Triple Negative Breast Cancer Dataset Research Articles

Integrated multi-cohorts for characterizing the immunogenomic landscape and predicting drug response in triple-negative breast cancer.

Triple-negative breast cancer (TNBC) is the breast cancer subtype with the highest fatality rate, and it seriously threatens women's health. Recent studies found that the level of immune cell infiltration in TNBC was associated with tumor progression and prognosis. However, due to practical constraints, most of these TNBC immune infiltration studies only used a small number of patient samples and a few immune cell types. Therefore, it is necessary to integrate more TNBC patient samples and immune cell types to comprehensively study immune infiltration in TNBC to contribute to the prognosis and treatment of patients. In this study, 12 TNBC datasets were integrated and an extensive collection of 182 gene sets with immune-related signatures were included to comprehensively investigate tumor immune microenvironment of TNBC. A single sample gene set enrichment analysis was performed to calculate the infiltration score of each immune-related signature in each patient, and an immune-related risk scoring model for TNBC was constructed to accurately assess patient prognosis. Significant differences were found in immunogenomic landscape between different immune risk subtypes. In addition, the immunotherapy response and chemotherapy drug sensitivity of patients with different immune risk subtypes were also analyzed. The results showed that there were significant differences in these characteristics. Finally, a prediction model for immune risk subtypes of TNBC patients was constructed to accurately predict patients with unknown subtypes. Based on the aforementioned findings, we believed that the immune-related risk score constructed in this study can assist in providing personalized medicine to TNBC patients.

DenVar: density-based variation analysis of multiplex imaging data.

Multiplex imaging platforms have become popular for studying complex single-cell biology in the tumor microenvironment (TME) of cancer subjects. Studying the intensity of the proteins that regulate important cell-functions becomes extremely crucial for subject-specific assessment of risks. The conventional approach requires selection of two thresholds, one to define the cells of the TME as positive or negative for a particular protein, and the other to classify the subjects based on the proportion of the positive cells. We present a threshold-free approach in which distance between a pair of subjects is computed based on the probability density of the protein in their TMEs. The distance matrix can either be used to classify the subjects into meaningful groups or can directly be used in a kernel machine regression framework for testing association with clinical outcomes. The method gets rid of the subjectivity bias of the thresholding-based approach, enabling easier but interpretable analysis. We analyze a lung cancer dataset, finding the difference in the density of protein HLA-DR to be significantly associated with the overall survival and a triple-negative breast cancer dataset, analyzing the effects of multiple proteins on survival and recurrence. The reliability of our method is demonstrated through extensive simulation studies. The associated R package can be found here, https://github.com/sealx017/DenVar. Supplementary data are available at Bioinformatics Advances online.

Mutational Analysis of Triple-Negative Breast Cancer Using Targeted Kinome Sequencing.

PurposeTriple-negative breast cancer (TNBC) does not have defined therapeutic targets and is currently treated with chemotherapy only. Kinase dysregulation triggers cancer cell proliferation and metastasis and is a crucial therapeutic target for cancer. In this study, targeted kinome sequencing of TNBC tumors was performed to assess the association between kinome gene alterations and disease outcomes in TNBC.MethodsA kinome gene panel consisting of 612 genes was used for the targeted sequencing of 166 TNBC samples and matched normal tissues. Analyses of the significantly mutated genes were performed. Genomic differences between Asian and non-Asian patients with TNBC were evaluated using two Asian TNBC datasets (from Seoul National University Hospital [SNUH] and Fudan University Shanghai Cancer Center [FUSCC]) and three non-Asian TNBC datasets (The Cancer Genome Atlas [TCGA], METABRIC, and Gustave Roussy). The prognostic value of kinome gene mutations was evaluated using tumor mutational burden (TMB) and oncogenic pathway analyses. Mutational profiles from the TCGA were used for validation.ResultsThe significantly mutated genes included TP53 (60% of patients), PIK3CA (21%), BRCA2 (8%), and ATM (8%). Compared with data from non-Asian public databases, the mutation rates of PIK3CA p.H1047R/Q were significantly higher in the SNUH cohort (p = 0.003, 0.048, and 0.032, respectively). This was verified using the FUSCC dataset (p = 0.003, 0.078, and 0.05, respectively). The TMB-high group showed a trend toward longer progression-free survival in our cohort and the TCGA TNBC cohort (p = 0.041 and 0.195, respectively). Kinome gene alterations in the Wnt pathway in patients with TNBC were associated with poor survival in both datasets (p = 0.002 and 0.003, respectively).ConclusionComprehensive analyses of kinome gene alterations in TNBC revealed genomic alterations that offer therapeutic targets and should help identify high-risk patients more precisely in future studies.

Gene Dosage Analysis on the Single-Cell Transcriptomes Linking Cotranslational Protein Targeting to Metastatic Triple-Negative Breast Cancer.

Many recent efforts have been put into the association between expression heterogeneity and different cell types and states using single-cell RNA transcriptome analysis. There is only limited understanding of gene dosage effects for the genetic heterogeneity at the single-cell level. By focusing on concordant copy number variation (CNV) and expression, we presented a computational framework to explore dosage effect for aggressive metastatic triple-negative breast cancer (TNBC) at the single-cell level. In practice, we collected CNV and single-cell expression data from the same patients with independent technologies. By focusing on 47,198 consistent copy number gains (CNG) and gene up-regulation from 1145 single cells, ribosome proteins with important roles in protein targeting were enriched. Independent validation in another metastatic TNBC dataset further prioritized signal recognition particle-dependent protein targeting as the top functional module. More interesting, the increased ribosome gene copies in TNBC may associate with their enhanced stemness and metastatic potential. Indeed, the prioritization of a well-upregulated functional module confirmed by high copy numbers at the single-cell level and contributing to patient survival may indicate the possibility of targeted therapy based on ribosome proteins for TNBC.

Accurate Segmentation of Nuclear Regions with Multi-Organ Histopathology Images Using Artificial Intelligence for Cancer Diagnosis in Personalized Medicine.

Accurate nuclear segmentation in histopathology images plays a key role in digital pathology. It is considered a prerequisite for the determination of cell phenotype, nuclear morphometrics, cell classification, and the grading and prognosis of cancer. However, it is a very challenging task because of the different types of nuclei, large intraclass variations, and diverse cell morphologies. Consequently, the manual inspection of such images under high-resolution microscopes is tedious and time-consuming. Alternatively, artificial intelligence (AI)-based automated techniques, which are fast and robust, and require less human effort, can be used. Recently, several AI-based nuclear segmentation techniques have been proposed. They have shown a significant performance improvement for this task, but there is room for further improvement. Thus, we propose an AI-based nuclear segmentation technique in which we adopt a new nuclear segmentation network empowered by residual skip connections to address this issue. Experiments were performed on two publicly available datasets: (1) The Cancer Genome Atlas (TCGA), and (2) Triple-Negative Breast Cancer (TNBC). The results show that our proposed technique achieves an aggregated Jaccard index (AJI) of 0.6794, Dice coefficient of 0.8084, and F1-measure of 0.8547 on TCGA dataset, and an AJI of 0.7332, Dice coefficient of 0.8441, precision of 0.8352, recall of 0.8306, and F1-measure of 0.8329 on the TNBC dataset. These values are higher than those of the state-of-the-art methods.

Prune-1 drives polarization of tumor-associated macrophages (TAMs) within the lung metastatic niche in triple-negative breast cancer

SummaryM2-tumor-associated macrophages (M2-TAMs) in the tumor microenvironment represent a prognostic indicator for poor outcome in triple-negative breast cancer (TNBC).Here we show that Prune-1 overexpression in human TNBC patients has positive correlation to lung metastasis and infiltrating M2-TAMs. Thus, we demonstrate that Prune-1 promotes lung metastasis in a genetically engineered mouse model of metastatic TNBC augmenting M2-polarization of TAMs within the tumor microenvironment. Thus, this occurs through TGF-β enhancement, IL-17F secretion, and extracellular vesicle protein content modulation.We also find murine inactivating gene variants in human TNBC patient cohorts that are involved in activation of the innate immune response, cell adhesion, apoptotic pathways, and DNA repair. Altogether, we indicate that the overexpression of Prune-1, IL-10, COL4A1, ILR1, and PDGFB, together with inactivating mutations of PDE9A, CD244, Sirpb1b, SV140, Iqca1, and PIP5K1B genes, might represent a route of metastatic lung dissemination that need future prognostic validations.

Gravitational search algorithm: a comprehensive analysis of recent variants

Gravitational search algorithm is a nature-inspired algorithm based on the mathematical modelling of the Newton’s law of gravity and motion. In a decade, researchers have presented many variants of gravitational search algorithm by modifying its parameters to efficiently solve complex optimization problems. This paper conducts a comparative analysis among ten variants of gravitational search algorithm which modify three parameters, namely Kbest, velocity, and position. Experiments are conducted on two sets of benchmark categories, namely standard functions and CEC2015 functions, including problems belonging to different categories such as unimodal, multimodal, and unconstrained optimization functions. The performance comparison is evaluated and statistically validated in terms of mean fitness value and convergence graph. In experiments, IGSA has achieved better precision with balanced trade-off between exploration and exploitation. Moreover, triple negative breast cancer dataset has been considered to analysis the performance of GSA variants for the nuclei segmentation. The variants performance has been analysed in terms of both qualitative and quantitive with aggregated Jaccard index as performance measure. Experiments affirm that IGSA-based method has outperformed other methods.

Identification of SCARA5 Gene as a Potential Immune-Related Biomarker for Triple-Negative Breast Cancer by Integrated Analysis.

Triple-negative breast cancer (TNBC) refers to breast cancer without significant expression of estrogen receptor (ER), progesterone receptor (PR), or human epidermal growth factor receptor 2. We sought to identify the hub genes and find the potential progression mechanism of TNBC as well as immunotherapeutic targets. First, we screened the overlapped hub genes of Immune and Stromal, and the tumor mutation burden gene sets, as well as the The Cancer Genome Atlas (TCGA)-TNBC data set. Among these hub genes, we performed gene set enrichment analysis (GSEA) and gene set variation analysis analyses to recognize and evaluate the hub genes. Moreover, immune cell infiltration was evaluated by the CIBERSORT algorithm and single-sample GSEA. In addition, the expression and methylation of scavenger receptor class A member 5 (SCARA5) in TNBC were verified by quantitative PCR and methylation-specific PCR. Also, MTT and transwell assays were used to assess the biological function of SCARA5 in TNBC. SCARA5 and CMA1 were listed, and they mainly participated in cancer-related signaling pathways and immune-related signaling pathways. Interestingly, SCARA5 was closely associated with tumor purity and immune cell infiltration. Moreover, we found that SCARA5 was significantly downregulated and hypermethylation was in the promoter of SCARA5 in TNBC tissues. Our study showed the role of SCARA5 in proliferation and migration of TNBC, and suggested that SCARA5 can potentially serve as a candidate immunotherapy target for TNBC.

A Six-Epithelial-Mesenchymal Transition Gene Signature May Predict Metastasis of Triple-Negative Breast Cancer.

PurposePathological complete response (pCR) to neoadjuvant chemotherapy (NACT) is associated with favourable outcomes of patients with triple-negative breast cancer (TNBC). However, a proportion of TNBC patients with the residual disease do not relapse and achieve long-term survival. The aim of this study was to identify biomarkers that predict clinical outcomes in these patients.Patients and MethodsA retrospective series of 10 TNBC patients who displayed non-pCR to NACT were included in the discovery cohort. Total RNA from pre-NACT core biopsies and paired surgical specimens were subjected to the Affymetrix Human Transcriptome Array. Gene set enrichment analysis (GSEA) was used to identify signal pathways and gene signatures associated with metastasis. The Cox proportional hazard model and Kaplan–Meier survival curves were employed to assess the prognostic value of the identified signature in two independent TNBC datasets included in Gene Expression Omnibus (GEO).ResultsThe epithelial–mesenchymal transition (EMT) pathway was markedly more enriched in pre- (NES = 1.92; p.adjust = 0.019) and post-NACT samples (NES = 2.02; p.adjust = 0.010) from patients who developed metastasis after NACT. A subset of 6 EMT genes including LUM, SFRP4, COL6A3, MMP2, CXCL12, and HTRA1 were expressed constantly at higher levels in samples from patients who progressed to metastatic disease. The potential of the 6-EMT gene signature to predict TNBC metastasis after NACT was validated with a GEO dataset (HR=0.36, p=0.0008, 95% CI: 0.200–0.658). Moreover, the signature appeared of predictive value in another GEO dataset of TNBC patients who received surgery followed by adjuvant chemotherapy (HR = 0.46, 95% CI: 0.225–0.937).ConclusionExpression analysis of the 6-EMT gene signature at diagnosis may be of predictive value for metastasis in TNCB patients who did not achieve pCR to NACT and for patients treated with surgery in combination with adjuvant therapy.

WNT/β–catenin pathway activation correlates with the increase of tumor-associated macrophages in triple negative breast cancer (TNBC).

e12564 Background: Triple negative breast cancer (TNBC) occurs in about 20% of all breast carcinomas. Because only a fraction of TNBCs responding to immune checkpoint blockade show a pre-existing T cell-inflamed tumor microenvironment (TME), it is critical to understand the mechanisms of T-cell exclusion. Tumor-cell intrinsic activation of the WNT/β–catenin pathway, overexpressed in 30% of human breast cancers, is linked to a T-cell excluded TME. In β–cateninhigh TNBC, however, the quality of the myeloid compartment has not been evaluated. Methods: A total of seventy-five, early-stage, untreated, TNBC patients was assessed (patient cohorts approved by IRB). β–catenin expression was detected by IHC and scored as high, intermediate, and low. The presence of T cells, tumor-associated macrophages (TAMs) and LAMP-expressing dendritic cells (LAMP+ DCs) was assessed by IHC using aCD3, aCD68, aCD163, and aLAMP, respectively. Public TNBC datasets TCGA (N = 157) and METABRIC (N = 319) were interrogated for correlations between β–catenin- and immune-associated genes. Results: Three patient groups (N = 25/group) were identified according to the negative, medium and high intracellular expression of β–catenin. As opposed to β–cateninlow TNBC, the β–cateninhigh group displayed significantly lower CD3+ T cells (median 5% ±7.37 SD vs median 30% ± 18.28 SD, p < 0.0001) and LAMP+ DCs (median 1% ± 2.515 SD vs median 10% ± 7.038 SD, p < 0.0001). The β–cateninlow group was enriched in lymphocyte-predominant TNBC. For the first time, we show that the immune-suppressive, CD68+CD163+ TAMs were strongly accumulated in the β–cateninhigh group (median 20% ± 12.20 SD vs median 5% ± 6.831 SD, p < 0.0001). The interrogation of the public TNBC datasets TCGA and METABRIC confirmed that – after patient statification according to the expression level of a WNT/β–catenin gene-signature (i.e. MMP7, SFRP1, WNT10A, WNT16, WNT9B) – multiple TAM-associated genes – identified by our group in a single-cell RNAseq dataset – were strongly upregulated in WNT/β–cateninhigh signature, highlighting the role of the WNT/β–catenin signaling pathway not only in T-cell exclusion but also in selective TAM accumulation. Conclusions: Immune-suppressive TAMs are accumulated in β–cateninhigh, T-cell excluded TNBCs emphasizing the importance of tumor-intrinsic factors in shaping the quality of the immune infiltrate.

Abstract P2-07-07: Exploring novel therapeutic target molecules for metaplastic breast carcinoma using comprehensive genome-wide gene expression analyses

Abstract Breast cancer subtypes have been classified based on the gene expression profiles, and the specifically effective treatments for each subtype have been developed and applied in clinical situations. However, there is no effective treatment for metaplastic breast carcinomas, which are pathological tissue types with poor prognosis. Metaplastic breast carcinomas are almost resistant to existing chemotherapy because it has not been identified the specific therapeutic target molecules. Recently, Lehmann et al. reported that breast cancers with chondroid metaplasia are categorized in Mesenchymal (M) type, which is one subtype out of six triple negative breast cancer (TNBC) subtypes based on drug susceptibility. From the knowledge on this report, we hypothesized that integrative analyses using the gene expression profile data sets obtained from various types of breast cancers with or without cartilage formation may enable us to identify the therapeutic target genes for metaplastic breast carcinomas, especially for breast cancers with chondroid metaplasia. Bioinformatics analyses were performed using the gene expression profile data sets of TNBC, which are registered in Gene Expression Omnibus (GEO). When compared with metaplastic breast carcinomas (28 data sets), 200 genes were specifically fluctuated over 2-fold in breast cancer with chondroid metaplasia (8 data sets). On the other hand, M type (11 data sets) specific 578 genes were identified over 56 TNBC data sets. In addition, 57 genes were overlapped between chondoroid metaplastic 200 genes and M type 578 genes. Therefore, these 57 overlapped genes might be considered as possible poor prognostic factors in breast cancer with chondroid metaplasia. As a result of Gene Ontology analysis on these 57 genes, skeletal system development and extracellular matrix structural constituent were significantly enriched in Biological process and Molecular function, respectively. Also, SCRG1, SOX8 and SOX10, which are genes related to cartilage differentiation, were included. Next, we focused 14 genes because these genes could serve as therapeutic targets. Then, we validated the expression levels of these 14 genes using tissues obtained from surgically resected human breast cancer with chondroid metaplasia dividing into three parts, the invasive ductal carcinoma, the chondroid metaplasia and normal mammary gland. As a result, SCRG1, CD86, HCLS1, EPSTI1, LYZ, and SLA were significantly increased in chondroid metaplasia part, suggesting that these validated genes were identified as possible therapeutic targets for breast cancer with chondroid metaplasia. Further study will help us to investigate a novel and effective therapeutic strategy for metaplastic breast carcinomas. Citation Format: Yamashita M, Kamei Y, Kusakabe E, Yamasawa H, Aoki R, Taguchi K, Nishiyama K, Murakami A, Sawada Y, Takada Y, Imai Y. Exploring novel therapeutic target molecules for metaplastic breast carcinoma using comprehensive genome-wide gene expression analyses [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 P2-07-07.

Therapeutic Rationale to Target Highly Expressed CDK7 Conferring Poor Outcomes in Triple-Negative Breast Cancer.

Triple-negative breast cancer (TNBC) patients commonly exhibit poor prognosis and high relapse after treatment, but there remains a lack of biomarkers and effective targeted therapies for this disease. Here, we report evidence highlighting the cell-cycle-related kinase CDK7 as a driver and candidate therapeutic target in TNBC. Using publicly available transcriptomic data from a collated set of TNBC patients (n = 383) and the METABRIC TNBC dataset (n = 217), we found CDK7 mRNA levels to be correlated with patient prognosis. High CDK7 protein expression was associated with poor prognosis within the RATHER TNBC cohort (n = 109) and the METABRIC TNBC cohort (n = 203). The highly specific CDK7 kinase inhibitors, BS-181 and THZ1, each downregulated CDK7-mediated phosphorylation of RNA polymerase II, indicative of transcriptional inhibition, with THZ1 exhibiting 500-fold greater potency than BS-181. Mechanistic investigations revealed that the survival of MDA-MB-231 TNBC cells relied heavily on the BCL-2/BCL-XL signaling axes in cells. Accordingly, we found that combining the BCL-2/BCL-XL inhibitors ABT-263/ABT199 with the CDK7 inhibitor THZ1 synergized in producing growth inhibition and apoptosis of human TNBC cells. Collectively, our results highlight elevated CDK7 expression as a candidate biomarker of poor prognosis in TNBC, and they offer a preclinical proof of concept for combining CDK7 and BCL-2/BCL-XL inhibitors as a mechanism-based therapeutic strategy to improve TNBC treatment. Cancer Res; 77(14); 3834-45. ©2017 AACR.

CCL5 expression and tumor infiltrating immune cells in triple negative breast cancer.

11553 Background: CCL5 is a chemo-attractant of regulatory T cells, promoting tumor immune avoidance and related to a poor outcome in several malignancies; however, in triple negative breast cancer (TNBC), it is related to a better outcome. Our aim was to evaluate the correlation between CCL5 and tumor infiltrating immune cells and their prognosis value. Methods: We evaluated 72 TNBC patients with residual disease after neoadjuvant chemotherapy with matched data of tumor infiltrating lymphocytes (TILs) count and CCL5 expression (profiled with NanoString). CCL5 expression levels were log2 transformed and median centered. Correlation between TILs (log2 transformed) and CCL5 was evaluated with the Spearman’s rank test. Cox PH model was used to investigate the effect of CCL5 (median as cutoff) and TILs ( < 20% and ≥20%) in distant-recurrence free survival. We used the CIBERSORT platform to evaluate the immune cells composition according to the expression of CCL5 (higher vs. lower or equal than median) in 3 independent TNBC datasets (GSE25066, GSE58812 and GSE76124). Results: There was a significant correlation between TILs and residual tumor size (P = 0.017) and CCL5 (ρ = 0.347, P = 0.003). In univariate analysis, TILs (HR = 0.276, 95%IC: 0.128-0.593; P = 0.001) and CCL5 (HR = 0.401; 95%CI: 0.206-0.781; P = 0.007) were both associated with outcome. In a multivariate analysis with CCL5 expression and TILs count, TILs was the only significant marker with a P = 0.008 (HR = 0.336; 95%CI: 0.150-0.753), in contrast to CCL5 (HR = 0.573; 95%CI: 0.285-1.154; P = 0.124). CIBERSORT analysis suggested that high CCL5 expression is associated with recruitment of CD8 cells (13% v 6%, P < 0.001; 6% v 1%, P < 0.001 and 12% v 8%, P = 0.003), activated CD4 memory T cells (4% vs. 2%, P < 0.001; 5% vs. 0%, P < 0.001; 3% vs. 0%, P < 0.001) and Macrophages M1 (9% vs. 7%, P = 0.022; 13% vs. 8%, P = 0.005; 11% vs. 5%, P < 0.001) in GSE25066, GSE58812 and GSE76124 datasets, respectively. Conclusions: TILs was the stronger and more significant prognostic immunological marker, even than CCL5 expression. High CCL5 expression was associated with enrichment of CD8 cells, activated CD4 memory T cells and Macrophages M1. Role of these cells in TNBC should be explored more deeply.

Abstract PD1-03: Multivariate analysis of subtype and gene expression signatures predictive of pathologic complete response (pCR) in triple-negative breast cancer (TNBC): CALGB 40603 (Alliance)

Abstract Support: U10CA180821, U10CA180882 Background: The addition of either carboplatin (Cb) or bevacizumab (Bev) to standard neoadjuvant chemotherapy (NACT) increases pCR rates in TNBC overall and in the dominant subset of basal-like cancers (Sikov et al, JCO 2015; Sikov et al, SABCS 2014). Multigene expression signatures more accurately reflect tumor biology for response prediction and prognosis than individual gene expression. We evaluated the ability of multivariate analysis of gene expression signatures to create predictive models for achievement of pCR in TNBC. Methods: RNA sequencing was successful on 389 pretreatment samples from patients with available pCR data, and used to assign PAM50 subtype and calculate gene signatures scores for 489 published expression signatures. Elastic net, a penalized regression model for high dimensional variable selection, was used to select features associated with pCR in all TNBC and in the basal-like subset. Models were derived in a training set (2/3 of samples) and validated in a separate test set (1/3). A separate model was derived using 196 TNBC samples from patients treated only on the standard NACT +/- Cb arms for application to external TNBC neoadjuvant data sets not treated with Bev. Results: Consistent with our prior partial data set, 343 (88%) of the cancers were classified basal-like, in whom the in breast pCR rate was 54%; the remainder were classified normal-like (n=32) or HER2-enriched (n=14) with a non-basal pCR rate of 56%. Elastic Net analysis in all TNBC generated a model of 23 signatures and treatment assignment with 68% sensitivity and 64% specificity. The area under the curve was 0.64 (p-value=0.0019). Nineteen modules, including immune cell signatures (Th1, NK, IgG), immunoglobulin variable region expression, addition of Cb and Bev and expression of genes at regions 15q25, 17p11.2-13.3, and 8p22 were positively associated with response. The latter two regions are associated with aggressive breast cancer, and while not part of the 17p13 signature, this region contains TP53, a gene important in TNBC. Six modules were associated with resistance, including luminal progenitor, TGFB, NOTCH, FOS/JUN, 8p amplicon, and eosinophil signatures. When limited to basal-like samples, a model including 32 modules and addition of Cb and Bev was generated, with 62.3% sensitivity and 59.1% specificity. Seventeen features were selected in both models. Omitting Bev-treated patients, a model using just the gene expression signatures was developed. The predictive value of this model will be assessed using an external cohort of TNBC patients treated with neoadjuvant docetaxel and Cb (NCT01560663) and results presented. Conclusions: Multivariate analysis of gene expression signatures derived from pretreatment samples enabled the construction of models to predict achievement of pCR in TNBC. These models performed well on our test set, and will be assessed for their predictive ability in other TNBC data sets. If validated by future analyses, this could help us identify patients likely to achieve pCR with standard NACT and may benefit from the addition of agents such as Cb or Bev. ClinicalTrials.gov Identifier: NCT00861705. Citation Format: Hoadley KA, Hyslop T, Fan C, Berry DA, Hahn O, Tolaney SM, Sikov WM, Perou CM, Carey LA. Multivariate analysis of subtype and gene expression signatures predictive of pathologic complete response (pCR) in triple-negative breast cancer (TNBC): CALGB 40603 (Alliance) [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 PD1-03.

Abstract P3-07-25: Improved clinical outcomes on enzalutamide observed in patients with PREDICT AR+ triple-negative breast cancer: prognosis or prediction?

Abstract Background: A novel genomic signature that identifies androgen receptor (AR)-driven disease (PREDICT AR) is being developed for its ability to select patients (pts) who may benefit from enzalutamide (ENZA), a potent AR inhibitor.1 In a phase 2 study of triple-negative breast cancer (TNBC) pts treated with ENZA (NCT01889238) ≈ half of pts had PREDICT AR+ TNBC, with improved clinical outcomes in this sub-population.2 Because pts with hormonally driven (estrogen-receptor/progesterone-receptor[ER/PR]+) BC are thought to have a better prognosis, we investigated whether the outcomes observed in pts with PREDICT AR+ TNBC could be explained by prognosis alone. Methods: PREDICT AR was assessed on advanced TNBC from 118 ENZA-treated pts.2 Because the duration of treatment (Tx) typically decreases with each subsequent Tx, exploratory analyses included median progression-free survival (mPFS) on ENZA vs duration of Tx prior to ENZA (used as a PFS surrogate) in pts who had received ≤1 prior Tx (data cutoff 24Mar15). Baseline characteristics by PREDICT AR status were also evaluated. An independent non-ENZA-Tx TNBC data set (after neo/adjuvant Tx)3 was probed for PREDICT AR status and clinical outcomes. Results: ENZA-Tx pts with PREDICT AR+ TNBC (n=56; 47%) were older (66 years vs 52 years), had a longer disease-free interval (DFI) (37 vs 20 months), more bone involvement (57% vs 24%), were more likely to have had prior ER/PR+ BC (30% vs 8%), and a longer median duration on the first-line Tx (29 vs 19 weeks[wks]) vs pts with PREDICT AR- TNBC (n=62;53%). Table 1. mPFS from ENZA-Tx pts and duration of prior Tx by PREDICT AR status and line of Tx.PREDICT AR StatusTx prior to ENZAmPFS on ENZA (wks)Median duration of prior Tx (wks)PREDICT AR+ (n=26)No prior Tx (n=10)32-- 1 prior Tx (n=16)4034PREDICT AR- (n=36)No prior Tx (n=11)8-- 1 prior Tx (n=25)1317 The prevalence of PREDICT AR+ primary TNBC from the non-ENZA-Tx dataset was 51% (n=182). DFI after neo/adjuvant therapy was not statistically different (p=0.605) between pts with PREDICT AR+ vs PREDICT AR- TNBC, and pathologic complete response rates were 23% vs 38% in PREDICT AR+ vs PREDICT AR- TNBC, respectively. Conclusion: mPFS in pts with PREDICT AR+ TNBC who received first-line ENZA was longer than first-line Tx duration in those who received prior Tx (32 vs 29 weeks). In pts with PREDICT AR+ TNBC who received second-line ENZA, mPFS was longer than the median duration of their first-line Tx (40 vs 34 weeks). This trend was not observed in pts with PREDICT AR- TNBC. Key demographic differences did not explain these outcomes, suggesting benefit from ENZA. Duration of first-line non-ENZA-Tx was shorter in pts with PREDICT AR- TNBC vs pts with PREDICT AR+ TNBC who received second-line ENZA (17 vs 34 weeks), suggesting a better prognosis in pts with PREDICT AR+ advanced TNBC; however, PREDICT AR status was not associated with DFI following neo/adjuvant treatment for primary TNBC. PREDICT AR may identify a hormonally driven subset of pts with advanced TNBC who have a better prognosis and who may benefit from ENZA. 1. Parker J et al. J Clin Oncol 2015;33(suppl): abstr 1083. 2. Traina TA et al. J Clin Oncol 2015;33(suppl): abstr 1003. 3. Hatzis C et al. JAMA 2011;305:1873-1881. Citation Format: Miller K, Krop I, Schwartzberg L, Abramson V, Garcia-Estevez L, Eakle J, Nanda R, Yardley D, Ademuyiwa F, Chan S, Crown J, Danso M, Gelmon K, Ma L, Martinez-Janez N, Gradishar W, Steinberg J, Tudor IC, Uppal H, Paton VE, Parker J, Hudis CA, Traina TA. Improved clinical outcomes on enzalutamide observed in patients with PREDICT AR+ triple-negative breast cancer: prognosis or prediction?. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-07-25.

Abstract B48: Clues to the causes of the abundance of triple-negative breast cancer in women of African descent

Abstract Background: Triple negative breast cancer (TNBC) is over-represented in indigenous African women and in women of African descent. The majority of women diagnosed with breast cancer in Nigeria, Uganda and Kenya have TNBC. Studies of TNBC in the US consistently report the percentage of TNBC to be 10-24% greater in African-Americans (AA) than in Caucasians (C). The abundance of TNBCs in both African and African-American women has yet to be explained and the incidence rates suggest that there may be a genetic predisposition to this particularly aggressive form of breast cancer. Methods: Epithelial cell lines were established from normal, healthy breast tissue donated to the Susan G. Komen for the Cure® Tissue Bank at the IU Simon Cancer Center. Six cell lines from AA donors were matched on the basis of age, menopausal status and Gail-risk score to six cell lines from C donors. RNA was extracted from the cells and gene expression assayed using the Illumina HumanWG6 v3 chip. Significant differences in gene expression were identified using PADE (PAirwise Differential Expression; Expression Analysis, Durham, NC). Results: 1552 genes were differentially expressed (p≤ 0.05). The largest number of genes over-expressed in AAs mapped to the long arm of chromosome 1. Copy number gains have been observed in this region in multiple TNBC data sets. Many of these genes are in the 1q21 region, a region frequently rearranged in epithelial tumors as well as soft tissue and bone. Analysis of differentially expressed genes using MetaCore from Thomson Reuters revealed that the over-expressed genes are responsible for skin development, cell adhesion/cell junctions, and functions of the innate immune system. Pattern recognition is in the vanguard of the innate immune system. The pattern recognition pathways mediated by RAGE (receptor for advanced glycation end products) and the Toll-like receptors are activated in the normal AA epithelium. The expression of genes in these pathways, specifically S100A12, S100A8, S110A9, and TLR5, is known to increase in the presence of Salmonella and Shigella. MDA5 recognizes rotavirus infection. LYST and DUSP4 are active in the control of Leishmania infection. LCN2 disrupts iron acquisition, which protects against a number of pathogens including Salmonella, Klebsiella and mycobacteria. Conclusions: The results of the above gene expression analysis are best understood in light of our recently published findings using these same cells (Sauder et al, BMC Cell Biology, 2014). Epithelial cells grown from mammary explant culture are basal-like and phenotypically plastic. They can assume a wide range of cell types depending on the cues provided by the microenvironment. When grown in basement membrane matrix or on human dermal fibroblasts, these cells produce a stratified squamous epithelium. Although the cells used for the gene expression analysis were grown in plastic tissue culture flasks and grew as a monotonous monolayer, the data suggest that they are, to some extent, transcriptionally “squamous”. For example, they express involucrin and the genes that code for the corneodesmosome proteins: corneodesmin, desmoglein and desmocollin. The skin of humans living in sub-Saharan Africa has evolved to provide optimal barrier protection. It specifically protects against water loss and, pathogenic microorganisms and parasites. Mammary glands are ectodermal appendages of the epithelium and they may share similar immune defense mechanisms. The increased expression of genes involved in innate immunity in AA breast epithelial cells is of significant interest. We hypothesize that a subclinical, chronic inflammatory state is incited within basal breast cells in response to a pathogen that is endemic to Africa and overrepresented in low income communities in the US. This persistent and unresolved innate immunity predates the development of TNBC and can thus participate in its initiation and promotion. Citation Format: Candice A.M. Sauder, Jillian E. Koziel, MiRan Choi, Brittney-Shea Herbert, Susan E. Clare. Clues to the causes of the abundance of triple-negative breast cancer in women of African descent. [abstract]. In: Proceedings of the Seventh AACR Conference on The Science of Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; Nov 9-12, 2014; San Antonio, TX. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2015;24(10 Suppl):Abstract nr B48.

Consistent Principal Components from Triple Negative Breast Cancer Expression Profiles Yields Agent Specific Predictors of Chemotherapy Response.

Abstract Genome scale expression profiling is likely to yield improved methods for optimizing therapy for individual breast cancer patients. Microarray based predictors of response to chemotherapy have been described for ER-positive tumors, but supervised methods haven't yielded highly accurate predictors of response in triple-negative breast cancer (TNBC). The problem of deriving predictive or prognostic classifiers is not trivial, and care must be taken to avoid pitfalls such as data set-specific noise and overfitting.To overcome these problems, we developed a two-stage unsupervised method to derive consistent metagenes from multiple data sets. In the first stage, we used principal component analysis to identify aspects of gene expression profiles that are consistently observed in multiple starting data sets. We then identified a subset of genes that best captured the information derived from these components. In the second stage, we factorized this subset of genes and derived metagenes applicable to any external data set. With this method, overfitting is eliminated by virtue of the unsupervised approach, and data set-specific noise is reduced by selecting only components presented in multiple data sets.We first applied our method to four TNBC data sets not obtained in a neoadjuvant setting and derived four metagenes. We then applied these metagenes to another four independent data sets annotated with response to various neoadjuvant therapies in order to assess their predictive performances. Two of the metagenes were associated with response: one predicted pathological complete response (pCR) in the two TFAC cohorts but not in the other two cohorts (Fig. a); the other was associated with pCR to TFAC, residual disease (RD) to epirubicin single agent, but not with response to FEC (Fig. b). Notably, these metagenes derived in an unsupervised manner were more strongly predictive than were classifiers derived using conventional supervised methods.As conclusion, we report and validate an unsupervised strategy as a robust and powerful alternative to supervised methods for deriving predictors for response to chemotherapy from microarray data. If confirmed in additional independent cohorts, these metagenes may lead to clinical tests to optimize therapy for individual TNBC patients.Figure 1. ROC curves indicate the association between two metagenes and agent-specific pathological complete response to neoadjuvant therapy. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 3163.