Abstract P2-08-12: Development and validation of prognostic gene signatures for basal-like breast cancer and high grade serous ovarian cancer

Background: Basal-like breast cancer (BLBC) have poor prognosis. Molecular similarities have been reported between BLBC and high grade serous ovarian cancer (HGSOC). To date, there have been no prognostic biomarkers specifically developed for BLBC or HGSOC that can provide risk stratification and inform treatment selections. In this study, we utilized RNA-seq data available from The Cancer Genome Atlas (TCGA) project to develop molecular signatures for risk stratification in BLBC, and further validated these signatures in HGSOC RNA-seq data from TCGA.

Methods: Raw count of RNA-seq data were downloaded from TCGA for 190 BLBC and 374 HGSOC patients. The datasets were annotated with 56963 Ensembl gene IDs. Excluding 31375 gene IDs with no greater than 10 counts in at least 90% of the samples, totally 25228 unique Ensembl gene IDs were used. Progression-free interval (PFI) is the primary study endpoint. Analyses of differentially expressed genes were performed using 3 bioconductor packages: DESEq2, edgeR and voom/limma. Signatures based on commonly identified genes among the 3 analytic methods were established using weighted linear combination of gene expression levels. Their performance was evaluated in the BLBC and HGSOC datasets using Kaplan-Meier survival analysis with log-rank tests and Cox proportional hazard regressions.

Results: Among 190 TNBC patients, 18 had recurrences within 2 years and 40 showed no recurrences for at least 5 years. These patients were used as recurrent vs. non-recurrent cases for differential expression analysis. 307 and 343 genes were differentially expressed based on adjusted p value threshold 0.05 and 0.01 in DESeq2 and edgeR analysis, respectively. voom/limma identified no genes differentially expressed based on adjusted p values, but 228 genes had unadjusted p values < 0.01 and were used in the following analysis. Taken together, 63, 58 and 21 genes were commonly identified by DESeq2/edgeR, DESeq2/limma and edgeR/limma analysis, respectively. All 3 signatures were able to significantly stratify the TNBC full dataset (n=190) using either 20-, 50- or 80-percentile as the cut-points. When evaluated in HGSOC patients using 80-percentile cut-point, both 63- and 58-gene signatures were able to significantly stratify patients into different risk groups (HR 2.16, 95% CI: 1.4-3.34, p < 0.001; HR 2.06, 95% CI: 1.36-3.11, p < 0.001). Multivariate Cox regression adjusting for age, grade and stage showed 63- and 58-gene signatures remained to be statistically significant in stratifying HGSOC patients (p = 0.0005 and 0.001, respectively).

Conclusion: Gene signatures were specifically identified to prognosticate BLBC patients based on RNA-seq data from TCGA project. Which were able to classify HGSOC patients into differential risk groups. With further validations, these signatures may provide additional prognostic tools for clinicians to better manage triple-negative breast cancer that mostly overlap with BLBC, and HGSOC patients who are difficult-to-treat currently.

Disclaimer The views expressed in this article are those of the authors and do not reflect the official policy of the department of Army/Navy/Air Force, Department of Defense, or U.S. government.

Citation Format: Hu H, Zhang Y, Liu J, Raj-Kumar P-K, Yang H, Lee M, Kovatich AJ, Shriver CD. Development and validation of prognostic gene signatures for basal-like breast cancer and high grade serous ovarian cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-08-12.

Abstract P5-06-02: Organochlorine pesticide residues in human breast tissue and their relationships with clinical and pathological characteristics of breast cancer

Agricultural pesticides are abundant environmental contaminants worldwide, prompting interest in studying their possible detrimental health effects. We examined organochlorine residues by quadrant (n = 245) in breast adipose tissues from 51 women with various stages of breast health to determine patterns of bioaccumulation within the breast and to assess relationships with patient clinical characteristics. Three organochlorine residues — 2,2-bis(p-chlorophenyl)-1,1-dichloroethylene (p,p′-DDE), hexachlorobenzene (HCB), and mirex — assayed by high resolution gas chromatography were abundant in breast tissue. p,p′-DDE (745 ± 1054 ng/g lipid) was the most prevalent residue, comprising 97.5% of the total chemical burden. Mean levels of p,p′-DDE and HCB were significantly correlated (P < 0.001) with patient age at mastectomy, and levels of p,p′-DDE were correlated (P < 0.05) with BMI. Pesticide concentrations did not differ significantly by breast quadrant and were not different in the quadrant(s) where the primary tumor was located compared to other cancer-free quadrants. In invasive cancer patients, organochlorine levels differed significantly based on clinical characteristics of the primary carcinoma, including stage, grade, ER status, and HER2 status, indicating that body burden of organochlorines may influence the development of specific subtypes of breast cancer. Potentially carcinogenic organochlorines were present at high levels within the human breast warranting further research to determine the impact of organochlorines in the etiology of breast cancer.

The opinions or assertions contained herein are the private ones of the author/speaker and are not to be construed as official or reflecting the views of the Department of Defense, the Uniformed Services University of the Health Sciences or any other agency of the U.S. Government.

Citation Format: Ellsworth RE, Kostyniak PJ, Chi L-H, Shriver CD, Costantino NS, Ellsworth DL. Organochlorine pesticide residues in human breast tissue and their relationships with clinical and pathological characteristics of breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P5-06-02.

Abstract P4-01-26: Circulating cell-free DNA in serum as a marker for the early detection of tumor recurrence in breast cancer patients

Background: Quantitative estimation of circulating cell-free DNA (cfDNA) isolated from serum by noninvasive procedures can serve as a potential biomarker for the early detection of many cancers. However, a simple, straightforward technique is unavailable to estimate the cfDNA in clinical labs. Moreover, the prognostic value of cfDNA in patients with breast cancer (BrCa) is currently under debate. The aim of this study was to develop a simple yet effective quantitative method for measuring the cfDNA in serum and to eventually investigate the relationship between cfDNA and the occurrence of recurrence in BrCa patients.

Methods: A total of 240 patient cases (n=240) were selected and are comprised of different subtypes of breast cancer patients and control individuals. We selected 21 serum samples from patients which showed recurrence after 4-7 years of disease-free survival. For the compare studies, each of the recurrent and non-recurrent serum samples was incubated with the SYBR Green I (2 μM). A standard graph was also made with known DNA concentration to calculate the amount of cfDNA in these recurrent and non-recurrent serum samples. Additionally, a comparative study was also performed with the serum of patients with non-recurrent BrCa versus healthy patients.

Results: We develop a simple fluorescent based measuring technique which can easily estimate the cfDNA in one step. SYBR Green binds to DNA, and as a result, the fluorescence of SYBR Green increases substantially. Global Wilcoxon analyses were performed to compare the cfDNA amount between non-recurrent and recurrent patients. There is a significant difference in fluorescent intensities between recurrent patients' samples versus non-recurrent patients which are directly proportional to the cfDNA levels. The amount of cfDNA is higher in recurrent patient (ratio is 1.3 up; p= 0.03; AUC=0.76) compared to similar non-recurrent patients. While we compared the fluorescence data between normal/healthy patients versus non-recurrent is turned out as non-significant (healthy to non-recurrent ratio = 1.03; p= 0.20, AUC=0.61).

Conclusion: In this current study, we developed a straightforward one-step technique to measure the amount of cfDNA in serum, which can easily translate into a clinical diagnostic tool. To the best of our knowledge, this is the first report which demonstrates serum cfDNA as an early detection marker for recurrent breast cancer patients. The relatively high level of cfDNA in the serum of recurrent breast cancer patients compared to non-recurrent breast cancer patients indicates an uncovered circulating genetic information which triggers the cancer recurrence pathway to relapse cancer in the near future.

Citation Format: Bera A, Eidelman O, Russ E, Landa A, Karaian J, Eklund M, Hu H, Pollard HB, Shriver CD, Srivastava M. Circulating cell-free DNA in serum as a marker for the early detection of tumor recurrence in breast cancer patients [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P4-01-26.

Abstract P4-09-14: Analysis of breast cancer recurrence using gene set enrichment analysis

Background: Even after successful treatment of primary breast tumors, there is a continued risk of recurrence. The risk varies between subtypes and there are ongoing efforts that aim to improve prediction of such risks for individual patients. Detection of subclinical metastases might be achieved by biomarkers in blood. In this study, we profiled protein expression in blood plasma from patients with known clinical outcome (recurrence vs no recurrence) to identify prognostic markers of breast cancer recurrence.

Methods: The subjects and specimens were made available through the Clinical Breast Care Project using IRB-approved protocols. We analyzed blood plasma samples taken at the time of diagnosis from consented patients who subsequently relapsed (33 cases) as well as those with no disease recurrence (31 controls). Based on hormone receptor and lymph node status the samples were grouped as: ER-/HER2- (17 cases/15 controls), ER+/LN+ (10/10) and ER+/LN- (6/6). We used aptamer-based SOMAscan assay platform to study the expression of 1252 proteins. We analyzed the protein expression data by using their coding genes in order to apply the Gene Set Enrichment Analysis method (GSEA v.2, Broad Institute). Pathway databases of KEGG, REACTOME, BIOCARTA and C4 collection were used. Significant gene sets were called at 5% FDR, and overlaps and low coverage gene sets (Tags <70%) were removed. Statistical analysis and clustering were done using R.

Results: Unsupervised clustering showed some difference in signal in the ER+/LN- group. Even though there was a lack of significantly differentiated proteins between the cases and controls of this group, many significant gene sets were identified. After applying the cutoff filters and removing the overlaps, there were 5 gene sets enriched with the pathway collection, involved in B-cell receptor signaling, mRNA metabolism, tight junction and SCF-KIT signaling. Similarly, 9 gene sets from the MORF compendium were differentially expressed with the C4 collection and included neighborhood genes of NME2, ACTG1, EIF3S2, AP2M1, DAP3, UBE2I, NPM1, AATF and NPM1. In contrast, neither differentially expressed proteins nor gene sets were identified from the ER+/LN+ and ER-/HER2- groups. Since the sample size of the ER+/LN- group was small, we conducted a similar analysis by randomly choosing 6 case and control samples in the other two groups respectively. There were still no differentially expressed proteins or gene sets identified above the specified cutoff parameters.

Conclusion: Using plasma protein expression data we identified underlying gene sets differentially expressed between ER+/LN- patients who had cancer recurrence and no recurrence. Many genes in these sets were already known biomarkers (e.g. PTEN, AKT1, STAT3, SET etc.). These results can be used for understanding patterns of recurrence in different cancer subtypes. Further research is needed to estimate the clinical significance of these gene products.

The views expressed in this article are those of the author and do not reflect the official policy of the Department of Army/Navy/Air Force, the Department of Defense, or U.S. Government.

Citation Format: Praveen Kumar A, Kovatich AJ, Biancotto A, Cheung F, Davidson-Moncada JK, Kvecher L, Liu J, Ru Y, Kovatich AW, Deyarmin B, Fantacone-Campbell JL, Hooke JA, Raj Kumar PK, Rui H, Hu H, Shriver CD. Analysis of breast cancer recurrence using gene set enrichment analysis [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 P4-09-14.

Abstract P4-07-01: Assessment of the hereditary component in 94 cancer predisposition genes to triple negative breast cancer

Background: In women with triple negative breast cancer (TNBC) unselected for age or family history, 8-14% and 5% of patients harbor germline mutations in BRCA1 and BRCA2, respectively. Diagnosis of TNBC <60 years of age is one of the NCCN criterion for genetic testing of BRCA1 and BRCA2. The contribution of germline mutations in other cancer predisposition genes to TNBC, is, however, not well-studied.

Methods: TNBC was classified as tumors with <1% positively staining cells for ER and PR and HER2 = 0+, 1+ or 2+/not amplified. Genomic DNA was isolated from blood samples and targeted sequencing was performed using the TruSight Cancer panel (Illumina). Pathogenic mutations were identified using VariantStudio and classified as pathogenic, uncertain significance (VUS) or benign using ClinVar.

Results: 196 female patients diagnosed with TNBC 2001-2014 had genomic DNA available. Average age at diagnosis was 52.8 years (range 34.1-83.4 years). The majority of patients were of European (66%) or African (31%) American ancestry; 26% had a family history and 13% had died of disease with an average time to death of 2.81 years. Twenty-three (12%) of women with TNBC had pathogenic mutations in breast cancer genes BRCA1 (n=14), BRCA2 (n=5), PALB2 (n=1) and CHEK2 (n=3), two women had mutations in the colon cancer genes MUTYH, one had a mutation in the ovarian cancer gene BRIP1, and an additional three women had pathogenic mutations in cancer predisposition genes FANCD2, SDHB and XPC. An additional 42 women had VUS in 20 genes, including one in BRCA1 and 5 in BRCA2.

Discussion: Although the majority of pathogenic mutations in this cohort of women with TNBC were in the BRCA1 and BRCA2 genes (10%), panel testing allowed for the detection of mutations in other breast (2%), colon (1%), ovarian (1%) and other cancer (2%) predisposition genes. Panel testing thus identifies genes other than BRCA1/2 associated with increased risk of TNBC and may incidentally identify women who would benefit from enhanced surveillance for other cancers.

The opinions or assertions contained herein are the private ones of the author/speaker and are not to be construed as official or reflecting the views of the Department of Defense, the Uniformed Services University of the Health Sciences or any other agency of the U.S. Government.

Citation Format: Ellsworth RE, Lovejoy LA, Shriver CD. Assessment of the hereditary component in 94 cancer predisposition genes to triple negative 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 P4-07-01.

Abstract P2-06-04: Use of principal component analyses to select ER-balanced subset for gene centering in PAM50 subtyping

Background: PAM (Prediction Analysis of Microarray) 50 is an established gene expression-based algorithm to classify breast tumors into basal-like, HER2-enriched, luminal A (LA), and luminal B (LB) subtypes. Clinical subtyping is mainly based on immunohistochemistry (IHC) assays of estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (Her2) and Ki67 classifying tumors into triple-negative (ER-/PR-/Her2-), Her2+ (ER-/PR-/Her2+), LA (ER+/Her2-/Ki67-), LB1 (ER+/Her2-/Ki67+) and LB2 (ER+/Her2+). These two subtyping methods do not completely match even on comparable subtypes. Nevertheless, the ER-balanced subset for gene-centering in PAM50 subtyping was selected based on clinical status. Here we explored the possibility of using principal component analyses and iterative PAM50 call to refine the selection of an ER balance subset to improve consistency between these methods focusing on LB calls which is more aggressive than LA tumors.

Methods: Normalized gene expression data was obtained from TCGA research network for 712 primary tumors which had IHC status available for ER, PR and Her2. Since Ki67 status was not available LA and LB was discriminated for ER+ cases with Her2- and Her2+ respectively. In house RNA-Seq dataset had 118 primary tumors and were drawn from the Clinical Breast Care Project where breast cancer patients were consented using an IRB-approved protocol. Tumors were selected and processed by laser microdissection. RNA was extracted from tissues using the Illustra triplePrep kit (GE Healthcare). Paired-end mRNA sequencing was performed using the Illumina HiSeq platform. Sequenced reads were processed using PERL based pipeline utilizing PRINSEQ, GSNAP and HTSeq. Principal component analysis (PCA) was done using R. Wilcoxon rank sum test was used for statistical significance (p<0.05).

Results: In both datasets, the PCA map grouping of cases does not perfectly reflect the clinical subtypes. This motivated us to select ER balance subset based on the PC1 separation and IHC subtype. The resulting PAM50 subtypes on PCA map distinguished Basal and LA as two well separated components. Using all of Basal and equal number of LA cases for ER balance subset for PAM50 resulted in increased LB call and a better consistency with IHC LB calls. Among 712 cases in TCGA LB numbers increased from 142 in initial PAM50 call to 203 in ER balanced refined PAM50 call. We noticed that there was significantly higher (p-value = 4.414e-11) MKI67 expression for the 39 cases switch from LA to LB between PAM50 calls. Similar trend was observed in our in-house dataset where majority of the IHC-LB1 cases was called as LB in PAM50. The new method increased LB call from 22 to 27 which in-turn increased consistency between molecular and clinical subtypes from 73 to 79 out of the total of 118 cases.

Conclusion: We show that an iterative PAM50 call coupled with PCA for selection of ER balance set potentially enhanced the consistency of the LB calls with clinical subtyping and that the tumors switched from LA to LB have high MKI67 expression.

The views expressed in this article are those of the author and do not reflect the official policy of the Department of Army/Navy/Air Force, the Department of Defense, or U.S. Government.

Citation Format: Raj-Kumar P-K, Liu J, Kovatich AJ, Kvecher L, Shriver CD, Hu H. Use of principal component analyses to select ER-balanced subset for gene centering in PAM50 subtyping [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 P2-06-04.

Abstract P1-03-05: Not presented

This abstract was not presented at the symposium.

Abstract P3-16-01: Using the new pan-cancer clinical data resource (TCGA-CDR) to identify breast cancer genomic correlates associating with different survival outcome endpoints

Introduction The Cancer Genome Atlas (TCGA) generated abundant high quality molecular data, however its relatively short-term patient follow-up limited its immediate clinical utility. We led a PanCanAtlas effort to systematically collate, integrate, and quality check the large body of acquired clinicopathologic data, generated 4 primary clinical outcome endpoints for each case, and created a new Pan-Cancer Clinical Data Resource (TCGA-CDR) for public use. We report here on the utility and validity of this TCGA-CDR in relating breast cancer (BC) genomic information to survival endpoints.

Methods Clinicopathologic data from all data files were integrated and processed. Overall survival (OS), disease-specific survival (DSS, an approximation), progression-free interval (PFI), and disease-free interval (DFI)were derived.Tests of the adequacy of the follow-up intervals for each endpoint were performed, and quality evaluation of these endpoints was established by their comparison with different clinical features. As a case study we compared each survival endpoint for significant association (FDR <0.2) with chromosomal aneuploidy.

Results The 4 endpoints were derived for 1097 TCGA BC cases having a median follow-up time of 27.7 months. Median times to events/censorship for OS, DSS, PFI, and DFI were 41.8/25.0, 32.6/26.0, 26.0/25.0, and 25.4/25.0 months respectively. PFI and DFI passed tests for adequate follow-up times; OS and DSS partially passed the same tests signaling some caution with their use in genomic associations.

Using the endpoints, outcomes of patients with ER+ and ER- tumors were compared, along with those of patients with low (I&II) and high (III&IV) stage breast tumors. Univariate analyses suggested patients with ER+ tumors had significantly better survival than patients with ER- tumors when using PFI (p=0.005), DFI (p=0.001), and DSS (p=0.009), with OS not reaching significance (p=0.09). Patients with low stage tumors showed significantly better outcomes than patients with high stage tumors for each endpoint (p<0.001). The 4 endpoints were also evaluated for their significant associations with chromosomal arm aneuploidy. Adjusted for patient age and AJCC stage, tumors with a loss of 8q and 8p (p=0.019, FDR=0.37) had worse PFI; and those with loss of 8q, 20q, and 8p had worse DFI. Tumors with gain of 11q or loss of 14, 7q, 12q, 18q, 20q, 3p, 7p, 8p, 18p, and 20p had worse OS. In contrast, tumors with loss of 16q had better DSS, while those with loss of 3q, 12q, 17q, 18q, 19q, 20q, 3p, 8p, 12p, 18p, 19p, and 20p had worse DSS. The finding that 8p loss associated with worse survival for all 4 endpoints, while 18p loss associated with worse OS and DSS, agrees with literature reports.

Conclusion These findings confirm that PFI and DFI, as extracted from the TCGA-CDR, are valid and appropriate BC survival endpoints, while OS and DSS may be recommended with some caution when employing TCGA data to evaluate new relationships between breast cancer genomic abnormalities and clinical outcomes.

The views expressed in this article are those of the author and do not reflect the official policy of the Department of Army/Navy/Air Force, the Department of Defense, or U.S. Government.

Citation Format: Liu J, Lichtenberg T, Hoadley KA, Cherniack A, Poisson L, Kovatich AJ, Benz C, Thorsson V, TCGA PanCanAtlas Research Network, Shriver CD, Hu H. Using the new pan-cancer clinical data resource (TCGA-CDR) to identify breast cancer genomic correlates associating with different survival outcome endpoints [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 P3-16-01.

Abstract P3-08-11: Contribution of germline mutations in cancer predisposition genes to tumor etiology in women diagnosed with invasive breast cancer before 40 years

Background: Although breast cancer in young women (YW) accounts for <10% of diagnoses annually, tumors in young patients have more aggressive characteristics and higher mortality rates. The cost of breast cancer, including treatment costs, physical and psychosocial effects, and lost productivity, is higher in YW than older patients. Improved understanding of etiology of breast cancer in YW is critical to developing effective prevention strategies.

Methods: All patients diagnosed before 40 years were identified. Family history was classified as average (No first or second degree relatives with breast or ovarian cancer or 1 second degree relative with breast cancer diagnosed >50 years), moderate (1 first degree relative with breast cancer, 2 first or second degree relatives with breast cancer diagnosed >50 years or 1 first or second degree relative with ovarian cancer) or strong (>1 first or second degree relative with bilateral breast cancer, breast and ovarian cancer or male breast cancer, >2 first or second degree relatives with breast cancer before age 50, breast and ovarian cancer in different relatives, ovarian cancer at any age or >3 first or second degree relatives with breast cancer at any age). Genomic DNA was isolated from blood samples and targeted sequencing was performed using the TruSight Cancer panel (Illumina). Pathogenic mutations were identified using VariantStudio.

Results: Seven percent (132/1950) of patients enrolled in the CBCP were diagnosed <40 years. Of these, 7% had a strong family history). TruSight sequencing was completed for 63 women for whom genomic DNA was available: five patients had pathogenic BRCA2 mutations (1813dupA, 5849del4, 999del5, Q2491X, Y3098X), all ER+ tumors, and seven patients had BRCA1 mutations (187delAG, 448insA, 943ins10, E84X, Q1313X, E1535X) all in triple negative breast cancers (TNBC). A pathogenic CHEK2 I157T was detected in an African American woman with TNBC. Variants of unknown significant were also detected in APC, ATM, BRCA1, BRCA2, CHEK2, CDH1, ERCC4, FANCA and PMS2 and heterozygote mutations detected in autosomal recessive genes BLM and RECQL4.

Discussion: Pathogenic mutations were found in 21% of young women with breast cancer with an additional 22% harboring potentially pathogenic mutations. BRCA1 mutations were associated with triple negative breast tumors in individuals with moderate to strong family history and BRCA2 mutations were associated with ER+ tumors in young women without strong family histories. These data demonstrate that although genetic predisposition may account for 21-43% of tumors, >50% of tumors in young women are not attributable to genetic causes, and identification of those non-genetic factors is critical to reduce the burden of breast cancer in this population.

Citation Format: Ellsworth RE, Rummel SK, Shriver CD. Contribution of germline mutations in cancer predisposition genes to tumor etiology in women diagnosed with invasive breast cancer before 40 years [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 P3-08-11.

Abstract P5-05-02: NEDD9 promotes breast cancer metastasis by regulating mitochondrial functions

NEDD9 has been characterized as a metastasis-promoting gene in various cancer cells including breast. We previously reported that NEDD9 promotes malignant phenotypes of breast cancer cells through distinct and non-overlapped domains. For example, the FAT (Focal Adhesion Targeting) domain of NEDD9 promotes cancer cell growth, while the SH-domain facilitates cell migration. These results suggest that NEDD9 promotes tumor metastasis by enhancing dissemination and growth in the tumor-host microenvironments through distinct and non-overlapped domains. Thus, targeting functions of NEDD9 is a promising approach for breast cancer therapies.

In order to further characterize NEDD9-mediated breast cancer growth, we performed yeast-two hybrid (Y2H) screening to identify proteins that associate with the FAT domain of NEDD9. Using the FAT domain constructed in pGBKKT7 (Clonetech, CA) as a bait to screen library of human fibroblast (Clonetech, CA), we identified several proteins that associate with the domain. They are small GTPases (i.e. RAB11a and ARF4), cytoskeletal proteins (i.e. Nexilin), and cytosolic proteins (i.e. HAX-1). Among of these potential partner proteins, we focused on the interaction between NEDD9 and HAX-1 in breast cancer cells. Co-immunoprecipitation assays confirm the molecular complex of NEDD9-HAX-1 in both SK-Br3 and SUM149 cells. Importantly, p130cas, which harbors similar domain structures with NEDD9, was not precipitated with NEDD9, suggesting a specific interaction between NEDD9 and HAX-1. Given the fact of NEDD9 as a key metastasis promoting gene, these results suggest that NEDD9-HAX-1 plays a key role breast cancer metastasis by facilitating growth in microenvironments.

While the biological function are not clear at present, previous studies demonstrated that HAX-1 localizes in mitochondria in breast cancer cells, Indeed, we demonstrated that NEDD9 was found in both cytosol and mitochondria fractions in malignant breast cancer cell MDA-MB-231, but not non-metastatic HCC38. These results suggest the presence of NEDD9-HAX1 complex in mitochondria and this complex may facilitate breast cancer metastasis. In addition to HAX-1, several mitochondrial proteins such as EFG1, DCTN6, and MMADHC were found in the Y2H screening system as described above. These results suggest that NEDD9 facilitates breast cancer metastasis through regulating multiple pathways including signaling pathways and mitochondrial functions, thus serving as a promising therapeutic target for cancer patients including breast.

The view expressed in this article are those of the author and do not reflect the official policy of the Department of Defense, or U.S.Government.

Citation Format: Iida J, Dorchak J, Slavik J, Clancy R, Cutler ML, Shriver CD. NEDD9 promotes breast cancer metastasis by regulating mitochondrial functions [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 P5-05-02.

Abstract P1-05-11: Comprehensive comparison of breast cancer molecular portraits by African and European ancestry in the cancer genome atlas

Background: African American breast cancer patients have worse survival rates than European American patients. Although racial differences in the distribution of breast cancer intrinsic subtype are known, it is unclear if there are other inherent genomic differences contributing to this racial outcome disparity.

Methods: We defined patient race based on genomic ancestry and compared multiple molecular features of breast cancer between 154 black and 776 white patients in The Cancer Genome Atlas (TCGA). We examined the contribution of these molecular features to survival outcomes using Cox proportional hazards models. We also estimated the heritability of breast cancer subtypes using a mixed effect model.

Results: Compared to whites, black patients had higher odds of basal-like (odds ratio=3.80, p<0.001) and HER2-enriched (odds ratio=2.22, p=0.027) breast cancers in reference to luminal A subtype. Beyond differences in relative frequency of intrinsic subtypes, black and white patients had distinct gene expression, protein expression, and somatic mutation landscapes. However, the majority of these molecular differences were eliminated after adjusting for subtype; in the subtype-adjusted models, we found 142 genes, 16 methylation probes, 4 copy number segments, 1 protein, and no somatic mutation were differentially expressed or present between black and white patients. Using the top 40 differentially expressed genes, we built a race-enriched gene signature, which had excellent capacity of distinguishing breast tumors from black versus white patients (c-index=0.852 in the validation dataset). We also estimated the heritability of breast cancer subtype (basal vs. non-basal) to be 0.436 (p=1.5x10-14) and showed that two genetic variants (rs1078806 in FGFR2, rs34084277 in BABAM1) were associated with intrinsic subtype and can partially explain racial differences in subtype frequencies.

Conclusion: On the molecular level, once intrinsic subtype frequency differences are accounted for, there are few genomic or proteomic differences observed between blacks and whites. More than 40% of breast cancer subtype frequency differences may be due to genetic ancestry. These results suggest that future studies are warranted to investigate genetic and non-genetic factors that contribute to the development and progression of breast cancer subtypes in order to reduce racial disparity.

Citation Format: Huo D, Hu H, Rhie SK, Gamazon ER, Cherniack AD, Liu J, Yoshimatsu TF, Pitt JJ, Hoadley KA, Troester M, Ru Y, Lichtenberg T, Sturtz LA, Shelley CS, Mills GB, Laird PW, Shriver CD, Perou CM, Olopade OI. Comprehensive comparison of breast cancer molecular portraits by African and European ancestry in the cancer genome atlas [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 P1-05-11.

Abstract PD6-01: Analysis of breast cancer in young women in the department of defense (DOD) database

Background: Women under the age of 40 account for approximately 7% percent of breast cancer patients. Breast tumors from young women are often ER-negative, occur in African-American patients, and have other indicators of high risk: yet, multivariate analyses demonstrated that young age is an independent predictor of poor outcome. Due to the unique nature of the patient population served by DOD, a disproportionate number of breast cancer cases in young women are seen. We compare the characteristics, treatment, and outcomes of young patients diagnosed with breast cancer with those of older patients.

Methods: The databases of the Military Health System Repository and the DOD Central Registration were used to identify female breast cancer patients treated at DOD facilities between 1998 and 2007. Information on demographics, breast cancer stage at diagnosis, definitive surgical treatments, systemic treatment, recurrence rate and overall survival was analyzed by age groups at the time of diagnosis (less than 40 years old, 40 to 49 years, and 50 years or older) using X2 testing with significance defined as p< 0.05.

Results: We identified 10,066 women who were diagnosed with invasive breast cancer at DOD facilities between 1998 and 2007, of which 11.3% (1139) were less than 40 years old at diagnosis. 53% of this young cohort were white, 25% were African-American and 8% were Hispanic (14% undisclosed). The percentage of breast cancer among African-American women in the young cohort was higher than in the older cohorts (19.3% in 40-49yo and 10.6% in ≥50yo). High-grade tumors were significantly more frequent in the younger cohort when compared to the older group (49.5% vs 34.7% and 25.2%, p<0.001). <40yo most commonly presented with Stage II disease (45.3%) at diagnosis, while older groups were mostly diagnosed with Stage I disease (41.6% and 52.4%). The most common subtype of breast cancer across ages was ER+ disease, however, <40yo group had proportionally less ER+ (49% vs 61% and 67.3%, P<0.001). There was a higher rate of bilateral mastectomies among the young women (18.4% vs 9.1% and 5.0%, p<0.0001). Independently of the stage of disease, chemotherapy was given significantly more frequently to <40y (90.43%) and 40-49yo (81.44%) than ≥50yo (53.71%). The 10-year overall survival of younger women was similar to the ≥50yo cohort, despite intensive treatment.

Discussion: This study is one of the largest retrospective studies of women under 40 years old with breast cancer. Younger women with invasive breast cancer had more aggressive tumors presenting at higher stages. In this group with good access to healthcare, younger women still had a similar overall survival rate to older women despite receiving more aggressive treatment and potentially having fewer comorbidities than the older group.

Citation Format: Zimmer AS, Gatti-Mays M, Soltani S, Lipkowitz S, Steeg PS, Zhu K, Perkins JG, Hu H, Shao S, Brown D, Shriver CD. Analysis of breast cancer in young women in the department of defense (DOD) database [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 PD6-01.

Abstract P4-10-08: Can a diagnosis of invasive breast cancer effectively motivate patients to follow healthy lifestyles?

Background: Survival rates for patients diagnosed with invasive breast cancer have increased dramatically, yet survivors often face a host of adverse health effects. Factors such as obesity, physical inactivity and tobacco use may contribute to decreased survival and quality of life. Behavioral risk factors in patients with and without breast cancer were evaluated to determine whether a diagnosis of invasive disease was sufficient motivation to modify lifestyle choices.

Methods: The dataset included female patients diagnosed between 2001-2011with malignant (n=421) or benign (n=230) breast disease and who had baseline and >1-year follow-up information available. Changes in body mass index (BMI), fat intake, exercise frequency, alcohol and tobacco use, caffeine consumption, hormone replacement therapy (HRT) use and frequency of breast self-exam (BSE) were assessed. Random coefficients models were used to examine longitudinal effects of an invasive diagnosis on healthy behaviors. P<0.05 was used to define significance.

Results: At diagnosis, patients with invasive cancer were significantly older (59 years), more likely to consume >7 glasses of alcohol/week (7%) but less likely to be using HRT (3%) than those with benign disease (50 years, 3% and 11%, respectively). At baseline, a majority of both invasive and benign patients were overweight, non (current) smokers, and consumed a high fat, highly caffeinated diet and exercised <90 minutes/week, and >50% of both groups performed BSE at least once/month. Exercise, BMI, and caffeine, alcohol and fat intake did not change over time in either invasive or benign groups. Smoking decreased to a similar extent in both invasive and benign patients. In contrast, compliance with monthly BSE increased and HRT use decreased significantly at each time point in the invasive patient group, with no corresponding changes in patients with benign disease.

Conclusions: These data support the critical importance of providing education and recommendations about engaging in healthy behaviors by the clinical staff. The two behaviors that improved significantly in patients with invasive breast cancer are both addressed by the clinical staff at Walter Reed National Military Medical Center at the time of diagnosis: with nurse navigators providing education about proper BSE and physicians recommending discontinuation of HRT. Failure to significantly change other behaviors suggests that a diagnosis of breast cancer is not a sufficient motivating factor for the patient to adopt healthier lifestyle choices without provision of education and resources. Our data suggest a need for increased health-related behavioral counseling and support systems to successfully modify personal behaviors, thus improving the health and quality of life of breast cancer survivors.

Citation Format: Ellsworth RE, Costantino N, Toro AL, Shriver CD, Ellsworth DL. Can a diagnosis of invasive breast cancer effectively motivate patients to follow healthy lifestyles?. [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 P4-10-08.

Abstract P4-09-14: PhosphohistoneH3 as a prognostic marker in breast cancer: High expression is associated with younger age, triple negative subtype, and disease specific survival

BACKGROUND PhosphohistoneH3 (PPH3) is an emerging marker in breast cancer and has been linked to both patient survival and age. Phosphorylation of HistoneH3 is an important step during the cell cycle leading to proper compaction of the chromatin during late G2 and early mitosis. Here we assessed the use of PPH3 as a prognostic marker within a group of invasive breast cancers in the Clinical Breast Care Project (CBCP).

METHODS CBCP participants and their samples were collected following IRB-approved, HIPAA-compliant protocols. Samples from 157 CBCP patients were selected for tissue whole section immunohistochemistry (IHC), using antibodies to PPH3, ER, PR, Ki67, and Her2. For each sample, staining of PPH3 was assessed across 5 high powered microscope fields and was considered positive if there was on average >2 stained cells per field. ER and PR were considered positive when there was >5% nuclear staining, and Ki67 was positive when there was >15% nuclear staining. Her2 was considered positive with an IHC score of 3+ or 2+ with a FISH score above 2.2. The samples were subtyped as Luminal A (LA: ER+/HER2-/Ki67-), Luminal B1 (LB1: ER+/HER2-/Ki67+), Luminal B2 (LB2: ER+/HER2+), Her2+ (ER-/PR-/HER2+), and Triple Negative (TN: ER-/PR-/HER2-). PPH3 was tested for associations with age and subtype using a stratified univariate Wilcoxon rank-sum analysis and a multivariate analysis controlling for subtype. To test the efficacy of PPH3 as a prognostic marker, Kaplan-Meier curves for disease specific survival were analyzed and the cox proportional hazard regression model was calculated. Further analysis addressing population demographics and additional cancer characteristics is ongoing.

RESULTS Wilcoxon analysis revealed an association between higher PPH3 levels and younger age (P=.0038). Subtype was also found to be associated with PPH3, with the TN subtype 6.26 times more likely to have higher PPH3 expression than LA (P=.005). The association with age was confirmed by repeating the analysis and stratifying into non-TN subtypes (P=.05) and TN only subtype (P=.017). Non-TN subtypes positive for PPH3 expression had median age of 53.18 at diagnosis and 63.29 for negative PPH3 expression; TN subtypes that were positive for PPH3 had a median age of 50.44 and 72.9 for negative PPH3. Multivariate analysis with age and subtype as the variables also supported these results (age P=.017; TN vs LA P=.022). Disease specific survival analysis showed that a shorter survival time was associated with positive PPH3 protein levels (P=0.03; hazard ratio=6.97).

CONCLUSIONS High expression of PPH3 is associated with a younger age, poorer survival rate, and the TN subtype. These results corroborate the use of PPH3 as a prognostic marker for breast cancer patients.

The views expressed in this article are those of the author and do not reflect the official policy of the Department of Defense, or U.S. Government.

Citation Format: Craig J, Kovatich AJ, Hooke JA, Kvecher L, Liu J, Fantacone-Campbell JL, Rui H, Shriver CD, Hu H. PhosphohistoneH3 as a prognostic marker in breast cancer: High expression is associated with younger age, triple negative subtype, and disease specific survival. [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 P4-09-14.

Abstract P2-05-16: Tumor-associated glycans as key molecules to promote growth of triple-negative breast cancer cells

Enhanced invasion and migration into the surrounding tissues are hallmarks of the malignancy of tumor cells. To successfully metastasize, a cancer cell has to detach from the primary tumor, invade into surrounding tissues, and intravasate into blood or lymphatic vessels. These processes are composed of complex mechanisms involving tumor recognition, degradation of extracellular matrix (ECM) proteins and migration into tissue. Triple negative (TN) breast cancers are defined by a lack of expression of estrogen, progesterone, and HER2 receptors. It is widely recognized that TN breast cancers have a poorer prognosis than any other subtype of breast cancer. Given the lack of effective targeted therapies for TN breast cancer patients, understanding of the mechanisms of migration and invasion of these tumors will provide insight into developing novel approaches to lower the mortality from TN breast cancer.

Previous studies demonstrated that NEDD9 plays a key role facilitating progression and metastasis of various tumor cells including breast. We previously demonstrated that NEDD9 plays a critical role in promoting migration and growth of MDA-MB-231. In order to further characterize the mechanisms of NEDD9-mediated cancer migration and growth, we established stable cell lines expressing NEDD9 using HCC38 as a parental cell line which expresses low level of endogenous NEDD9. Microarray studies demonstrated that enzymes (CHST11, CHST15, and CSGALNACT1) involved in biosynthesis of chondroitin sulfate (CS) but not heparan sulfate (HS) were markedly upregulated in HCC38(NEDD9) compared to control HCC38(Vector) cells. These results suggest that NEDD9 regulates specific structures of tumor-associated glycans such as chondroitin sulfate. Core proteins of CD44 and Serglycin were markedly upregulated in HCC38(NEDD9) cells compared to HCC38(Vector) cells, while those of Syndecan-1, Syndecan-2, and Versican were downregulated in HCC38(NEDD9). Immunofluorescence studies using specific antibody, GD3G7, confirmed the enhanced expression of CS-E subunit in HCC38(NEDD9). Immunoprecipitation and western blotting analysis demonstrated that CS-E was attached to Serglycin and CD44 core proteins. We demonstrated that removing CS by chondroitinase ABC significantly inhibited anchorage-independent growth of HCC38(NEDD9) in methylcellulose. Importantly, the fact that GD3G7 significantly inhibited colony formation of HCC38(NEDD9) cells suggest that CS-E subunit plays a key role in this process. Furthermore, treatment of HCC38(NEDD9) cells with chondroitinase ABC or GD3G7 significantly inhibited mammosphere formation. Exogenous addition of CS-E enhanced colony formation and mammosphere formation of HCC38 parental and HCC38(Vector) cells. These results suggest that NEDD9 regulates the synthesis and expression of tumor associated glycocalyx structures including CS-E, which plays a key role in promoting and regulating breast cancer progression metastasis and possibly stem cell phenotypes.

The opinion and assertions contained herein are the private views of the authors and are not to be construed as official or as representing the views of the Department of the Army or the Department of Defense.

Citation Format: Iida J, Dorchak J, Clancy R, Slavik J, Cutler ML, Shriver CD. Tumor-associated glycans as key molecules to promote growth of triple-negative breast cancer cells. [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 P2-05-16.

Abstract P3-13-03: Multicenter clinical trial of percutaneous laser ablation for early stage primary breast cancer. Results of 49 cases with radiographic and pathological correlation

Introduction: Percutaneous laser ablation of early stage primary breast cancer remains investigational. A multicenter, international clinical trial (NCT01478438) was completed to determine feasibility of this technique.

Methods: Patients with a single focus of biopsy proven infiltrating ductal carcinoma measuring 20 mm or less by pre-ablation MRI were treated by image-guided percutaneous laser ablation. A laser diode source (805 nominal nanometer wavelength) was used to perform the thermal ablation. Thermal sensors placed at the periphery of the tumor measured achievement of predefined temperature levels, indicating successful ablation. The patients were evaluated by post-ablation mammogram, ultrasound and MRI at 4 weeks post-ablation, after which they underwent surgical excision. Pathology specimens were evaluated by hematoxylin & eosin, CK 8/18, Ki-67 and estrogen receptor staining.

Results: Forty-nine of the 61 enrolled patients (ages 42-77, mean age 64 years) undergoing percutaneous laser ablation have finished protocol analysis and are reported in this series. Ablation was considered complete by the treating physician in all cases. The mean tumor size was 11.3 mm. The mean laser time was 15.7 minutes. There were no serious adverse events. Seven patients (14%) reported mild adverse events (pain, blisters, lump). Post-ablation cell viability was determined by MRI and by changes in CK 8/18, Ki67 and estrogen receptor staining. A post-ablation discordance between MRI and pathology was found in evaluation of 4 patients (8%). Three patients (6%) were considered "false negative" with a post-ablation residual tumor burden of less than 2mm which was not detected by MRI. One patient (2%) had a complete pathologic ablation but positive MRI ("false positive"). One patient (2%) had adjacent residual DCIS, visible in retrospect on the pre-ablation MRI and was considered a screening failure. Eight patients (16%) were found to have residual invasive cancer by both post-ablation MRI and pathologic analysis. Complete ablation was confirmed in 36 patients (73%) when evaluated by both post-ablation MRI and pathologic analysis.

Conclusion: Percutaneous laser ablation holds promise as an alternative to lumpectomy in the treatment of early stage breast cancer. There is a strong correlation (92%) between findings on post-ablation MRI and changes in CK 8/18, Ki67 and estrogen receptor staining in this series. Additional trials are necessary to determine the long-term curative potential of this technique.

Citation Format: Schwartzberg BS, Abdelatif OMA, Lewin JM, Bernard JM, Brehm JL, Bu-Ali HM, Cawthorn SJ, Chen-Seeto M, Feldman SM, Govindarajulu S, Jones LI, Juette A, Kavia S, Maganini RO, Pain SJ, Shere MH, Shriver CD, Smith SG, Valencia A, Whitacre EB, Whitney R. Multicenter clinical trial of percutaneous laser ablation for early stage primary breast cancer. Results of 49 cases with radiographic and pathological correlation. [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 P3-13-03.

Steroid induction of therapy-resistant cytokeratin-5-positive cells in estrogen receptor-positive breast cancer through a BCL6-dependent mechanism

Therapy resistance remains a major problem in estrogen receptor-α (ERα)-positive breast cancer. A subgroup of ERα-positive breast cancer is characterized by mosaic presence of a minor population of ERα-negative cancer cells expressing the basal cytokeratin-5 (CK5). These CK5-positive cells are therapy resistant and have increased tumor-initiating potential. Although a series of reports document induction of the CK5-positive cells by progestins, it is unknown if other 3-ketosteroids share this ability. We now report that glucocorticoids and mineralocorticoids effectively expand the CK5-positive cell population. CK5-positive cells induced by 3-ketosteroids lacked ERα and progesterone receptors, expressed stem cell marker, CD44, and displayed increased clonogenicity in soft agar and broad drug-resistance in vitro and in vivo. Upregulation of CK5-positive cells by 3-ketosteroids required induction of the transcriptional repressor BCL6 based on suppression of BCL6 by two independent BCL6 small hairpin RNAs or by prolactin. Prolactin also suppressed 3-ketosteroid induction of CK5+ cells in T47D xenografts in vivo. Survival analysis with recursive partitioning in node-negative ERα-positive breast cancer using quantitative CK5 and BCL6 mRNA or protein expression data identified patients at high or low risk for tumor recurrence in two independent patient cohorts. The data provide a mechanism by which common pathophysiological or pharmacologic elevations in glucocorticoids or other 3-ketosteroids may adversely affect patients with mixed ERα+/CK5+ breast cancer. The observations further suggest a cooperative diagnostic utility of CK5 and BCL6 expression levels and justify exploring efficacy of inhibitors of BCL6 and 3-ketosteroid receptors for a subset of ERα-positive breast cancers.

Abstract P4-06-03: Assays on core biopsies and surgically resected tumors may result in different subtyping of the invasive breast cancer from the same patient

Background Core biopsies (CBs) are often used for biomarker expression assays to determine the treatment regimen. However, a number of other clinically important analyses (e.g. OncoType Dx), are performed on surgically resected tumors (SRTs). A previous study has shown that biomarkers ER, PR, and Ki67 expressed higher in CBs than in SRTs. Here we analyze how this difference impacts the subtyping of ER+ breast tumors.

Methods Female patients enrolled in the Clinical Breast Care Project (CBCP) from a civilian site were selected for this study, where expression of ER, PR, HER2, and Ki67 were assayed by IHC in a reference lab on CBs; the same 4 assays were performed on SRTs by a CBCP central lab. Both labs are CLIA-certified. Patients treated with neoadjuvant chemotherapy and those with multiple tumors were excluded. 167 cases were identified for this study to compare assays performed on CBs and SRTs from the same patients. ER and PR were positive if >1% nuclear staining, HER2 was negative if IHC = 0 or 1+, positive if IHC = 3+, and for IHC = 2+ FISH was used for the final call. Ki67 was positive if > = 15% nuclear staining. LA was ER+/HER2-/Ki67-, LB1 was ER+/HER2-/Ki67+, and LB2 was ER+/HER2+. For histologic grades, only readings from the central lab on SRTs were used. Statistical analyses were performed using SAS.

Results This analysis confirmed that Ki67, ER, and PR showed higher percent nuclear staining in CBs than in SRTs from the same patients. The difference for Ki67 was more striking and unidirectional. ER and PR cases clustered at the upper percent levels. Histograms with a bin-width of 15% show a peak at 15% for Ki67 difference between CBs and SRTs, whereas the peaks for ER and PR differences were at 0%. McNemar's (or Exact McNemar’s) test showed significant differences between the binary status calls for Ki67 (p = 3.2E-15) and ER (p = 0.012), but not for PR (p = 0.65). Assays on CBs and SRTs resulted in different subtype calls for the cases (Table 1). Grade distributions were different between LA and LB (p<0.001 for both CB- and SRT-based subtypes, Chi-Square or Fisher's Exact test), but not so between LB1 and LB2 (p = 0.23 for CB, 0.31 for SRT). However, SRT-based LB1 cases concentrate more on higher grades compared to CB-based cases (p = 0.048).

Table 1. ER+ subtypes based on IHC assays (from CBs and SRTs) and corresponding grades (from SRTs) CBSRTSubtypeG1G2G3G1G2G3LA2126034518LB11435342820LB2036032

Discussion On IHC assays, Ki67 expression is strikingly higher in CBs than in SRTs, and ER expression is also higher in CBs than in SRTs. This directly resulted in more LB than LA subtypes based on CBs. SRT-based LB1 cases concentrate more on higher grades compared to CB-based cases, which is more consistent with the observation that LB subtypes have worse outcomes. A limitation of this study is that technical differences between the labs may contribute to the observed differences between CBs and SRTs. Further studies need to be performed to determine whether SRT should also be assayed in addition to CB for treatment regimen decision-making.

The views expressed in this abstract are those of the authors and do not reflect the official policy of the Department of Defense, or US Government.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-06-03.

Abstract P4-04-16: Highly variably expressed exons (HVEE) of cancer genes and survival disparity in human breast cancer

Background: RNA-Seq is a powerful technology for accurately quantifying the transcriptome of human cancers. The human breast cancer section of the TCGA has generated RNA-Seq data for over 950 samples from primary tumors and adjacent normal tissues. Analysis of these data has provided insights into transcript isoforms and gene fusions related to the etiology of breast cancer. These rich data no doubt contain other insights into the expression of cancer genes and how it relates to cancer clinical outcomes.

Methods: TCGA RNA-Seq data were downloaded from TCGA data portal. A total of 712 samples were available when this project started, including data from 607 primary breast tumors of which 108 were Basal-like, 316 Luminal A, 139 Luminal B, and 54 HER2-enriched. Data from 65 adjacent normal tissues were also included to be used as controls. We examined the exon-level expression for a group of 137 strongly cancer related genes that can be grouped into 12 pathways (Vogelstein et al. 2013). We focused on exons that display a significant (p <0.05) difference in their expression variation, compared to the other exons of the same gene (> 2 fold changes in terms of standard deviation based on bootstrap resampling). We refer to these exons as HVEEs (Highly variably expressed exons). Using both unsupervised and supervised algorithms (e.g. non-negative matrix factorization), we classified the samples from various subtypes based on the expression patterns of the HVEEs and looked for patterns that associate with clinical outcomes.

Results: For many cancer genes, such as BRCA2 and PTEN, the distribution of within-gene exon expression is similar across all tumor samples and the control group. Subtype-specific amplification of oncogenes (e.g. ERBB2 in the HER2 subtype) and lower expression of cancer genes (e.g. AR in the basal subtype) have also been observed. We observed that exon expression across genes and across samples has in general small variations. Out of the 2,153 exons from the 137 cancer genes, we identified only 41 genes with HVEEs. Most of these genes have only one HVEE but a small number of them have multiple HVEEs (48 in total). Interestingly, HVEEs in 31 of these genes are consistent across the four breast cancer subtypes, as well as the normal group. However, the medians of such HVEE expression across the normal samples are in general significantly lower than those across the tumor samples in any tumor subtypes. Further, using the exon expression data of only the HVEEs, we identified distinct clusters for the basal-like subtype. The 108 basal-like subtype samples can be classified into two clusters with 41 and 67 samples, respectively, based on the expression level of the HVEEs from the above mentioned 31 genes. Overall survival for patients in these two clusters trends toward significance (log-rank test p = 0.058).

Discussion: The consistency of HVEEs across breast cancer and the normal samples and the ability of the HVEE expression level to stratify basal-like breast cancer in a clinically meaningful way is intriguing. We are extending the study to a larger dataset for validation.

The views expressed in this abstract are those of the authors and do not reflect the official policy of the Department of Defense, or US Government.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-04-16.

Abstract P4-06-04: Effect of genomic heterogeneity on breast cancer progression and metastatic spread

Purpose: DNA fingerprinting has revealed discordant patterns of chromosomal alterations in primary breast tumors (PBT) compared to matched metastatic axillary lymph node tumors (MLNT); however, whether these genetic differences reflect the timing of dissemination of cells with metastatic potential from the primary tumor, intratumoral heterogeneity of the primary breast tumor, or independent seeding of lymph nodes with metastatic cells remains unclear.

Patients and Methods: From the 30 node-positive patients evaluated in this study, allelic imbalance (AI) data was generated using 52 microsatellite markers from 5-19 areas of each PBT as well as from available MLNT. Data were analyzed using pairwise correlations, ANOVA and PHYLIP.

Results: The frequency of genomic changes was significantly higher (P<0.001) in PBT areas (13%) than MLNT (9%). No two PBT areas had identical patterns of AI and the percent concordance (PC) of AI events between tumor regions ranged from 0-65% (average 33%). Of the 196 MLNT from 28 patients, PC between MLNT from the same patient ranged from 0-88% (average = 32%). Neither the overall frequency of AI nor the PC differed significantly between sentinel (SLN) and non-SLNs. Phylogenetic analysis revealed that within patients, many MLNT appeared to be descended from different areas of the PBT, but patterns of descent were complex.

Conclusions: Both PBT and MLNT are characterized by extensive molecular heterogeneity, MLNT appear to originate from different areas of the PBT, and SLN metastases are not genomically more advanced than non-SLN metastases. These data suggest that metastatic dissemination may be influenced by both spatial and temporal factors, with cells with metastatic potential colonizing lymph nodes throughout the development of the PBT, and that SLN metastases do not appear to be a source of metastatic cells for non-SLN but rather MLNT arise by independent colonization.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-06-04.

Abstract P3-07-13: Evaluation of epidemiological and molecular differences in African American and Caucasian women with triple negative breast cancer

Background: Breast cancer's (BC) higher mortality rates in African American women (AAW) have been attributed to the higher frequency of triple negative breast cancer (TNBC), an aggressive tumor subtype, in young AAW. What is not known is whether TNBC represents a single disease entity or differs in etiology, molecular characteristics or outcomes by population.

Methods: Demographic, pathological and survival data from AAW (n = 62) and CW (n = 98) with TNBC were analyzed using chi-square analysis, Student's t-tests, and log-rank tests. RNA from laser microdissected TNBC from a subset of patients was hybridized to HG U133A 2.0 microarrays and data analyzed with a FDR <0.05, >2-fold change to define significance.

Results: The frequency of TNBC compared to all BC was significantly higher in AAW (26%) compared to CW (8%), however, significant survival and pathological differences were not detected. The expression of CRYBB2 was ∼4-fold higher in tumors from AAW; these expression levels were able to classify just 64% of AAW patients correctly. Among demographic characteristics, AAW consumed significantly lower amounts of caffeine and alcohol, were less likely to breastfeed and more likely to be obese.

Conclusions: These data suggest that TNBC in AAW is not a unique disease compared to TNBC in CW. Rather, higher frequency of TNBC in AAW may, in part, be attributable to the effects of lifestyle choices. Because these risk factors are modifiable, they provide new opportunities for the development of risk reduction strategies that may effectively decrease survival disparities by preventing the development of TNBC in AAW.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P3-07-13.

Abstract P1-08-20: Increased risk of hormone therapy failure in breast cancers expressing low phospho-Stat5: Validation of quantitative immunofluorescence assay parameters

Previous analyses of three breast cancer cohorts revealed that loss of phospho-Stat5 in breast cancer is associated with significantly elevated risk of hormone therapy failure (1, 2). Nuclear localized tyrosine phosphorylated Stat5 (Nuc-pYStat5) may therefore have clinical value as a predictive marker. Analysis of two of the three previously reported anti-estrogen treated patient cohorts used pathologist scoring of diaminobenzidine (DAB) chromogen-stained Stat5. However the third cohort, analyzed by quantitative immunofluorescence analysis (QIF) on the Genoptix/HistoRx AQUA platform, revealed a greater hazard ratio than the cohorts analyzed by pathologist DAB-scoring. To extend and validate these observations, we applied the Nuc-pYStat5 cutpoint derived in our previous study (2) to an independent cohort of anti-estrogen-treated breast cancer patients using two distinct QIF software platforms, AQUA and Definiens Tissue Studio. Tissue Studio relies on supervised machine learning and multiparametric features of a high-resolution whole slide image to identify cancer cell regions, while AQUA software relies on costaining of a tumor marker to identify cancer cell regions. The two QIF platforms produced highly concordant Nuc-pYStat5 levels (R2 linear = 0.96, P<0.001, N = 344) and confirmed a significant elevated risk of failing antiestrogen therapy in patients whose tumors had lost Nuc-pYStat5 (Hazard ratio 3.6; 95% CI 1.8-7.4; P<0.02; N = 98). On both QIF platforms, Nuc-pYStat5 remained an independent marker after multivariate adjustment for standard pathology parameters, including ER/PR, HER2, age, node status and grade, with a hazard ratio of 5.8 (95% CI 1.3-22.2; P = 0.02; N = 52). High concordance between Nuc-pYStat5 levels produced by the two QIF platforms held up in a second independent dataset of more than 300 breast cancer specimens (R2 linear = 0.97, P<0.001, N = 382). Nuc-pYStat5 levels by the two QIF methods remained highly concordant across the entire dynamic range in both patient cohorts. Furthermore, high concordance was also observed between replicate QIF analyses of Nuc-pYStat5 on serial tumor microarray sections stained in the same run on an automated immunostainer (Concordance Correlation Coefficient (CCC) = 0.96; 95% CI 0.96-0.97). Modest inter-assay staining variation (CCC = 0.84; 95% CI 0.82-0.87) for Nuc-pYStat5 when serial tumor microarrays were stained on different runs several days apart could be corrected for by normalization procedures (CCC = 0.94; 95% CI 0.92-0.95). This progress supports the utility of QIF analysis of Nuc-pYStat5 levels in human breast cancer and further documents the potential value of Nuc-pYStat5 as a predictive marker of response to antiestrogen therapy. The study confirms that further retrospective and prospective validation studies are warranted.

References:

1) Yamashita et al. Stat5 expression predicts response to endocrine therapy and improves survival in estrogen receptor-positive breast cancer. Endocr Relat Cancer. 2006;13:885-93.

2) Peck et al. Loss of nuclear localized and tyrosine phosphorylated Stat5 in breast cancer predicts poor clinical outcome and increased risk of antiestrogen therapy failure. J Clin Oncol. 2011;29:2448-58.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P1-08-20.

Abstract P4-06-09: HER2+ and HER2- luminal B subtypes have similar overall survival and histologic grade distributions

Background There are multiple subtypes in invasive breast cancers (IBCs). Immunohistochemistry (IHC)-based assays using ER, PR, HER2, and Ki67 for subtyping has been developed. However, association between such subtypes and treatment outcomes and histology is not completely known, and are impacted by dataset-to-dataset and pathologist-to-pathologist variations. We report an analysis on these problems, as a pilot study of a project involving 5,000 patients and 250 protein biomarkers.

Methods Patients were enrolled for the Clinical Breast Care Project from a military site with data collected per IRB-approved protocols, from 2000 to 2010. Total of 215 female IBC cases were included in this study, with surgically resected tumors (SRT) assayed for ER, PR, HER2, and Ki67 by IHC in a central CLIA-certified lab following clinical guidelines where applicable. All slides were reviewed by a single experienced breast pathologist. ER and PR was positive if nuclear staining was >5%. HER2 was negative if IHC = 0 or 1+ and positive if IHC = 3+; For IHC = 2+, the FISH result determined the final call. Ki67 was positive if nuclear staining was > = 15%. For IBC subtypes, LA was ER+/HER2-/Ki67-; Two LB subtypes were defined, with LB1 being ER+/HER2-/Ki67+ and LB2 being ER+/HER2+; Her2+ was ER-/PR-/HER2+; TN was ER-/PR-/HER2-. Statistical analyses were performed using SAS, Kaplan-Meier estimate and log-rank test were used for survival analysis and the follow-up period was 10 years with a median of 4.6 years. Chi-Square test was used for categorical data analysis supplemented by Fisher's Exact test as appropriate.

Results 204 of the 215 cases were classified into subtypes of LA (n = 74, 7 deceased), LB1 (n = 53, 4 deceased), LB2 (n = 14, 1 deceased), Her2+ (n = 14, 1 deceased), and TN (n = 49, 16 deceased). Despite the low number of events in some subtypes, there was a significant difference in overall survival between the 5 subtypes of IBCs defined here (p = 0.0023), with TN cases showing the least favorable outcome. No difference was observed in outcome between LB1 and LB2 (p = 0.86). Overall, Ki67+ cases trended toward worse outcomes (p = 0.08), which was also observed in TN (p = 0.17) but not other subtypes. Histologic grades were significantly different among the 5 subtypes (p = 6.25E-20); 96% of LA cases were G1 or G2, over 80% of LB1 and LB2 cases were G2 or G3, and all Her2+ and 93% of TN cases were G2 or G3. Within the luminal subtypes, grade distribution for LA cases was significantly different from that for LB cases (p<0.0001) but there was no difference between LB1 and LB2 cases (p = 0.95).

Discussion In this cohort where all IHC and pathology slides were reviewed by a single pathologist, we used cell proliferation marker Ki67 to help classify luminal IBCs into LA, LB1 (HER2-), and LB2 (HER2+). Overall survival analysis result for all cases was consistent with the literature, Ki67+ cases trended toward worse outcomes, and no outcome difference was identified between LB1 and LB2. Histologic grade distributions in different subtypes were consistent with the literature; we further found no difference between LB1 and LB2 subtypes.

The views expressed in this abstract are those of the authors and do not reflect the official policy of the Department of Defense, or US Government.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-06-09.

Abstract P2-05-21: Molecular subtypes of invasive breast cancers show differential expression of the proliferation marker Aurora Kinase A (AURKA)

Background: Invasive breast cancer (IBC) has been classified into four major subtypes based on gene expression profiling. The luminal A subtype (LA) has the best prognosis, when compared to luminal B (LB), HER2+, and basal-like (Basal). Ki67 by gene expression or immunohistochemistry (IHC) is commonly used as a proliferation index. The function of Ki67 in proliferation remains unknown. AURKA (STK15) is known to play an important role in mitosis, and is a component of the 21-gene recurrence score of the Oncotype Dx. With multiple platforms of molecular data available from hundreds of IBC tissues in The Cancer Genome Atlas project (TCGA), we sought to study the association of AURKA with different IBC subtypes and explore its use as a proliferation marker in IBCs.

Methods: Gene expression (Agilent, log2 transformed), relative DNA copy number (CN, Affymetrix SNP 6.0), and exome sequence mutation (Illumina) data for 459 IBC cases were downloaded from the TCGA data portal. PAM50 classification results of all samples were obtained from the TCGA breast cancer AWG group and included 203 LA, 113 LB, 51 HER2+, 84 Basal-like, and 8 Normal-like which were not used in this study due to the low numbers. Kruskal-Wallis tests were used to evaluate the differences among four subtypes on AURKA expression and CN, followed by Wilcoxon Mann-Whitney test with Bonferroni adjustment for pairwise analyses. Pearson's Correlation Coefficient was used for correlation analyses. All statistical analyses were performed using SAS and R, and two-sided, p values <.05 were considered statistically significant.

Results: There was a significant difference among IBC subtypes, in gene expression as well as in CN (p values < 0.0001). AURKA mRNA levels were significantly lower in LA (mean±SD, −2.61±0.63) compared to LB (−1.45±0.78), HER2+ (−1.38±0.61), and Basal (−1.26±0.62) subtypes (p values all < 0.0001). No significant difference was detected between other subtype pairs. In CN analysis, Basal (0.09±0.22) was lower than HER2+ (0.32±0.308, p < 0.0002) and LB (0.33±0.41, p < 0.0001), and LA (0.14±0.28) is lower than HER2 (p < 0.0016) and LB (p < 0.0001), but no other significant CN difference between the subtypes were found. The means and SDs are provided for reference only. No correlation of p53 mutation status and AURKA expression were observed. However, AURKA gene expression level is correlated with MKI67 gene expression (R = 0.69, p < 2.2e−16), and its correlation with PAM50 proliferation score is even higher (R = 0.80, p < 2.2e−16).

Discussion: Using the TCGA data we observed that the mean gene expression level of AURKA is significantly lower in LA than the other IBC subtypes, by more than 50% (note the log2 transformation). This differential expression is not completely due to CN changes (especially for the Basal subtype). There is a strong association with other tumor cell proliferation markers such as the MKI67 gene and the PAM50 proliferation score. We are using computational and laboratorial studies to better understand the role of AURKA in the etiology of invasive breast cancers.

The views expressed in this article are those of the author and do not reflect the official policy of the Department of Defense, or U.S. Government.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P2-05-21.

Abstract P6-02-08: Molecular drivers of adipogenotoxicosis in breast tumor-associated adipose

Background: Having long been thought to function only as an inert energy storage depot, the role of adipose tissue in tumorigenesis has been largely ignored; however, adipose is an active endocrine organ that can directly influence tumor growth. Improved understanding of the role of adipose in tumorigenesis is crucial given the association between obesity and breast cancer risk in post-menopausal women, increasing rates of obesity and use of autologous fat transfer in breast reconstruction.

Methods: Adipose, adjacent to and distant from (>3 cm from the closest tumor margin) invasive breast tumors, was laser microdissected from 20 post-menopausal women, and from 22 post-menopausal women with non-malignant breast disease. Gene expression data were generated using U133A 2.0 microarrays. Data were analyzed to identify significant patterns of differential expression between adipose classes at the individual gene and molecular pathway level. Gene expression differences were validated using qRT-PCR in an additional set of 29 specimens.

Results: SPP1, RRM2, MMP9 and PLA2G7 were expressed at >3-fold (P < 0.01) higher levels in adjacent adipose compared to distant adipose from the same breast. A number of immune response genes including MARCO, FABP7, ELF5, MYBPC1, MMP7, CLDN8, HLA-DQB1 and HLA-DQA1 were differentially expressed in distant adipose compared to adipose from non-malignant breasts. The most significant gene expression differences were detected between tumor-adjacent and non-malignant adipose with >3-fold higher expression of EGFL6 and ITGB2 and >3-fold lower levels of PIP, which are involved in growth, proliferation, and cellular adhesion in adjacent compared to non-malignant adipose. Pathway analysis revealed that immune response differs between non-malignant, distant and tumor-adjacent adipose with an enhanced B- and T-cell response detected in adjacent compared to distant or non-malignant adipose. Inflammatory response as well as DNA transcription and replication pathways were differentially expressed in distant compared to non-malignant adipose.

Conclusions: Gene expression levels differ in breast adipose depending on presence of and proximity to tumor cells. Adipose adjacent to the tumor demonstrated the largest immune response, supporting the idea of adipogenotoxicosis, which through pro-inflammatory and genotoxic responses, promotes tumor development. In addition, genes involved in cellular proliferation, degradation of the extracellular matrix and angiogenesis are differentially expressed in adjacent compared to distant or non-malignant adipose, thus tumor-adjacent adipose may be contributing to the growth and invasion of the primary tumor. These data thus suggest that adipose is not an inert component of the breast microenvironment but plays an active role in tumorigenesis.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P6-02-08.

Abstract P5-01-07: Fibroadenomatoid changes are more prevalent in middle-aged women and have a positive association with invasive breast cancer

Background: The role of benign breast diseases (BBDs) in the development of invasive breast cancers (IBCs) has been studied for many years. Some BBDs have been studied comprehensively (e.g., fibrocystic changes (FCC)) while less is known about other BBDs (e.g., fiboadenomatoid changes (FAC)). FAC has been considered by some researchers as a precursor of fibroadenoma (FA). Conclusions from different studies vary, partially due to different interpretation methods and diagnostic criteria when multiple hospitals and pathologists were involved. In this study, we used subjects in the Clinical Breast Care Project (CBCP) from a military medical center where pathology slides were reviewed by a single breast pathologist to study FAC, FA, and FCC in comparison to the published literature.

Methods: Subjects were enrolled in the study following IRB-approved, HIPAA-compliant protocols. All the clinicopathologic data are available from the CBCP data warehouse (DW4TR). In the CBCP, FCC is composed of 4 components: stromal fibrosis, cysts, apocrine metaplasia, and sclerosing adenosis. Two modeling studies were performed. i) For the BBDs and IBC association study, two groups of subjects were identified: 1136 subjects diagnosed with “Benign” or “Atypical” diseases, and 619 cases diagnosed with IBCs. A logistic regression model was developed for the prediction of IBCs by the 3 BBDs and 2 well-established risk factors (RF): age (younger, <=40; middle-aged, 41–60; older, >60) and race (Caucasian, African American, Asian, and other). ii) For the RF association study with the BBDs, 6 additional RFs reported to be associated with these BBDs were identified from the literature: current use of oral contraceptives, number of live births, education, body mass index, hormonal replacement therapy, and IBC family history. These 8 RFs were used to develop a logistic regression model for each of the BBDs. The analyses were performed in SAS.

Results: In the first study, age and race were confirmed as RFs for IBCs. FAC was positively associated with IBC (OR = 3.04, 95%CI=2.06 to 4.50). FA was negatively associated with IBC, and the level of the association was stronger in women without FCC (OR = 0.15, 95%CI=0.08 to 0.28), compared to women with FCC (OR = 0.40, 95%CI=0.24 to 0.65). FCC was not significantly associated with IBC. Results from the second study indicated that, age was significantly associated with FAC (p = 0.015), specifically the middle-aged women were more likely to have FAC compared to younger women (OR = 2.03, 95%CI=1.23 to 3.34), while the older women were at a non-significantly increased risk. Trends of association with FAC were also noted for the number of live birth (p = 0.095), ethnicity (p = 0.096), and current oral contraceptive pill use (p = 0.077). The FCC model results were in general consistent with the literature, and we also confirmed that age was negatively associated with the diagnosis of FA.

Discussion: Our study was consistent with FCC findings in the literature. We observed that FAC was positively associated with IBC, whereas FA was negatively associated. Also, FAC occurred more often in middle-aged women while FAs occurrence was higher in younger women. Our results suggest that FAC and FA may be two different diseases.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P5-01-07.

P1-06-24: Nuclear Localization of Stat5a Predicts Response to Antiestrogen Therapy and Prognosis of Clinical Breast Cancer Outcome

Nuclear-localized and tyrosine-phosphorylated Stat5 has been reported as a favorable prognostic marker and predictor of response to antiestrogen therapy in breast cancer. Phospho-Stat5 antibodies do not distinguish between phosphorylated Stat5a and the closely related Stat5b, but Stat5a is considered more critical for normal mammary development than Stat5b. The purpose of this study was to determine whether levels of nuclear-localized Stat5a protein (Nuc-Stat5a) were prognostic of clinical outcome or predictive of antiestrogen response. Stat5a was detected by traditional diaminobenzidine-chromogen immunohistochemistry (IHC) and pathologist scoring or by quantitative immunofluorescence in five archival cohorts of breast cancer. Levels of nuclear-localized Stat5a (Nuc-Stat5a) were evaluated by pathologist scoring of whole tissue sections detected by IHC or automated quantitative analysis (AQUA) of immunofluorescently-labeled tissue microarrays. Levels of Nuc-Stat5a were reduced in invasive breast cancer tissues and lymph node metastases compared to normal tissue and ductal carcinoma in situ when quantified by AQUA (Material I; n=180). Tissues from patients not treated with adjuvant therapy or treated with antiestrogen monotherapy were analyzed according to Nuc-Stat5a status for breast cancer-specific survival (CSS) and time to recurrence (TTR) using univariate and multivariate statistical models, adjusting for clinical features including tumor grade, size, lymph node and ER, PR and Her2 status. In two prognostic cohorts of node-negative breast cancer patients, low expression of Nuc-Stat5a, detected by standard IHC (Material II; n=223) or quantitative analysis (Material III; n=198), was prognostic of poor breast cancer outcome as measured by univariate and multivariate CSS (Material II/III) and TTR (Material II). CSS and TTR analysis of two independent materials of tumors from patients treated with antiestrogen monotherapy and analyzed by standard IHC (Material IV; n=73) or quantitative immunofluorescence (Material V; n=97) indicated that patients whose tumors expressed low levels of Nuc-Stat5a were at a greater than 4-fold risk of antiestrogen therapy failure when adjusted for hormone receptor status and clinical features (multivariate CSS: Material IV HR=4.3 (1.2,15.6), p=0.03; Material V HR=5.0 (1.87,13.06), p=0.001). In conclusion, loss of Nuc-Stat5a is a promising independent marker of poor breast cancer prognosis in node-negative, non-adjuvant treated breast cancer patients. Additionally, Nuc-Stat5a may be a useful clinical tool to predict tumor response to antiestrogen therapy.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P1-06-24.

P1-09-01: Effect of Obesity on Gene Expression in Invasive Breast Tumors

Background: Obesity is a risk factor for breast cancer in postmenopausal women and weight gain after diagnosis is associated with decreased survival and less favorable clinical characteristics such as greater tumor burden, higher grade, and poor prognosis, regardless of menopausal status. Despite the negative impact of obesity on clinical outcome, molecular mechanisms influencing breast cancer prognosis remain elusive.

Methods: Postmenopausal women with invasive breast cancer (n=20) in the Clinical Breast Care Project who were obese at diagnosis (BMI ≥30) were matched by age, stage, ER status, HER2 status, and tumor grade to women who maintained a healthy weight (BMI = 18.5−24.9). Sections of pure invasive carcinoma were obtained by laser microdissection from flash-frozen or OCT embedded sections. Gene expression data using U133 2.0 microarrays was generated to compare genome-wide patterns of expression in breast tumors between obese and normal-weight patients. Data were analyzed using Partek Genomics Suite.

Results: Principal Component Analysis accurately clustered the specimens into two groups, healthy weight or obese. Using a false-discovery rate (FDR) p-value <0.05 with a minimum 2.0-fold change of expression, 119 probes from 106 genes were differentially expressed, with 35 genes expressed at significantly higher levels and 71 at significantly lower levels in tumors from obese women compared to those from healthy-weight patients. Hierarchical clustering using these 119 probes reliably partitioned all samples as healthy-weight or obese. Genes differentially expressed include lower expression of several members of the histone cluster 1 family, APOD and IGF1, and higher expression of FABP7, members of the major histocompatibility complex and LTF in tumors from obese women.

Conclusions: Invasive breast tumors from obese women are genetically different from those of healthy-weight women. Altered gene expression may affect tumor cell proliferation and survival, contributing to aggressive phenotypes and altered immune response. These underlying molecular differences may contribute to the less favorable prognosis observed in obese women with breast cancer.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P1-09-01.

P1-03-06: Fibroadenomatoid Changes Have a Higher Occurrence Rate in Middle-Aged Benign Breast Disease Patients with the Trend Retained in Cancer Patients

Background: Fibroadenoma (FA) is a common benign breast lesion known to have a high incidence rate in younger women. There are controversial reports whether FA elevates the risk of developing breast cancers. In clinical practice, FA may be surgically removed due to multiple reasons making it complicated to study its impact on the development of breast cancers that have a higher incidence rate in older women. Fibroadenomatoid change (FAC), also known as fibroadenomatous hyperplasia, is an uncommon lesion with histologic features similar to that of FA but lacking well-defined borders and usually discovered incidentally on breast biopsy specimens. FAC is not surgically targeted. The Walter Reed Army Medical Center, through the Clinical Breast Care Project, has enrolled over 2000 subjects undergoing a biopsy; all the pathology was reviewed by a single pathologist. These subjects provide an opportunity to study the age-dependent pattern of FAC in different patient populations.

Methods: Subjects were enrolled following IRB-approved protocols with data collected through two comprehensive questionnaires, a Core Questionnaire and a Pathology Checklist. A total of 1964 female subjects were identified for this study, including 1135 benign/atypical, 192 in situ, and 637 invasive cancer patients. Patients were divided into three age groups: <=45 years, 46–65 years, and >=66 years. Chi-Square test in the SAS was used for statistical analysis.

Results: As shown in the table, FA occurrence rate decreases significantly with increasing age in benign disease patients. FAC, on the other hand, shows a significantly higher occurrence rate in middle-aged patients with benign findings, and this trend is retained in the invasive or in situ cancer populations. FAC rate is also significantly higher in patients with cancer (invasive, or invasive and in situ combined) compared to benign patients in each age group with p-values ranging from 0.0001 to 0.019 (not shown).

P5-05-03: The Effect of Breastfeeding on Molecular Characteristics of Invasive Breast Cancer

INTRODUCTION: Breastfeeding has been associated with an overall decreased risk of developing breast cancer and the protective effect of breastfeeding has been associated with decreased risk of triple negative and estrogen-responsive breast cancer. Although the mechanism by which breastfeeding provides a protective effect is not well-defined, physical changes in the mammary epithelium reflecting maximal differentiation, may be involved.

METHODS: The database of the Clinical Breast Care Project was queried to identify all patients born between 1946 and 1964 (“baby boomers”) with invasive breast cancer who had given birth to at least one living child. Clinicopathologic characteristics were compared between patients who breastfed for at least six cumulative months and those who never breastfed using chi-square and Fisher's exact tests. RNA was isolated after laser-microdissection of 24 pairs of breastfeeding and non-breastfeeding tumors matched by stage, subtype and grade. Gene expression data was generated using HG U133A 2.0 microarrays and analyzed using the Partek Genomics Suite using a FDR<0.05, 2-fold expression difference to define significance.

RESULTS: Of the 1,008 parous women with invasive breast cancer in the database, 325 (33%) breastfed at least 6 months, 165 (16%) breastfeed less than 6 months, and 518 (51%) never breastfed. Women who breastfed <6 months were not included in additional analyses. Age at menarche and first birth and number of children did not differ between groups; age at diagnosis was significantly (P<0.0001) lower (54.5 years) in women who breastfed compared to those who did not (61.0 years). No differences were detected between patients who breastfed and those who did not for ethnicity or for any tumor characteristics, including tumor stage, grade or size, hormone, HER2 or lymph node status, or subtype. At the molecular level, however, 21 probes representing 19 genes were differentially expressed between women who breastfed and those who did not. Eleven genes, including HAPLN1, had significantly higher and 8 genes, including LGALS7, had significantly lower expression in tumors from women who did not breastfeed.

CONCLUSIONS: Despite the matching of tumor samples, and lack of differences in pathological characteristics between women who breastfed ≥6 months and those who did not breastfeed, differentially expressed genes were identified between tumors from the two groups. Many of the differentially expressed genes have been associated with prognosis, especially in ER negative breast tumors. These data suggest that breastfeeding does alter the underlying molecular characteristics of invasive breast tumors, and may reflect alterations in exposure to estrogen levels during breastfeeding.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P5-05-03.

P3-03-03: Congruence between Patterns of microRNA Expression and Histologic Grading of Invasive Breast Carcinomas

Background

Histologic grading may be used as an indicator of prognosis in breast cancer; patients with low-grade carcinomas have ∼85% ten-year survival compared to just 45% survival in patients with high-grade disease. Although useful for risk stratification, assigning nuclear grade is subjective, and a large proportion of carcinomas are classified as intermediate-grade with uncertain prognosis, thus limiting clinical utility. MicroRNAs (miRNAs) regulate gene expression and serve an important role in breast cancer development. In this study we examined miRNA expression profiles in low-grade and high-grade breast carcinomas to determine if miRNA expression is associated with pathological classifications of tumor grade.

Methods: Breast tumors were obtained from 69 patients enrolled in the Clinical Breast Care Project. Samples were partitioned into low-grade (n=30) or high-grade (n=39) categories using the Nottingham Histologic Score. Following laser microdissection of frozen tissue sections, miRNA was isolated from pure populations of breast tumor cells and hybridized to Affymetrix GeneChip® miRNA arrays containing over 800 human miRNA probes. Expression profiles were analyzed with Partek Genomics Suite using the miRNA Expression Module.

Results: We identified 30 unique miRNAs that showed differential expression at a False Discovery Rate (FDR) p<0.05 between low-grade and high-grade breast carcinomas. Gene targets for these miRNAs function primarily in metabolism and cell communication. Expression of hsa-miR-18a and hsa-miR-572 was significantly different between histologic grades at an FDR p<1×10−8 and hierarchical clustering based on these miRNAs correctly classified 97% (29/30) of low-grade and 90% (35/39) of high-grade tumors. miR-18a has been shown to inhibit ER signaling and promote cellular differentiation, while the role of miR-572 in breast carcinogenesis is not well known.

Conclusions: Dysregulation of miRNAs may accompany changes in cellular morphology typically used in histologic classification of breast carcinomas. Patterns of miRNA expression may improve reproducibility and clinical utility of tumor grading and may prove useful for prediction of recurrence and survival for patients with intermediate-grade carcinomas.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P3-03-03.

P2-01-09: Gene Expression Alterations in the Lymph Node Microenvironment in Response to Successful Metastatic Colonization

Background: Breast stroma is known to play an active role in tumorigenesis, undergoing both phenotypic and molecular changes to facilitate and promote tumor development and growth. The metastatic microenvironment also plays a role in successful colonization; however, the genetic changes in these secondary microenvironments associated with metastasis are not well described.

Methods: Women with invasive breast cancer with at least one lymph node with macrometastases and one lymph node with no detectable metastases were identified from the Clinical Breast Care Project. Lymph node tissue was microdissected from both the metastatic lymph node microenvironment and negative nodes and hybridized to U133A 2.0 gene expression arrays. Differential expression was detected using Partek® Genomics Suite™6.5 using a cutoff of P<0.001, >2-fold change.

Results: Nineteen genes were differentially expressed between negative lymph nodes and lymph node tissue microdissected from lymph nodes with metastatic tumors. Eleven genes, including EPCAM, KRT19 and MUC1 were expressed at significantly higher levels in lymph node tissue from metastatic lymph nodes while eight genes, such as CXCL2 and CXCL5, were expressed at significantly higher levels in negative lymph nodes. Results have been validated in external sample sets for AZGP1, CLEC4M, CXCL2, EPCAM, MUC1, PIP and TFPI.

Conclusions: Lymph node tissue differs in gene expression between those harboring metastatic tumors and those without metastasis. Genes expressed at higher levels in lymph nodes with macrometastases are involved in tumorigenesis, suggesting that like the breast stromal microenvironment, the metastatic microenvironment undergoes crosstalk with the tumor cells. In addition, a cluster of genes involved in immune function are expressed at lower levels in metastatic lymph nodes, suggesting that suppression of proper immune response may be required for successful metastatic colonization.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P2-01-09.

P3-06-06: Comparison of Gene Expression Profiles of Lymph Node Positive and Lymph Node Negative ER Positive Breast Tumors in Pre- and Postmenopausal Women

Background

Breast cancer is the most common female cancer in US and is the second leading cause of cancer related death in women. Metastases are the primary cause of cancer morbidity and mortality. Axillary lymph node (LN) status has long been used as a prognostic factor for breast cancer. The molecular mechanisms that control LN metastasis remains poorly understood. To better understand the various genes and regulatory pathways that drive breast cancer LN metastasis, we compared the gene expression profiles between breast tumors that have metastasized to the LNs and those which have not in pre- and postmenopausal women.

Material and Methods: Tumor cells were isolated from the primary tumors (ER+) of postmenopausal node positive (PMNP; N=20), postmenopausal node negative (PMNN; N=19), premenopausal node positive (PRNP; N=18) and premenopausal node negative (PRNN; N=16) women using laser capture microdissection. RNA was isolated using the RNAqueous®-micro kit (Ambion, Austin, TX). Total RNA was converted to Biotin-labelled aRNA using two rounds of amplification with MessageAmp II aRNA amplification kit (Applied Biosystems, Foster City, CA). The aRNA concentration was determined by Nanodrop 1000 and the quality was assessed with a Bioanalyzer. The aRNA was fragmented and hybridized to Human Genome U133 Plus 2.0 GeneChip (Affymetrix, Santa Clara, CA). Microarray raw data were analyzed using a variety of R programming packages for probe density processing, background correction, normalization, quality control/quality assessment, and calculation of gene expression value, etc. To identify differentially expressed genes, Wilcoxon rank sum test with FDR (false discovery rate) control was performed for pair-wise comparison between different groups. Functional analyses were performed on the identified statistically significant differentially expressed genes to search for the functional categories and pathways in which they are involved and further understand their potential roles in breast cancer metastatic process.

Results: Multivariate data mining (hierarchical clustering analysis and principal component analysis, etc) revealed that in postmenopausal women, the node positive and node negative women are well separated while this was not the case in premenopausal women. Further analysis of the PMNN and PMNP groups to identify differentially expressed genes (with at least a 1.5 fold difference) at FDR =0.1 showed that 232 genes were upregulated and 470 genes were downregulated in PMNP vs PMNN groups. Gene function analysis revealed that genes down regulated in the PMNP group compared to PMNN are related to extracellular matrix, cell adhesion, EGF-like pathway, cytoskeleton etc, while the over-expressed genes are related to cell cycle and cell division, chromosome condensation, etc.

Discussion: The ability to differentiate lymph node positive cases from lymph node negative cases in ER+ breast cancer based on transcriptional profiling may have an impact on the clinical management of ER+ breast cancer cases. Having transcriptional profiles that identify ER+ tumors likely to have poor outcomes would suggest more aggressive treatment for such patients.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P3-06-06.

P3-05-02: Subtype-Specific Co-Occurrence of Atypical Hyperplasia and In Situ Carcinoma with Invasive Breast Cancers

Background

Atypical ductal hyperplasia (ADH), lobular carcinoma in situ (LCIS), and ductal carcinoma in situ (DCIS) are considered risk factors for the development of invasive breast cancer (IBC). The co-occurrence of these lesions with IBC may provide insights into cancer initiation and development. IBC subtypes have distinct clinicopathological features. A clinically practical IHC-based subtyping classification has been developed based on the expression of ER, PR, HER2, and Ki67, defining Luminal A (LA), Luminal B (LB), HER2+, and Triple Negative (TN) subtypes. The Walter Reed Army Medical Center (WRAMC), through the Clinical Breast Care Project (CBCP), has enrolled over 500 IBC subjects with single pathologist review and central lab analysis. The co-occurrence of ADH, LCIS, and DCIS will be studied in relation to IBC subtypes.

Methods: Subjects were enrolled following IRB-approved protocols. IBC patients enrolled at WRAMC were selected and their clinical and pathology data were reviewed. ER and PR positivity is defined as > 5% nuclear staining. The HER2 result is negative if the IHC=0 or 1+ and positive if IHC=3+. For IHC=2+, the FISH result determines the final HER2 status. Ki67 is considered positive if nuclear staining is >= 15%. For IBC subtypes, LA is ER+/HER2−/Ki67-; LB is either ER+/HER2−/Ki67+, or ER+/HER2+; HER2+ is ER-/PR-/HER2+; TN is ER-/PR-/HER2−. Statistical analysis was performed using SAS, and the Chi-Square test was used for categorical data analysis supplemented by the Fisher's Exact test where appropriate. For age analysis, ANOVA was performed with Bonferroni adjustment for multi-pair t-test.

Results: A total of 459 IBC patients were identified and categorized into LA (41.6%), LB (27.7%), HER2+ (10.2%), and TN (20.5%). Many of the previously reported subtype-specific characteristics were confirmed. Age at diagnosis varied by subtype (p=0.0034) with LA being the oldest (Mean±SD=59.9+12.5 years) and TN the youngest (54±12.6 years, p=0.0048). Ethnicity distribution of African American (AA) relative to Caucasian American patients varied significantly in subtypes with AA=18% in LA, 31% in LB, 32% in Her2+, and 42% in TN (p=0.0008). The grade, the AJCC stage and its components T and N were all significantly different among the subtypes (p ranges from <0.0001 to 0.0020). The grades and stages were consistently lowest for LA, highest for HER2+ and TN. We further found that the co-occurrence of ADH, DCIS, and LCIS with IBC were subtype-specific with the following distributions: ADH—LA (25.1%), LB (18.9%), HER2+ (0%), and TN (6.4%) (p<0.0001, n=78); DCIS—LA (63.4%), LB (76.4%), HER2+ (80.9%), and TN (58.5%) (p=0.0039, n=311); LCIS—LA (36.7%), LB (19.7%), HER2+ (4.3%), and TN (6.4%) (p<0.0001, n=103).

Discussion: By including Ki67 in IHC-based IBC subtyping we confirmed many subtype-specific clinico-pathological characteristics in the CBCP WRAMC population. We further report subtype-specific co-occurrences of ADH, DCIS, and LCIS. These co-occurrence patterns may reveal distinct developmental mechanisms between the different subtypes of IBC.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P3-05-02.

Abstract P3-13-02: Ethnicity Difference of Benign Breast Diseases in Breast Cancer and Non-Cancer Patients

Background: Breast cancer (BC) is a heterogeneous disease. Ethnicity differences in BC for Caucasian (CA) and African American (AA) women have been reported, but there is no report on ethnicity differences in benign breast diseases (BBDs) though it is known that BBDs may be precursors or risk factors of BCs. In the Clinical Breast Care Project (CBCP), a comprehensive characterization of BBDs performed on breast biopsies makes it possible to conduct a study on the ethnicity difference of BBDs between CA and AA, in the Cancer and Non-Cancer groups respectively.

Method: CA and AA patients undergoing a biopsy were selected from CBCP, totaling 1,963. A Pathology Checklist is available, reporting 131 pathologic conditions including 83 BBDs. In addition, a Core Questionnaire covering information such as demographics, medical history, risk factors, life style, etc. is completed. The Cancer group is composed of 731 CAs and 170 AAs (including in situ, invasive, and malignant NOS), and the Non-Cancer group is composed of 748 CAs and 314 AAs. BBDs with a frequency of <1% in each group were removed from the study leaving a total of 26 BBDs analyzed for the Cancer and 25 BBDs for the Non-Cancer groups. Pearson's Chi-square test was used to analyze statistical significance.

Results: The Cancer group showed 6 BBDs significantly associated with the ethnicity; 2 were more frequent in CA, i.e., fat necrosis (5.5% CA vs. 0.6% AA, p =0.011), and mild intraductal hyperplasia (6.8% CA vs. 2.4% AA, p= 0.041). The other 4 BBDs were more frequently observed in AA, which were cysts (47.2% CA vs 56.5% AA, p=0.036), multiple papillomas (8.3% CA vs 15.9% AA, p=0.005), moderate intraductal hyperplasia (18.1% CA vs 28.2% AA, p=0.004), and fibroadenomatoid nodule (5.3% CA 11.8% AA, p= 0.004). These BBDs were not significantly different in the Non-Cancer group between the two ethnicities. Six BBDs were significantly associated with the ethnicity in the Non-Cancer group. Four of them were more frequent in CA, including duct ectasia (9.5% CA vs. 3.2% AA, p=0.001), microcalcifications (35.3% CA vs. 27.4% AA, p=0.015), fibocystic changes (61.0% CA vs. 44.6% AA, p= 1.30E-06), and sclerosing adenosis (21.7% CA vs. 13.4% AA, p=0.002). The other 2 BBDs were more frequent in AA, i.e., sclerosing papilloma (0.7% CA vs 2.6% AA, p=0.025), and fibroadenoma (20.9% CA vs 29.9% AA, p=0.002). These BBDs were not significantly different in the Cancer group between the two ethnicities, for example microcalcifications were 50.1% in CA and 55.9% in AA.

Discussion: Multiple papillomas and moderate intraductal hyperplasia are moderate risk factors for BC, and in the Cancer group they were more frequently detected in AA than in CA. Microcalcification as a BC risk factor did not show ethnicity difference in the Cancer group, but was detected more frequently in CA in the Non-Cancer group. It is interesting that not a single BBD was found to be significantly associated with the ethnicity (AA and CA) across the Cancer and the Non-Cancer groups. Thus, the ethnicity difference of BBDs in AA and CA reported here not only suggests possible ethnicity-specific BC risk factors but also generates new hypotheses for future studies.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P3-13-02.

Abstract P3-13-04: The Role of PSPHL in Breast Cancer in African American Women

Background: Younger age at diagnosis and poorer prognosis in African American women (AAW) with breast cancer have led to a number of studies evaluating gene expression differences in tumors from AAW compared to those from Caucasian women (CW). Studies have found that PSPHL expression is significantly higher not only in breast but also in prostate and endometrial tumors from African Americans compared to Caucasians. SNP rs6700 in the 3’ UTR of PSPHL has been found to correlate with expression. Here, we investigated how rs6700 affects PSPHL expression levels in AAW and CW with and without breast cancer.

Methods: Genomic DNA was isolated from 74 women with (35 AAW and 39 CW) and 37 women without breast cancer (23 AAW and 14 CW). rs6700

was genotyped using TaqMan SNP Genotyping assay C__314922_10.

PSPHL expression levels in breast tissue were measured in these same women using TaqMan Gene Expression assay Hs00863464_m1 using RNA isolated from laser microdissected breast tumor cells or normal breast tissue from women with or without breast cancer, respectively.

Results: PSPHL expression in breast tumors from AAW was significantly higher than in CW (P = 6 x 10-6). The median fold change in expression in AAW was 42.2; while only 2 (6%) AAW had no detectable expression of PSPHL, 21/39 (54%) CW did not express PSPHL. The minor allele frequency (T) in AAW with breast cancer was 0.41 compared to 0.06 in CW with breast cancer. In both AAW and CW, median levels of PSPHL expression were lowest (15.45 in AAW, 0 in CW) in women with the C/Cgenotype. In addition, all patients whose tumors had no detectable PSPHL expression had the C/C genotype. In women without breast cancer, the C/C genotype was also correlated with lower PSPHL expression and was more frequent among CW compared to AAW (79% and 39%, respectively).

Conclusions: PSPHL is more highly expressed in both diseased and normal breast tissue from AAW compared to CW and expression of PSPHL correlated with rs6700 genotypes. Although the function of PSPHL is unknown, it may function in cellular proliferation. Higher expression in AAW may, therefore, promote the development of tumors in AAW. It must be noted, however, that the frequency of the C allele is significantly lower in the African American population as a whole (53% vs. 92%); therefore, differences described here may reflect population differences rather than any specific contribution to breast cancer etiology. However, due to the higher expression of PSPHL in African Americans with other hormone-dependent cancers, its role in tumorigenesis requires further study.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P3-13-04.

Abstract P4-07-09: The Impact of Matrix Metalloproteinase-1 Promoter 1G/2G Polymorphism on Breast Diseases

Background: Matrix Metalloproteinase-1 (MMP-1) is a ubiquitously expressed interstitial collagenase. Overexpression of MMP-1 has a role in initiating mammary tumorigenesis by degrading stroma and by releasing growth factors. A single guanine insertion polymorphism in the MMP-1 promoter creates the binding site, 5'-GGAA-3', for the Ets transcription factor, and increases transcription of MMP-1. The MMP-1 2G polymorphism is linked to early onset, increased risk or aggressiveness of several cancers. Its relationship with other potential markers in invasion and metastasis of breast cancer is unknown.

Experimental Design: To study the impact of the 2G polymorphism on breast cancer we analyzed the genotypes of 109 patients (52 invasive breast cancer [IBC], 29 ductal carcinoma in situ [DCIS], 13 atypical ductal hyperplasia [ADH] and 15 benign breast disease). Immunohistochemical (IHC) data for MMP-1, HER2, ER/PR and P53 from these donors were also analyzed. IHC results for MMP-1 were scored as 0 (no expression) or increasing expression of+1, +2 or +3. Data were analyzed using Pearson's chi-square test to identify statistical significance.

Results: A significantly higher number of patients in the IBC group expressed high MMP-1 (+2 and +3; p <0.001) while the benign group had the least number of patients expressing higher MMP-1 (score +3; p = 0.0075). In the IBC group, among patients with low levels of MMP-1 (+1), 57% had the 1G/1G phenotype, and among those expressing high levels of MMP-1 (+2 and +3), over 70% were 1G/2G heterozygotes or 2G/2G homozygotes. The 2G allele frequency in the ADH group was 0.62 and these patients had higher MMP-1 expression (+2 and +3). Further analyses of HER2, ER/PR and P53 in relation to the MMP-1 polymorphism within the IBC group showed MMP-1 allelic variation in Her-2 positive group was significantly different compared with Her-2 negative group (p = 0.039), with a distribution curve shifted to a greater frequency of 2G homozygosity. A similar result was also observed in P53 positive group when compared with P53 negative group (p = 0.043).

Conclusions: 1) In the IBC group, the 2G insertion polymorphism contributes to MMP-1 over expression. 2) Increased expression of MMP-1 in ADH and higher 2G allele frequency are consistent with the hypothesis that increased MMP-1 2G polymorphism plays a role in initiation of ADH through up regulation of MMP-1 expression. 3) Earlier studies show prognostic role for the coexistence of increased expression of HER2 and P53 in breast cancer. Our observation of a significant increase in the 2G homozygotes in HER2 and P53 positive patients supports a prognostic role for this polymorphism and suggests its possible association with other breast cancer markers. Thus, the MMP-1 2G polymorphism may both contribute to breast disease onset and serve as a prognostic marker for breast cancer.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P4-07-09.

Abstract P6-04-10: Genetic Discrimination of Aggressive from Indolent DCIS

Background: Treatment options for DCIS vary from surgical excision with or without radiation and/or chemopreventive therapy, or mastectomy. Intuitively, more aggressive treatment options should lead to improved survival rates, however, studies have shown no difference in breast cancer mortality between women treated with wide excision only versus those with excision plus radiation and treatments can be costly, lengthy and associated with side effects. To avoid over-treating women with indolent disease, while intensively treating women with aggressive disease, new molecular tools must be developed to supplement pathological information to classify DCIS lesions and predict clinical outcome.

Methods: Formalin-fixed paraffin-embedded (FFPE) pure DCIS biopsy specimens were collected from the pathology archives of the Anne Arundel Medical Center. Samples included those with poor prognosis characterized by either recurrence of DCIS or progression to invasive cancer (n=7) and those good prognosis, having ≥5-year disease-free survival (n=10). RNA was isolated after laser-microdissection of pure tumor cells and hybridized to Breast Cancer DSA™ microarrays (Almac Diagnostics). S-way ANOVA was used to account for batch effects and then Support Vector Machine (SVM) was used to identify candidate genes effective at discriminating good from poor prognosis DCIS. Pathway analysis was performed using MetaCore (GeneGeo).

Results: 328 genes were found to be differentially expressed between good and poor prognosis specimens (P<0.01). Preliminary analysis with SVM found that a 70-gene candidate signature from these 328 genes wasoptimal under the tested conditions for discriminating favorable from poor prognosis DCIS. This candidate signature included genes such as MEF2C, PTK2 and ZBTB2. Pathway analysis revealed that genes involved in cytoskeleton modeling, apoptosis and survival, DNA damage repair and cell adhesion are expressed at lower levels in poor prognosis DCIS while those involved in cell cycle, immune response and cell proliferation are expressed at higher levels.

Conclusions: While studies have attempted to identify molecular profiles associated with aggressive DCIS by comparing DCIS co-occurring with invasive disease to pure DCIS, to our knowledge, this is the first study that identified a candidate molecular signature of prognosis in pure DCIS. Although many of the 70 genes found to differ between favorable and poor prognosis DCIS have not been previously associated with breast cancer or have unknown function, MEF2C and PTK2 have been implicated in invasion and migration, while ZBTB2 is a master regulator of p53 and stimulates cellular proliferation. These data demonstrate aggressive DCIS do differ from indolent DCIS at the genetic level and that these differences may be useful in developing molecular tools to classify DCIS lesions and guide appropriate treatment.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P6-04-10.

Abstract P3-01-04: Differential Gene Expression Analysis among Post-Menopausal Caucasian Invasive Breast Cancer, Benign and Normal Subjects

Background: Breast Cancer (BC) is the second leading cause of cancer death among women in United States. There have been studies aiming to develop a blood-based BC detection assay, but most were focused on comparing blood samples between BC and normal subjects. Including benign disease patients in the study is critical in developing a BC detection assay. In addition, BC is a heterogeneous disease with distinct characteristics being associated with ethnicity and menopausal status. Thus we designed a study to analyze gene expression, in peripheral blood samples from invasive BC, benign disease and normal subjects, stratified by menopausal status and ethnicity. Here we report the results from Caucasian postmenopausal women.

Method: Subjects were selected from the Clinical Breast Care Project (CBCP). Microarray data using Affymetrix GeneChip Human Genome U133 plus 2.0 arrays, of peripheral blood samples from pathologically confirmed invasive BC (n= 17) and benign patients (n= 17) were compared to those from normal subjects (n=17). All subjects were postmenopausal Caucasian women, matched for age. Using GenespringGX 11.0 software, the data were normalized, and QC on hybridization and internal controls were performed before filtering out probesets with the lowest 20% signal intensity in each array. Asymptotic t-test with FDR correction was used and significant pathways were obtained.

Results: Comparison of gene expression in invasive BC patients (mean±SD = 63.7±7.0 years old) vs. normal subjects (mean±SD = 63.2±7.4 years) identified 1102 significantly different genes, satisfying the thresholds of corrected p < 0.05 and Fold Change (FC) > 1.5. Of them, 1003 genes were up regulated and 99 genes were down regulated in BC patients. Comparison of gene expression in benign disease women (mean±SD = 61.4±10.4 years) vs. normal women showed 1320 significantly different genes (corrected p < 0.05, FC > 1.5), with 1121 genes being up regulated and 199 genes being down regulated in benign patients. Of the top 10 genes with highest FC values, 6 genes (MBNL1, FAR1, MDM4, ITGA4, RAB8B, and EXOC5) were common in both analyses and were up regulated. NOTCH pathway was up regulated in both benign (p=0.038) and invasive (p=0.042) groups. Similarly, IL-7 pathway was up regulated in both benign (p=0.047) and invasive (p=0.030) patients. TCR pathway (p=7.36E-04) was up regulated in benign group only. The benign vs. invasive BC subjects did not show significant differential gene expression.

Discussion: Our results provide a list of differentially expressed genes that are mostly up regulated in invasive and benign patients vs. normal subjects. NOTCH pathway is involved in cell-cell communication and angiogenesis. IL7 pathways play an important role in immune system response, cell proliferation and cell survival signaling. TCR pathway is highly significant only in benign patients and stimulation of TCR pathway induces a signaling cascade that ultimately results in activation of induced cell death. This activation could be early body response to prevent cancer development in benign subjects. When more microarray analyses are completed for this study, we hope to obtain a better understanding of the possibility of developing a blood-based BC detection assay.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P3-01-04.

An Approach To Correlate the Temporal Information To Facilitate Specimen Selection in the Breast Cancer Research Project

Background: Temporal information management is very important in translational research. In the Clinical Breast Care Project (CBCP), the information on subjects and their specimens may be collected at multiple time points using multiple instruments.All such information is stored in an in-house data warehouse. Currently, 4000+ subjects have been enrolled in the study following HIPAA-compliant IRB-approved protocols with 35,000+ specimens collected. Some of the patient's information is static but other data are time dependent. As a result, selecting samples for experiments is a challenge due to complicated temporal relationships between samples and information collected through various instruments.Methods and Results: In the CBCP, the clinical information, blood, and solid tissues of a subject may be collected at multiple time points, associated with the completion of a Core Questionnaire (CQ) for clinical information, and/or a Pathology Checklist (PC) for pathology and sample information. We have designed and implemented an algorithm to use a set of pre-defined flags to precisely describe each sample related to patient's clinical and pathology information in the temporal domain. Five categories (flags) were created to describe the relationship between the sample date (SD) and the CQ date based on whether SD is within 60 days of the CQ date or there is missing data or not. The relationship between blood samples and pathology information is more complicated. Within 90 days, any of the 15 surgical procedures might be performed on a patient and blood samples might be collected before, at the time of, or between any procedures. For some experiments, it is crucial to select blood samples taken before tumor is impacted or severely impacted. Thus, we defined a dozen categories to describe the relationship between the SD and the procedure date (PD), including when the SD is earlier than any PD, equals to the first PD, or between certain procedures. Using these flags we have characterized the relationships between SDs and CQ dates, and between SDs and PDs for all the samples and all the subjects, and stored all the information into two relational tables. The temporal criteria for sample selection are now represented by the relationships between these flags, and can be implemented through several filtering processes. The described algorithm drastically reduces the time needed for precise sample selection from several days for manual efforts to several hours.Discussion: We are in the process of developing a general data model for temporal information management. The method described here is a transitional solution that fulfills our current needs. As an initial effort some of the thresholds for categorizing different temporal conditions are arbitrary, and we are validating them with experimental results for future improvement. Nonetheless, this algorithm has greatly enhanced the efficiency of our subject and specimen selection for wet bench experiments. The same principle can be applied to the future temporal data model solution, for CBCP and other human disease studies.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 4173.

Genetic Characterization of Columnar Cellular Lesions and Atypical Ductal Hyperplasia of the Breast

Aims: Columnar cell lesions (CCL) and atypical ductal hyperplasia (ADH) frequently coexist and share molecular changes with in situ and invasive components, suggesting that CCL and ADH may be precursors to breast cancer. These conclusions are, however, largely based on studies examining CCL and/or ADH from patients diagnosed with more advanced disease. Thus, we assessed allelic imbalance (AI) in pure CCL or ADH specimens to characterize molecular changes in early breast lesions.Methods and results: DNA samples were obtained from laser microdissected lesions from CCL with (n=42) or with ADH (n=31) without concurrent in situ or invasive disease. AI was assessed at 26 chromosomal regions commonly altered in breast cancer. The average AI frequency was 6.2% (range 0-20%) in CCL and 6.1% (range 0-25%) in ADH with no significant difference in levels of AI between CCL and ADH. The highest levels of AI were on chromosomes 8q24 and 17q21 in ADH (23%) and CCL (15%), respectively.Conclusions: In conjunction with low overall levels of AI, chromosomal alterations that characterize low- and high-grade in situ and invasive disease were not frequent in pure CCL and ADH. Thus, pure lesions are not genetically advanced and are molecularly distinct from synchronous lesions.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 5156.

Converting Paper Medical Records to Electronic Version To Support Breast Cancer Translational Research and Clinical Practice

Background: The healthcare system has extremely large volumes of patients' paper medical records (PMRs) scattered throughout various medical facilities. Currently the industry is transitioning to Electronic Medical Records (EMR). Although each source of information is equally important, a complete longitudinal health record is rarely available or currently attainable. The goal of this effort is to convert the existing PMRs into searchable electronic equivalents and merge them with existing EMRs to create a more complete longitudinal health record to support clinical care and biomedical research.Method: Using a subset of PMRs for subjects enrolled in the Clinical Breast Care Project at Walter Reed Army Medical Center, (WRAMC), we are developing an automated method to digitize and index the records, extract the biomedical information and prepare the data for delivery to clinicians and researchers. The electronic records were loaded into a database for immediate access by clinicians. To support translational research, MRs need to be de-identified, and information extracted and loaded into a research database; we used ten MRs to develop the operational method. Several methods were tested in the de-identification process: 1) manually, by striking out the protected health information (PHI) on paper before it was digitized and 2) “electronically”, by redacting the record electronically on the computer after it was digitized. We are in the process of testing automated data extraction tools and natural language processors to automatically de-identify and extract data from the EMR.Result: We quickly realized that only 10% of PMRs existed onsite at WRAMC; remaining MRs were held by the patients or other medical facilities. Approximately 300 PMRs, containing 66,600 pages, were scanned and digitized for this project. We have created a successful digitization process, which includes creating PDFs with hidden and searchable information. We have compared the effort and accuracy of various redaction methods. To de-identify 10 records, it took ∼10 hours manually and ∼20 hours electronically. It took longer electronically because of the preparations to ensure the removed information could not be retrieved. We expect that automated redaction tools will greatly reduce that effort. We also found that the electronic method had a 99% accuracy compared to 96% for the paper method. A portal prototype to allow access to medical records by clinicians and researchers is currently being tested and evaluated.Discussion: Conversion of non-searchable data into an explorable and computable format will enable clinicians to acquire needed information more conveniently in their clinical care including treatment plan development. Similarly, properly de-identified, complete MRs will serve as a rich source of clinical information to support translational research. Although the method we are developing will initially satisfy the CBCP need for clinical service and research, it will be further developed into a full solution to expand into other disease condition fields.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 5123.

Proteomic Discrimination of Well- from Poorly-Differentiated Breast Carcinomas

Background: Pathological grade is a useful prognostic factor for stratifying breast cancer patients into favorable (well-differentiated tumors) and less favorable (poorly-differentiated tumors) outcome groups. The current system of tumor grading, however, is highly subjective and a large proportion of tumors are characterized as intermediate-grade, making determination of optimal treatments difficult.Methods: Primary breast tumor specimens from patients diagnosed with well- (n=27) and poorly-differentiated (n=51) invasive ductal carcinoma were obtained from patients enrolled in the Clinical Breast Care Project. Frozen tissues were sectioned and mounted on gold coated MALDI target plates for protein expression profiling. Hematoxylin and eosin (H&E) stained slides were prepared from serial sections for histological characterization. MALDI matrix was deposited as individual spots on the tissue sections in a histology directed manner to assay specific areas and tissue types of interest. Mass spectral data were then acquired from multiple sites across each tissue section.Results: 129 features were observed in well-differentiated and 132 in poorly-differentiated tumors. While the majority of features detected were similar between the two groups, 6 protein features were expressed at significantly lower and 12 at significantly higher levels in the poorly-differentiated tumors, including increased expression of Calgranulin A and Calgizzarin.Conclusions: Protein expression differences detected here suggest that well- and poorly-differentiated invasive breast tumors are molecularly distinct diseases and that these protein changes may contribute to the structural integrity of the tumor cell. In particular, calgranulin A and calgizzarin are members of the S100 protein family, and function in processes such as cell proliferation and differentiation. Further refinement of this differentiation protein signature may not only improve our understanding of the biological processes involved with tumor grade but provide pathologists with new molecular tools to classify breast tumors and reduce the subjectivity associated with current grading criteria.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 6126.

Role for CD44 in Enhancing Invasion, Migration, and Growth of Triple Negative (TN) Breast Cancer Cells

Triple negative (TN) breast cancers are defined by a lack of expression of estrogen, progesterone, and her-2/neu receptors. It is widely recognized that TN breast cancers have a poorer prognosis than any other subtypes of breast caner. Given the lack of effective targeted therapies for TN breast cancer patients, understanding of the mechanisms of growth and invasion in the tissues provides insight into developing novel approaches to lower the mortality from TN breast cancer.Neoplastic epithelial cells in breast carcinomas interact with various components in the tissue microenvironment including extracellular matrix (ECM) and mesenchymal cells. Recent studies identified CD44 as a metastasis-related molecule with multiple functions by promoting cell-cell and cell-ECM interactions. CD44 is an integral transmembrane protein encoded by a single 20-exon gene. In the standard form (CD44s), 10 of the 20 exons are translated. Multiple variant isoforms exist (CD44v1-10) which arises from alternate mRNA splicing of the remaining 10 exons. In contrast to the ubiquitous expression of the standard form of CD44, splice variants are highly restricted in their expression in normal or malignant tissues. Indeed, CD44 variants containing v3, v5, v6, v7-8, v10 exons are expressed in malignant breast cancer tissues. However, there is limited information regarding the biological functions of these exons to promote tumor invasion and metastasis. The goal of this study is to evaluate specific exon(s) of CD44 expressed on TN breast cancer cells for promoting tumor progression and metastasis.In order to approach this goal, we utilized three TN cell lines (HCC38, HCC1937, and HCC1806) as model systems to evaluate CD44 in regulating invasion, migration, and growth in extracellular matrix (ECM) environments. Reverse transcriptase (RT)-PCR analysis using exon specific primers indicate that these cells expressed CD44v8-v10 and CD44s. We demonstrated that an inhibitory antibody against exon v10 of CD44 significantly inhibited b1 integrin-mediated migration and invasion into Matrigel and type I collagen gel. Importantly, this antibody also inhibited three dimensional (3D) growth which is a b1 integrin-independent process. The significant inhibition of these processes was also achieved when a FLAG-fusion exon v10 peptide (FLAG-v10, in which FLAG is tagged at the N-terminal of the peptide) was used as an inhibitor, implying that this exon would function to assemble molecular complexes on TN breast cancer cells that facilitate invasion, migration, and growth. Thus, these results suggest that generation of small synthetic molecules that block the functions of exon v10 of CD44 is promising approaches to inhibit invasion and metastasis of TN breast cancer cells.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 6161.

Microarray Data Analysis Using Peripheral Blood Samples Suggests Differential Enrichment of Signaling Pathways between Breast Cancer Patients and Normal Subjects

Background: Breast cancer is the second most common cancer among women in the United States and the second leading cause of cancer death in women. Early and accurate diagnosis is crucial for better treatment and reduction of mortality. Diagnostics based on blood samples are being developed for many diseases including a variety of cancers.Gene expression profiling has been widely used to investigate the mechanisms of tumorigenesis with the goal of developing novel treatment strategies. We have applied microarray technology to study the gene expression signatures in blood samples from patients with invasive breast cancer, with benign breast disease and from disease free (normal) control groups to find potential diagnostic patterns. We have identified differentially expressed genes in each group and investigated the pathways these genes are involved in and how these are regulated.Material and Methods: Human blood samples from 102 invasive breast cancer patients, 57 patients with benign breast disease and 102 normal controls were collected for this study. Patients are enrolled in the Clinical Breast Care Project (CBCP) following HIPAA-compliant IRB approved protocols, with proven breast pathology diagnosis categories. Affymetrix HG-U133 Plus 2 GeneChips were used to investigate the gene expression profile. Microarray experiments were performed following Affymetrix protocols. CEL files were then analyzed with RMA algorithm for the calculation of gene expression matrix. Differentially expressed genes were identified between different groups using Wilcoxon rank sum test. We also applied Gene Set Enrichment Analysis (GSEA) to the entire gene expression profiles to investigate the gene sets or pathways enriched in different groups.Results: We identified about 2,051 differentially expressed genes between normal and invasive groups (p < 0.001 and Fold Change > 1.2). Furthermore, about 445 and 51 genes were identified in normal vs. benign and benign vs. invasive groups respectively. We performed GSEA using 395 gene sets from pathway databases, initially focusing on normal vs. invasive groups. 123 gene sets are highly expressed in the normal group and 272 gene sets are highly expressed in the invasive group. For the normal group, 10 gene sets are significantly enriched and 22 gene sets are significantly enriched in the invasive group (p < 0.05). These results show that some important pathways are down regulated in the invasive group, such as, the B-cell antigen receptor pathway, the BCR signaling pathway, the T-cell signal transduction pathway, the IL4 receptor pathway, the PIP3 pathway, the ERK pathway, etc. The pathways up regulated in the invasive group are oxidative phosphorylation pathway, ATP synthesis, etc.Discussion: Our results suggest that a list of genes differentially expressed in different groups and they may be used to compose cancer marker panels which can be integrated with currently clinical procedures for cancer diagnosis. We also find some crucial pathways enriched in each group. Most of the pathways downregulated in the blood of patients in the invasive group are related to the immune response. In contrast, most of pathways upregulated in the blood of patients in the invasive group are associated with metabolism.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 3024.

Differential Benign Breast Disease Co-Occurrence with Cancer in Caucasian and African American Women

Background: Breast Cancers (BCs) in Caucasian (CA) and African American (AA) women have different characteristics. Recently, there have been reports, mostly focusing on Caucasians that Benign Breast Diseases (BBDs) may be risk factors for BCs. There are also a few reports of different co-occurring patterns of BBDs with BCs between the two populations. In the Clinical Breast Care Project (CBCP) more than 4,000 subjects have been enrolled following HIPAA-compliant, IRB-approved protocols, with more than 35,000 biospecimens collected. For patients requiring biopsies, an expanded pathology report for research is completed. The occurrences of any of the 131 pathologic conditions, including 66 BBDs, are recorded. All the results are reviewed by the same pathologist. The CBCP provides a uniformed pathology dataset for a comprehensive characterization of the association of BBDs with BCs.Method: The diagnoses were made from potentially multiple biopsies obtained over a short period of time, mostly within 60 days. CBCP subjects with BBD diagnosis and ethnicity information from their most recent visit were selected, giving a dataset totaling 1479 CA and 484 AA women. We studied the association between BBDs and BCs (including in situ, invasive, and malignant NOS) in these two groups using the chi-square test.Results: In both populations 6 BBDs are positively associated with BCs, including atypical ductal hyperplasia (ADH) (4% vs 15% for AA p<3.8E-05, meaning 4% cancer-free cases with ADH compared to 15% cancer cases co-occurring with it; 7% vs 16% for CA p<1.9E-08), microcalcifications (27% vs 56% for AA p<1.2E-09, 35% vs 53% for CA p<8.0E-12), multiple (peripheral) papillomas (6% vs 16% for AA p<0.0005, 4% vs 8% for CA p<0.0002), columnar cell hyperplasia (9% vs 21% for AA p<0.0003, 6% vs 18% for CA p<3.5E-12), and moderate intraductal hyperplasia (IDH) (14% vs 28% for AA p<0.0002, 14% vs 18% for CA p<0.03). On the contrary, 3 BBDs are negatively associated with BCs including fibroadenoma (30% vs 10% for AA p<4.9E-07; 21% vs 7% for CA p<1.0E-13), and mild IDH (10% vs 2% for AA p<0.003; 13% vs 7% for CA p<0.0002). Eight BBDs differentially co-occur with BCs between AA and CA. BCs in AA are associated with sclerosing adenosis (13% vs 31%, p<4.5E-06), fibrocystic changes (45% vs 61%, p<0.001), fibroadenomatoid nodule (6% vs 12%, p<0.042), and cysts (42% vs 57%, p<0.004). BCs in CA are positively associated with columnar cell hyperplasia with atypia (3% vs 8% p<3.6E-05), atypical lobular hyperplasia (1% vs 5%, p<6.0E-05), and radiation changes (0% vs 1.4% p<0.004), but negatively associated with duct ectasia (10% vs 2%, p<7.9E-10).Discussion Our findings are in accordance with established BC risk factors such as ADH and moderate IDH for both ethnic groups. We further found that 8 BBDs differentially co-occur with BCs between AA and CA groups, and interestingly the 4 significant BBDs detected in the AA group typically show >10% increased co-occurrence rate in cancer cases compared to cancer-free cases, whereas the 4 BBDs detected in the CA group almost always show a <10% co-occurrence rate overall. More research is undergoing to understand the implications of this distinct co-occurrence pattern.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 3066.

Admixture mapping to identify breast cancer susceptibility loci in African American women

Abstract #2090

Background: The incidence of breast cancer in young women is higher in African American (AAW) compared to Caucasian (CW) and is associated with an aggressive phenotype, including high-grade, ER/PR and HER2 negative status and p53 positive characteristics. The aggressive tumor behavior and pathological features of AAW are also seen in women from the Bite of Biafra, an area of Western Africa that was the predominant source of the American slave trade. Similar tumor characteristics and shared ancestry suggest that population-specific molecular factors contribute to the aggressive tumor phenotypes associated with AAW. Therefore, a method of admixture mapping is proposed for this study. Admixture mapping is also known as mapping by admixture linkage disequilibrium. It assumes in a recent admixed population, if certain phenotype or the prevalence of a disease is mainly caused by one of the ancestral groups, the genetic variant can be located by the elevated ancestry from that group.
 Methods: Pathological characterization was performed histology specimens from 91 self-described AAW diagnosed with invasive breast cancer enrolled in the Clinical Breast Care Project. Genomic DNA was isolated and SNP data generated using the Affymetrix Mapping 100K arrays. African ancestry was determined using maximum likelihood estimator and chromosomal regions associated with disease identified using AncestryMap and AdmixMap.
 Results: The average age of diagnosis was 51 years. 25% of patients had poorly-differentiated, triple negative tumors. All 91 patients had significant African ancestry, with a mean level of Nigerian ancestry of 0.3857. Results from both AncestryMap and AdmixMap suggest loci on chromosomal regions 5q32.2, 9q13.1 and 17p13.3 are associated with the development of breast cancer.
 Discussion: Recently, admixture mapping was used to successfully identify a prostate cancer gene to chromosome 8q24 in African American men. Using a similar approach, we identified three chromosomal regions associated with the presence of invasive breast cancer in AAW with significant Western African ancestry. Of note, loss of 9q13 has been associated with poor survival, and allelic imbalance of 17p13.3 has been associated with loss of hormone receptor expression and poor prognosis. Thus, the use of admixture mapping may aid in the identification of genes involved in the development of the aggressive form of breast cancer associated with AAW.

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 2090.

Identification of breast cancer metastasis initiation and virulence genes

Abstract #2049

Background: Metastasis involves processes such as invasion, angiogenesis, intravasation, extravasation and eventual proliferation in the target organ. Metastasis initiation genes are those involved in dissemination of cells from the primary tumor while metastasis virulence genes contribute to metastatic colonization, specifically at the secondary site. Identification of metastasis initiation and virulence genes is critical to improve our understanding of and the ability to effectively treat metastatic breast cancer.
 Methods: A 51-gene signature was previously identified using RNA from 20 primary breast and matched metastatic lymph node tumors using the U133 2.0 microarrays (Affymetrix). Real-time PCR was performed for each assay using TaqMan assays (Applied Biosystems) using RNA from 25 pairs of primary breast and matched metastatic lymph node tumors. Data was evaluated using Mann-Whitney testing.
 Results: Gene expression levels were significantly different for 40 genes including 25 with P<0.001 and 15 with P<0.05. Thirteen of these genes had significantly higher levels and 27 had lower levels of expression in metastatic tumors. Ten genes had >50-fold difference in expression including WNT2, KRT14, TAC1, COL11A1, MMP13, GRP, and KERA with higher expression in primary tumors and EPHA3, PAX5 and NTS with higher expression in the metastasis.
 Discussion: A breast cancer metastasis signature involving 40 genes has been identified and validated. Genes expressed at higher levels in primary breast tumors are largely involved in degradation of the extracellular matrix (ECM), likely enabling cells with metastatic potential to disseminate. Decreased expression of these genes in the metastatic tumors suggests that once tumor cells have disseminated, invasion into foreign tissues does not require active tissue remodeling. Rather, cells that have successfully metastasized are characterized by the expression of genes involved in transcription, metabolism and immune response, potentially blocking cellular differentiation and providing cells with proliferation and survival advantages. These data not only improve our understanding of the biological processes involved in successful metastatic but provide new targets to arrest tumor cells dissemination and metastatic colonization.

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 2049.

Fingerprinting genomic heterogeneity in primary breast carcinomas and among sentinel lymph node metastases: implications for clinical management of breast cancer patients

Abstract #1059

Background: Sentinel lymph node (SLN) status is a key prognostic factor for breast cancer patients. The SLN hypothesis postulates that cancer cells initially spread from the primary tumor to the first-draining (sentinel) node(s) before reaching higher echelon nodes. Although SLN biopsy is clinically useful for staging patients, little is known about the genomic heritage of cells that define the genetic makeup and clinical behavior of nodal metastases. We used allelic imbalance (AI) coupled with high-density SNP genotyping to examine the extent of genetic heterogeneity in primary tumors and relationships among metastases in sentinel and non-sentinel axillary lymph nodes in 30 patients with node positive breast cancer.
 Material and Methods: Pathologically positive nodes were identified by H&E/IHC, while sentinel nodes were localized by standard scintigraphic techniques. For each patient, nodal metastases and 6-15 areas from multiple paraffin blocks of the primary carcinoma were isolated by microdissection. AI was assessed by microsatellite typing at 26 chromosomal regions; genome-wide copy number variation was examined using Affymetrix GeneChip Human Mapping 500K arrays and the Genotyping Consol. The genomic heritage of sentinel and non-sentinel nodal metastases in relation to multiple areas of the primary tumor was assessed by hierarchical clustering and phylogenetic analyses.
 Results: Extensive genomic heterogeneity was observed in primary tumors and among nodal metastases (0-44% in primaries, 0-36% in metastases). Overall levels of variation were significantly higher (P<0.01) in primary tumors (17%) than metastases (10%), but did not differ between sentinel (10.2%) and non-sentinel (10.0%) metastases. Within patients, many (but not all) metastases appeared to be descended from different areas of the primary tumor, but patterns of descent were complex. As a group, sentinel node metastases were not genomically distinct from other axillary node metastases, and there was no evidence that SLN metastases are seeded by cells that colonize the axilla early in tumorigenesis.
 Discussion: Although metastases develop in a sequential manner that maintains the known capability of SLN biopsy to accurately stage the axilla, metastases appear to develop from genomically-diverse progenitor cells descended from different areas of the primary tumor. Genomic diversity in primary tumors and metastases may be an important factor in prognosis and treatment resistance in breast cancer patients.

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 1059.