Search

Synchronous early-stage endometrioid endometrial carcinomas (EECs) and endometrioid ovarian carcinomas (EOCs) are associated with a favorable prognosis and have been suggested to represent independent primary tumors rather than metastatic disease. We subjected sporadic synchronous EECs/EOCs from five patients to whole-exome massively parallel sequencing, which revealed that the EEC and EOC of each case displayed strikingly similar repertoires of somatic mutations and gene copy number alterations. Despite the presence of mutations restricted to the EEC or EOC in each case, we observed that the mutational processes that shaped their respective genomes were consistent. High-depth targeted massively parallel sequencing of sporadic synchronous EECs/EOCs from 17 additional patients confirmed that these lesions are clonally related. In an additional Lynch Syndrome case, however, the EEC and EOC were found to constitute independent cancers lacking somatic mutations in common. Taken together, sporadic synchronous EECs/EOCs are clonally related and likely constitute dissemination from one site to the other.

Aims The aims of this study were to perform a whole-exome sequencing analysis of a breast cylindroma and to investigate the role of molecular analyses in the differentiation between breast cylindroma, a benign tumour that displays MYB expression, and CYLD gene mutations, and its main differential diagnosis, the breast solid-basaloid adenoid cystic carcinoma, a malignant tumour that is characterized by the presence of the MYB–NFIB fusion gene and MYB overexpression. Methods and results A 66-year-old female underwent quadrantectomy after an irregular dense shadow was discovered in the right breast at the screening mammogram. Histologically, the tumour displayed features suggestive of a solid-basaloid variant of adenoid cystic carcinoma with a differential diagnosis of cylindroma. Fluorescence in situ hybridization, reverse transcription–polymerase chain reaction, immunohistochemistry and whole-exome sequencing revealed absence of the MYB–NFIB fusion gene, low levels of MYB protein expression and a clonal somatic CYLD splice site mutation associated with loss of heterozygosity of the wild-type allele. Conclusions The results of the histological, immunohistochemical and molecular analyses were consistent with a diagnosis of breast cylindroma, providing a proof-of-principle that the integration of histopathological and molecular approaches can help to differentiate between a low-malignant potential and a benign breast tumour of triple-negative phenotype.

Purpose Lobular carcinoma in situ (LCIS) has been proposed as a non-obligate precursor of invasive lobular carcinoma (ILC). Here we sought to define the repertoire of somatic genetic alterations in pure LCIS and in synchronous LCIS and ILC using targeted massively parallel sequencing. Methods DNA samples extracted from microdissected LCIS, ILC and matched normal breast tissue or peripheral blood from 30 patients were subjected to massively parallel sequencing targeting all exons of 273 genes, including the genes most frequently mutated in breast cancer and DNA repair-related genes. Single nucleotide variants and insertions and deletions were identified using state-of-the-art bioinformatics approaches. Results The constellation of somatic mutations found in LCIS (n = 34) and ILC (n = 21) were similar, with the most frequently mutated genes being CDH1 (56% and 66%, respectively), PIK3CA (41% and 52%, respectively) and CBFB (12% and 19%, respectively). Among 19 LCIS and ILC synchronous pairs, 14 (74%) had at least one identical mutation in common, including identical PIK3CA and CDH1 mutations. Paired analysis of independent foci of LCIS from 3 breasts revealed at least one common mutation in each of the 3 pairs ( CDH1 , PIK3CA , CBFB and PKHD1L1 ). Conclusion LCIS and ILC have a similar repertoire of somatic mutations, with PIK3CA and CDH1 being the most frequently mutated genes. The presence of identical mutations between LCIS–LCIS and LCIS–ILC pairs demonstrates that LCIS is a clonal neoplastic lesion, and provides additional evidence that at least some LCIS are non-obligate precursors of ILC.

Background HER3 activating mutations have been shown in preclinical models to be oncogenic and ligand-independent, but to depend on kinase-active HER2. Patients and methods Whole-exome sequencing of the primary HER2-negative breast cancer and its HER2-negative synchronous liver metastasis from a 46-year-old female revealed the presence of an activating and clonal HER3 G284R mutation. Results HER2 dual blockade with trastuzumab and lapatinib as third-line therapy led to complete metabolic response in 2 weeks and confirmed radiological partial response after 8 weeks. Following the resection of the liver metastasis, the patient remains disease-free 40 weeks after initiation of the HER2 dual blockade therapy. Immunohistochemical analysis demonstrated a substantial reduction of phospho-rpS6 and phospho-AKT in the post-therapy biopsy of the liver metastasis. Discussion This is the first-in-man evidence that anti-HER2 therapies are likely effective in breast cancers harboring HER3 activating mutations.

LBA11516 Background: ctDNA assays can noninvasively assess tumor burden and biology by identifying tumor-derived somatic alterations. For broad applicability, including early cancer detection, an unprecedented high-intensity approach (ultra-deep sequencing of plasma cell-free DNA (cfDNA) with broad genomic coverage) is needed to address intra-patient and population-level heterogeneity. We present initial results with this approach in patients (pts) with metastatic breast (BC), non-small cell lung (NSCLC), and castration-resistant prostate cancer (CRPC). Methods: Blood and tissue were prospectively collected w/in 6 wks with no intervening therapy change from pts with de novo or progressive cancer. cfDNA and white blood cell (WBC) genomic DNA from each pt were sequenced with a 508-gene panel (2 Mb; >60,000X raw depth). cfDNA variant calling used molecular barcoding for error correction and filtering for WBC variants. Tissue was sequenced using the MSK-IMPACT assay (410 genes, 1.4 Mb, >500X depth) blinded to plasma/WBC sequencing. Variants were classified as clonal or subclonal based on tumor sequencing in BC and NSCLC. Results: Of 161 eligible pts, 124 (39 BC, 41 NSCLC, and 44 CRPC) were evaluable for concordance. In tissue, 864 variants were detected across the 3 tumor types, with 627 (73%) also detected in plasma: single nucleotide variants/indels - 75%, fusions - 67%, and copy number alterations - 58%. In 90% of pts, at least 1 of the variants detected in tumor tissue was also detected in plasma: BC - 97%, NSCLC - 85%, CRPC - 84%. Most actionable mutations detected in tissue were also detected in plasma (54/71, 76%; SNVs only: 28/31, 90%). A subset of driver mutations (eg. in ESR1, PIK3CA, ERBB2, EGFR) were observed in plasma but not tissue. Clonal variants in tissue were more likely to be detected in plasma than subclonal variants (p

11526 Background: A large fraction of cfDNA fragments are derived from hematopoietic sources. Somatic alterations in cfDNA can be tumor-derived but also could represent somatic changes associated with clonal hematopoiesis. We performed deep sequencing of both plasma cfDNA and matched white blood cell (WBC) genomic DNA (gDNA) to determine the contribution of clonal hematopoiesis to the variants observed in cfDNA. Four cohorts were investigated: metastatic breast (BC), non-small cell lung (NSCLC), castration-resistant prostate cancer (CRPC), and non-cancer participants (pts). Methods: Metastatic cancer pts with de novo or progressive disease were prospectively enrolled. Non-cancer pts were blood bank donors. Plasma cfDNA and matched WBC gDNA were sequenced using a targeted 508-gene panel (2 Mb) to > 60,000X raw depth. Variant calling used a novel pipeline that employed molecular barcoding for error suppression followed by de novo assembly and graph-based variant calling. Results: Of 151 metastatic cancer pts (48 BC, 49 NSCLC, 54 CRPC), median age was 64 (30-87) with 53% female and 33% treatment naive. Of 47 non-cancer pts, median age was 61 (20-78) with 51% female. Analysis of cfDNA identified 1072 variants (AF > 0.1%, > 2 mutant reads, passing bioinformatic quality filters) which were also detected in WBC gDNA as non-germline ( < 35% allele frequency [AF]) non-synonymous variants. For these cfDNA variants, AF ranged from 0.1-14.4% and correlated with AF in WBC gDNA (r2 = 0.47, p < 0.001). Mutated genes were consistent with clonal hematopoiesis, with the most frequently mutated genes being DNMT3A, TET2, PPM1D, and TP53 (215, 77, 45, and 36 variants, respectively). For both cancer and non-cancer pts (age > 45), median number of overlapping variants was 5 per pt (range 0-22). The number of WBC gDNA and cfDNA variants per individual was positively associated with age (p < 0.001) in both cancer and non-cancer pts (interaction p = 0.08). Conclusions: Somatic cfDNA variants are frequently derived from clonal hematopoiesis and increase with age. Accurate assessment of somatic alterations in cfDNA should account for this phenomenon to distinguish between tumor-derived and WBC-derived variants.

Purpose: Paired primary breast cancers and metachronous metastases after adjuvant treatment are reported to differ in their clonal composition and genetic alterations, but it is unclear whether these differences stem from the selective pressures of the metastatic process, the systemic therapies, or both. We sought to define the repertoire of genetic alterations in breast cancer patients with de novo metastatic disease who had not received local or systemic therapy. Experimental Design: Up to two anatomically distinct core biopsies of primary breast cancers and synchronous distant metastases from nine patients who presented with metastatic disease were subjected to high-depth whole-exome sequencing. Mutations, copy number alterations and their cancer cell fractions, and mutation signatures were defined using state-of-the-art bioinformatics methods. All mutations identified were validated with orthogonal methods. Results: Genomic differences were observed between primary and metastatic deposits, with a median of 60% (range 6%–95%) of shared somatic mutations. Although mutations in known driver genes including TP53, PIK3CA, and GATA3 were preferentially clonal in both sites, primary breast cancers and their synchronous metastases displayed spatial intratumor heterogeneity. Likely pathogenic mutations affecting epithelial-to-mesenchymal transition–related genes, including SMAD4, TCF7L2, and TCF4 (ITF2), were found to be restricted to or enriched in the metastatic lesions. Mutational signatures of trunk mutations differed from those of mutations enriched in the primary tumor or the metastasis in six cases. Conclusions: Synchronous primary breast cancers and metastases differ in their repertoire of somatic genetic alterations even in the absence of systemic therapy. Mutational signature shifts might contribute to spatial intratumor genetic heterogeneity. Clin Cancer Res; 23(15); 4402–15. ©2017 AACR.

Primary testicular diffuse large B cell lymphoma (PTL) is an aggressive malignancy that occurs in the immune-privileged anatomical site of the testis. We have previously shown that structural genomic rearrangements involving the MHC class II transactivator CIITA and programmed death ligands (PDLs) 1 and 2 are frequent across multiple B cell lymphoma entities. Specifically in PTL, we found rearrangements in the PDL locus by fluorescence in situ hybridization (FISH). However, breakpoint anatomy and rearrangement partners were undetermined, while CIITA rearrangements had not been reported previously in PTL. Here, we performed bacterial artificial chromosome capture sequencing on three archival, formalin-fixed, paraffin-embedded tissue biopsies, interrogating 20 known rearrangement hotspots in B cell lymphomas. We report novel CIITA, FOXP1 and PDL rearrangements involving IGHG4, FLJ45248, RFX3, SMARCA2 and SNX29. Moreover, we present immunohistochemistry data supporting the association between PDL rearrangements and increased protein expression. Finally, using FISH, we show that CIITA (8/82; 10%) and FOXP1 (5/74; 7%) rearrangements are recurrent in PTL. In summary, we describe rearrangement frequencies and novel rearrangement partners of the CIITA, FOXP1 and PDL loci at base-pair resolution in a rare, aggressive lymphoma. Our data suggest immune-checkpoint inhibitor therapy as a promising intervention for PTL patients harbouring PDL rearrangements.

BACKGROUND: False positive findings introduced by analytical noise in sequencing and bioinformatics pipelines constitute a challenge for accurate massively parallel sequencing (MPS). Reports of intratumor genomic heterogeneity based on MPS rarely estimate the impact of false positive mutation calls. The purpose of this study was to measure apparent genomic heterogeneity in different regions of the same tumor and to assess the technical noise in variant calling in replicate sequencing of the same DNA. METHODS: Three anatomically distinct biopsies were obtained from 3 different regions of 11 breast cancers (33 samples) including 6 low/intermediate grade, estrogen receptor (ER)-positive and 5 high-grade, triple-negative (TNBC) cancers. DNA from 8 different biopsies was split and independently processed on different days to obtain technical replicates. The NimbleGen SeqCap EZ Exome Library preparation method was used for exome capture and paired-end sequencing of 75 base pair fragments was performed on Illumina HiSeq 2000. Read alignment and variant calling were performed with BWA and GATK haplotype caller. Concordance in variant calls and in minor allele frequencies (MAF) was assessed in the 3 biopsies of the same tumor and 8 technical replicates. We adjusted for uneven sequence coverage and analyzed known germline variants from dbSNP, known cancer related variants from COSMIC separately from novel variants (i.e. not previously reported in dbSNP or COSMIC). We estimated intratumor genomic heterogeneity of genes after removing alterations identified in areas where mapping is difficult and variant calls that had low analytical reliability in the technical replicates. RESULTS: The mean coverage was over 150X and > 90% of target regions had ≥ 20X coverage. We validated the specificity (98.2%) and sensitivity (86.7%) of the variant calling pipeline on the GIAB reference data. The concordance for germ line SNPs and variants in COSMIC in technical replicates was 94.9% and 92.7%, respectively. Novel variants had very low concordance, 55.9%, in technical replicates. The concordance between MAF estimates from the technical replicates was high (0.974, 0.957 and 0.969 for single nucleotide variations, insertions and deletions, respectively). The concordance for germline SNPs and COSMIC variants in pairwise comparisons of biopsies from the same tumor was 93.2% and 91.1%. For known variants, lower concordance was observed in TNBC (88.3%-98.5%) compared to ER-positive tumors (93.6%-98.8%, P CONCLUSION: We observed heterogeneity to be slightly greater in high-grade, ER-negative compared to low-grade, ER-positive breast cancers. Differences in variants observed in multiple biopsies of the same tumor are only slightly greater than those expected by technical noise alone. Citation Format: Weiwei Shi, Anees Chagpar, Tingting Jiang, Donald R Lannin, Brigid Killelea, Nina Horowitz, Raymond Lim, James Platt, Charlotte KY Ng, Vikram B Wali, Britta Weigelt, Jorge S Reis-Filho, Christos Hatzis, Lajos Pusztai. Reliability of whole exome sequencing for assessing intratumor heterogeneity from breast tumor biopsies [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr PD3-4.

Introduction: Metaplastic breast carcinoma (MBC) is an aggressive histologic type of breast cancer, which preferentially displays a triple-negative phenotype. These tumors are characterized by the presence of malignant cells exhibiting differentiation towards squamous epithelium or mesenchymal elements, including spindle, chondroid, osseous and rhabdoid differentiation. Unlike other rare histologic types of breast cancer such as adenoid cystic and secretory carcinomas, which are underpinned by the MYB-NFIB and ETV6-NTRK3 fusion genes respectively, pathognomonic genetic alterations have not been identified in MBC. It has been suggested, however, that the frequency of PIK3CA somatic mutations would be significantly higher in MBCs than in other forms of triple-negative disease. Here we sought to characterize the mutational landscape of MBCs by means of high-depth whole exome sequencing analysis. Material and Methods: Twenty-one triple-negative MBCs were retrieved from the authors’ institutions. Representative sections from frozen blocks were microdissected to ensure tumor cell content greater than 50%. DNA samples extracted from microdissected tumor and matched peripheral blood leukocytes were subjected to high-depth (250x) whole exome sequencing on an Illumina GAIIx or HiSeq2000. Somatic point mutations were called using MuTect and somatic insertions and deletions (indels) were called using Strelka, Varscan2 and Haplotype Caller. Potentially pathogenic mutations were predicted using computational algorithms including PolyPhen-2, Mutation Taster, Mutation Assessor, CHASM and FATHMM. Significantly mutated genes were identified using MutSigCV. Pathway and network enrichment analysis of mutations was performed with Ingenuity Pathway Analysis and HOTNET. The genomic landscape of MBCs was compared with that of triple-negative breast cancers (TNBCs) analyzed as part of The Cancer Genome Atlas project. Results: A mean of 135 somatic non-synonymous point mutations and indels were identified per MBC. The most frequently mutated gene was TP53, found in 12/21 cases (57%), and the only significantly mutated gene as defined by MutSigCV (q Conclusion: The majority (57%) of MBCs harbored non-synonymous mutations affecting TP53. While the frequencies of mutations affecting recurrently mutated genes in MBCs are similar to those found in other forms of TNBCs, MBCs significantly more frequently harbor mutations affecting PI3K pathway-related genes than TNBCs of no special type. Citation Format: Charlotte KY Ng, Britta Weigelt, Salvatore Piscuoglio, Y Hannah Wen, Maria R De Filippo, Luciano G Martelotto, Rachael Natrajan, Raymond Lim, Edi Brogi, Larry Norton, Anne Vincent-Salomon, Jorge S Reis-Filho. Mutational landscape of metaplastic breast carcinomas [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P2-03-08.

Background: Solid papillary carcinoma with reverse polarization (SPCRP) is a rare breast cancer subtype, whose most striking morphologic feature is the presence of back-to-back columnar epithelial cells with nuclei situated in an apical rather than in the normal basal location. We sought to characterize the morphologic and genetic landscape of SPCRP and determine whether it represents a distinct subtype of breast cancer underpinned by highly recurrent disease-specific genetic alterations. Material and Methods: Archival sections of 13 SPCRPs were subjected to microdissection. DNA samples extracted from microdissected tumor and matched normal samples were subjected to whole exome sequencing (n = 2) on an Illumina HiSeq2000, and DNA samples from tumors were further subjected to targeted massively parallel sequencing (n = 1), SNaPshot profiling (n = 7) and/or Sanger sequencing. State-of-the-art bioinformatics methods were used to define somatic mutations. Non-malignant breast epithelial MCF10A cells with and without somatic PIK3CA H1047R knock-in were used to assess the functional impact of the mutations identified by sequencing of SPCRPs in two- and three-dimensional cell culture systems. Results: Ten of 13 (77%) SPCRPs were found to harbor IDH2 R172 hotspot mutations, and eight of these cases had concurrent pathogenic mutations affecting PIK3CA or PIK3R1, in particular PIK3CA H1047R hotspot mutations. Functional monolayer and three-dimensional cell culture studies demonstrated that IDH2 R172 and PIK3CA mutations constitute likely drivers of this tumor and contribute to its unusual nuclear polarization phenotype. Conclusions: Our results suggest that SPCRP is a distinct subtype of breast cancer underpinned by IDH2 hotspot mutations in conjunction with mutations affecting canonical genes of the PI3K pathway. Given that IDH2 hotspot mutations have not been described in breast cancer to date, these may be used as a diagnostic ancillary marker for SPCRPs. Furthermore, IDH2 mutations may serve as a potential therapeutic target in SPCRP patients with disseminated disease. Citation Format: Sarah Chiang, Britta Weigelt, Huei-Chi Wen, Salvatore Piscuoglio, Luciano G. Martelotto, Charlotte K Y Ng, Jörg Balss, Gabrielle Baker, Kimberly S. Cole, Andreas von Deimling, Julie M. Batten, Jonathan D. Marotti, Hwei-Choo Soh, Benjamin L. McCalip, Raymond S. Lim, Kalliopi P. Siziopikou, Randi Burke, Song Lu, Xiaolong Liu, Tarek Hammour, Edi Brogi, A John Iafrate, Jorge S. Reis-Filho, Stuart J. Schnitt. Solid papillary carcinoma with reverse polarization are driven by IDH2 and PI3K pathway mutations. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 138.

Introduction: Lobular carcinoma in situ (LCIS) is considered both a risk factor and non-obligate precursor of invasive breast cancer. We sought to determine the genomic landscape of LCIS and the mutational processes involved in the clonal evolution and progression from LCIS to ductal carcinoma in situ (DCIS) and invasive lobular carcinoma (ILC). Methods: Patients with a history of LCIS undergoing therapeutic or prophylactic mastectomy were prospectively enrolled in an IRB approved protocol. Frozen tissue blocks were collected, screened for lesions of interest (LCIS, DCIS, ILC, invasive ductal carcinomas (IDC)) and subjected to microdissection and DNA/RNA extraction. Matched germline DNA was available for all cases. Whole exome sequencing was performed on a HiSeq2000 and data were aligned to the reference human genome and processed using GATK. Single nucleotide variants (SNVs) and small insertions/deletions were identified using MuTect and Varscan, respectively. Purity and ploidy estimates were calculated using ABSOLUTE. Clonal frequencies were estimated using Pyclone and the clonal structure of each sample was reconstructed using SubcloneSeeker. Shannon index and Simpson index metrics were used to calculate heterogeneity levels. Mutational signatures were defined according to their mutational trinucleotide context, and the expression levels of APOBEC gene family members were assessed by quantitative reverse transcription (qRT)-PCR. Results: 30 LCIS, 10 ILCs, 7 DCIS and 5 IDCs from 15 patients qualified for data analysis. CDH1 was the most frequently mutated gene and found to be targeted by mutations in 26 LCIS samples (23 somatic, 3 germline). The repertoire of somatic mutations in LCIS was similar to that of luminal A breast cancers, with the exception of the significantly higher frequency of CDH1 mutations and the lower prevalence of TP53 mutations. ILCs were clonally related to at least one LCIS in 10 patients, and in 3/7 patients, DCIS was clonally related to at least one LCIS. Clonal composition analysis revealed that the presence of a minor clone(s) in LCIS, and the levels of intra-tumor genetic heterogeneity were significantly higher in LCIS clonally related with DCIS/ILC than in LCIS unrelated to DCIS/ILC. In two cases, a minor LCIS subclone constituted the major clone in the associated DCIS/ILC. A comparative analysis of the mutational signatures in the truncal and branch mutations of these cases revealed that whilst the truncal mutations displayed an aging signature, branch mutations were enriched for the APOBEC signature. qRT-PCR analysis demonstrated that cases displaying the APOBEC signature also harbored significantly higher levels of APOBEC3B expression than samples with the aging signature. Conclusions: LCIS displays intra-lesion genetic heterogeneity, and the progression from LCIS to DCIS or ILC may involve the selection of clones resulting from distinct mutational processes during clonal evolution. Our findings also suggest that cytodine deamination driven by the overexpression of APOBEC3B may drive the progression of LCIS to DCIS/ILC in a subset of cases. Citation Format: Reis-Filho JS, Schizas M, Piscuoglio S, Sakr RA, Ng CKY, Lim RS, Carniello JVS, Towers R, Martelotto L, Giri DD, de Andrade VP, Viale A, Solit DB, Weigelt B, King TA. Lobular carcinoma in situ displays intra-lesion genetic heterogeneity and its progression to invasive disease involves clonal selection and variations in mutational processes. [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 S4-04.

Background: Germline mutations in the BRCA1 and BRCA2 genes lead to hereditary breast cancers that are defective in homologous recombination (HR) repair and sensitive to DNA damaging agents. HR deficiency (HRD) also occurs in sporadic breast cancers, but its incidence and etiology are unclear. Genomic signatures of HRD recently were employed as biomarkers in clinical trials with modest success. We posited that sporadic breast cancers displaying functional HR deficiency would harbor genetic alterations affecting HR DNA repair genes. To test this hypothesis, we applied a functional assay to define lack of competent HR DNA repair and sequenced the exomes of consecutive sporadic breast cancers. Methods: We developed an assay to assess the ability of cancer cells to localize RAD51 into sub nuclear foci in response to ex-vivo irradiation (IR) in fresh sporadic breast cancer tissue specimens from 60 patients. RAD51 focus formation was compared between mock and IR conditions to determine relative fold induction. Twenty-nine tumors with sufficient DNA underwent whole-exome sequencing. Structural genomic signatures of HRD (i.e. large state transitions (LST), telomeric imbalance (NtAI), loss-of-heterozygositiy (LOH)) and a previously reported mutational signature related to BRCA1/2 hereditary breast cancers were assessed. HR deficient tumors were defined as those with both RAD51 foci defects and a genomic signature of HR deficiency (LST>15 or presence of a BRCA mutational signature). Somatic, germ-line, and copy number changes in HR genes were investigated. BRCA1 methylation status was determined. Results: Seventeen of 60 (28%) tumors displayed defective RAD51 recruitment following ex-vivo IR (RAD51-DEF). RAD51-DEF was seen in all breast cancer subtypes , including 7 of 33 (21%) ER+/HER2-, 4 of 14 (29%) HER2+, and 6 of 13 (48%) triple-negative cases. Of the 29 sequenced tumors, 13 (45%) were RAD51-DEF and 16 (55%) were competent for inducing RAD51 foci. LST was elevated in 10 tumors (LST >15) and associated with RAD51-DEF (p=0.02), whereas NtAI (p=0.10) and LOH (p=0.052) did not show a significant association with RAD51-DEF. The BRCA1/2 mutational signature was evident in 4 tumors, all were RAD51-DEF (p=0.03) and 2 were BRCA2 mutated. Nine of 29 (31%) sequenced tumors were determined to have HRD by the RAD51 assay and presence of a genomic scar. Eight of these 9 (88%) cases with HRD had a genetic alteration of both alleles of a bona fide HR gene due to a pathogenic mutation (somatic or germline) coupled with loss of heterozygosity or a homozygous deletion compared to 1 (5%) tumor without HRD (p Conclusion: Combined functional and genomic analyses of breast tumors demonstrated that genetic loss of an HR gene may underpin HRD in sporadic breast cancers. Our findings warrant further comprehensive genetic assessment (somatic, germline, and copy number) of HR genes as potential biomarker for HR-directed therapies. Citation Format: Powell SN, Riaz N, Mutter RW, Ng CKY, Delsite R, Piscuoglio S, King TA, Martelotto L, Sakr R, Brogi E, Edelweiss M, Lim R, Higginson D, Weigelt B, Lee W, Reis-Filho JS. A functional assay for homologous recombination (HR) DNA repair and whole exome sequencing reveal that HR-defective sporadic breast cancers are enriched for genetic alterations in DNA repair genes. [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 S4-03.

Introduction: Metaplastic breast carcinoma (MBC) is a rare form of triple-negative breast cancer (TNBC), accounting for approximately 0.2%-5% of all invasive breast cancers. These tumors are characterized by the presence of neoplastic cells displaying differentiation towards squamous epithelium or mesenchymal elements. MBCs are reported to have an aggressive clinical behavior, to exhibit a worse prognosis and to respond less frequently to conventional chemotherapy regimens than common forms of TNBCs. In this study, we sought to define whether morphologically distinct subgroups of MBCs would be underpinned by distinct gene expression or copy number profiles, and whether MBCs, akin to other special histologic types of TNBC (e.g. secretory carcinoma and adenoid cystic carcinoma), would be underpinned by a highly recurrent fusion gene. Methods: RNA and DNA samples were extracted from microdissected frozen MBCs (5 squamous, 5 spindle and 7 chondroid) and subjected to gene expression profiling using the Illumina Human HT-12 v4 platform and gene copy number profiling using the Affymetrix Human SNP 6.0 arrays, respectively. Genes differentially expressed between MBC subtypes were identified using SAM, and functional annotation of these genes was performed using Ingenuity Pathway Analysis. Intrinsic molecular subtypes were determined using the PAM50 and claudin-low intrinsic gene lists. In addition, all cases were subjected to paired-end massively parallel RNA-sequencing (Illumina GAIIx). Putative expressed fusion transcripts were identified using a validated algorithm (i.e. ChimeraScan), and confirmed by means of RT-PCR. Results: MBCs with spindle cell morphology were all classified as of claudin-low intrinsic subtype, whereas MBCs with chondroid or squamous cell metaplasia were classified as of normal breast-like, basal-like or claudin-low subtypes, suggesting that these morphologic subgroups are heterogeneous. Unsupervised analysis of microarray and RNA-sequencing gene expression data further demonstrated that MBCs with spindle cell differentiation displayed distinctive transcriptomic profiles, and formed clusters distinct from those enriched for MBCs with chondroid and squamous cell metaplasia. MBCs with spindle cell morphology preferentially expressed regulators of epithelial-to-mesenchymal transition including lower expression of E-cadherin and EpCAM. At the genomic level, MBC subtypes displayed patterns of gene copy number alterations similar to those of common forms of TNBCs from The Cancer Genome Atlas, and no significant differences were found among the distinct MBC subtypes. Nine in-frame fusion genes, TBL1XR1-PIK3CA, WAPL-CDHR1, MAP2K3-HMGCLL, PARG-BMS1, FN1-ICAM1, TNKS1BP1-SPARC, AAK1-ARNT2, MBTPS1-TCEANC2 and PSMA6-SHMT1 were identified and validated in the index cases, however none of these was found to be recurrent in the cases analyzed in this study. Conclusion: MBC subtypes, despite harboring similar patterns of gene copy number alterations, display significant transcriptomic differences, which may account for their distinct histologic features. Our findings also demonstrate that unlike other histologic special types of TNBC, MBCs are not underpinned by a highly recurrent expressed fusion gene. Citation Format: Piscuoglio S, Ng CKY, Cowell CF, Mariani O, Martelotto L, Natrajan R, Lim RS, Maher CA, Vincent-Salomon A, Weigelt B, Reis-Filho JS. Genomic and transcriptomic heterogeneity in metaplastic breast carcinomas. [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 P6-03-10.

Background: The analysis of cell-free tumor DNA (ctDNA) from plasma has been heralded as a non-invasive technique for disease monitoring and as a means to overcome the challenges posed by intra-tumor genetic heterogeneity. ctDNA levels have been shown to correlate with tumor burden in breast cancer patients. Hence, we sought to define whether massively parallel sequencing of cell-free plasma DNA would capture the entire repertoire of somatic mutations present in the primary tumors and/ or metastases from patients with oligometastatic breast cancer. Methods: Frozen diagnostic biopsies from primary tumors and their distant metastases were obtained from five prospectively accrued treatment-naïve patients with stage IV breast cancer at presentation (1 estrogen receptor (ER)+/HER2+, 2 ER+/HER2-, 2 ER-/HER2+). A second, independent formalin-fixed paraffin-embedded (FFPE) diagnostic biopsy was obtained from the primary tumor and metastasis from 4 patients. Plasma samples were obtained from all patients. DNA samples from microdissected frozen tumors and peripheral blood, as well as plasma from one patient, were subjected to high-depth whole exome sequencing. DNA samples from all biopsies (frozen/FFPE), plasma and peripheral blood were subjected to targeted capture massively parallel sequencing, with baits for all somatic mutations detected by whole exome sequencing and all exons of the 100 genes most frequently mutated in breast cancer. Driver mutations were defined by state-of-the-art bioinformatic methods and literature search. Results: We identified and confirmed a median of 54 (range 25-75) and 53 (range 26-85) non-synonymous mutations in the primary tumors and metastases from the 5 cases analyzed, respectively. By sequencing the plasma DNA to a median depth of 248x (range 92-431x), state-of-the-art mutation callers revealed 0-4 mutations (0%-8% of mutations) per patient, and direct interrogation of the sequencing data, based on prior knowledge of the mutations present in the lesions, resulted in the identification of 2-18 mutations (3%-38% of mutations) per patient. Of the bona fide driver mutations, 2/3 TP53 mutations, 0/1 PIK3CA hotspot mutation, 0/1 BRCA2 frameshift mutation, 0/1 GATA3 frameshift mutation and 0/1 ERBB3 activating mutation were captured in the plasma DNA. A SMAD4 pathogenic mutation and a TCF7L2 truncating mutation were found in two diagnostic biopsies of metastatic lesions but not in two biopsies of the primary tumors in one patient each. Whilst the SMAD4 mutation was detected in the plasma DNA from the respective patient, the TCF7L2 mutation was not. Of the 62 mutations restricted to the primary tumors (0-42 per patient) and 74 restricted to the metastatic tumors (1-41 per patient), 4 and 7, respectively, were captured in the plasma DNA. Conclusions: Massively parallel sequencing assessment of plasma DNA allows for the identification of mutations found in primary tumors and/ or their metastases, however, only a subset of these could be detected at up to 431x depth. These observations suggest that current approaches for whole exome or targeted massively parallel sequencing may not be sufficient to capture the genetic heterogeneity of breast cancers in patients with oligometastatic disease. Citation Format: Ng CKY, Bidard F-C, Piscuoglio S, Lim RS, Pierga J-Y, Cottu P, Vincent-Salomon A, Viale A, Norton L, Sigal B, Weigelt B, Reis-Filho JS. Capturing intra-tumor genetic heterogeneity in cell-free plasma DNA from patients with oligometastatic breast cancer. [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 P2-01-02.

Recently, the landscape of single base mutations in diffuse large B-cell lymphoma (DLBCL) was described. Here we report the discovery of a gene fusion between TBL1XR1 and TP63, the only recurrent somatic novel gene fusion identified in our analysis of transcriptome data from 96 DLBCL cases. Based on this cohort and a further 157 DLBCL cases analyzed by FISH, the incidence in de novo germinal center B cell–like (GCB) DLBCL is 5% (6 of 115). The fusion appears exclusive to GCB and was not seen in 138 non-GCB cases examined (P = .008, Fisher exact test) but was present at low incidence in follicular lymphoma (1 of 81). In all 7 cases identified, the 3′ end of the fusion consists of exons 4 and onwards of TP63. The recurrence, subtype enrichment, and the remarkably conserved nature of the TP63 portion of the fusion suggest an important functional role in the lymphomas that harbor this event.

INTRODUCTION: Lobular carcinoma in situ (LCIS) is considered both a risk factor and non-obligate precursor of low-grade estrogen receptor-positive breast cancer. We sought to define the mutational repertoire, subclone complexity and heterogeneity of LCIS, and whether invasive lobular carcinomas (ILCs) would stem from specific subclones within a LCIS. METHODS: Patients with a history of LCIS undergoing therapeutic or prophylactic mastectomy were prospectively enrolled in an IRB approved protocol. Frozen tissue blocks were collected, screened for lesions of interest (LCIS, ductal carcinoma in situ (DCIS), ILC, invasive ductal carcinoma (IDC)) and subject to microdissection and DNA extraction. Matched germline DNA was available for all cases. Whole exome sequencing was performed on a HiSeq2000 (Illumina) and data were aligned to the reference human genome hg19 and processed using GATK. SNVs were called using MuTect, and indels were called using a combination of Varscan and Strelka. Purity and ploidy estimates were calculated by ABSOLUTE. Clonal frequencies were estimated using Pyclone. RESULTS: 30 LCIS, 10 ILCs, 6 IDCs and 7 DCIS from 15 patients qualified for data analysis, resulting in 18 LCIS-ILC pairs, 22 LCIS-LCIS pairs, 12 LCIS-DCIS pairs, and 14 LCIS-IDC pairs for comparison. 9/18 (50%) LCIS-ILC pairs and 8/22 (36%) LCIS-LCIS pairs were clonally related, supported by several shared mutations (median 18, range 7-81 for LCIS-ILC; median 14, range 5-22 for LCIS-LCIS). All related LCIS-ILC pairs and 6/8 related LCIS-LCIS pairs shared a pathogenic CDH1 mutation; 75% of related LCIS-ILC pairs also shared a PIK3CA hotspot mutation. 7/12 (58%) LCIS-DCIS pairs were found to be clonally related but the number of shared mutations was generally lower than that found in LCIS-ILC pairs (median 9, range 2-11). No evidence of a clonal relationship was found in any of the LCIS-IDC pairs tested. Clonal composition analysis revealed that samples of LCIS display intra-lesion genetic heterogeneity in the form of the presence of a minor clone in 70% of cases. In one case, the LCIS minor subclone (∼15%) constituted the major clone in the ILC and in another case the LCIS minor subclone constituted the major clone in the associated DCIS. The majority of the clonally related lesions were located in the same quadrant of the breast, however evidence of clonality was found in 5 LCIS-LCIS and 3 LCIS-DCIS pairs located in separate quadrants of the breast. CONCLUSIONS: Intra-lesion genetic heterogeneity is a common phenomenon in LCIS. The dominant clone of a LCIS may not always be the clone directly involved in the progression to malignancy. The spatial relationships of clonally related lesions in this study suggest that anatomy does not always infer clonality, as lesions located in separate quadrants of the breast may be clonally related. Citation Format: Michail Schizas, Rita A. Sakr, Britta Weigelt, Charlotte KY Ng, Jose Victor S. Carniello, Dilip Giri, Salvatore Piscuoglio, Luciano G. Martelotto, Russell Towers, Victor P. Andrade, Raymond Lim, David B. Solit, Jorge S. Reis-Filho, Tari A. King. Whole exome sequencing reveals heterogeneity within lobular carcinoma in situ (LCIS) and clonal selection in the progression to malignant lesions. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2971. doi:10.1158/1538-7445.AM2015-2971

Background: The metastatic process may constitute a biological bottleneck and may select for clones with genetic alterations that are distinct from those found in the dominant clone of the primary tumor. Here we sought to define whether the repertoires of somatic mutations found in biopsies of primary tumors would differ from those of their respective metastatic lesions in patients who did not receive any form of systemic therapy. Methods: Flash frozen diagnostic biopsies from primary tumors and their distant metastases were obtained from 9 prospectively accrued treatment-naïve patients with stage IV breast cancer at presentation (3 estrogen receptor (ER)+/HER2+, 2 ER+/HER2-, 2 ER-/HER2+ and 2 ER-/HER2-). Metastatic lesions from the liver (5), bone (2), skin (1) or contra-lateral axillary lymph node (1) were biopsied. An additional formalin-fixed paraffin embedded (FFPE) biopsy of each primary tumor and metastasis, as well as plasma from 5 patients, were obtained. DNA from microdissected frozen samples and peripheral blood, as well as plasma from one patient, was subjected to high-depth whole exome sequencing and gene copy number profiling. DNA from all biopsies (frozen/FFPE) and plasma was subjected to targeted capture massively parallel sequencing for all single nucleotide variants (SNVs) and insertions and deletions (indels) found by exome sequencing and all exons of the 100 most frequently mutated genes in breast cancer. Driver mutations were defined by state-of-the-art bioinformatic methods and literature curation. Results: In all cases, we detected founder genetic events in the modal population of primary tumors and their respective metastatic lesions (12-146 non-synonymous SNVs and 1-16 indels). Substantial genetic differences were observed between primary and their metastasis from these therapy-naïve patients, with a median of 9 non-synonymous mutations (range 0-42, median 9.6% of mutations) restricted to the primary tumor and not found in either biopsies of the metastases, and a median of 19 (range 1-73, median 20.2%) non-synonymous mutations restricted to the metastasis and not found in either biopsies of the primary tumors. The repertoire of mutations restricted to the metastatic lesions was unique to each case; however, we observed an enrichment of genes involved in the epithelial-to-mesenchymal transition (e.g. TCF7L2, SMAD4, KRIT1 and L1CAM). Plasma DNA analysis revealed that only 3.2%-38.3% of the mutations found in the primary tumor or its metastatic deposit could be detected in plasma. Conclusions: Therapy-naïve primary breast cancers and their metastatic deposits differ in their repertoire of genetic alterations, even when two biopsies of each primary tumor and metastatic lesions are analyzed. The assessment of targetable genetic alterations in single biopsies from primary tumors may not be sufficient for the optimal selection of precision medicine-based therapies. Citation Format: Charlotte K. Y. Ng, Francois-Clement Bidard, Salvatore Piscuoglio, Raymond S. Lim, Jean-Yves Pierga, Paul Cottu, Anne Vincent-Salomon, Agnes Viale, Larry Norton, Brigitte Sigal, Britta Weigelt, Jorge S. Reis-Filho. Genetic heterogeneity and distinct driver mutations in synchronous primary and metastatic breast cancers from therapy-naïve patients. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3889. doi:10.1158/1538-7445.AM2015-3889

Background: Male breast cancer (MaBC) accounts for Methods: Subtyping of the 64 MaBCs included in this study was performed by means of immunohistochemistry using the definitions described in the latest St. Gallen’s consensus report. DNA extracted from microdissected tumor and adjacent normal tissue were subjected to massively parallel targeting sequencing of all exons of 273 genes most frequently mutated in FBCs or directly related to DNA repair. Somatic mutations were defined using a combination of MuTect, SomaticSniper, MutationSeq and Haplotype Caller. Selected mutations were validated with Sequenom MassARRAY. CNAs were identified using Varscan2 and GISTIC2.0. Pathway and network analysis of mutations/CNAs was performed using Ingenuity Pathway Analysis and HOTNET. The genomic landscape of MaBCs was compared with that of FBCs of the same subtype analyzed as part of The Cancer Genome Atlas project. Results: All MaBCs were ER-positive and HER2-negative. Using the St. Gallen’s criteria, 37.5% and 62.5% were classified as luminal A-like or luminal B-like, respectively. The genes most frequently mutated in MaBCs were PIK3CA, GATA3, FLG and PLEC, with PIK3CA being the only significantly mutated gene as defined by MutSigCV (q=0.003). CNA analysis revealed recurrent gains of 1q and 8q and loss of 16q. GISTIC2.0 identified significantly recurrent high-level amplifications in 1q25.3, 8p11 (FGFR1, ZNF703), 8q24.3 (DEPTOR), 17q23 (PPM1D) and 15q26 (IGF1R) and deletions in 11q22 (ATM) and 21q22.12 (RUNX1). HOTNET analysis of the genes mutated and/or targeted by gene amplifications in MaBCs revealed significantly altered subnetworks involving genes related to DNA repair, PI3K and FGF signaling pathways. The most frequently mutated genes in luminal A-like MaBCs were PIK3CA and MLL3, while those of luminal B-like MaBCs were PIK3CA and GATA3. MLL3 and GATA3 mutations were only found in luminal A-like MaBCs (p=0.039) and luminal B-like MaBCs (p=0.048), respectively. Although the mutational landscapes of MaBCs and luminal FBCs were qualitatively similar, PIK3CA and TP53 were less frequently mutated in MaBCs (p Conclusions: MaBCs are preferentially of luminal subtype and are characterized by recurrent mutations in PIK3CA, GATA3, FLG and PLEC. Genetic alterations in MaBCs often target DNA repair and FGF signaling pathways. The known drivers of luminal FBCs appear to be less frequently altered in MaBCs. Given these important differences between MaBCs and FBCs, caution should be exercised in the extrapolation of biologic and clinical implications from studies in FBCs to the management of MaBCs. Citation Format: Salvatore Piscuoglio, Melissa Murray, Charlotte KY Ng, Elena Guerini Rocco, Luciano G Martelotto, Francois-Clement Bidard, Carey A Eberle, Nicola Fusco, Rita A Sakr, Leticia De Mattos-Arruda, Raymond Lim, Timour Baslan, James Hicks, Tari A King, Edi Brogi, Larry Norton, Britta Weigelt, Clifford A Hudis, Jorge S Reis-Filho. The genomic landscape of male breast cancers [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr S6-06.

INTRODUCTION: Epidemiologic data suggest that lobular carcinoma in situ (LCIS) is a risk factor for invasive breast cancer, whereas analyses of copy number aberrations between synchronous LCIS and invasive lobular carcinoma (ILC) have provided evidence to support its role as a non-obligate precursor of ILC. Using fresh frozen (FF) tissue samples, we sought to define the repertoire of somatic mutations in LCIS and ILC, and to compare the mutational repertoire of synchronous LCIS and ILC, and synchronous foci of LCIS. METHODS: Patients with a biopsy proven history of LCIS, with or without ILC, undergoing surgery were prospectively enrolled on an IRB approved study and FF tissue blocks were collected. Representative lesions of LCIS and ILC were microdissected and DNA was extracted using QIAamp® DNA Micro kit (Qiagen®). Matched normal germline DNA or FF normal ducts were available for all cases. At least 50 ng of DNA was used for hybridization capture using a bait library (NimbleGen®) comprising 273 genes, either recurrently mutated in breast cancer or associated with DNA repair. Targeted massively parallel sequencing was performed on a HiSeq2000 (Illumina®). Sequencing data were aligned to the reference human genome hg19 and processed by GATK. Single nucleotide variants (SNVs) were defined on the basis of three mutation callers (i.e. Mutect, Mutationseq and GATK Haplotypecaller); insertions and deletions were identified using the GATK Haplotypecaller. RESULTS: 47 samples of LCIS and 21 samples of ILC from 30 patients (with matched normal samples) were qualified for data analysis. The constellation of somatic mutations found in LCIS and ILC was similar, with the most frequently mutated genes being PIK3CA, CDH1, HRNR and MAP3K1 found in 25% and 52%, 23% and 38%, 15% and 19% and 10% and 9% of samples, respectively. Paired analysis of synchronous LCIS and ILCs (n=21) revealed that PIK3CA and CDH1 mutations were shared in 24% and 28% of cases respectively. Overall, in 8 (38%) pairs, the LCIS and ILC had one identical mutation in common and in 7 (33%) pairs, there were two or more common identical mutations. In 5 pairs, there were no common mutations. Paired analysis of two independent foci of LCIS from 9 breasts revealed one common mutation in each of 4 (44%) pairs (CDH1, RUNX1, SPEN, ERBB2). Interestingly, the point mutation in ERBB2 (L755S) was found in two LCIS foci from a patient who underwent prophylactic mastectomy. CONCLUSION: The mutational spectrum of LCIS is highly similar to that of ILC. PIK3CA mutations were the most common somatic mutations in both LCIS and ILC. Further, in this study, 71% of LCIS-ILC pairs shared at least one identical mutation, providing strong circumstantial evidence that LCIS is a non-obligate precursor of ILC. The finding of shared mutations among independent foci of LCIS also raises questions about the clonal origin of these lesions. Citation Format: Rita A. Sakr, Jose V. Scarpa, Michael Schizas, Dilip Giri, Marina De Brot, Russell Towers, Charlotte KY Ng, Raymond Lim, Victor P. Andrade, Britta Weigelt, Jorge S. Reis-Filho, Tari A. King. Targeted capture next generation sequencing of fresh frozen lobular carcinoma in situ and invasive lobular cancer identifies a common repertoire of mutations. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 927. doi:10.1158/1538-7445.AM2014-927

Introduction Primary mediastinal large B-cell lymphoma (PMBCL) is an aggressive malignancy commonly diagnosed in young adult females. In recent years, mutational and gene expression profiling has established genotypic and phenotypic similarity of PMBCL with both classical Hodgkin and diffuse large B-cell lymphoma (DLBCL). In-depth analyses of genomes and transcriptomes have highlighted several inactivating mutations (SOCS1, TP53), chromosomal amplifications (2p, 9p, Xp, Xq) and translocations (CIITA) thought to be integral in establishing and/or maintaining the PMBCL phenotype. Programmed death ligands (PDL) 1 (CD274) and 2 (PDCD1LG2), which are located on chromosome 9p24.1, are two emerging genes of interest that have been shown to be altered in PMBCL and can induce T-cell anergy by binding to the receptor, programmed death 1. Here, we describe the recurrence of chromosomal rearrangements of the PDL locus in various B-cell lymphomas and explore the association of these rearrangements with transcript levels. Methods To establish the frequency of CD274 and PDCD1LG2 aberration, we conducted fluorescence in situ hybridization (FISH) on 551 clinical samples and 20 established cell lines using in-house break-apart probes. Epstein-Barr virus encoded RNA in situ hybridization was also carried out on the clinical cohort. The clinical cases, sourced from the British Columbia Cancer Agency’s Centre for Lymphoid Cancer tissue repository, consisted of 125 PMBCLs, 216 DLBCLs, 130 primary DLBCL of the central nervous system (PCNSL), 12 nodular lymphocyte predominant Hodgkin lymphomas (NLPHL) and 68 follicular lymphomas (FL) with diagnoses based on the WHO classification. The DLBCL cohort could be further subdivided into 134 nodal DLBCLs and 82 testicular DLBCLs (T-DLBCL). Quantitative real-time PCR (qRT-PCR) was subsequently conducted on 17 cell lines and a clinical sub-cohort of 76 samples, for which fresh-frozen material was available, to determine the effect of mutations on transcript expression. We then characterized the PDL aberrations of two clinical PMBCL cases and three cell lines (DEV, L-428, L-1236), at base pair resolution, by applying the bioinformatic tools, nFuse, deFuse and destruct to both newly produced and previously published whole genome (WGS) and whole transcriptome (RNA-seq) libraries. Results FISH revealed a PDL locus (9p24.1) break-apart frequency of 20% (25/125) in PMBCL. There were no differences in any known clinical parameters or frequency of Epstein-Barr virus positivity between positive and negative PDL break-apart cases. Break-apart frequencies in other malignancies were calculated to be 3% in DLBCL, 7% in T-DLBCL and 1% in PCNSL; no positive cases were identified in either NLPHL or FL. The proportion of break-apart positive cases was significantly higher in PMBCL as compared to the other lymphomas surveyed (P < 0.05). Further, in agreement with the published literature, we observed an amplification frequency of the PDL locus in 36% (45/125) of PMBCLs. qRT-PCR established that PDCD1LG2 transcript levels were significantly higher in cases with 9p24.1 locus rearrangements compared to copy number neutral (P = 0.0003), gain (P = 0.001) and amplified cases (P = 0.005). Likewise, CD274 transcript levels were significantly higher in rearranged cases compared to copy number neutral cases (P = 0.03). Following the analysis of WGS and RNA-seq libraries, we were able to characterize four novel fusion transcripts involving the 9p24.1 locus: PDCD1LG2-NRG1 (PMBCL clinical case), PDCD1LG2-IGHV7-81 (L-1236), CIITA-PDCD1LG2 (DEV) and KIAA1432-CLDN14 (L-428). Aberrations involving both NRG1 and CIITA have previously been implicated in breast cancer and B-cell lymphomas, respectively. We also identified a translocation in another PMBCL clinical case with breakpoints in the intergenic spaces near LRMP and CD274, though this rearrangement did not produce a fusion transcript. Conclusion Taken together, our findings show that rearrangement of the PDL locus is recurrent in PMBCL, characteristic of PMBCL and leads to overexpression of PDL transcripts. Given the well-referenced function of PDLs in repressing the anti-tumor response, these data suggest that targeting the PDL axis in a subgroup of B-cell lymphomas holds clinical promise. Disclosures: No relevant conflicts of interest to declare.

Abstract 295 Introduction: Diffuse large B-cell lymphoma (DLBCL) is the most common type of aggressive non-Hodgkin lymphoma (NHL), accounting for approximately 30–40% of all new lymphoma cases. While standard therapy using rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) has significantly increased the survival of DLBCL patients, approximately one third of DLBCL patients still remain unresponsive to or relapse after standard treatment. Further investigation into the genomic architecture of DLBCL will contribute to elucidating the causes of the poor outcomes in this subgroup of patients. While the copy number and the gene expression profiles of DLBCL specimens have been well described as separate analyses, a large-scale high resolution integration of both orthologous measurements has yet to be reported. The integration of these two data types in a clinically well-annotated cohort of DLBCL is crucial as it can potentially distinguish driver from passenger genomic aberrations and reveal associations with clinical outcome. Methods and Patients: Affymetrix SNP 6.0 microarrays were used to ascertain the copy number profiles in 151 pretreatment biopsies of DLBCL that were representative of the population of DLBCL patients treated at the British Columbia Cancer Agency. Clinical outcome data were available for all 151 patients with 142 patients receiving R-CHOP or R-CHOP-like treatment. Matching RNA-seq libraries were used to quantitate the gene expression levels in 91 samples. The SNP 6.0 pre-processing method cRMAv2 was used to generate raw probe intensities that were then normalized to 1258 HapMap3 SNP 6.0 arrays. Copy number state calls were predicted using HMM-Dosage. RNA-seq data were aligned using the split-read aware aligner GSNAP and gene expression values were generated using the metric reads per kilobase of transcript per million mapped reads. DriverNet analyses were utilized to predict functionally relevant driver genes and outcome correlations in R-CHOP treated patients were performed using Cox regression and the log-rank test. Results: The copy number landscape derived from the SNP 6.0 microarrays revealed previously reported large scale chromosomal deletions in chromosome 6p and amplifications in chromosomes 3, 7 and 18. By integrating the gene expression with copy number data, we found that gene copy number was correlated with its own gene expression (classified as being cis-correlated) in 23.5% of genes. In addition, we investigated copy number aberrations which were highly correlated with gene expression across the genome (classified as trans-correlated). This analysis revealed aberration hotspots in genomic locations 3q26-q28 (TBL1XR1, BCL6, TP63), 17p12 (NCOR1, MAP2K4), 18q11.1-q11.2 (RBBP8) and 22q11.21 (BID, IL17RA) suggesting that these hotspots regulate important pathways that may contribute to the pathogenesis of DLBCL. We identified previously reported focal amplifications (e.g. REL) and deletions (e.g. B2M, CDKN2A). Moreover, we identified novel focal deletions, including homozygous deletions, in chromatin modifying genes: LCOR (7.9%), RCOR1 (9.9%), and NCOR1 (17.9%), all of which were cis-correlated and were validated using fluorescence in situ hybridization. DriverNet analyses identified RCOR1 deletions as one of the main driver alterations. RCOR1 deletions were also found to be associated with progression-free survival (5-year progression-free survival: deleted 40% vs. non-deleted 75%, p=0.0188). Discussion: Our systematic integration of SNP 6.0 and RNA-seq data confirmed findings of previous studies and also revealed novel genomic aberration hotspots and highly focal and frequent deletions in chromatin modifying genes. Results derived from our large-scale high resolution data set indicate the feasibility and efficacy of integrative genomic analyses in revealing novel and pathogenetically relevant genomic aberrations in lymphoid cancers. The discovery of the association of RCOR1 deletions with progression-free survival suggests that RCOR1 deletions could be used as a prognostic marker and might indicate a molecular phenotype that can be targeted by novel therapeutic agents in DLBCL. Disclosures: No relevant conflicts of interest to declare.

Background HER2 is overexpressed and amplified in approximately 15% of invasive breast cancers, and is the molecular target and predictive marker of response to anti-HER2 agents. In a subset of these cases, heterogeneous distribution of HER2 gene amplification can be found, which creates clinically challenging scenarios. Currently, breast cancers with HER2 amplification/overexpression in just over 10% of cancer cells are considered HER2-positive for clinical purposes; however, it is unclear as to whether the HER2-negative components of such tumors would be driven by distinct genetic alterations. Here we sought to characterize the pathologic and genetic features of the HER2-positive and HER2-negative components of breast cancers with heterogeneous HER2 gene amplification and to define the repertoire of potential driver genetic alterations in the HER2-negative components of these cases. Results We separately analyzed the HER2-negative and HER2-positive components of 12 HER2 heterogeneous breast cancers using gene copy number profiling and massively parallel sequencing, and identified potential driver genetic alterations restricted to the HER2-negative cells in each case. In vitro experiments provided functional evidence to suggest that BRF2 and DSN1 overexpression/amplification, and the HER2 I767M mutation may be alterations that compensate for the lack of HER2 amplification in the HER2-negative components of HER2 heterogeneous breast cancers. Conclusions Our results indicate that even driver genetic alterations, such as HER2 gene amplification, can be heterogeneously distributed within a cancer, and that the HER2-negative components are likely driven by genetic alterations not present in the HER2-positive components, including BRF2 and DSN1 amplification and HER2 somatic mutations. Electronic supplementary material The online version of this article (doi:10.1186/s13059-015-0657-6) contains supplementary material, which is available to authorized users.

Background Massively parallel sequencing studies have led to the identification of a large number of mutations present in a minority of cancers of a given site. Hence, methods to identify the likely pathogenic mutations that are worth exploring experimentally and clinically are required. We sought to compare the performance of 15 mutation effect prediction algorithms and their agreement. As a hypothesis-generating aim, we sought to define whether combinations of prediction algorithms would improve the functional effect predictions of specific mutations. Results Literature and database mining of single nucleotide variants (SNVs) affecting 15 cancer genes was performed to identify mutations supported by functional evidence or hereditary disease association to be classified either as non-neutral (n = 849) or neutral (n = 140) with respect to their impact on protein function. These SNVs were employed to test the performance of 15 mutation effect prediction algorithms. The accuracy of the prediction algorithms varies considerably. Although all algorithms perform consistently well in terms of positive predictive value, their negative predictive value varies substantially. Cancer-specific mutation effect predictors display no-to-almost perfect agreement in their predictions of these SNVs, whereas the non-cancer-specific predictors showed no-to-moderate agreement. Combinations of predictors modestly improve accuracy and significantly improve negative predictive values. Conclusions The information provided by mutation effect predictors is not equivalent. No algorithm is able to predict sufficiently accurately SNVs that should be taken forward for experimental or clinical testing. Combining algorithms aggregates orthogonal information and may result in improvements in the negative predictive value of mutation effect predictions. Electronic supplementary material The online version of this article (doi:10.1186/s13059-014-0484-1) contains supplementary material, which is available to authorized users.