modulation of aromatase inhibitor resistance by miRNAs in breast cancer

The blockade of E2 signaling by either tamoxifen or aromatase inhibitors (AI) is the standard treatment of patients with estrogen receptor (ER)-positive breast cancer. However, acquired resistance to antiestrogen therapies remains a big clinical challenge. The in vitro breast cancer models MCF-7:5C and MCF-7:2A mimic clinical AI resistance in that they grow under tamoxifen, but can rapidly regress with E2 due to the reconfiguration of survival signaling known as E2-inducible apoptosis. This principle is currently explored in clinical trials in order to develop strategies to overcome endocrine resistance. To better understand the biology of AI resistance we previously ascertained a wide range of alterations of stress related pathways including the accumulation of endoplasmic reticulum stress, oxidative stress, and inflammatory stress that occur prior to E2-induced apoptosis (Ariazi et al. 2011; Fan et al. 2013, 2015; Sweeney et al. 2014). In this current work we investigated miRNA expression profiles of the 5C and 2A models (both compared to MCF-7:WS8 reference) in order to further elucidate the molecular scenario that characterizes the AI resistance phenotypes and their susceptibility to E2-induced apoptosis. Using Affymetrix GeneChip miRNA2.0 arrays we identified 184 miRNAs differentially expressed between 2A and 5C (FC > 1.5 or < 1/1.5, P < 0.05). Of these, 30 miRNAs were specific for 2A, 99 for 5C, and 55 overlapped. Common miRNAs include upregulated oncogenic miRNAs of the miR-17-92 cluster (Chr. 13q31) and the paralogous 106a-363 cluster (Chr. X), as well as downregulated tumor suppressive miRNAs such as miR-342-5p. Thirty-four miRNAs that cluster at Chr. 14q32 were overexpressed in 5C cells and highly correlate with the downregulated miR-99a and miR-125b (Chr. 21q21). At the clinical level based on data from The Cancer Genome Atlas (miRNA-Seq v. 3.1.17.0), low expression of miR-31 was associated with poor (HR = 3.0, 95% CI: 1.9-4.8; Padj = 8.7E-5), and low miR-222 expression was associated with better outcome (HR = 0.3, 95% CI: 0.1-0.6; Padj. = 4.4E-3). As miR-31 expression is high and miR-222 expression is low in 5C cells compared to the MCF-7:WS8 reference, we suggest protective roles. Functions of miRNAs were deciphered via analyses of their predicted targets (CLIP-confirmed). Using KEGG and GO databases functional enrichment analyses of 5C and 2A specific miRNA sets revealed pathways associated with cell proliferation for both models including insulin, mTOR, and ErbB signaling as well as immune response and metabolism associated pathways. While the 2A specific miRNA set revealed additional metabolic pathways, the 5C specific miRNA set points to pathways involved in apoptosis. Thus, we confirmed the biological processes inherent to AI resistance and provide critical evidence for miRNA profiles as an important regulatory principle in these AI resistance models.

Citation Format: Reiner Hoppe, Ping Fan, Florian Büttner, Stefan Winter, Heather Cunliffe, V. Craig Jordan, Hiltrud B. Brauch. Modulation of aromatase inhibitor resistance by miRNAs in breast cancer. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 1073.

The influence of clinical and genetic factors on patient outcome in small cell carcinoma of the ovary, hypercalcemic type

OBJECTIVE

Small cell carcinoma of the ovary, hypercalcemic type (SCCOHT) is an aggressive tumor, with long term survival at ~30% in early stage disease. SCCOHT is caused by germline and somatic SMARCA4 mutations, but the effect of the mutation type on patients remains unknown. Furthermore, the rarity of SCCOHT has resulted in varied treatment, with no standardized protocols. We analyzed 293 cases to determine the effect of treatment modalities and SMARCA4 mutations on patient diagnosis and outcome.

METHODS

In 293 SCCOHT patients we collected information on age and stage at diagnosis, treatment modality (surgery, chemotherapy, radiotherapy, and/or high-dose chemotherapy with autologous stem cell rescue (HDC-aSCR)), SMARCA4 mutation origin (germline/somatic), and overall survival. Cox analysis and log-rank tests were performed on 257 cases with available survival data.

RESULTS

The strongest prognostic factors were stage at diagnosis (p=2.72e-15) and treatment modality (p=3.87e-13). For FIGO stages II-IV, 5-year survival was 71% for patients who received HDC-aSCR, compared to 25% in patients who received conventional chemotherapy alone following surgery (p=0.002). Patients aged ≥40 had a worse outcome than younger patients (p=0.04). Twenty-six of 60 tested patients carried a germline SMARCA4 mutation, including all patients diagnosed <15years; carriers presented at a younger age than non-carriers (p=0.02).

CONCLUSIONS

Stage at diagnosis is the most significant prognostic factor in SCCOHT and consolidation with HDC-aSCR may provide the best opportunity for long-term survival. The large fraction of SMARCA4 germline mutations carriers warrants genetic counseling for all patients.

Abstract LB-175: Deep clonal profiling identifies distinct mechanisms of heterogeneity and evolution in breast cancer

Background: Breast tumors exhibit intratumor heterogeneity resulting in targeted therapy resistance and other challenges in disease management. To address the sources of heterogeneity, we performed a unique, in-depth analysis of clonal architecture in primary chemoradiation-naïve breast cancers. We combined DNA content-based flow cytometry and ploidy analysis with aCGH and whole exome next-generation sequencing (NGS) in multiple biopsies from the tumors and involved lymph nodes (LNs).

Material and methods: We sorted nuclei from distinct populations of diploid, tetraploid, and aneuploid cells in surgical tumor samples from three chemoradiation-naïve patients. Each sorted tumor cell population was interrogated with aCGH and whole exome NGS. In Patient #1, we sorted and interrogated the genomes of tumor populations from 12 fresh frozen sections morphologically mapped from within a HER2+, ER+, PR- primary invasive ductal carcinoma (IDC) of histological grade 3 with LN involvement and 2-3 sections from 2 out of 5 LNs. In Patient #2, 11 morphologically mapped fresh frozen sections were analyzed from a grade 2, ER+, PR+, HER2+, BRCA2 mutant LN- IDC. In parallel, matching samples were processed for IHC assays. In Patient #3 biopsies from a grade 2 ER+, PR+ primary tumor were profiled.

Results: The 18 primary and LN biopsies from Patient #1 fell into 6 distinct ploidy groups albeit with aberrant but homogenous aCGH profiles, characterized by SARC amplification and homozygous deletions in ROBO1 and ROBO2. In contrast a dominant ploidy was identified throughout 10 biopsies in Patient #2 but with heterogeneous aCGH profiles. Patient #3 had a clonal homozygous deletion in Numb in each of 4 tumor biopsies. Mutation profiles obtained through exome sequencing further confirmed that ploidy was the main driver in Patient #1 whereas copy number aberrations played the key role in Patient #2 with the BRCA2 mutation (R3129X). A dendrogram based on exome variant calls of the aneuploid populations in Patient #1 strongly correlated with ploidy group and further revealed the specific clonal population characterized by a 5N ploidy and homozygous mutations in TP53 and PIK3CA as the progenitor to the ploidies present in the distant LNs. Strikingly, patients #1 and #2 were HER2 wild type across their clonal populations, contradicting IHC staining in a single core biopsy.

Conclusions:Rather than inferring the presence of distinct tumor cell populations, our flow-sorting based approach of first identifying clonal populations and then interrogating their genomes, provides an objective method of exploring the sources and clinical significance of tumor heterogeneity. Our approach of clonal analysis has broad implications in the study of tumor heterogeneity and the identification of drivers in breast and other solid tumors that can advance more effective treatment and clinical management of patients with this disease.

Citation Format: MARIACARLA ANDREOZZI, Princy Francis, Elizabeth Lenkiewicz, Mia Champion, Brady Laughlin, Karen Anderson, Heather Cunliffe, Ann E. McCullough, Michael T. Barrett, Barbara Pockaj. Deep clonal profiling identifies distinct mechanisms of heterogeneity and evolution in breast cancer. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr LB-175. doi:10.1158/1538-7445.AM2015-LB-175

Abstract LB-202: The rare, highly malignant small cell carcinoma of the ovary displays common inactivating germline and somatic mutations in the tumor suppressor SMARCA4

Small Cell Carcinoma of the Ovary Hypercalcemic Type (SCCOHT) is a rare and highly aggressive malignancy that affects children and young women at a mean age of 24 (range 14 months - 58 years). SCCOHT is refractory to standard of care therapy for ovarian cancer, with ∼75% mortality within 18 months of diagnosis. The early age of onset of SCCOHT and reports of familial occurrence in some cases, strongly suggest an underlying hereditary etiology. To understand the molecular pathogenesis of SCCOHT, we performed next-generation genomic sequencing on a series of tumor and germline samples from SCCOHT patients. This analysis revealed germline and somatic inactivating mutations in SMARCA4, a subunit of the SWI/SNF chromatin-remodeling complex, in 75% (9/12) of SCCOHT patients. Moreover, immunohistochemical (IHC) analysis of 15 tumors revealed that 87% (13/15) of tumors lacked SMARCA4 protein. The high prevalence of SMARCA4 mutations in SCCOHT has not been previously reported in other, more common ovarian carcinomas. We therefore examined the expression of SMARCA4 protein in 300 ovarian carcinomas of different histologies by IHC and found SMARCA4 protein loss in only 6 tumors. In addition, the BIN-67 SCCOHT cell line, which harbors 2 splice site mutations in SMARCA4, showed complete absence of SMARCA4 protein by Western blot while representative cell lines from 4 other ovarian carcinoma subtypes as well as immortalized granulosa cells (SVOG) and adult granulosa tumor cells (KGN) all maintained SMARCA4 expression. The prevalence of germline and sporadic SMARCA4 mutations as well as frequent SMARCA4 protein loss in SCCOHTs implicates this gene as a tumor suppressor in this cancer and more broadly suggests a role for the SWI/SNF complex in its pathogenesis. In addition to providing evidence to the pathogenesis of SCCOHT, this finding provides the opportunity to develop treatment approaches for SCCOHT based on targeting vulnerabilities of SMARCA4-deficient cells.

Citation Format: Pilar Ramos, Anthony Karnezis, David Craig, Aleksandar Sekulic, Megan Russell, William Hendricks, Michael Barrett, Karey Shumansky, Yidong Yang, Sohrab Shah, Leah Prentice, Marco Marra, Jeffrey Kiefer, Victoria Zismann, Troy McEachron, Bodour Salhia, Joseph Pressey, John Farley, Stephen Anthony, Richard Roden, Heather Cunliffe, David Huntsman, Jeffrey Trent. The rare, highly malignant small cell carcinoma of the ovary displays common inactivating germline and somatic mutations in the tumor suppressor SMARCA4. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr LB-202. doi:10.1158/1538-7445.AM2014-LB-202

Increased expression of folate receptor-α (FRA) in triple-negative breast cancer: A potential therapeutic target

1037

Background: Folate receptor alpha (FRA the product of the FOLR1 gene) has been identified as a potential prognostic and therapeutic target in a number of cancers. A correlation has been shown between intense expression of FRA in breast tumors and poor prognosis, yet little is known about FOLR1/FRA expression across clinically relevant breast cancer subtypes. Methods: 131 breast cancer tumors including 4 benign, 33 ER+, 26 HER2+, and 68 triple negative (TN) were constructed into tissue microarrays (TMAs). FRA expression was analyzed by immunohistochemistry (IHC) using a high affinity FRA antibody. Tumor membrane staining intensity was scored by a pathologist as negative (0), weak (1+), moderate (2+) and strong (3+). The percent of cells within each tissue core stained at each intensity was recorded to calculate an H-score. The H-score is a weighted score that captured both the proportion of positive staining and intensity for each tumor. H-score values can range from zero (no membrane staining) to a maximum of 300 (100% membrane staining at 3+). H-scores for each patient sample were averaged over 3 TMA cores. The mean H-scores for each tumor subtype and the percentage of 3+ staining in >30% of tumor cells were compared by a Mann-Whitney test. The distribution of FOLR1 mRNA was completed using a TCGA RNA-seq dataset from 691 breast tumors classified as ER+, HER2+ and TN. FOLR1 levels of TN versus ER+ and HER2+ were compared by a Mann-Whitney test. Results: The mean H-score for the benign tumors was 0, ER+ (13.31), HER2+ (39.36), TN (119.02). The median H-score for the benign tumors was 0, ER+ (0), HER2+ (7.5), TN (127.5). The TN tumors mean and median H scores were significantly higher than benign, ER+ or HER2+ (p<0.001). The largest percentage of 3+ staining in >30% of tumor cells was observed in TN tumors (36.7%) and lowest in ER+ tumors (0%) (p<0.001). TN tumors had significantly higher levels of FOLR1 mRNA compared to ER+ and HER2+ subtypes (p<0.0001). Conclusions: Our data indicate that expression of FRA is highly prevalent in TN tumors and is supported by FOLR1 mRNA levels. Anti-FRA therapy may represent an important therapeutic intervention in TNBC who to this point have no active targeted treatment options.

The St. Gallen Prize Lecture 2011: evolution of long-term adjuvant anti-hormone therapy: consequences and opportunities

The successful translation of the scientific principles of targeting the breast tumour oestrogen receptor (ER) with the nonsteroidal anti-oestrogen tamoxifen and using extended durations (at least 5 years) of adjuvant therapy, dramatically increased patient survivorship and significantly enhanced a drop in national mortality rates from breast cancer. The principles are the same for the validation of aromatase inhibitors to treat post-menopausal patients but tamoxifen remains a cheap, life-saving medicine for the pre-menopausal patient. Results from the Oxford Overview Analysis illustrate the scientific principle of "longer is better" for adjuvant therapy in pre-menopausal patients. One year of adjuvant therapy is ineffective at preventing disease recurrence or reducing mortality, whereas five years of adjuvant tamoxifen reduces recurrence by 50% which is maintained for a further ten years after treatment stops. Mortality is reduced but the magnitude continues to increase to 30% over a 15-year period. With this clinical database, it is now possible to implement simple solutions to enhance survivorship. Compliance with long-term anti-hormone adjuvant therapy is critical. In this regard, the use of selective serotonin reuptake inhibitors (SSRIs) to reduce severe menopausal side effects may be inappropriate. It is known that SSRIs block the CYP2D6 enzyme that metabolically activates tamoxifen to its potent anti-oestrogenic metabolite, endoxifen. The selective norepinephrine reuptake inhibitor, venlafaxine, does not block CYP2D6, and may be a better choice. Nevertheless, even with perfect compliance, the relentless drive of the breast cancer cell to acquire resistance to therapy persists. The clinical application of long-term anti-hormonal therapy for the early treatment and prevention of breast cancer, focused laboratory research on the discovery of mechanisms involved in acquired anti-hormone resistance. Decades of laboratory study to reproduce clinical experience described not only the unique mechanism of selective ER modulator (SERM)-stimulated breast cancer growth, but also a new apoptotic biology of oestradiol action in breast cancer, following 5 years of anti-hormonal treatment. Oestradiol-induced apoptotic therapy is currently shown to be successful for the short-term treatment of metastatic ER positive breast cancer following exhaustive treatment with anti-hormones. The "oestrogen purge" concept is now being integrated into trials of long-term adjuvant anti-hormone therapy. The Study of Letrazole Extension (SOLE) trial employs "anti-hormonal drug holidays" so that a woman's own oestrogen may periodically purge and kill the nascent sensitized breast cancer cells that are developing. This is the translation of an idea first proposed at the 1992 St. Gallen Conference. Although tamoxifen is the first successful targeted therapy in cancer, the pioneering medicine is more than that. A study of the pharmacology of tamoxifen opened the door for a pioneering application in cancer chemoprevention and created a new drug group: the SERMs, with group members (raloxifene and lasofoxifene) approved for the treatment and prevention of osteoporosis with a simultaneous reduction of breast cancer risk. Thus, the combined strategies of long-term anti-hormone adjuvant therapy, targeted to the breast tumour ER, coupled with the expanding use of SERMs to prevent osteoporosis and prevent breast cancer as a beneficial side effect, have advanced patient survivorship significantly and promise to reduce breast cancer incidence.

PD03-07: Breast Cancer Heterogeneity and Treatment Resistance: Clues from Metaplastic Tumors

At late stage, nearly all breast cancers are heterogeneous and refractory to treatment, like metaplastic breast cancer is at an early stage. These rare carcinomas are highly aggressive and de-differentiated. They are enriched for mesenchymal and stem cell features and essentially fail current therapies. As metaplastic tumors provide a time-compressed picture of breast cancer progression early on, understanding these tumors will yield insight into mechanisms that drive breast cancer into advanced stages and treatment resistance.

To investigate a genetic basis for heterogeneity in metaplastic breast cancer, we established a progression model comprising three cell lines. The cell lines were derived from a primary tumor, a local recurrence and a pleural effusion of a 40-year old patient. The primary tumor was a stage III invasive metaplastic, triple negative, inflammatory breast cancer, resected after neoadjuvant chemotherapy (capecitabine and taxotere, then adriamycin and one cycle of bevacizumab). The local recurrence, biopsied seven months post mastectomy, developed after the patient received adjuvant carboplatin and gemcitabine for 3 cycles and then radiation to the chest wall. At this time, the patient had lung metastases and was treated with taxol and bevacizumab yielding a mixed response. Local invasive growth continued and a malignant pleural effusion developed four months later. Analyzing the genetic and molecular characteristics of this progression model in vitro, its tumorigenicity and metastasis in vivo, and interrogating lead findings in a growing collection of metaplastic tumors helps us to dissect the genetic heterogeneity in breast cancer, and potentially to identify the cell types that drive disease progression and treatment resistance. Our gene expression analyses and genomic evaluations identified epithelial to mesenchymal transition (EMT) as a key characteristic in the progression and treatment resistance of this cancer. Major changes in cytoskeletal genes, chemokines and their receptors, amplification of drug transporter proteins, metalloproteinases and matrix proteins seen with increasing motility and invasiveness along with recruitment of host inflammatory responses in the in vivo model, loss of chromosomal regions harboring known and putative tumor suppressors, and deletions of genes encoding proteins for metabolic inactivation of sex hormones in the breast tissue, along with specific loss of clusters of desmosomal genes are guiding our understanding of metaplastic breast cancer progression. The results provide insight into the development, the extremely invasive nature, and treatment resistance of these tumors. Our collaborative network of clinicians, pathologists, translational genomic researchers and bioinformatics specialists will enable us to identify and prioritize genetic events as disease drivers, prognostic biomarkers of disease progression, and determinants of treatment resistance. Our goal is to identify molecular and functional targets for effective therapy and evaluate them in the clinic. Lessons learned from metaplastic breast cancer will improve our understanding of breast cancer progression in general, and could translate into effective treatments for advanced breast cancer where current standard of care is failing.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr PD03-07.

In vitro assessment of the inflammatory breast cancer cell line SUM 149: Discovery of two single nucleotide polymorphisms in the RNase L gene

196

Background: Inflammatory breast cancer (IBC) is a rare, aggressive form of breast cancer, accounting for 5% of breast cancers diagnosed annually in the United States. Understanding the distinct biology of IBC could help provide novel treatment targets. We sought to evaluate whether or not the IBC cell lines SUM 149 and SUM 190 demonstrated evidence of viral infection.

Methods: We performed single nucleotide polymorphism (SNP) genotyping for 2 variants of the ribonuclease (RNase) L gene that have been correlated with the risk of prostate cancer due to a possible viral etiology. We also performed proliferation assays; developed dose response curves to assess the treatment effect of interferon-alpha (IFN-a); and assayed for evidence of the putative human mammary tumor virus (HMTV, which has been implicated but not definitively associated with IBC,) in the DNA and RNA of SUM 149 cells.

Results: According to our allelic discrimination SNP assay, 2/2 IBC cell lines were homozygous for the 462 and 541 variants, whereas 0/10 non-IBC cell lines were homozygous positive for the 462 variant (p = 0.015) and 2/10 non-IBC cell lines contained homozygous alleles for the 541 variant (p = 0.52). We also found a dose and time-dependent decrease in the proliferation of SUM 149 IBC cells treated with IFN-a. In contrast, non-IBC cell lines did not show a dose-response decrease in cell proliferation. Our reverse transcriptase polymerase chain reaction (RT-PCR) and Southern blot analysis for the env/LTR and late LTR sequences of the putative HMTV revealed no evidence of the putative viral genome.

Conclusions: We discovered 2 SNPs, in the 462 and 541 variants of the RNase L gene, that were homozygously mutated in IBC cell lines but the 462 variant was absent in non-IBC lines. Our discovery of these mutated SNPs present in IBC cell lines suggests a possible genetic risk factor for IBC. In our study, the IBC cell line SUM 149 demonstrated a direct and specific response to treatment with IFN-a, an antiviral agent. We noted no evidence of HMTV infection in that cell line. Further studies of the prevalence and significance of the RNase L 462 and 541 variants in human IBC tissue specimens are warranted to validate our in vitro findings.

Exploiting the apoptotic actions of oestrogen to reverse antihormonal drug resistance in oestrogen receptor positive breast cancer patients

The ubiquitous application of selective oestrogen receptor modulators (SERMs) and aromatase inhibitors for the treatment and prevention of breast cancer has created a significant advance in patient care. However, the consequence of prolonged treatment with antihormonal therapy is the development of drug resistance. Nevertheless, the systematic description of models of drug resistance to SERMs and aromatase inhibitors has resulted in the discovery of a vulnerability in tumour homeostasis that can be exploited to improve patient care. Drug resistance to antihormones evolves, so that eventually the cells change to create novel signal transduction pathways for enhanced oestrogen (GPR30+OER) sensitivity, a reduction in progesterone receptor production and an increased metastatic potential. Most importantly, antihormone resistant breast cancer cells adapt with an ability to undergo apoptosis with low concentrations of oestrogen. The oestrogen destroys antihormone resistant cells and reactivates sensitivity to prolonged antihormonal therapy. We have initiated a major collaborative program of genomics and proteomics to use our laboratory models to map the mechanism of subcellular survival and apoptosis in breast cancer. The laboratory program is integrated with a clinical program that seeks to determine the minimum dose of oestrogen necessary to create objective responses in patients who have succeeded and failed two consecutive antihormonal therapies. Once our program is complete, the new knowledge will be available to translate to clinical care for the long-term maintenance of patients on antihormone therapy.

Primary renal hypoplasia in humans and mice with PAX2 mutations: evidence of increased apoptosis in fetal kidneys of Pax2(1Neu) +/- mutant mice

PAX2 mutations cause renal-coloboma syndrome (RCS), a rare multi-system developmental abnormality involving optic nerve colobomas and renal abnormalities. End-stage renal failure is common in RCS, but the mechanism by which PAX2 mutations lead to renal failure is unknown. PAX2 is a member of a family of developmental genes containing a highly conserved 'paired box' DNA-binding domain, and encodes a transcription factor expressed primarily during fetal development in the central nervous system, eye, ear and urogenital tract. Presently, the role of PAX2 during kidney development is poorly understood. To gain insight into the cause of renal abnormalities in patients with PAX2 mutations, kidney anomalies were analyzed in patients with RCS, including a large Brazilian kindred in whom a new PAX2 mutation was identified. In a total of 29 patients, renal hypoplasia was the most common congenital renal abnormality. To determine the direct effects of PAX2 mutations on kidney development fetal kidneys of mice carrying a Pax2 (1Neu)mutation were examined. At E15, heterozygous mutant kidneys were approximately 60% of the size of wild-type littermates, and the number of nephrons was strikingly reduced. Heterozygous 1Neu mice showed increased apoptotic cell death during fetal kidney development, but the increased apoptosis was not associated with random stochastic inactivation of Pax2 expression in mutant kidneys; Pax2 was shown to be biallelically expressed during kidney development. These findings support the notion that heterozygous mutations of PAX2 are associated with increased apoptosis and reduced branching of the ureteric bud, due to reduced PAX2 dosage during a critical window in kidney development.

Assessment of antibiotic aerosol generation using commercial jet nebulizers

The performance of 14 commercial jet nebulisers has been assessed; Unineb, Suremist (Unimed (UK) Ltd), Micro-Cirrus (Intersurgical Ltd), Pulmo-Neb (DeVilbiss Health Care UK Ltd), Side-Stream (Medic-Aid Ltd), Micro-Neb III (Lifecare Ltd), RespirGard (Marquest Medical Products Inc), Aeromist, Venticaire (S and W Vickers Ltd), Up-Draft II, Ava-Neb, Up-Draft (Hudson Respiratory Care Inc), Bennett/Twin, Raindrop (Puritan-Bennett Corporation). The units were operated with a high flow compressor (Maxi III, Medix Ltd) at 101/min. Performance was assessed by measuring the fraction of the initial mass of drug released as an aerosol and nebulisation time for initial drug volume of 2-6mls, and the mass median diameter and mass fraction of the aerosol in particles < 5.17 microns diameter. The Side-Stream nebuliser gave the best performance, although incorporation of a filter to trap exhaled antibiotic may prove difficult. The Micro-Cirrus generated a particularly fine aerosol. The Raindrop nebuliser performed well, while the Up-Draft II nebulised efficiently but was associated with extended nebulisation times which may limit its utility.

Measurement and prediction of drug solvent losses by evaporation from a jet nebuliser

The evaporative losses of solvent (water) from a commercial jet nebuliser (Unineb, Unimed Ltd) were estimated and found to be approximately 14% of total nebuliser output. A simple analysis of the nebulisation process then enabled the amount of the drug colomycin released as potentially respirable aerosol to be predicted from measurements of the total nebulisation time. Good agreement was found between the predicted (P) and measured (M) values such that P = 1.07 M (r2 = 0.98). The analysis also indicates that the proportion of the drug released as aerosol will increase as nebulisation proceeds provided nebuliser output is continuous and that for a given initial mass of drug a larger amount of drug is released as an aerosol if the volume of administration used in the nebuliser is as large as possible.