Comprehensive molecular profiling of advanced/metastatic olfactory neuroblastomas

Olfactory neuroblastoma (ONB) is a rare, locally aggressive, malignant neoplasm originating in the olfactory epithelium in the nasal vault. The recurrence rate of ONB remains high and there are no specific treatment guidelines for recurrent/metastatic ONBs. This study retrospectively evaluated 23 ONB samples profiled at Caris Life Sciences (Phoenix, Arizona) using DNA sequencing (Sanger/NGS [Illumina], n = 15) and gene fusions (Archer FusionPlex, n = 6), whole genome RNA microarray (HumanHT-12 v4 beadChip, Illumina, n = 4), gene copy number assays (chromogenic and fluorescent in situ hybridization), and immunohistochemistry. Mutations were detected in 63% ONBs including TP53, CTNNB1, EGFR, APC, cKIT, cMET, PDGFRA, CDH1, FH, and SMAD4 genes. Twenty-one genes were over-expressed and 19 genes under-expressed by microarray assay. Some of the upregulated genes included CD24, SCG2, and IGFBP-2. None of the cases harbored copy number variations of EGFR, HER2 and cMET genes, and no gene fusions were identified. Multiple protein biomarkers of potential response or resistance to classic chemotherapy drugs were identified, such as low ERCC1 [cisplatin sensitivity in 10/12], high TOPO1 [irinotecan sensitivity in 12/19], high TUBB3 [vincristine resistance in 13/14], and high MRP1 [multidrug resistance in 6/6 cases]. None of the cases (0/10) were positive for PD-L1 in tumor cells. Overexpression of pNTRK was observed in 67% (4/6) of the cases without underlying genetic alterations. Molecular alterations detected in our study (e.g., Wnt and cKIT/PDGFRA pathways) are potentially treatable using novel therapeutic approaches. Identified protein biomarkers of response or resistance to classic chemotherapy could be useful in optimizing existing chemotherapy treatment(s) in ONBs.

Multi-platform molecular profiling of a large cohort of glioblastomas reveals potential therapeutic strategies

Glioblastomas (GBM) are the most aggressive and prevalent form of gliomas with abysmal prognosis and limited treatment options. We analyzed clinically relevant molecular aberrations suggestive of response to therapies in 1035 GBM tumors. Our analysis revealed mutations in 39 genes of 48 tested. IHC revealed expression of PD-L1 in 19% and PD-1 in 46%. MGMT-methylation was seen in 43%, EGFRvIII in 19% and 1p19q co-deletion in 2%. TP53 mutation was associated with concurrent mutations, while IDH1 mutation was associated with MGMT-methylation and TP53 mutation and was mutually exclusive of EGFRvIII mutation. Distinct biomarker profiles were seen in GBM compared with WHO grade III astrocytoma, suggesting different biology and potentially different treatment approaches. Analysis of 17 metachronous paired tumors showed frequent biomarker changes, including MGMT-methylation and EGFR aberrations, indicating the need for a re-biopsy for tumor profiling to direct subsequent therapy. MGMT-methylation, PR and TOPO1 appeared as significant prognostic markers in sub-cohorts of GBM defined by age. The current study represents the largest biomarker study on clinical GBM tumors using multiple technologies to detect gene mutation, amplification, protein expression and promoter methylation. These data will inform planning for future personalized biomarker-based clinical trials and identifying effective treatments based on tumor biomarkers.

Abstract P6-07-27: Characterization of neuroendocrine breast carcinomas for biomarkers of therapeutic options

Introduction: Neuroendocrine breast carcinoma (NBC) is an extremely rare type of cancer, constituting less than 0.1% of all breast tumors, without specific treatment options. We investigated a biomarker database for frequency of molecular markers that may guide personalized treatment choices for these patients.

Materials and Methods: Molecular profiles of 40 breast carcinomas with neuroendocrine features [ER/PR+ (n=20), HER2+ (n=1) and TNBC (n=14)] were assessed (all female patients, mean age: 60.3 years, range: 39-83 years). Gene expression (Illumina DASL microarray platform), protein expression (IHC), gene amplification (ISH) and next-generation sequencing (NGS; TruSeq Illumina platform) were performed.

Results: 57% of NBCs were positive for hormone receptors (ER/PR), 40% were triple negative TNBC and 3% HER2+ subtypes. Therapeutic biomarkers (IHC) that may guide chemotherapies (and used in other primary sites neuroendocrine tumors) included: high TOP2A (85%) for etoposide or doxorubicin, low TS (57%) for 5-fluorouracil and low ERCC1 (45%) for cisplatin. Additional biomarkers for chemotherapy included: high TOPO1 (60%) for irinotecan, low RRM1 (48%) for gemcitabine and low MGMT (57%) for temozolomide. Biomarkers associated with available targeted therapies included: PTEN loss (39%), positive ALK (33%), cKit (30%), EGFR (29%), AR (26%) and PDGFRA (17%). No gene amplifications were detected in cMET, EGFR, or TOP2A. Targeted sequencing analysis of 47 genes detected variants in TP53, PIK3CA, ERBB4 and APC genes. Gene expression data (included somatostatin receptor gene family- SSTR1/2/3/4/5) was available for 5 patients, for which 3/5 patients exhibited overexpression of at least one SSTR gene.

Conclusions: Molecular profiling by a multiplatform approach reveals potential personalized therapy options for this very rare breast cancer subtype. With recent success of somatostatin analogs for other neuroendocrine tumors, the overexpression of SSTR gene family in NBC is worthy of further investigation.

Citation Format: Gatalica Z, Feldman R, Ghazalpour A, Vranic S. Characterization of neuroendocrine breast carcinomas for biomarkers of therapeutic options [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-07-27.

Multiplatform molecular profiling identifies potentially targetable biomarkers in malignant phyllodes tumors of the breast

Malignant phyllodes tumor is a rare breast malignancy with sarcomatous overgrowth and with limited effective treatment options for recurrent and metastatic cases. Recent clinical trials indicated a potential for anti-angiogenic, anti-EGFR and immunotherapeutic approaches for patients with sarcomas, which led us to investigate these and other targetable pathways in malignant phyllodes tumor of the breast. Thirty-six malignant phyllodes tumors (including 8 metastatic tumors with two cases having matched primary and metastatic tumors) were profiled using gene sequencing, gene copy number analysis, whole genome expression, and protein expression. Whole genome expression analysis demonstrated consistent over-expression of genes involved in angiogenesis including VEGFA, Angiopoietin-2, VCAM1, PDGFRA, and PTTG1. EGFR protein overexpression was observed in 26/27 (96%) of cases with amplification of the EGFR gene in 8/24 (33%) cases. Two EGFR mutations were identified including EGFRvIII and a presumed pathogenic V774M mutation, respectively. The most common pathogenic mutations included TP53 (50%) and PIK3CA (15%). Cases with matched primary and metastatic tumors harbored identical mutations in both sites (PIK3CA/KRAS and RB1 gene mutations, respectively). Tumor expression of PD-L1 immunoregulatory protein was observed in 3/22 (14%) of cases. Overexpression of molecular biomarkers of increased angiogenesis, EGFR and immune checkpoints provides novel targeted therapy options in malignant phyllodes tumors of the breast.

Abstract 3928: Caveolin-1: Beyond a marker for basal-like breast cancers

Introduction: Caveolin-1 (Cav1) is associated with basal-like triple-negative (ER-/PR-/Her2-) breast cancers (TNBC). Its biological contribution to this subtype has not been fully explored and controversy persists regarding the molecular role of Cav1 in carcinogenesis. Experimental Procedures: Thirty-four TNBC (17 Cav1+/17 Cav1-) patients molecularly-profiled with a commercial assay (Caris Life Sciences, AZ) were evaluated retrospectively. Cav1 status was determined by immunohistochemistry (caveolin-1 polyclonal; ≥2+ ≥50%). The majority of specimens (28/34) used for profiling were from primary breast sites and contained ≥50% neoplastic cells. The transcriptomes were profiled using Illumina's HumanHT-12 microarray (v4). Data were normalized using mean normalization procedure. Differential expression analysis was performed using R's Limma package. Pathway analysis was carried out using R's signaling pathway impact analysis (SPIA) package with 69 cancer, immunity, and cell signaling related KEGG pathways. Results: Using a cutoff of two-fold and adjusted p-value of 0.05, we identified 954 genes differentially expressed between Cav1+/- TNBC patients. Included in these were 31 genes which were found to be up-regulated by over five- fold and 3 genes down-regulated by over five fold in Cav1+ TNBC. Genes of notable interest for their role in cell signaling, cell adhesion, tumor invasion and metastasis, included an up-regulation of TGFBR2, SPARC, integrins (ITGA11, ITGB5, ITGBL1), cell adhesion proteins (LAMB3, COL5A3) and molecules which facilitate tumor invasion (LAMB3, MMP1, MMP2, MMP9). In addition, genes found to be down-regulated in Cav1+ patients and notable for their roles in promoting epithelial-mesenchymal-transition (EMT) included Claudin 3(CLD3) and CA125/MUC16 (Mucin 16). We also detected an approximately two-fold down-regulation of CDKN2A in Cav1+ patients. Using SPIA pathway analysis, 12 pathways were found to be differentially activated in Cav1+ vs. Cav1- TNBC. The most differentially activated pathways were the focal adhesion pathway (p = 4.51E-18), PI3k-Akt signaling pathway (p = 2.01E-6) and TGF-β and MAPK signaling pathways (p = 0.005, 0.014, respectively). Conclusions: Differential gene expression patterns and pathway analyses provide evidence for distinct profiles for gene expression between Cav1+/- TNBC. Cav1+ TNBC patients exhibit up-regulation of genes important for cell signaling, extracellular matrix remodeling and tumor invasion, and down-regulation of genes that may facilitate EMT and loss of cell cycle control. The focal adhesion pathway, as well as TGF-β, PI3K and MAPK signaling pathways, were identified as differentially activated among Cav1+/- TNBC. Taken together, these data support the role of Cav1+ in identifying a subtype of TNBC that may have a greater risk for invasion and metastasis. The correlation of this subtype with prognosis and drug response should be investigated in future studies.

Citation Format: Rebecca A. Feldman, Zoran Gatalica, Semir Vranic, Ryan Bender, Sandeep Reddy, Anatole Ghazalpour. Caveolin-1: Beyond a marker for basal-like breast cancers. [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 3928.

Mutations in the Kinase Domain of the HER2/ERBB2 Gene Identified in a Wide Variety of Human Cancers

The HER2 (official name ERBB2) gene encodes a membrane receptor in the epidermal growth factor receptor family amplified and overexpressed in adenocarcinoma. Activating mutations also occur in several cancers. We report mutation analyses of the HER2 kinase domain in 7497 histologically diverse cancers. Forty-five genes, including the kinase domain of HER2 with HER2 IHC and dual in situ hybridization, were analyzed in tumors from 7497 patients with cancer, including 850 breast, 770 colorectal, 910 non-small cell lung, 823 uterine or cervical, 1372 ovarian, and 297 pancreatic cancers, as well as 323 melanomas and 2152 other solid tumors. Sixty-nine HER2 kinase domain mutations were identified in tumors from 68 patients (approximately 1% of all cases, ranging from absent in sarcomas to 4% in urothelial cancers), which included previously published activating mutations and 13 novel mutations. Fourteen cases with coexisting HER2 mutation and amplification and/or overexpression were identified. Fifty-two of 68 patients had additional mutations in other analyzed genes, whereas 16 patients (23%) had HER2 mutations identified as the sole driver mutation. HER2 mutations coexisted with HER2 gene amplification and overexpression and with mutations in other functionally important genes. HER2 mutations were identified as the only driver mutation in a significant proportion of solid cancers. Evaluation of anti-HER2 therapies in nonamplified, HER2-mutated cancers is warranted.

Expanding role of gut microbiota in lipid metabolism

PURPOSE OF REVIEW

This article highlights recent advances in the emerging role that gut microbiota play in modulating metabolic phenotypes, with a particular focus on lipid metabolism.

RECENT FINDINGS

Accumulating data from both human and animal studies demonstrate that intestinal microbes can affect host lipid metabolism through multiple direct and indirect biological mechanisms. These include a variety of signaling molecules produced by gut bacteria that have potent effects on hepatic lipid and bile metabolism and on reverse cholesterol transport, energy expenditure, and insulin sensitivity in peripheral tissues. Additionally, host genetic factors can modulate the abundance of bacterial taxa, which can subsequently affect various metabolic phenotypes. Proof of causality for identified microbial associations with host lipid-related phenotypes has been demonstrated in several animal studies, but remains a challenge in humans. Ultimately, selective manipulation of the gut microbial ecosystem for intervention will first require a better understanding of which specific bacteria, or alternatively, which bacterial metabolites, are appropriate targets.

SUMMARY

Recent discoveries have broad implications for elucidating bacterially mediated pathophysiological mechanisms that alter lipid metabolism and other related metabolic traits. From a clinical perspective, this newly recognized endocrine organ system can be targeted for therapeutic benefit of dyslipidemia and cardiometabolic diseases.

Abstract P4-09-19: Comprehensive multiplatform molecular profiling identifies potentially targetable biomarkers in malignant phyllodes tumors of the breast

Introduction: Malignant phyllodes tumors are rare breast malignancies (0.1% of all breast tumors) with limited effective treatment options for recurrent and metastatic disease. Recent trials indicated a potential for anti-angiogenic therapy in soft tissue sarcomas, which led us to investigate these pathways.

Materials and Methods: Thirty-five malignant phyllodes tumors (including two cases with matched primary and metastatic tumors) were profiled using gene sequencing (Next-generation and Sanger), gene copy number analysis (in-situ hybridization), whole genome RNA expression, and protein expression (immunohistochemical assay).

Results: RNA microarray assay showed consistent over-expression of genes involved in angiogenesis including VEGFA, Angiopoietin2, VCAM1, PDGFRA, PTTG1, and CYP3A5 in all cases analyzed (n=5). No mutations in KDR (VEGFR2) were detected (0/26). EGFR protein overexpression was observed in 25/26 (96%) of cases with amplification of the EGFR gene in 8 cases (33%). EGFR gene mutations were identified in 2 cases (8%) including one case with presumed pathogenic V774M mutation and one case with EGFRvIII mutation. The most common mutations included those of TP53 (50%) and PIK3CA (15%) while other mutations (BRCA1, BRCA2, RET, CDH1, MLH1, ATM) were rare affecting single phyllodes cases. Two cases with matched primary and metastatic cancers harbored the same mutations in both sites (PIK3CA/KRAS and RB1 gene mutations, respectively).

Conclusions: Comprehensive multiplatform profiling approach to phyllodes tumors identifies various molecular alterations of which some are potentially actionable. Our data suggests that anti-angiogenic therapy may also be effective in patients with malignant phyllodes tumor. Evaluation of EGFR pathway discovered consistent protein over-expression but rare activating mutations, which necessitates refinement in patient selection targeting these pathways.

Citation Format: Gatalica Z, Vranic S, Ghazalpour A, Xiu J, Ocal I, McGill J, Bender R, Discianno E, Sanati S, Reddy S, Pockaj B. Comprehensive multiplatform molecular profiling identifies potentially targetable biomarkers in malignant phyllodes tumors of the breast. [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 P4-09-19.

Abstract P3-07-47: Comprehensive profiling of metaplastic breast carcinoma reveals frequent over-expression of PD-L1

Background: Metaplastic breast carcinoma (MBC) is a rare subtype of breast carcinoma less responsive to conventional chemotherapy relative to usual breast carcinomas (UBCs) such as ductal and lobular subtype. In molecular terms MBC usually clusters with triple negative breast cancer (TNBC), but MBCs portray a worse prognosis in comparison to TNBCs. Published studies investigating MBCs for specific biomarkers of therapy response are rare and limited by the methodological approaches.

Methods: This study included 132 patients with 38 histologically confirmed MBCs and 94 UBCs. Amongst the 94 UBCSs, 44 were estrogen receptor positive, 33 were triple negative and 17 were HER2 positive. Direct sequencing analysis was performed on genomic DNA isolated from formalin-fixed paraffin-embedded tissue (FFPE) using the Illumina MiSeq Next Generation Sequencing platform (NGS). Immunohistochemistry for PD-L1 (SP142, Spring Bioscience), PD-1 (NAT105, Ventana) and EGFR (31G7, Life Technologies) was performed using automated procedures. Additionally, mutation analysis for EGFRvIII was performed on RNA extracted from FFPE tissue.

Results: At the genomic level, numerous cases of MBC had multiple genomic alterations with the most frequent genetic mutation in TP53 gene (14/24), similar to the TNBC controls (17/33). BRCA1 mutations were detected in 2/10 cases. Potentially actionable mutations were rare and included PIK3CA gene. Importantly, PD-L1 expression on cancer cells was detected in significantly higher proportion of MBCs (37%) than in the UBC cohort (6%) or TNBC control (14%) (p=3.7x10-5 and p=0.03, respectively). PD-1 positive tumors infiltrating lymphocytes (TILS) varied greatly in MBC (0 to >50/mm2). Over-expression of EGFR was frequent in MBCs (62%); however no mutations in the gene including EGFRvIII were detected.

Conclusion: Comprehensive profiling of a large cohort of this rare carcinoma highlighted predominance of TP53 mutations, wild type EGFR gene expression, a distinct increase in proportion of PD-L1 expression in carcinoma cells, and PD-1 expression in TILS. The latter properties can be exploited in clinical trials utilizing immune check point inhibitors.

Citation Format: Gatalica Z, Joneja U, Ghazalpour A, Swensen J, Feldman R, Cai F, Chen W, Xiao N, Reddy S, Palazzo J. Comprehensive profiling of metaplastic breast carcinoma reveals frequent over-expression of PD-L1. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-07-47.

Comprehensive profiling of renal medullary and collecting duct carcinomas

572

Background: Renal medullary carcinoma (RMC) is an aggressive malignancy affecting predominantly young African Americans with sickle cell trait (SCT) or disease (SCD), while a pathologically similar collecting duct carcinoma (CDC) affects patients without sickle cell trait. Clinical responses to chemotherapy and IL-2 in RMC/CDC are poor and novel therapies are needed. Methods: 9 patients with RMC (ages 13-58 y. o., all male) and 15 patients (ages 26-74 y. o., M:F = 13:2) with collecting duct carcinoma (CDC) were studied. Expression of PD-L1 was evaluated with 2 monoclonal antibodies (SP142 and SP263) and tumor infiltrating lymphocytes (TIL) were evaluated for PD1 expression (MRQ-22 antibody) using immunohistochemistry (IHC). Additional studies included ALK protein expression (D5F3 antibody), gene translocation (break apart FISH), next generation sequencing (NGS), and microsatellite instability (MSI). Results: Cancer cell PD-L1 expression above the threshold ( ≥ 2+, ≥ 5%) was seen in 7/9 RMC and 5/13 CDC cases. Concordance between 2 PD-L1 antibodies was 94.4%. PD-1+ TIL were absent in 6/18 cases and variably present in 12/18 cases (from 1 to > 15 TIL/40x power field). No MSI was detected in any of the cases tested (0/6). No case expressed ALK protein, but one case of CDC showed ALK gene re-arrangement. Mutations were identified in SMARCB1, FH, TP53 (3x), ATM, BRCA2, CHEK2 (2x), NF2 (3x), SETD2, and CDKN2A. No mutations in VHL or KDR were detected. One patient with RMC (and SCT) achieved complete clinical remission after treatment with bevacizumab plus paclitaxel. Conclusions: RMC and CDC strongly express PD-L1 in the majority of cases (12/22), suggesting that these patients may benefit from targeting the PD-L1/PD1 interaction. The absence of MSI in these cancers indicates a different mechanism of PD-L1 upregulation from colorectal carcinomas. Consistent with our previous study that showed frequent activation of (pseudo)hypoxia-induced pathways in RMCs (Human Pathology 2011;42:1979), we describe a case of RMC successfully treated with anti-VEGF therapy.

The Genetic Landscape of Hematopoietic Stem Cell Frequency in Mice

Prior efforts to identify regulators of hematopoietic stem cell physiology have relied mainly on candidate gene approaches with genetically modified mice. Here we used a genome-wide association study (GWAS) strategy with the hybrid mouse diversity panel to identify the genetic determinants of hematopoietic stem/progenitor cell (HSPC) frequency. Among 108 strains, we observed ∼120- to 300-fold variation in three HSPC populations. A GWAS analysis identified several loci that were significantly associated with HSPC frequency, including a locus on chromosome 5 harboring the homeodomain-only protein gene (Hopx). Hopx previously had been implicated in cardiac development but was not known to influence HSPC biology. Analysis of the HSPC pool in Hopx^−/− mice demonstrated significantly reduced cell frequencies and impaired engraftment in competitive repopulation assays, thus providing functional validation of this positional candidate gene. These results demonstrate the power of GWAS in mice to identify genetic determinants of the hematopoietic system.

•

Genetic variation across mouse strains influences hematopoietic stem cell frequency

•

This variation can be exploited for genome-wide association studies

•

Hopx is a regulator of hematopoietic stem/progenitor cell function

•

This approach can be used to identify genetic determinants of other stem cell systems

Programmed cell death 1 (PD-1) and its ligand (PD-L1) in common cancers and their correlation with molecular cancer type

Cancer cells expressing PD-1 ligands (PD-L1/PD-L2) inhibit immune-modulatory T-cell activation facilitating disease progression. Preliminary clinical trials exploring interruption of PD-1/PD-L1 signaling showed benefit in several cancer types. We analyzed the distribution of PD-1-positive tumor-infiltrating lymphocytes (TIL) and cancer cells' expression of PD-L1 in a molecularly profiled cohort of 437 malignancies (380 carcinomas, 33 sarcomas, and 24 melanomas). We showed that the presence of PD-1(+) TILs significantly varied among cancer types (from 0% in extraskeletal myxoid chondrosarcomas to 93% in ovarian cancer), and was generally associated with the increased number of mutations in tumor cells (P = 0.029). Cancer cell expression of PD-L1 varied from absent (in Merkel cell carcinomas) to 100% (in chondro- and liposarcomas), but showed the inverse association with the number of detected mutations (P = 0.004). Both PD-1 and PD-L1 expression were significantly higher in triple-negative breast cancers (TNBC) than in non-TNBC (P < 0.001 and 0.017, respectively). Similarly, MSI-H colon cancers had higher PD-1 and PD-L1 expression than the microsatellite stable tumors (P = 0.002 and 0.02, respectively). TP53-mutated breast cancers had significantly higher PD-1 positivity than those harboring other driver mutations (e.g., PIK3CA; P = 0.002). In non-small cell lung cancer, PD-1/PD-L1 coexpression was identified in 8 cases (19%), which lacked any other targetable alterations (e.g., EGFR, ALK, or ROS1). Our study demonstrated the utility of exploring the expression of two potentially targetable immune checkpoint proteins (PD-1/PD-L1) in a substantial proportion of solid tumors, including some aggressive subtypes that lack other targeted treatment modalities.

Oncocytoma-like renal tumor with transformation toward high-grade oncocytic carcinoma: a unique case with morphologic, immunohistochemical, and genomic characterization

Renal oncocytoma is a benign tumor with characteristic histologic findings. We describe an oncocytoma-like renal tumor with progression to high-grade oncocytic carcinoma and metastasis. A 74-year-old man with no family history of cancer presented with hematuria. Computed tomography showed an 11 cm heterogeneous multilobulated mass in the right kidney lower pole, enlarged aortocaval lymph nodes, and multiple lung nodules. In the nephrectomy specimen, approximately one third of the renal tumor histologically showed regions classic for benign oncocytoma transitioning to regions of high-grade carcinoma without sharp demarcation. With extensive genomic investigation using single nucleotide polymorphism-based array virtual karyotyping, multiregion sequencing, and expression array analysis, we were able to show a common lineage between the benign oncocytoma and high-grade oncocytic carcinoma regions in the tumor. We were also able to show karyotypic differences underlying this progression. The benign oncocytoma showed no chromosomal aberrations, whereas the high-grade oncocytic carcinoma showed loss of the 17p region housing FLCN (folliculin [Birt-Hogg-Dubé protein]), loss of 8p, and gain of 8q. Gene expression patterns supported dysregulation and activation of phosphoinositide 3-kinase (PI3K)/v-akt murine thymoma viral oncogene homolog (Akt), mitogen-activated protein kinase (MAPK)/extracellular-signal-regulated kinase (ERK), and mechanistic target of rapamycin (serine/threonine kinase) (mTOR) pathways in the high-grade oncocytic carcinoma regions. This was partly attributable to FLCN underexpression but further accentuated by overexpression of numerous genes on 8q. In the high-grade oncocytic carcinoma region, vascular endothelial growth factor A along with metalloproteinases matrix metallopeptidase 9 and matrix metallopeptidase 12 were overexpressed, facilitating angiogenesis and invasiveness. Genetic molecular testing provided evidence for the development of an aggressive oncocytic carcinoma from an oncocytoma, leading to aggressive targeted treatment but eventual death 39 months after the diagnosis.

Abstract LB-171: Concordance between protein expression and mRNA expression in the large cohort of cancer patients

Introduction: Immunohistochemical (IHC) analysis of protein expression in clinical specimens has been invaluable in diagnosing, monitoring and guiding treatment in cancer patients. This approach, however, has its limitations including the difficulty to obtain robust and reliable antibodies, the semi quantitative nature of the assay, subjectivity of intensity call, and the lack of specificity to detect clinically relevant isoform expression in tumor specimen. In light of these limitations, RNA expression may be a useful replacement or a complementary approach to IHC in clinical settings. In this study, we compared the concordance between IHC and mRNA expression levels as measured by microarray using a large cohort of cancer patients spanning multiple lineages.

Methods: 10246 patient samples that included 40 tumor types were utilized in this study. Transcript and protein expression levels were measured for ESR1, AR, KIT, ERBB2 (Her2), RRM1, and ERCC1. Transcript levels were measured by Illumina's HumanHT-12 microarray (v4), and IHC was performed using the following antibodies: SP1 (ESR1), AR27 (AR), 4B5 (Her2), 8F1 (ERCC1), anti-CD117 polyclonal (KIT), and 10526-1-AP (RRM1).

Results: We observed a significant correlation between ERBB2, ESR1, and KIT with IHC data. Of the 3 probes on the microarray measuring ERBB2, only one probe mapping to the 3’ end of the two ERBB2 isoforms showed significant association with IHC data. For KIT, IHC and mRNA expression correlated well across, thymic carcinomas, poorly differentiated neuroendocrine tumors, melanoma, GIST, and head and neck cancer. For ESR1, significant correlations were observed for uterine sarcoma, endometrial carcinoma, and breast adenocarcinoma. Cervical cancers and ovarian surface epithelial carcinomas showed diminished association, and there was a lack of association in NSCLC and neuroendocrine carcinoid tumors. For AR, we observed strong correlation only in prostate tumors. ERCC1 and RRM1 mRNA expressions did not correlate well with protein expression in any of the tumor lineages examined.

Conclusion: These observations suggest that measuring mRNA via microarray has the potential to serve as a surrogate for IHC in clinical setting. The mixed associations observed for some of the genes exhibited lineage specificity suggesting that the relation between microarray results and IHC may be driven by the tissue-specific expression of isoforms, differential expression of genes in the stromal cells, and the technical limitations of each platform including the specificity of either the probes used to measure mRNA or antibodies used to measure protein expression. As such, combining the two platforms in the clinical setting may provide a more complete assessment of gene expression in tumor samples.

Citation Format: Anatole Ghazalpour, Wangjuh Sting Chen, Wenhsiang Wen, Zoran Gatalica, Ryan Bender. Concordance between protein expression and mRNA expression in the large cohort of cancer patients. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr LB-171. doi:10.1158/1538-7445.AM2014-LB-171

Genetic regulation of mouse liver metabolite levels

We profiled and analyzed 283 metabolites representing eight major classes of molecules including Lipids, Carbohydrates, Amino Acids, Peptides, Xenobiotics, Vitamins and Cofactors, Energy Metabolism, and Nucleotides in mouse liver of 104 inbred and recombinant inbred strains. We find that metabolites exhibit a wide range of variation, as has been previously observed with metabolites in blood serum. Using genome‐wide association analysis, we mapped 40% of the quantified metabolites to at least one locus in the genome and for 75% of the loci mapped we identified at least one candidate gene by local expression QTL analysis of the transcripts. Moreover, we validated 2 of 3 of the significant loci examined by adenoviral overexpression of the genes in mice. In our GWAS results, we find that at significant loci the peak markers explained on average between 20 and 40% of variation in the metabolites. Moreover, 39% of loci found to be regulating liver metabolites in mice were also found in human GWAS results for serum metabolites, providing support for similarity in genetic regulation of metabolites between mice and human. We also integrated the metabolomic data with transcriptomic and clinical phenotypic data to evaluate the extent of co‐variation across various biological scales.

Biomarker analysis of glioblastoma and potential implications for therapy

2041

Background: Glioblastoma multiform (GBM), the most aggressive CNS cancer, has limited effective therapeutic options, with underlying molecular heterogeneity contributing to the differences in treatment response. Our study was designed to interrogate biomarkers from a large cohort of GBM patients to seek therapeutic implications. Methods: Data was analyzed from 570 high grade astrocytoma patients (vast majority GBM) who received tumor profiling at Caris Life Sciences from 2009 to 13. Test methodologies included IHC, FISH, CISH, Sanger SEQ, MGMT promoter methylation and NextGen SEQ (Illumina TruSeq). Results: In our patient cohort, 59% had MGMT methylation and 70% had negative MGMT IHC, predicting potential response to temozolomide. Protein expression for ERCC1, TOPO1, and TS was 53% negative, 49% positive, and 37% negative, indicating potential benefit from cisplatin, irinotecan and fluorouracil, respectively. Drug pumps PGP and MRP1, were positive by IHC in 10% and 67% of patients, suggesting possible resistance to etoposide, vinca alkaloids and methotrexate. For targeted therapies, c-Kit IHC was positive in 7% patients, mutated in 6% and PDGFRA IHC was positive in 27%, indicating potential benefit from imatinib and other TKI’s. RAS/RAF and PIK3CA/mTOR pathway activation was also noted with BRAF, KRAS, PIK3CA mutations and PTEN loss, observed in 8%, 3%, 7% and 10% of cases, respectively. TS negativity was seen in 91% of MGMT methylated patients and in 37% of MGMT unmethylated patients (p = 0.025, student’s t-test), revealing a possible novel combination therapy of fluoropyrimidines with temozolomide for a select cohort. Similarly, biomarker profiles of molecular subgroups defined by EGFR amplification (44% in our cohort) and IDH1, p53 mutations will be analyzed for therapeutic implications. Conclusions: By profiling tumor biomarkers from a large cohort of GBM patients using validated assays in a single facility, we demonstrate the vast molecular heterogeneity of GBM and highlight the importance of individualized therapy based on a patient’s unique tumor profile. Incorporating a comprehensive biomarker analysis into clinical management of this aggressive cancer allows for an informed selection of more effective therapies.

The systems genetics resource: a web application to mine global data for complex disease traits

The Systems Genetics Resource (SGR) (http://systems.genetics.ucla.edu) is a new open-access web application and database that contains genotypes and clinical and intermediate phenotypes from both human and mouse studies. The mouse data include studies using crosses between specific inbred strains and studies using the Hybrid Mouse Diversity Panel. SGR is designed to assist researchers studying genes and pathways contributing to complex disease traits, including obesity, diabetes, atherosclerosis, heart failure, osteoporosis, and lipoprotein metabolism. Over the next few years, we hope to add data relevant to deafness, addiction, hepatic steatosis, toxin responses, and vascular injury. The intermediate phenotypes include expression array data for a variety of tissues and cultured cells, metabolite levels, and protein levels. Pre-computed tables of genetic loci controlling intermediate and clinical phenotypes, as well as phenotype correlations, are accessed via a user-friendly web interface. The web site includes detailed protocols for all of the studies. Data from published studies are freely available; unpublished studies have restricted access during their embargo period.

Potential impact of different methodologies for detecting EGFR mutations in selecting chemotherapy for patients with lung cancer

e19006

Background: Mutation analysis of the kinase domain of EGFR (exons 18-21) is a standard recommended procedure for patients diagnosed with non-small cell lung cancer (NSCLC). NSCLC patients whose tumor harbors certain EGFR mutations have notable responses to EGFR inhibitors. Several different methodologies are available as published assays or commercially available kits and include Sanger Sequencing, allele specific PCR (ASP) and restriction fragment length polymorphism (RFLP). Differences in the design of these three assays dictate which clinically relevant mutations will be detected. Methods: Sanger Sequencing of EGFR found 518 potentially clinically actionable mutations and 45 variants of unknown significance of the 4307 samples tested. To assess which clinically relevant EGFR mutations will be detected by the ASP and RFLP, we performed an in silico analysis of all observed mutations against the design specification of the ASP and RFLP assays. The RFLP assay used in this assessment was developed in our laboratory and is designed to detect all G719 mutations, all exon 19 deletions, all exon 20 insertions and the specific mutations T790M, L858R, L861R and L861Q. A commercially available ASP kit designed to detect 29 mutations in the kinase domain was also analyzed in this study. Results: Based on the performance characteristics assumed by the RFLP and ASP assays, the analytical sensitivities for detecting clinically actionable mutations of the two methodologies are 98.8% and 86.7% respectively, when compared to Sanger Sequencing. It is notable that the RFLP assay missed the S768I mutation (0% detected) whereas the assay did not detect some G719 mutations, exon 19 deletions, and exon 20 insertions (76.3, 86.6 and 35.3% detected respectively). Conclusions: Sequencing is still the preferred method of mutation detection in EGFR for NSCLC as it is the most comprehensive. However, if tumor nuclei are limited a more sensitive method than sequencing is required; EGFR mutation detection by RFLP would identify more of the potentially clinically relevant mutations than ASP. The ASP method would produce false negative results in 13.3% of patients expected to respond to EGFR inhibitors.

Unraveling inflammatory responses using systems genetics and gene-environment interactions in macrophages

Many common diseases have an important inflammatory component mediated in part by macrophages. Here we used a systems genetics strategy to examine the role of common genetic variation in macrophage responses to inflammatory stimuli. We examined genome-wide transcript levels in macrophages from 92 strains of the Hybrid Mouse Diversity Panel. We exposed macrophages to control media, bacterial lipopolysaccharide (LPS), or oxidized phospholipids. We performed association mapping under each condition and identified several thousand expression quantitative trait loci (eQTL), gene-by-environment interactions, and eQTL "hot spots" that specifically control LPS responses. We used siRNA knockdown of candidate genes to validate an eQTL hot spot in chromosome 8 and identified the gene 2310061C15Rik as a regulator of inflammatory responses in macrophages. We have created a public database where the data presented here can be used as a resource for understanding many common inflammatory traits that are modeled in the mouse and for the dissection of regulatory relationships between genes.

Juxtaglomerular cell tumor: a morphological, immunohistochemical and genetic study of six cases

Juxtaglomerular cell tumors (JGCTs) are rare tumors characterized by renin synthesis, hyperaldosteronism and hypertension. A curious immunohistochemical overlap between JGCT and gastrointestinal stromal tumor (GIST) including the expression of vimentin, CD34, CD117, α-smooth muscle actin was previously reported, prompting us to further investigate JGCT and its phenotypic and molecular genetic characteristics. Virtual karyotyping showed gain of chromosomes 3, 4, 10, 13, 17 and 18 in one JGCT, and fluorescence in situ hybridization (FISH) study confirmed this multiple gain pattern. Additionally, loss of chromosome 9 was observed in four of six cases analyzed with FISH. A whole genome expression analysis revealed 415 up-regulated (including renin, and CD117) and 325 down-regulated genes between the 2 cases. The study confirmed earlier reports on the gain of chromosomes 4 and 10, and provided further evidence of up-regulation of the genes located on these 2 chromosomes. For the first time our study indicated the importance of the loss of chromosome 9 and loss of expression of several tumor suppressor genes located on this chromosome as possible pathogenetic events important in development of JGCT.

Hybrid mouse diversity panel: a panel of inbred mouse strains suitable for analysis of complex genetic traits

We have developed an association-based approach using classical inbred strains of mice in which we correct for population structure, which is very extensive in mice, using an efficient mixed-model algorithm. Our approach includes inbred parental strains as well as recombinant inbred strains in order to capture loci with effect sizes typical of complex traits in mice (in the range of 5% of total trait variance). Over the last few years, we have typed the hybrid mouse diversity panel (HMDP) strains for a variety of clinical traits as well as intermediate phenotypes and have shown that the HMDP has sufficient power to map genes for highly complex traits with resolution that is in most cases less than a megabase. In this essay, we review our experience with the HMDP, describe various ongoing projects, and discuss how the HMDP may fit into the larger picture of common diseases and different approaches.

Theranostic profiling of Ewing sarcoma (ES) and desmoplastic small round cell tumors (DSRCT)

e20507

Background: Ewing Sarcoma (ES)/PNET and Desmoplastic Small Round Cell Tumor (DSRCT) are sarcomas with distinct chromosomal translocations involving the EWS gene (predominately EWS-FLI1 and EWS-WT1; respectively). Their diagnosis and treatment has been difficult due to the rarity, diverse clinical presentation, overlapping histologic features and genetic complexity (Taylor BS et al, 2011). Therefore, novel approaches in clinical management are warranted. Methods: Seventeen cases (9 ES and 8 DSRCT) were analyzed using a commercial molecular profiling service (CarisTargetNow, Caris Life Sciences, Phoenix, AZ). The whole genome transcriptome analysis (29285 transcripts) was performed using HumanHT-12 beadChip (Illuminia Inc, San Diego, CA) and comparison to pooled soft tissue reference sample. Additionally, a select number of chemotherapy-predictive (theranostic) biomarkers were evaluated using immunohistochemistry, FISH, and DNA sequencing. Results: We observed 160 commonly up and 357 commonly down regulated genes between ES and DSRCT in transcriptome analysis. Cell cycle signaling, DNA replication and E2F mediated pathway genes were most commonly up regulated. In addition, higher expression of SOX-2, a recently identified cancer stem cell marker (Riggi et al, 2010), were observed in DSRCT than in ES, suggesting EWS-WT1 translocation might result in reprogramming of DSRCT to express cancer stem cells. Above threshold expression of TOP2A and TOPO1was observed in approximately 50% of all cases. Additional theranostic biomarkers (ERCC1, TS, SPARC and MGMT) showed significant inter-individual variations. No KRAS mutations or EGFR gene amplification were observed in any case. Conclusions: 1. Our transcriptome analyses might provide future therapeutic targets in cell cycle regulation, DNA replication, receptor TKI pathways and stem cell reprogramming. 2. Using multimodality test approaches, we confirmed and refined the benefits in both tumor types of individualized therapeutic assessment including predicted susceptibilities to anthracycline, irinotecan, platinum analogs, fluorouracil, and temozolomide.

Therapeutic target expression in advanced prostate cancer patients submitted for comprehensive molecular profiling

e15209

Background: Approximately 30,000 men die from advanced PCa annually. The prognosis of this disease stresses a need for novel therapeutic and clinical strategies. Methods: We analyzed targeted genes and proteins in 141 patient specimens using IHC (for as many as 29 different biomarkers), FISH for EGFR, ALK, c-Myc, Her2neu and TOP2A, sequencing for BRAF, c-ckit, EGFR, KRAS, and PIK3CA, and RNA expression microarray. Therapeutic associations from the published literature were then assigned based on the differentially expressed biomarkers. Due to amount of tumor tissue present in the specimen, physician test requests and other factors, the number of biomarkers interrogated varied from case to case. Results: Therapeutic associations based on IHC results included:) 92% HER2 negative (potential lack of benefit from HER2 targeted therapy); 85% AR high (potential benefit from anti-androgen therapy); 79% ERCC1 negative (potential benefit from platinum-based therapies); 70% TS negative (potential benefit from fluorouracil),54% PTEN negative (suggesting activation of PIK3CA pathway), 66% RRM1 low (potential benefit from gemcitabine); 64% TOPO2A low (potential lack of benefit from anthracyclines); 44% TOPO1 high (potential benefit from irinotecan) 34% MGMT high (potential lack of benefit from temozolomide); 28% SPARC high (potential benefit from nab-paclitaxel). Potential therapeutic associations based on FISH and Sequencing included: 82% HER2 negative by FISH (potential minimal benefit to HER2 targeted therapy), 13% EGFR positive by FISH and a majority 90% and 81% wild-type for KRAS (exons 2, 3) and BRAF (exons 11, 15), respectively, indicating potential benefit from EGFR targeted therapies. Interestingly, two drug pumps, PGP and MRP1 were 85% negative and 91% positive, respectively. Conclusions: Advanced PCa is an aggressive disease where a one-size-fit-all approach is not effective. Our data confirms that these tumors are heterogeneous at the molecular level and has important therapeutic implications. When developing biomarker enriched clinical trials or advanced treatment planning, a comprehensive molecular profile may provide important insights not otherwise considered.

Molecular profiling of uveal melanoma patients

10630

Background: Although uveal melanoma represents only 5% of all melanomas, it is the most common primary intraocular malignancy of the adult eye. Approximately 50% of patients will develop metastases which are resistant to medical interventions. There is a great need for improved therapy as the prognosis is poor for advanced stages. Our study was undertaken to investigate the presence of novel therapeutic targets. Methods: We analyzed 49 uveal melanoma patients with immunohistochemistry for 16 markers including cKIT, PDGFR, cMET, PTEN and IGF1R. Further, microarray analysis was performed on 29 samples using the Illumina platform. We also investigated amplification of EGFR and mutational analysis of cKIT, BRAF, KRAS and NRAS on a smaller patient subset. Results: Overexpression of KIT at the protein and RNA level was 74% (28 out of 38) and 45% (13 out of 29), respectively. Expression of cKIT did not correlate with gain-of-function cKIT mutations in any of the 34 samples tested. In our study, MET was overexpressed in 15 out of 17 cases at the RNA level and IGF1R was high in 4 out of 6 patients indicating poor prognosis. PTEN expression by IHC was present in 90% (36 out of 40) patients indicating the PI3K pathway is not activated in the majority of uveal melanoma patients. BRAF was wildtype in all 42 patients tested. Similarly, no KRAS or NRAS mutations were detected. Protein and RNA expression of PDGFR were low in our patients. MGMT was lost in 16 out of 40 patients at the protein level and 10 out of 29 patients at the RNA level. EGFR expression, copy number and protein levels were low in the patients tested. Conclusions: Our data on cKIT suggests that it is a promising target in uveal melanoma. Low expression of MGMT in about a third of our patients may indicate the likelihood of favorable response to dacarbazine and temozolomide. There are currently several clinical trials investigating various cKIT inhibitors, as well as temozolomide in advanced uveal melanoma patients. Our findings highlight the importance of molecular profiling uveal melanoma patients.

Frequency of TLE3 overexpression in breast carcinoma subtypes including a large cohort of triple-negative patients

1040

Background: The taxanes are an important class of agents for the treatment of a broad range of malignancies including breast cancer. They improve survival in patients with early stage and metastatic breast cancer. Transducin-like enhancer of split 3 (TLE3) is a transcriptional repressor which influences growth and microtubule stability and its expression has been implicated in response to taxane therapy in breast cancer. We investigated the tumor expression of TLE3 in breast cancer patients, including a large cohort of the triple negative subtype. Methods: We analyzed TLE3 (M-201), ER(1D5), PR(PgR636) and HER2/neu(Polyclonal) expression by immunohistochemistry in 978 breast cancer patients. Immunoreactivity was assessed by scoring the percentage of cells stained in each field and by the intensity of staining. Results: To sub-classify the 978 breast cancer patients, we utilized hormone receptors (ER and PR) and HER2 expression/amplification. Overall, 36% of the total breast cancer patients were hormone receptor positive, 15% were HER2 positive and 49% were triple negative. The percentage of triple negative patients was higher in our cohort, given the fact that molecular profiling services are used more frequently for this subtype. A total of 477 patients were triple negative of which 61% stained positive for TLE3 expression. Of the 150 HER2 positive patients, 73% stained positive for TLE3 expression as compared with 82% TLE3 positivity in the 351 hormone receptor positive patients. By pairwise comparison, the hormone receptor positive vs triple-negative subtype showed the highest statistical significance in ratios of TLE3 positives (p =2.5e-10). Conclusions: Our results show that TLE3 is over-expressed in the majority of HER2 positive and hormone receptor positive breast cancer patients. Interestingly, the frequency of over-expression of TLE3 was lowest in the triple negative subtype thereby making it more important to identify those patients in this group who are most likely to respond to taxanes prior to therapy. To our knowledge, this is the first study providing a comprehensive review of TLE expression in breast cancer subtypes.

Genetic locus on mouse chromosome 7 controls elevated heart rate

Elevated heart rate (HR) is a risk factor for cardiovascular diseases. The goal of the study was to map HR trait in mice using quantitative trait locus (QTL) analysis followed by genome-wide association (GWA) analysis. The first approach provides mapping power and the second increases genome resolution. QTL analyses were performed in a C3HeB×SJL backcross. HR and systolic blood pressure (SBP) were measured by the tail-cuff plethysmography. HR was ∼80 beats/min higher in SJL compared with C3HeB. There was a wide distribution of the HR (536-763 beats/min) in N2 mice. We discovered a highly significant QTL (logarithm of odds = 6.7, P < 0.001) on chromosome 7 (41 cM) for HR in the C3HeB×SJL backcross. In the Hybrid Mouse Diversity Panel (58 strains, n = 5-6/strain) we found that HR (beats/min) ranged from 546 ± 12 in C58/J to 717 ± 7 in MA/MyJ mice. SBP (mmHg) ranged from 99 ± 6 in strain I/LnJ to 151 ± 4 in strain BXA4/PgnJ. GWA analyses were done using the HMDP, which revealed a locus (64.2-65.1 Mb) on chromosome 7 that colocalized with the QTL for elevated HR found in the C3HeB×SJL backcross. The peak association was observed for 17 SNPs that are localized within three GABA(A) receptor genes. In summary, we used a combined genetic approach to fine map a novel elevated HR locus on mouse chromosome 7.

PD08-09: PTPN12 Gene Expression Signature in Triple Negative Breast Cancer Cohort

PTPN12 tyrosine phosphatase may play a role in tumor development/progression in triple negative breast cancer patients (TNP). The effects of PTPN12 appear to be mediated through several tyrosine kinase receptors including EGFR, HER2, and PDGFR-beta. We investigated the variability associated with PTPN12 transcript in the microarray gene expression data obtained from 105 TNP as determined by IHC for ER and PR and IHC and FISH for HER2 during our clinical molecular profiling on solid tumors. The mRNA levels of PTPN12 in our cohort was highly variable suggesting a complex genetic regulation of PTPN12 transcription in TNP patients. The highly variable nature of PTPN12 mRNA levels lead us to perform a correlation-based analysis of the transcriptome in TNP samples to gain insight into pathways and cellular processes associated with PTPN12 variation. Specifically, we quantile-normalized and performed two-dimensional hierarchical clustering of the 1000 top correlated genes with PTPN12 expression across the 105 TNP samples. We identified seven distinct gene clusters and three distinct patient subpopulations. The three distinct subtypes of TNP were comprised of low expressing PTPN12 (median log2 expression of 12.3), medium expressing PTPN12 (median log2 expression of 13.1), and high expressing PTPN12 (median log2 expression of 13.4). From the 7 gene clusters identified, 6 were positively correlated and one cluster (cluster 3) was negatively correlated with PTPN12 expression. Upon examining the genes within each cluster, we found that all contain unique set of genes related to cell proliferation, cell death, cell motility, cell cycle regulation and other cancer related pathways. From the 7 clusters, clusters 1, 6 and 7 had the highest fraction of such genes. The gene expression pattern and the gene content of these three clusters is as follows: Cluster 1 (208 genes) were genes that were highly expressed in all three TNP groups and the expression of the genes was highest in high PTPN12 expressing patients. Functional classification of genes by the DAVID bioinformatics tool at NCBI showed several genes related to cancer, including 19 MAP kinase signaling genes, and 15 genes involved in regulation of apoptosis. Cluster 6 (150 genes) contained genes showing very low expression in low-expressing TNPs, moderately expressed in medium expressing TNPs, and highly expressed in high-expressing TNPs. Cluster 6 had the highest fraction of cancer related genes including 7 mitosis and cell cycle check point genes, 6 cytoskeletal genes, 4 phosphatases and 6 kinases including KRAS, 4 DNA repair genes, 8 signal transduction genes, 3 positive regulator of apoptosis, and 22 genes involved in regulation of gene expression. Cluster 7 (166 genes) contained genes that were expressed at low level throughout the three TNP groups but the severity of under-expression was higher in TNP patients with low PTPN12 expression. In cluster 7 there were 5 genes belonging to the EGFR and FGF signaling pathways, and 6 genes involved in mitosis. It is worth noting that EGFR pathway has been postulated to crosstalk with PTPN12 protein. All together our results provide support for the involvement of PTPN12 in cancer development and highlights a promising therapeutic target for TNP patients.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr PD08-09.

Abstract B92: Differences in gene expression between primary chemotherapy-naive prostate carcinomas and hormone-refractory prostate carcinomas

Introduction: Despite the widespread use of prostate specific antigen screening for early detection, prostate cancer remains the second leading cause of cancer related death among men in the US. Metastatic, hormone refractory prostate cancer (HRPCa) is the end stage, lethal form of the disease. Defining the molecular mechanisms underlying the transition of an androgen responsive prostate cancer represents an important clinical problem. Currently, no effective therapies exist for end stage, hormone refractory disease. In this study, we investigated the differentially expressed genes in primary prostate tumor vs. hormone refractory prostate tumor in order to identify potentially important therapeutic targets.

Methods: Formalin fixed paraffin embedded tumor samples from 14 cases (8 primary, chemotherapy-naïve and 6 locally recurrent, hormonal therapy-refractory carcinomas) were analyzed for the whole genome RNA (29,285 transcripts) microarray analysis using Illumina cDNA-mediated annealing, selection, extension and ligation (DASL) process with the HumanHT-12 v4 beadChip (Illumina Inc., San Diego, CA). Additionally, a select number of chemotherapy-predictive (theranostic) biomarkers were analyzed using immunohistochemical methods.

Results: Transcriptome analysis identified 299 genes highly differentially expressed between the primary and recurrent groups (p-value &lt; 0.01, Student's t-test). Using these 299 genes, we performed two dimensional hierarchical clustering and were able to discover 8 clusters corresponding to 5 general patterns: Pattern 1-underexpressed in both groups but more so in recurrent than primary (2 clusters with 28 and 22 genes respectively); Pattern 2-slightly overexpressed in primary and underexpressed or no change in recurrent (1 cluster with 65 genes); Pattern 3-slightly under expressed/ no change in primary and overexpressed in recurrent (1 cluster, 12 genes); Pattern 4-overexpressed in primary and no change in recurrent (1 cluster, 25 genes); Pattern 5 - overexpressed in both groups but more so in recurrent than primary (3 clusters with 58, 32, and 57 genes in each cluster).

Individual gene analysis revealed upregulation of androgen receptor (AR) mRNA in recurrent HRPCa. This was, however, accompanied by a down regulation of regulator of G-protein signaling 2 (RGS2) and ras responsive element binding protein 1 (RREB1). Consistent down regulation of excision repair cross-complementing group 1 (ERCC1) and thymidylate synthase (TS) proteins were also observed in HRPCa.

Conclusion: Up-regulation of AR in HRPCa is associated with down regulation of androgen receptor signaling co-regulators (RGS2 and RREB1), providing a mechanism of androgen-independent activation of AR in HRPCa. Platinum based drugs may have potential benefit in treating of HRPCa due to the down regulation of nucleotide excision repair protein ERCC1, while fluoropyrimidines are potentially beneficial due to the lack of TS protein expression in HRPCa. The heat map derived signature readily distinguishes between prostate cancer specimens from men who were treatment naïve vs. hormone refractory cancers with 5 different expression patterns.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2011 Nov 12-16; San Francisco, CA. Philadelphia (PA): AACR; Mol Cancer Ther 2011;10(11 Suppl):Abstract nr B92.

Frequency distribution of SPARC in triple-negative breast cancer patients

37

Background: SPARC (secreted protein acid rich in cysteine) belongs to a group of extracellular matrix proteins and promotes adhesion of cells from the matrix. It plays an important role in tumor development in breast cancer and has a significant bearing on patient prognosis and long term survival. It is also known to predict response to nab-paclitaxel in certain tumor types including breast cancer. In 2005, FDA approved a solvent free formulation of paclitaxel for the treatment of metastatic breast cancer that utilizes albumin bound (nab) technology (nab-paclitaxel). Clinical studies have shown that nab-paclitaxel is significantly more effective than paclitaxel. Our study evaluated the frequency distribution of SPARC among triple negative breast cancer patients in which identification of a novel therapeutic target is warranted.

Methods: In a total of 951 breast cancer patients, we analyzed tumor SPARC expression by immunohistochemistry (IHC) using a monoclonal (R&D Systems) and a polyclonal antibody (Exalpha Biologicals). Immunoreactivity was assessed by scoring the percentage of cells stained in each field and by the intensity of staining. A cutoff point of 2+ and >30% stained tumor cells were considered as positive.

Results: From our analysis of 951 breast cancer patients profiled, a total of 165 patients (17%) were triple negative for ER, PR and HER2. Within this pathologic subtype, 29% patients stained positive with SPARC monoclonal antibody and 21% stained positive with SPARC polyclonal antibody. The correlation of SPARC tumor staining with hormone receptor status will be presented in detail.

Conclusions: We conclude that SPARC over-expression is a functionally important feature of a subset of triple negative breast cancer patients. The triple negative subset of tumors generally has a more aggressive clinical course and does not benefit from conventional targeted therapies. Our study suggests that nab-paclitaxel may serve as a therapeutic agent for the subset of triple negative patients that over-express SPARC. To the best of our knowledge, this is the first study involving a large patient pool in which SPARC has been investigated in a single clinical laboratory using standardized IHC with two different SPARC antibodies.

Comparative analysis of proteome and transcriptome variation in mouse

The relationships between the levels of transcripts and the levels of the proteins they encode have not been examined comprehensively in mammals, although previous work in plants and yeast suggest a surprisingly modest correlation. We have examined this issue using a genetic approach in which natural variations were used to perturb both transcript levels and protein levels among inbred strains of mice. We quantified over 5,000 peptides and over 22,000 transcripts in livers of 97 inbred and recombinant inbred strains and focused on the 7,185 most heritable transcripts and 486 most reliable proteins. The transcript levels were quantified by microarray analysis in three replicates and the proteins were quantified by Liquid Chromatography–Mass Spectrometry using O(18)-reference-based isotope labeling approach. We show that the levels of transcripts and proteins correlate significantly for only about half of the genes tested, with an average correlation of 0.27, and the correlations of transcripts and proteins varied depending on the cellular location and biological function of the gene. We examined technical and biological factors that could contribute to the modest correlation. For example, differential splicing clearly affects the analyses for certain genes; but, based on deep sequencing, this does not substantially contribute to the overall estimate of the correlation. We also employed genome-wide association analyses to map loci controlling both transcript and protein levels. Surprisingly, little overlap was observed between the protein- and transcript-mapped loci. We have typed numerous clinically relevant traits among the strains, including adiposity, lipoprotein levels, and tissue parameters. Using correlation analysis, we found that a low number of clinical trait relationships are preserved between the protein and mRNA gene products and that the majority of such relationships are specific to either the protein levels or transcript levels. Surprisingly, transcript levels were more strongly correlated with clinical traits than protein levels. In light of the widespread use of high-throughput technologies in both clinical and basic research, the results presented have practical as well as basic implications.

An old dogma in biology states that, in every cell, the flow of biological information is from DNA to RNA to proteins and that the latter act as a working force to determine the organism's phenotype. This model predicts that changes in DNA that affect the clinical phenotype should also similarly change the cellular levels of RNA and protein levels. In this report, we test this prediction by looking at the concordance between DNA variation in population of mouse inbred strains, the RNA and protein variation in the liver tissue of these mice, and variation in metabolic phenotypes. We show that the relationship between various biological traits is not simple and that there is relatively little concordance of RNA levels and the corresponding protein levels in response to DNA perturbations. In addition, we also find that, surprisingly, metabolic traits correlate better to RNA levels than to protein levels. In light of current efforts in searching for the molecular bases of disease susceptibility in humans, our findings highlight the complexity of information flow that underlies clinical outcomes.

Mouse genome-wide association and systems genetics identify Asxl2 as a regulator of bone mineral density and osteoclastogenesis

Significant advances have been made in the discovery of genes affecting bone mineral density (BMD); however, our understanding of its genetic basis remains incomplete. In the current study, genome-wide association (GWA) and co-expression network analysis were used in the recently described Hybrid Mouse Diversity Panel (HMDP) to identify and functionally characterize novel BMD genes. In the HMDP, a GWA of total body, spinal, and femoral BMD revealed four significant associations (-log10P>5.39) affecting at least one BMD trait on chromosomes (Chrs.) 7, 11, 12, and 17. The associations implicated a total of 163 genes with each association harboring between 14 and 112 genes. This list was reduced to 26 functional candidates by identifying those genes that were regulated by local eQTL in bone or harbored potentially functional non-synonymous (NS) SNPs. This analysis revealed that the most significant BMD SNP on Chr. 12 was a NS SNP in the additional sex combs like-2 (Asxl2) gene that was predicted to be functional. The involvement of Asxl2 in the regulation of bone mass was confirmed by the observation that Asxl2 knockout mice had reduced BMD. To begin to unravel the mechanism through which Asxl2 influenced BMD, a gene co-expression network was created using cortical bone gene expression microarray data from the HMDP strains. Asxl2 was identified as a member of a co-expression module enriched for genes involved in the differentiation of myeloid cells. In bone, osteoclasts are bone-resorbing cells of myeloid origin, suggesting that Asxl2 may play a role in osteoclast differentiation. In agreement, the knockdown of Asxl2 in bone marrow macrophages impaired their ability to form osteoclasts. This study identifies a new regulator of BMD and osteoclastogenesis and highlights the power of GWA and systems genetics in the mouse for dissecting complex genetic traits.

Gene networks associated with conditional fear in mice identified using a systems genetics approach

Background

Our understanding of the genetic basis of learning and memory remains shrouded in mystery. To explore the genetic networks governing the biology of conditional fear, we used a systems genetics approach to analyze a hybrid mouse diversity panel (HMDP) with high mapping resolution.

Results

A total of 27 behavioral quantitative trait loci were mapped with a false discovery rate of 5%. By integrating fear phenotypes, transcript profiling data from hippocampus and striatum and also genotype information, two gene co-expression networks correlated with context-dependent immobility were identified. We prioritized the key markers and genes in these pathways using intramodular connectivity measures and structural equation modeling. Highly connected genes in the context fear modules included Psmd6, Ube2a and Usp33, suggesting an important role for ubiquitination in learning and memory. In addition, we surveyed the architecture of brain transcript regulation and demonstrated preservation of gene co-expression modules in hippocampus and striatum, while also highlighting important differences. Rps15a, Kif3a, Stard7, 6330503K22RIK, and Plvap were among the individual genes whose transcript abundance were strongly associated with fear phenotypes.

Conclusion

Application of our multi-faceted mapping strategy permits an increasingly detailed characterization of the genetic networks underlying behavior.

A high-resolution association mapping panel for the dissection of complex traits in mice

Systems genetics relies on common genetic variants to elucidate biologic networks contributing to complex disease-related phenotypes. Mice are ideal model organisms for such approaches, but linkage analysis has been only modestly successful due to low mapping resolution. Association analysis in mice has the potential of much better resolution, but it is confounded by population structure and inadequate power to map traits that explain less than 10% of the variance, typical of mouse quantitative trait loci (QTL). We report a novel strategy for association mapping that combines classic inbred strains for mapping resolution and recombinant inbred strains for mapping power. Using a mixed model algorithm to correct for population structure, we validate the approach by mapping over 2500 cis-expression QTL with a resolution an order of magnitude narrower than traditional QTL analysis. We also report the fine mapping of metabolic traits such as plasma lipids. This resource, termed the Hybrid Mouse Diversity Panel, makes possible the integration of multiple data sets and should prove useful for systems-based approaches to complex traits and studies of gene-by-environment interactions.

Copy number variation influences gene expression and metabolic traits in mice

Copy number variants (CNVs) are genomic segments which are duplicated or deleted among different individuals. CNVs have been implicated in both Mendelian and complex traits, including immune and behavioral disorders, but the study of the mechanisms by which CNVs influence gene expression and clinical phenotypes in humans is complicated by the limited access to tissues and by population heterogeneity. We now report studies of the effect of 19 CNVs on gene expression and metabolic traits in a mouse intercross between strains C57BL/6J and C3H/HeJ. We found that 83% of genes predicted to occur within CNVs were differentially expressed. The expression of most CNV genes was correlated with copy number, but we also observed evidence that gene expression was altered in genes flanking CNVs, suggesting that CNVs may contain regulatory elements for these genes. Several CNVs mapped to hotspots, genomic regions influencing expression of tens or hundreds of genes. Several metabolic traits including cholesterol, triglycerides, glucose and body weight mapped to three CNVs in the genome, in mouse chromosomes 1, 4 and 17. Predicted CNV genes, such as Itlna, Defcr-1, Trim12 and Trim34 were highly correlated with these traits. Our results suggest that CNVs have a significant impact on gene expression and that CNVs may be playing a role in the mechanisms underlying metabolic traits in mice.

Validation of candidate causal genes for obesity that affect shared metabolic pathways and networks

A major task in dissecting the genetics of complex traits is to identify causal genes for disease phenotypes. We previously developed a method to infer causal relationships among genes through the integration of DNA variation, gene transcription, and phenotypic information. Here we validated our method through the characterization of transgenic and knockout mouse models of candidate genes that were predicted to be causal for abdominal obesity. Perturbation of eight out of the nine genes, with Gas7, Me1 and Gpx3 being novel, resulted in significant changes in obesity related traits. Liver expression signatures revealed alterations in common metabolic pathways and networks contributing to abdominal obesity and overlapped with a macrophage-enriched metabolic network module that is highly associated with metabolic traits in mice and humans. Integration of gene expression in the design and analysis of traditional F2 intercross studies allows high confidence prediction of causal genes and identification of involved pathways and networks.

Weighted gene coexpression network analysis strategies applied to mouse weight

Systems-oriented genetic approaches that incorporate gene expression and genotype data are valuable in the quest for genetic regulatory loci underlying complex traits. Gene coexpression network analysis lends itself to identification of entire groups of differentially regulated genes-a highly relevant endeavor in finding the underpinnings of complex traits that are, by definition, polygenic in nature. Here we describe one such approach based on liver gene expression and genotype data from an F(2) mouse inter-cross utilizing weighted gene coexpression network analysis (WGCNA) of gene expression data to identify physiologically relevant modules. We describe two strategies: single-network analysis and differential network analysis. Single-network analysis reveals the presence of a physiologically interesting module that can be found in two distinct mouse crosses. Module quantitative trait loci (mQTLs) that perturb this module were discovered. In addition, we report a list of genetic drivers for this module. Differential network analysis reveals differences in connectivity and module structure between two networks based on the liver expression data of lean and obese mice. Functional annotation of these genes suggests a biological pathway involving epidermal growth factor (EGF). Our results demonstrate the utility of WGCNA in identifying genetic drivers and in finding genetic pathways represented by gene modules. These examples provide evidence that integration of network properties may well help chart the path across the gene-trait chasm.

Integrating genetic and network analysis to characterize genes related to mouse weight

Systems biology approaches that are based on the genetics of gene expression have been fruitful in identifying genetic regulatory loci related to complex traits. We use microarray and genetic marker data from an F2 mouse intercross to examine the large-scale organization of the gene co-expression network in liver, and annotate several gene modules in terms of 22 physiological traits. We identify chromosomal loci (referred to as module quantitative trait loci, mQTL) that perturb the modules and describe a novel approach that integrates network properties with genetic marker information to model gene/trait relationships. Specifically, using the mQTL and the intramodular connectivity of a body weight-related module, we describe which factors determine the relationship between gene expression profiles and weight. Our approach results in the identification of genetic targets that influence gene modules (pathways) that are related to the clinical phenotypes of interest.

Integrating genetic and network analysis to characterize genes related to mouse weight

Systems biology approaches that are based on the genetics of gene expression have been fruitful in identifying genetic regulatory loci related to complex traits. We use microarray and genetic marker data from an F2 mouse intercross to examine the large-scale organization of the gene co-expression network in liver, and annotate several gene modules in terms of 22 physiological traits. We identify chromosomal loci (referred to as module quantitative trait loci, mQTL) that perturb the modules and describe a novel approach that integrates network properties with genetic marker information to model gene/trait relationships. Specifically, using the mQTL and the intramodular connectivity of a body weight-related module, we describe which factors determine the relationship between gene expression profiles and weight. Our approach results in the identification of genetic targets that influence gene modules (pathways) that are related to the clinical phenotypes of interest.

Complex inheritance of the 5-lipoxygenase locus influencing atherosclerosis in mice

We previously mapped a locus on chromosome 6 with a large effect (LOD > 6) on aortic lesion size in a (C57BL/6J x CAST/Ei) F(2) cross and identified arachidonate 5-lipoxygenase (5LO) as a candidate gene in this region. Subsequent studies with the 5LO knockout model showed effects on atherosclerosis and aortic aneurysms. We now report detailed genetic analysis of the chromosome 6 locus. We created a panel of overlapping and reciprocal subcongenic lines from the B6.CAST Ldlr(-/-) chromosome 6 congenic strain (CON6.Ldlr(-/-)) and analyzed aortic lesion size in different subcongenic lines. Our results revealed that there are at least two subregions, designated as Ath37 and Ath38 that affect the size of aortic lesions independently of 5LO. We also showed that homozygote 5LO null mice develop smaller atherosclerotic lesions. We conclude that the relation between the mouse chromosome 6 locus and atherosclerosis is complex and is due to at least two genes with large effects within this region. This complexity should be considered when interpreting results of knockout studies.

Hepatocellular carcinoma in Txnip-deficient mice

The molecular pathogenesis and the genetic aberrations that lead to the progression of hepatocellular carcinoma (HCC) are largely unknown. Here, we demonstrate that the thioredoxin interacting protein (Txnip) gene is a candidate tumor suppressor gene in vivo. We previously showed that the recombinant inbred congenic strain HcB-19 has a spontaneous mutation of the Txnip gene, and we now show that the strain has dramatically increased incidence of HCC, and that the HCC cosegregates with the Txnip mutation. Approximately 40% of the Txnip-deficient mice developed hepatic tumors with an increased prevalence in male mice. Visible tumors develop as early as 8 months of age. Histological analysis confirmed the morphology of HCC in the Txnip-deficient mice. Molecular markers of HCC, alpha-fetoprotein and p53, were increased in tumors of Txnip-deficient mice. The upregulation of p53 preceded tumor development; however, bromodeoxyuridine (BrdU) labeling of normal hepatic tissue of Txnip-deficient mice did not reveal increased cell proliferation. Finally, microarray analyses of tumor, non-tumor adjacent, and normal tissue of Txnip-deficient mice highlighted the genetic differences leading to the predisposition and onset of HCC. Our findings suggest that Txnip deficiency is sufficient to initiate HCC and suggest novel mechanisms in hepatocarcinogenesis.