Validation of Immunotherapy Response Score as Predictive of Pan-solid Tumor Anti-PD-1/PD-L1 Benefit

Immunotherapy response score (IRS) integrates tumor mutation burden (TMB) and quantitative expression biomarkers to predict anti-PD-1/PD-L1 [PD-(L)1] monotherapy benefit. Here, we evaluated IRS in additional cohorts. Patients from an observational trial (NCT03061305) treated with anti-PD-(L)1 monotherapy were included and assigned to IRS-High (-H) versus -Low (-L) groups. Associations with real-world progression-free survival (rwPFS) and overall survival (OS) were determined by Cox proportional hazards (CPH) modeling. Those with available PD-L1 IHC treated with anti-PD-(L)1 with or without chemotherapy were separately assessed. Patients treated with PD-(L)1 and/or chemotherapy (five relevant tumor types) were assigned to three IRS groups [IRS-L divided into IRS-Ultra-Low (-UL) and Intermediate-Low (-IL), and similarly assessed]. In the 352 patient anti-PD-(L)1 monotherapy validation cohort (31 tumor types), IRS-H versus IRS-L patients had significantly longer rwPFS and OS. IRS significantly improved CPH associations with rwPFS and OS beyond microsatellite instability (MSI)/TMB alone. In a 189 patient (10 tumor types) PD-L1 IHC comparison cohort, IRS, but not PD-L1 IHC nor TMB, was significantly associated with anti-PD-L1 rwPFS. In a 1,103-patient cohort (from five relevant tumor types), rwPFS did not significantly differ in IRS-UL patients treated with chemotherapy versus chemotherapy plus anti-PD-(L)1, nor in IRS-H patients treated with anti-PD-(L)1 versus anti-PD-(L)1 + chemotherapy. IRS associations were consistent across subgroups, including both Europeans and non-Europeans. These results confirm the utility of IRS utility for predicting pan-solid tumor PD-(L)1 monotherapy benefit beyond available biomarkers and demonstrate utility for informing on anti-PD-(L)1 and/or chemotherapy treatment.

Significance

This study confirms the utility of the integrative IRS biomarker for predicting anti-PD-L1/PD-1 benefit. IRS significantly improved upon currently available biomarkers, including PD-L1 IHC, TMB, and MSI status. Additional utility for informing on chemotherapy, anti-PD-L1/PD-1, and anti-PD-L1/PD-1 plus chemotherapy treatments decisions is shown.

Abstract 2171: A multivariate biomarker predicts sacituzumab govitecan response in solid tumors

Despite the recent clinical success of antibody drug conjugates (ADC) in oncology, predictive biomarkers are lacking, potentially limiting their impact. Herein, we evaluated the ability of candidate biomarkers alone and in combination to predict objective response rates observed in solid tumor patients treated with the TROP2-targeted ADC, sacituzumab govitecan (SG), as determined in the IMMU-12-01 basket trial. We leveraged available next generation sequencing (NGS)-based molecular profiling data from an independent advanced solid tumor cohort (n = 23,968) and developed a multivariate biomarker algorithm that produced biomarker positive rates correlating with the objective response rates (ORR) observed in IMMU-12-01. Candidate biomarkers evaluated included TROP2 gene expression, proliferation (by gene expression) and tumor cellularity. Notably, while TROP2 gene expression was highly correlated with protein expression across 45 tumor types (r = 0.93), TROP2 gene expression alone did not significantly predict ORR across 9 tumor types (r = 0.40, p = 0.29). In contrast, a biomarker algorithm combining TROP2 and proliferation by gene expression with tumor cellularity strongly predicted response both when using tumor type-specific biomarker rates in a discovery cohort (r = 0.83, p = 0.006) and in an independent validation cohort (r = 0.82, p = 0.007). These results indicate that the multivariate biomarker accounts for 67% of the variability observed in response rates and may thus identify patients likely to benefit from SG. Among tumor types with objective responses in IMMU-12-01, biomarker positive rates ranged from 9.9% in colorectal cancer to 57.4% in bladder cancer. Additional tumor types with biomarker positive rates >30% included cancers of the head and neck, cervix, salivary gland, skin (non-melanoma) and ovary, all with positive biomarker rates >30%. Interestingly, most tumor types had biomarker positive rates >5%, suggesting the potential for a tumor type-agnostic approach to patient selection. Considering SG and other ADC’s mechanism of action, a plausible model for response is that (1) higher target expression increases ADC drug delivery, (2) higher tumor cellularity increases ADC bystander effect and (3) higher tumor cell proliferation increases tumor cell death. In summary, we uncovered a novel biomarker algorithm capable of predicting SG response across solid tumors that may be generalizable to ADCs as a class, with the potential to further optimize use and maximize benefit.

Citation Format: Nickolay A. Khazanov, Laura E. Lamb, Daniel H. Hovelson, Kat Kwiatkowski, D. Bryan Johnson, Daniel R. Rhodes, Scott A. Tomlins. A multivariate biomarker predicts sacituzumab govitecan response in solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2171.

Abstract 968: Evaluation of Her2 RNA expression as a potential predictive biomarker for anti-Her2 therapy

Trastuzumab deruxtecan (Enhertu) is effective in "HER2 Low" breast cancer, defined by 1+ or 2+ expression by immunohistochemistry (IHC). Interest has now turned toward defining a sub-population of IHC 0+ tumors that may have HER2 expression below the limit of IHC detection/quantification and may thus also be responsive. We previously validated a high dynamic range HER2 RNA expression assay run as part of our comprehensive genomic profiling test, StrataNGS.

Herein, we evaluated the HER2 RNA expression data together with copy number and clinical outcome data from the Strata Clinical Molecular Database (SCMD) in advanced breast cancer (n = 3,063) and other advanced solid tumors (n = 26,715). As expected, HER2 gene expression was significantly higher in tumors with DNA amplification (>=6 copies; median: 13.9 vs. 10.0 in log2 units; p < 1e-100). Despite similar copy number levels in amplified breast vs. other cancers (median: 21.8 vs. 19.8 copies), HER2 expression levels were ~2-fold higher (median: 14.5 vs. 13.5; p = 1.3e-10). Similarly, HER2 expression levels were higher in non-amplified breast vs. other cancers (median: 10.7 vs. 9.9; p<1e-100), suggesting that DNA amplification and cell lineage affect HER2 expression. Using our previously validated HER2 threshold, among 75 eligible SCMD breast cancer patients treated with 1st or 2nd line systemic trastuzumab or pertuzumab containing therapy, HER2 RNA High patients (n=46, 59%) had significantly longer time to next therapy (TTNT) compared to HER2 RNA Not High patients (median TTNT 26.9 vs. 5.6 months, adjusted hazard ratio 0.31, p=0.005 when adjusted for 1st vs. 2nd line, pertuzumab inclusion, and inclusion of chemotherapy or hormonal therapy).

In patients with available IHC data (n = 388), HER2 RNA expression trended with IHC across the 0-3+ range, however, while 3+ tumors had distinctly high RNA expression (median: 14.4), 0-2+ tumors had lower expression with overlapping distributions (median: 10.5, 10.9, 11.5, respectively), suggesting that 0-2+ tumors do not represent distinct biological groups, but rather a continuum of low expression. We defined a HER2 RNA Low threshold (>10.6), corresponding to the top 75% of IHC 1-2+ breast cancers. Importantly, at this threshold, nearly half (44.1%) of 0+ breast cancers were also classified as HER2 RNA Low. Additionally, 25.8% of all non-breast solid tumors were classified as HER2 RNA Low. Given that HER2 RNA High predicted benefit from 1st generation anti-HER2 therapies, future studies should consider HER2 RNA Low as an alternative biomarker to Her2 IHC Low, with the opportunity to further expand trastuzumab deruxtecan use into the IHC 0+ breast cancer population and potentially to additional solid tumors.

Citation Format: Laura E. Lamb, Nickolay A. Khazanov, Daniel H. Hovelson, Kat Kwiatkowski, D. Bryan Johnson, Daniel R. Rhodes, Scott A. Tomlins. Evaluation of Her2 RNA expression as a potential predictive biomarker for anti-Her2 therapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 968.

Development and Validation of StrataNGS, a Multiplex PCR, Semiconductor Sequencing-Based Comprehensive Genomic Profiling Test

Despite widespread use in targeted tumor testing, multiplex PCR/semiconductor (Ion Torrent) sequencing-based assessment of all comprehensive genomic profiling (CGP) variant classes has been limited. Herein, we describe the development and validation of StrataNGS, a 429-gene, multiplex PCR/semiconductor sequencing-based CGP laboratory-developed test performed on co-isolated DNA and RNA from formalin-fixed, paraffin-embedded tumor specimens with ≥2 mm² tumor surface area. Validation was performed in accordance with MolDX CGP validation guidelines using 1986 clinical formalin-fixed, paraffin-embedded samples and an in-house developed optimized bioinformatics pipeline. Across CGP variant classes, accuracy ranged from 0.945 for tumor mutational burden (TMB) status to >0.999 for mutations and gene fusions, positive predictive value ranged from 0.915 for TMB status to 1.00 for gene fusions, and reproducibility ranged from 0.998 for copy number alterations to 1.00 for splice variants and insertions/deletions. StrataNGS TMB estimates were highly correlated to those from whole exome- or FoundationOne CDx-determined TMB (Pearson r = 0.998 and 0.960, respectively); TMB reproducibility was 0.996 (concordance correlation coefficient). Limit of detection for all variant classes was <20% tumor content. Together, we demonstrate that multiplex PCR/semiconductor sequencing-based tumor tissue CGP is feasible using optimized bioinformatic approaches described herein.

Real-World Performance of a Comprehensive Genomic Profiling Test Optimized for Small Tumor Samples

PURPOSE

Tissue-based comprehensive genomic profiling (CGP) is increasingly used for treatment selection in patients with advanced cancer; however, tissue availability may limit widespread implementation. Here, we established real-world CGP tissue availability and assessed CGP performance on consecutively received samples.

MATERIALS AND METHODS

We conducted a post hoc, nonprespecified analysis of 32,048 consecutive tumor tissue samples received for StrataNGS, a multiplex polymerase chain reaction (PCR)-based comprehensive genomic profiling (PCR-CGP) test, as part of an ongoing observational trial (NCT03061305). Sample characteristics and PCR-CGP performance were assessed across all tested samples, including exception samples not meeting minimum input quality control (QC) requirements (< 20% tumor content [TC], < 2 mm2 tumor surface area [TSA], DNA or RNA yield < 1 ng/µL, or specimen age > 5 years). Tests reporting ≥ 1 prioritized alteration or meeting TC and sequencing QC were considered successful. For prostate carcinoma and lung adenocarcinoma, tests reporting ≥ 1 actionable or informative alteration or meeting TC and sequencing QC were considered actionable.

RESULTS

Among 31,165 (97.2%) samples where PCR-CGP was attempted, 10.7% had < 20% TC and 59.2% were small (< 25 mm2 tumor surface area). Of 31,101 samples evaluable for input requirements, 8,089 (26.0%) were exceptions not meeting requirements. However, 94.2% of the 31,101 tested samples were successfully reported, including 80.5% of exception samples. Positive predictive value of PCR-CGP for ERBB2 amplification in exceptions and/or sequencing QC-failure breast cancer samples was 96.7%. Importantly, 84.0% of tested prostate carcinomas and 87.9% of lung adenocarcinomas yielded results informing treatment selection.

CONCLUSION

Most real-world tissue samples from patients with advanced cancer desiring CGP are limited, requiring optimized CGP approaches to produce meaningful results. An optimized PCR-CGP test, coupled with an inclusive exception testing policy, delivered reportable results for > 94% of samples, potentially expanding the proportion of CGP-testable patients and impact of biomarker-guided therapies.

Virus-positive Merkel Cell Carcinoma Is an Independent Prognostic Group with Distinct Predictive Biomarkers

PURPOSE

Merkel cell carcinoma (MCC) is an aggressive cutaneous neuroendocrine carcinoma that can be divided into two classes: virus-positive (VP) MCC, associated with oncogenic Merkel cell polyomavirus (MCPyV); and virus-negative (VN) MCC, associated with photodamage.

EXPERIMENTAL DESIGN

We classified 346 MCC tumors from 300 patients for MCPyV using a combination of IHC, ISH, and qPCR assays. In a subset of tumors, we profiled mutation status and expression of cancer-relevant genes. MCPyV and molecular profiling results were correlated with disease-specific outcomes. Potential prognostic biomarkers were further validated by IHC.

RESULTS

A total of 177 tumors were classified as VP-MCC, 151 tumors were VN-MCC, and 17 tumors were indeterminate. MCPyV positivity in primary tumors was associated with longer disease-specific and recurrence-free survival in univariate analysis, and in multivariate analysis incorporating age, sex, immune status, and stage at presentation. Prioritized oncogene or tumor suppressor mutations were frequent in VN-MCC but rare in VP-MCC. TP53 mutation developed with recurrence in one VP-MCC case. Importantly, for the first time we find that VP-MCC and VN-MCC display distinct sets of prognostic molecular biomarkers. For VP-MCC, shorter survival was associated with decreased expression of immune markers including granzyme and IDO1. For VN-MCC, shorter survival correlated with high expression of several genes including UBE2C.

CONCLUSIONS

MCPyV status is an independent prognostic factor for MCC. Features of the tumor genome, transcriptome, and microenvironment may modify prognosis in a manner specific to viral status. MCPyV status has clinicopathologic significance and allows for identification of additional prognostic subgroups.

Identification of TP53RK-Binding Protein (TPRKB) Dependency in TP53-Deficient Cancers

Tumor protein 53 (TP53; p53) is the most frequently altered gene in human cancer. Identification of vulnerabilities imposed by TP53 alterations may enable effective therapeutic approaches. Through analyzing short hairpin RNA (shRNA) screening data, we identified TP53RK-Binding Protein (TPRKB), a poorly characterized member of the tRNA-modifying EKC/KEOPS complex, as the most significant vulnerability in TP53-mutated cancer cell lines. In vitro and in vivo, across multiple benign-immortalized and cancer cell lines, we confirmed that TPRKB knockdown in TP53-deficient cells significantly inhibited proliferation, with minimal effect in TP53 wild-type cells. TP53 reintroduction into TP53-null cells resulted in loss of TPRKB sensitivity, confirming the importance of TP53 status in this context. In addition, cell lines with mutant TP53 or amplified MDM2 (E3-ubiquitin ligase for TP53) also showed high sensitivity to TPRKB knockdown, consistent with TPRKB dependence in a wide array of TP53-altered cancers. Depletion of other EKC/KEOPS complex members exhibited TP53-independent effects, supporting complex-independent functions of TPRKB. Finally, we found that TP53 indirectly mediates TPRKB degradation, which was rescued by coexpression of PRPK, an interacting member of the EKC/KEOPS complex, or proteasome inhibition. Together, these results identify a unique and specific requirement of TPRKB in a variety of TP53-deficient cancers. IMPLICATIONS: Cancer cells with genomic alterations in TP53 are dependent on TPRKB.

The CARMA3-Bcl10-MALT1 Signalosome Drives NFκB Activation and Promotes Aggressiveness in Angiotensin II Receptor-Positive Breast Cancer

The angiotensin II receptor AGTR1, which mediates vasoconstrictive and inflammatory signaling in vascular disease, is overexpressed aberrantly in some breast cancers. In this study, we established the significance of an AGTR1-responsive NFκB signaling pathway in this breast cancer subset. We documented that AGTR1 overexpression occurred in the luminal A and B subtypes of breast cancer, was mutually exclusive of HER2 expression, and correlated with aggressive features that include increased lymph node metastasis, reduced responsiveness to neoadjuvant therapy, and reduced overall survival. Mechanistically, AGTR1 overexpression directed both ligand-independent and ligand-dependent activation of NFκB, mediated by a signaling pathway that requires the triad of CARMA3, Bcl10, and MALT1 (CBM signalosome). Activation of this pathway drove cancer cell-intrinsic responses that include proliferation, migration, and invasion. In addition, CBM-dependent activation of NFκB elicited cancer cell-extrinsic effects, impacting endothelial cells of the tumor microenvironment to promote tumor angiogenesis. CBM/NFκB signaling in AGTR1⁺ breast cancer therefore conspires to promote aggressive behavior through pleiotropic effects. Overall, our results point to the prognostic and therapeutic value of identifying AGTR1 overexpression in a subset of HER2-negative breast cancers, and they provide a mechanistic rationale to explore the repurposing of drugs that target angiotensin II-dependent NFκB signaling pathways to improve the treatment of this breast cancer subset.Significance: These findings offer a mechanistic rationale to explore the repurposing of drugs that target angiotensin action to improve the treatment of AGTR1-expressing breast cancers. Cancer Res; 78(5); 1225-40. ©2017 AACR.

Genomic Profiling of Penile Squamous Cell Carcinoma Reveals New Opportunities for Targeted Therapy

Penile squamous cell carcinoma (PeSCCA) is a rare malignancy for which there are limited treatment options due to a poor understanding of the molecular alterations underlying disease development and progression. Therefore, we performed comprehensive, targeted next-generation sequencing to identify relevant somatic genomic alterations in a retrospective cohort of 60 fixed tumor samples from 43 PeSCCA cases (including 14 matched primary/metastasis pairs). We identified a median of two relevant somatic mutations and one high-level copy-number alteration per sample (range, 0-5 and 0-6, respectively). Expression of HPV and p16 was detectable in 12% and 28% of patients, respectively. Furthermore, advanced clinical stage, lack of p16 expression, and MYC and CCND1 amplifications were significantly associated with shorter time to progression or PeSCCA-specific survival. Notably, four cases harbored EGFR amplifications and one demonstrated CDK4 amplification, genes for which approved and investigational targeted therapies are available. Importantly, although paired primary tumors and lymph node metastases were largely homogeneous for relevant somatic mutations, we identified heterogeneous EGFR amplification in primary tumor/lymph node metastases in 4 of 14 cases, despite uniform EGFR protein overexpression. Likewise, activating HRAS mutations occurred in 8 of 43 cases. Taken together, we provide the first comprehensive molecular PeSCCA analysis, which offers new insight into potential precision medicine approaches for this disease, including strategies targeting EGFR.

Development and validation of a scalable next-generation sequencing system for assessing relevant somatic variants in solid tumors

Next-generation sequencing (NGS) has enabled genome-wide personalized oncology efforts at centers and companies with the specialty expertise and infrastructure required to identify and prioritize actionable variants. Such approaches are not scalable, preventing widespread adoption. Likewise, most targeted NGS approaches fail to assess key relevant genomic alteration classes. To address these challenges, we predefined the catalog of relevant solid tumor somatic genome variants (gain-of-function or loss-of-function mutations, high-level copy number alterations, and gene fusions) through comprehensive bioinformatics analysis of >700,000 samples. To detect these variants, we developed the Oncomine Comprehensive Panel (OCP), an integrative NGS-based assay [compatible with < 20 ng of DNA/RNA from formalin-fixed paraffin-embedded (FFPE) tissues], coupled with an informatics pipeline to specifically identify relevant predefined variants and created a knowledge base of related potential treatments, current practice guidelines, and open clinical trials. We validated OCP using molecular standards and more than 300 FFPE tumor samples, achieving >95% accuracy for KRAS, epidermal growth factor receptor, and BRAF mutation detection as well as for ALK and TMPRSS2:ERG gene fusions. Associating positive variants with potential targeted treatments demonstrated that 6% to 42% of profiled samples (depending on cancer type) harbored alterations beyond routine molecular testing that were associated with approved or guideline-referenced therapies. As a translational research tool, OCP identified adaptive CTNNB1 amplifications/mutations in treated prostate cancers. Through predefining somatic variants in solid tumors and compiling associated potential treatment strategies, OCP represents a simplified, broadly applicable targeted NGS system with the potential to advance precision oncology efforts.

Abstract A1-43: Targeted amplicon-based next-generation sequencing of routine solid tumor specimens to detect clinically relevant somatic alterations

Background: Although precision medicine approaches have revolutionized oncology, widespread adoption requires robust, inexpensive approaches enabling the targeted assessment of all relevant alteration classes from routine tissue samples.

Methods: Here we interrogated &gt;7,000 cancer exomes and transcriptomes, along with &gt;30,000 array based cancer genomes to identify recurrent somatic alterations (mutations, copy number alterations [CNAs] and gene fusions) across solid tumors. From this analysis, we developed and validated an integrated multiplexed PCR based Ion Torrent next generation sequencing panel (Oncomine Cancer Research Panel [OCP]) targeting the actionable somatic cancer genome optimized for 20ng of formalin-fixed, paraffin-embedded (FFPE) tissue isolated DNA/RNA.

Results: We validated the OCP using FFPE cell line mixtures, as well as a prospective cohort of 104 FFPE tumor specimens sent for concurrent clinical molecular testing, with &gt;97% sensitivity and specificity for the presence/absence of KRAS, EGFR, BRAF and ALK point mutations, indels or gene fusions in this molecular testing cohort. We also applied the OCP to 100 lung cancers, identifying known and novel alterations, including ALK and ROS1 gene fusions. Lastly, applying the OCP to 116 prostate cancers, including 50 previously treated samples, we recapitulated known molecular subtypes, observed distinct profiles according to previous treatment and obtained 100% concordance for isoform specific TMPRSS2:ERG gene fusion detection compared to qPCR. Additionally, OCP profiling supports a novel molecular subtype of prostate cancer defined by IDH1 R132 hotspot mutations and informed on resistance mechanisms in a pre- and post-treatment sample pair. Importantly, 44%, 35% and 9% of patients in the molecular testing, lung and prostate cancer cohorts, respectively, harbored additional alterations (beyond routine molecular testing) associated with FDA approved or NCCN guideline referenced therapies.

Conclusions: Through analysis of both DNA and RNA to assess the actionable somatic cancer genome, the validated OCP panel may have utility in both clinical and research settings.

Citation Format: Daniel H. Hovelson, Andrew S. McDaniel, Bryan Johnson, Andi K. Cani, Kate Rhodes, Paul D. Williams, Chia-Jen Liu, Santhoshi Bandla, Catherine S. Grasso, Michael J. Quist, Seth Sadis, Daniel R. Rhodes, Scott A. Tomlins. Targeted amplicon-based next-generation sequencing of routine solid tumor specimens to detect clinically relevant somatic alterations. [abstract]. In: Proceedings of the AACR Special Conference on Translation of the Cancer Genome; Feb 7-9, 2015; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(22 Suppl 1):Abstract nr A1-43.

Next-Gen Sequencing Exposes Frequent MED12 Mutations and Actionable Therapeutic Targets in Phyllodes Tumors

Phyllodes tumors are rare fibroepithelial tumors with variable clinical behavior accounting for a small subset of all breast neoplasms, yet little is known about the genetic alterations that drive tumor initiation and/or progression. Here, targeted next-generation sequencing (NGS) was used to identify somatic alterations in formalin-fixed paraffin-embedded (FFPE) patient specimens from malignant, borderline, and benign cases. NGS revealed mutations in mediator complex subunit 12 (MED12) affecting the G44 hotspot residue in the majority (67%) of cases spanning all three histologic grades. In addition, loss-of-function mutations in p53 (TP53) as well as deleterious mutations in the tumor suppressors retinoblastoma (RB1) and neurofibromin 1 (NF1) were identified exclusively in malignant tumors. High-level copy-number alterations (CNA) were nearly exclusively confined to malignant tumors, including potentially clinically actionable gene amplifications in IGF1R and EGFR. Taken together, this study defines the genomic landscape underlying phyllodes tumor development, suggests potential molecular correlates to histologic grade, expands the spectrum of human tumors with frequent recurrent MED12 mutations, and identifies IGF1R and EGFR as potential therapeutic targets in malignant cases.

IMPLICATIONS

Integrated genomic sequencing and mutational profiling provides insight into the molecular origin of phyllodes tumors and indicates potential druggable targets in malignant disease. Visual Overview: http://mcr.aacrjournals.org/content/early/2015/04/02/1541-7786.MCR-14-0578/F1.large.jpg.

Abstract 3173: High-throughput, systematic analysis of paired-end next-generation sequencing data to characterize the gene fusion landscape in cancer

Gene fusions encode oncogenic drivers in hematological and solid tumors and are often associated with dramatic clinical responses with the appropriate targeted agents. In principle, massively parallel paired-end sequencing can identify structural rearrangements in tumor genomes and transcriptomes. However, computational methods to identify gene fusions are varied, still evolving and largely trained on cell line data. We sought to develop systematic methods to characterize known oncogenic gene fusions and to discover novel gene fusions in cancer. RNASeq data for approximately 3,400 clinical cases from 16 cancer types was obtained from the Cancer Genomics Hub (CGHub) of The Cancer Genome Atlas (TCGA). We surveyed the performance of several gene fusion callers and chose two (deFuse and TopHat) for further method development. An analysis pipeline was developed and executed in parallel on a high-performance computing cluster. Filtering and annotation was conducted on the aggregated data as a post-processing step, to enable exploratory analyses of various filters. We optimized filtering approaches on datasets that included known standards (e.g., TMPRSS2-ERG in prostate adenocarcinoma, PML-RARA in acute myeloid leukemia, etc.) to enrich for these and other gene fusions with correct 5’-3’ orientation while excluding cases with ambiguous breakpoints and spanning reads, alignment errors, and read-through transcripts from adjacent genes. Predicted fusions were summarized based on the occurrence of unique genes participating in fusions with multiple partners and of unique gene pairs, each within specific diseases. Elevated expression was observed after the predicted breakpoint of the 3’ gene in cases positive for predicted fusions, and added important confirmatory evidence. Thus, we characterized the incidence and distribution of several known oncogenic gene fusions including EML4-ALK and CCDC6-RET while expanding the number of gene partners identified in combination with oncogenes such as ROS1. In addition to characterizing the incidence and distribution of 31 known gene fusions, we nominated over 100 novel gene fusion pairs. One example of a novel gene fusion susceptible to available targeted therapy was FGFR3-TACC3 in 4% of bladder cancer, 2% of squamous cell lung carcinoma, and 1% each of glioblastoma and head and neck squamous cell carcinoma. Computational methods are now poised to complement biochemical approaches in the definition of the gene fusion landscape in cancer.

Citation Format: Seth E. Sadis, Nickolay A. Khazanov, Armand R. Bankhead, Dinesh Cyanam, Paul D. Williams, Sean F. Eddy, Peter J. Wyngaard, Daniel R. Rhodes. High-throughput, systematic analysis of paired-end next-generation sequencing data to characterize the gene fusion landscape in cancer. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3173. doi:10.1158/1538-7445.AM2013-3173

Identification of targetable FGFR gene fusions in diverse cancers

Through a prospective clinical sequencing program for advanced cancers, four index cases were identified which harbor gene rearrangements of FGFR2, including patients with cholangiocarcinoma, breast cancer, and prostate cancer. After extending our assessment of FGFR rearrangements across multiple tumor cohorts, we identified additional FGFR fusions with intact kinase domains in lung squamous cell cancer, bladder cancer, thyroid cancer, oral cancer, glioblastoma, and head and neck squamous cell cancer. All FGFR fusion partners tested exhibit oligomerization capability, suggesting a shared mode of kinase activation. Overexpression of FGFR fusion proteins induced cell proliferation. Two bladder cancer cell lines that harbor FGFR3 fusion proteins exhibited enhanced susceptibility to pharmacologic inhibition in vitro and in vivo. Because of the combinatorial possibilities of FGFR family fusion to a variety of oligomerization partners, clinical sequencing efforts, which incorporate transcriptome analysis for gene fusions, are poised to identify rare, targetable FGFR fusions across diverse cancer types.

Abstract A19: A qPCR assay, OncoScore Colon, predicts resistance to cetuximab in formalin-fixed, paraffin-embedded colorectal cancer tissue independent of KRAS status

Gene expression modules derived from an unsupervised analysis of 20 independent microarray datasets comprising more than 2,000 colorectal cancer patients were identified. Each module represents a set of highly co-expressed genes related to an important aspect of underlying cancer variability. Modules containing genes related to epithelial and mesenchymal biology associated with sensitivity and resistance to EGFR family targeted inhibitors (gefitinib and lapatinib), respectively. In retrospective analysis of clinical samples, the epithelial-mesenchymal axis associated with cetuximab response in two independent patient cohorts. The first study was a Phase II clinical trial (Khambata-Ford et al., J Clin Oncol, 2007) with accompanying microarray data from pre-treatment metastatic colorectal tumor biopsies. Expression of the modules was determined by normalizing and averaging co-expressed module genes. Patients with a more epithelial and less mesenchymal module expression profile were enriched for cetuximab response. An independent cohort of patients was analyzed using module scores that were generated from a qPCR gene expression module test, OncoScore™ Colon, which quantifies modules by averaging three representative module genes relative to housekeeping genes using formalin-fixed-paraffin-embedded primary tumor samples. In these patients, presence of the mesenchymal module was significantly associated with a decrease in progression free survival. Notably, the status of the mesenchymal module was independent of KRAS mutation status—as KRAS mutations occurred in both mesenchymal module-positive and -negative patients. Further clinical studies are ongoing to continue to support the development of the OncoScore™ Colon assay and to further test the predictive capacity of the module with regards to cetuximab resistance and other MAPK pathway inhibitors. This study demonstrates the value of a gene expression module-based qPCR panel for stratifying colorectal cancer patients for treatment response, and suggests that our approach may have immediate utility for cetuximab treatment response prediction.

Abstract 3664: Breast cancer companion diagnostic platform based on objectively defined tumor co-expression patterns stratifies multiple clinical and therapeutic endpoints comparison to existing molecular subtyping definitions

Gene expression profiles of human breast tumors have greatly expanded our understanding of the genes and pathways that underlie breast cancer. Profiling studies have also supported a molecular classification of breast cancer. The resulting molecular subtypes Luminal, Basal-like, ERBB2+, and Normal-like were shown to have different prognostic and predictive characteristics. Related studies have led to a proliferation of multigene prognostic and predictive diagnostic tests. Two independent multigene tests, OncoType Dx and MammaPrint, have been shown to be helpful in predicting the risk of recurrence of patients with early stage breast cancer. Current multigene tests consistently prioritize the proliferation, estrogen receptor (ER), and ERBB2 pathways. An alternative approach to identifying key molecular variables within breast cancer is based on a definition of objectively defined tumor co-expression patterns. To this end, we defined co-expression patterns within 56 independent breast cancer molecular profiling datasets representing &gt;5,000 unique patients. We then performed a meta-analysis across datasets to define the most robust, consistently occurring co-expression patterns. These patterns, termed modules, recapitulate the proliferation, ER, and ERBB2 pathways, but also monitor expression of other important variables including core cancer cell growth pathways, immune signaling and microenvironment, and hallmark genomic aberrations. An important feature of co-expression patterns is that a small number of genes serve as an effective surrogate for each module. Thus, we created a single multigene qPCR test that measures the expression of 18 distinct breast cancer modules and validated the test for use with formalin-fixed paraffin-embedded (FFPE) tumor samples. In retrospective microarray scoring analyses with key clinical datasets, and with analysis of FFPE specimens from breast cancer cohorts, we demonstrate that breast cancer modules can be used to recapitulate the molecular subtypes of breast cancer and to have prognostic and predictive properties similar to the current multigene tests. Because they recapitulate existing molecular tests, while also reading out many additional axes of molecular variability, breast cancer modules provide a universal assay with broad application to companion diagnostics development.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 3664. doi:1538-7445.AM2012-3664

Urine TMPRSS2:ERG fusion transcript stratifies prostate cancer risk in men with elevated serum PSA

More than 1,000,000 men undergo prostate biopsy each year in the United States, most for "elevated" serum prostate-specific antigen (PSA). Given the lack of specificity and unclear mortality benefit of PSA testing, methods to individualize management of elevated PSA are needed. Greater than 50% of PSA-screened prostate cancers harbor fusions between the transmembrane protease, serine 2 (TMPRSS2) and v-ets erythroblastosis virus E26 oncogene homolog (avian) (ERG) genes. Here, we report a clinical-grade, transcription-mediated amplification assay to risk stratify and detect prostate cancer noninvasively in urine. The TMPRSS2:ERG fusion transcript was quantitatively measured in prospectively collected whole urine from 1312 men at multiple centers. Urine TMPRSS2:ERG was associated with indicators of clinically significant cancer at biopsy and prostatectomy, including tumor size, high Gleason score at prostatectomy, and upgrading of Gleason grade at prostatectomy. TMPRSS2:ERG, in combination with urine prostate cancer antigen 3 (PCA3), improved the performance of the multivariate Prostate Cancer Prevention Trial risk calculator in predicting cancer on biopsy. In the biopsy cohorts, men in the highest and lowest of three TMPRSS2:ERG+PCA3 score groups had markedly different rates of cancer, clinically significant cancer by Epstein criteria, and high-grade cancer on biopsy. Our results demonstrate that urine TMPRSS2:ERG, in combination with urine PCA3, enhances the utility of serum PSA for predicting prostate cancer risk and clinically relevant cancer on biopsy.

Abstract CN07-02: Target discovery with Oncomine: Discovering new targets and positioning old targets by integrative analysis of 10,000s of tumor genomes and transcriptomes

The identification of oncogenic protein kinases in cancer has prompted the development of several novel targeted therapies. In many cases, such therapies have elicited dramatic clinical responses in patients harboring genetic activation of the target. To identify novel targets and additional opportunities to apply existing targeted therapies, we undertook a systematic analysis of somatic mutations and DNA copy number alterations across 10,000+ of cancer patients from The Cancer Genome Atlas (TCGA), COSMIC and Oncomine™ databases. We applied multiple analysis strategies to filter genetic alterations for those most likely to exert a ‘driver’ oncogenic effect. For somatic mutation data, we prioritized likely oncogenes as those with frequent hotspot mutations and infrequent nonsense mutations. For copy number data, we prioritized likely oncogenes as those undergoing frequent, high-level focal amplifications resulting in concomitant over-expression in amplified cases. To identify additional opportunities to apply existing targeted therapies, we characterized 22 specific kinases that are targeted by approved or investigational therapies. Among the 22 kinases, nearly all showed evidence for recurrent hotspot mutations or high-level focal amplifications in at least a small fraction of tumors. This objective analysis confirmed well known associations between genetic alterations and particular cancer types, but also identified several novel associations, usually in small sub-populations, suggesting additional indications for existing therapies. In addition to this focused analysis, we also conducted a genome-wide survey of all genes with evidence for oncogenic alterations and the cancer types with most frequent alterations. While most of the top hits correspond to well-known oncogenes, several novel candidate oncogenes were prioritized.

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 CN07-02.

Abstract 2815: Urine TMPRSS2:ERG for prostate cancer risk stratification in men with elevated serum PSA

Background: Over 1,000,000 men undergo prostate biopsy each year in the U.S., most for “elevated” serum PSA. Given the lack of sensitivity and specificity, and unclear mortality benefit of PSA testing, methods to individualize management of elevated PSA are needed. We evaluated urine expression of TMPRSS2:ERG, a gene fusion occurring in 50% of prostate cancers, for risk-stratifying men presenting for biopsy.

Methods: TMPRSS2:ERG was measured by a clinical grade, transcription-mediated-amplification assay in prospectively collected whole-urine from 1,094 men undergoing biopsy at 10 academic and community clinics.

Findings: Urine TMPRSS2:ERG was associated with indicators of clinically significant cancer at biopsy and prostatectomy, including tumor size, high prostatectomy Gleason score and upgrading at prostatectomy. TMPRSS2:ERG in combination with urine PCA3, improved the multivariate PCPT risk calculator performance for predicting cancer on biopsy (AUC in test set, 0.79 vs. 0.64, p&lt;0.001). Using a three-class stratification, men in the highest and lowest TMPRSS2:ERG+PCA3 score groups had markedly different rates of cancer (69% vs. 21%, p&lt;0.001), clinically significant cancer by Epstein criteria (61% vs. 15%, p&lt;0.001) and high grade cancer (40% vs. 7%, p&lt;0.001) on biopsy.

Interpretation: Urine TMPRSS2:ERG, in combination with urine PCA3, enhances the utility of serum PSA for predicting prostate cancer and clinically relevant cancer on biopsy.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 2815. doi:10.1158/1538-7445.AM2011-2815

Abstract 3149: OncoPredictor: A systematic approach for predicting responsive cancer populations from large scale cell line screening

Investigational new drugs for cancer must demonstrate convincing preclinical efficacy and a compelling strategy to translate preclinical observations to the clinical setting. Personalized medicine approaches are gaining wider acceptance, and large scale cell line databases have demonstrated utility in identifying biomarkers of drug response that can inform clinical development strategies. However, the lack of an integrated platform to translate preclinical biomarker profiles to clinical populations limits the power of this approach. To solve this problem we developed, first, a cell line screening and genomic analysis pipeline that associates drug response across 200+ cell lines with mutation, DNA copy number, and gene expression biomarkers; and second, a parallel database of biomarker frequencies in clinical tumor samples, compiled from all available published genomic data. In the present study, we tested 8 targeted anti-cancer agents and identified cell line biomarkers representing each of the genomic data types – mutation, DNA amplification, and gene over-expression – and then assessed the distribution of these biomarkers across tumor samples. In each case the tumor populations predicted to be responsive by this unsupervised approach were validated by results from clinical trials. We also present an example of biomarker results leading to potential new indications for an approved drug.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 3149. doi:10.1158/1538-7445.AM2011-3149

Growth factor stimulation induces a distinct ER(alpha) cistrome underlying breast cancer endocrine resistance

Estrogen receptor α (ERα) expression in breast cancer is predictive of response to endocrine therapy; however, resistance is common in ERα-positive tumors that overexpress the growth factor receptor ERBB2. Even in the absence of estrogen, ERα can be activated by growth factors, including the epidermal growth factor (EGF). EGF induces a transcriptional program distinct from estrogen; however, the mechanism of the stimulus-specific response is unknown. Here we show that the EGF-induced ERα genomic targets, its cistromes, are distinct from those induced by estrogen in a process dependent on the transcription factor AP-1. The EGF-induced ERα cistrome specifically regulates genes found overexpressed in ERBB2-positive human breast cancers. This provides a potential molecular explanation for the endocrine therapy resistance seen in ERα-positive breast cancers that overexpress ERBB2. These results suggest a central role for ERα in hormone-refractory breast tumors dependent on growth factor pathway activation and favors the development of therapeutic strategies completely antagonizing ERα, as opposed to blocking its estrogen responsiveness alone.

Abstract A41: Somatic mutations in human cancer cell lines predict sorafenib response

It is now widely accepted that a spectrum of somatic alterations occur in human cancer and that particular alterations are associated with response or lack of response to anti-cancer therapies. Here, we undertook a systematic correlation analysis of somatic mutations, copy number aberrations and drug response across 300+ human cancer cell lines. We studied the RAF and VEGFR inhibitor, sorafenib, evaluating associations between sorafenib response and somatic mutations in 35 common cancer genes and high-level amplifications and homozygous deletions in more than 100 minimal common regions of gain and loss, which include many known oncogenes and tumor suppressor genes. Notably, we found that NRAS mutations, among other alterations, were significantly associated with sensitivity to sorafenib (odds ratio = 7, p-value = 0.003), whereas activating BRAF mutations were strongly associated with a lack of response (odds ratio = 4.9, p-value = 0.012). Specifically, 64% (7 of 11) of NRAS mutant cell lines were sensitive to sorafenib, while only 20% (34 of 170) of NRAS wild-type lines were sensitive. On the contrary, just 6.7% (2 of 30) of BRAF mutant lines were sensitive to sorafenib. Because NRAS mutations occur frequently in leukemia and no leukemia cell lines were included in the original analysis, we selected 5 NRAS mutant and 5 NRAS wild-type leukemia cell lines for further evaluation. Consistent with our previous observation, the four most sensitive leukemia cell lines were all NRAS mutant, each line exhibiting greater than 50% growth inhibition after 72 hours of exposure to 4μM sorafenib. These results confirm the original association and suggest that across both solid tumors and hematologic malignancies, NRAS mutations may identify a subset of cancers that are likely to respond sorafenib. In addition, this work demonstrates that systematic analysis of somatic alterations and drug response across a large collection of cancer cell lines is capable of discovering novel predictive biomarkers.

Citation Information: Clin Cancer Res 2010;16(7 Suppl):A41

Abstract A47: Somatic mutations in human cancer cell lines predict sorafenib response

It is now widely accepted that a spectrum of somatic alterations occur in human cancer and that particular alterations are associated with response or lack of response to anti-cancer therapies. Here, we undertook a systematic correlation analysis of somatic mutations, copy number aberrations and drug response across 300+ human cancer cell lines. We studied the RAF and VEGFR inhibitor, sorafenib, evaluating associations between sorafenib response and somatic mutations in 35 common cancer genes and high-level amplifications and homozygous deletions in more than 100 minimal common regions of gain and loss, which include many known oncogenes and tumor suppressor genes. Notably, we found that NRAS mutations, among other alterations, were significantly associated with sensitivity to sorafenib (odds ratio = 7, p-value = 0.003), whereas activating BRAF mutations were strongly associated with a lack of response (odds ratio = 4.9, p-value = 0.012). Specifically, 64% (7 of 11) of NRAS mutant cell lines were sensitive to sorafenib, while only 20% (34 of 170) of NRAS wild-type lines were sensitive. On the contrary, just 6.7% (2 of 30) of BRAF mutant lines were sensitive to sorafenib. Because NRAS mutations occur frequently in leukemia and no leukemia cell lines were included in the original analysis, we selected 5 NRAS mutant and 5 NRAS wild-type leukemia cell lines for further evaluation. Consistent with our previous observation, the four most sensitive leukemia cell lines were all NRAS mutant, each line exhibiting greater than 50% growth inhibition after 72 hours of exposure to 4uM sorafenib. These results confirm the original association and suggest that across both solid tumors and hematologic malignancies, NRAS mutations may identify a subset of cancers that are likely to respond sorafenib. In addition, this work demonstrates that systematic analysis of somatic alterations and drug response across a large collection of cancer cell lines is capable of discovering novel predictive biomarkers.

Citation Information: Mol Cancer Ther 2009;8(12 Suppl):A47.

Exploring clinical associations using '-omics' based enrichment analyses

Background

The vast amounts of clinical data collected in electronic health records (EHR) is analogous to the data explosion from the “-omics” revolution. In the EHR clinicians often maintain patient-specific problem summary lists which are used to provide a concise overview of significant medical diagnoses. We hypothesized that by tapping into the collective wisdom generated by hundreds of physicians entering problems into the EHR we could detect significant associations among diagnoses that are not described in the literature.

Methodology/Principal Findings

We employed an analytic approach original developed for detecting associations between sets of gene expression data, called Molecular Concept Map (MCM), to find significant associations among the 1.5 million clinical problem summary list entries in 327,000 patients from our institution's EHR. An odds ratio (OR) and p-value was calculated for each association. A subset of the 750,000 associations found were explored using the MCM tool. Expected associations were confirmed and recently reported but poorly known associations were uncovered. Novel associations which may warrant further exploration were also found. Examples of expected associations included non-insulin dependent diabetes mellitus and various diagnoses such as retinopathy, hypertension, and coronary artery disease. A recently reported association included irritable bowel and vulvodynia (OR 2.9, p = 5.6×10⁻⁴). Associations that are currently unknown or very poorly known included those between granuloma annulare and osteoarthritis (OR 4.3, p = 1.1×10⁻⁴) and pyloric stenosis and ventricular septal defect (OR 12.1, p = 2.0×10⁻³).

Conclusions/Significance

Computer programs developed for analyses of “-omic” data can be successfully applied to the area of clinical medicine. The results of the analysis may be useful for hypothesis generation as well as supporting clinical care by reminding clinicians of likely problems associated with a patient's existing problems.

The role of SPINK1 in ETS rearrangement-negative prostate cancers

ETS gene fusions have been characterized in a majority of prostate cancers; however, the key molecular alterations in ETS-negative cancers are unclear. Here we used an outlier meta-analysis (meta-COPA) to identify SPINK1 outlier expression exclusively in a subset of ETS rearrangement-negative cancers ( approximately 10% of total cases). We validated the mutual exclusivity of SPINK1 expression and ETS fusion status, demonstrated that SPINK1 outlier expression can be detected noninvasively in urine, and observed that SPINK1 outlier expression is an independent predictor of biochemical recurrence after resection. We identified the aggressive 22RV1 cell line as a SPINK1 outlier expression model and demonstrate that SPINK1 knockdown in 22RV1 attenuates invasion, suggesting a functional role in ETS rearrangement-negative prostate cancers.

A polycomb repression signature in metastatic prostate cancer predicts cancer outcome

The Polycomb Group (PcG) protein EZH2 is a critical component of a multiprotein complex that methylates Lys(27) of histone 3 (H3K27), which consequently leads to the repression of target gene expression. We have previously reported that EZH2 is overexpressed in metastatic prostate cancer and is a marker of aggressive diseases in clinically localized solid tumors. However, the global set of genes directly regulated by PcG in tumors is largely unknown, and thus how PcG mediates tumor progression remains unclear. Herein we mapped genome-wide H3K27 methylation in aggressive, disseminated human prostate cancer tissues. Integrative analysis revealed that a significant subset of these genes are also targets of PcG in embryonic stem cells, and their repression in tumors is associated with poor prognosis. By stepwise cross-validation, we developed a "Polycomb repression signature" composed of 14 direct targets of PcG in metastatic tumors. Notably, solid tumor subtypes in which this gene signature is repressed show poor clinical outcome in multiple microarray data sets of tumors including breast and prostate cancer. Taken together, our results show a fingerprint of PcG-mediated transcriptional repression in metastatic prostate cancer that is reminiscent of stem cells and associated with cancer progression. Therefore, PcG proteins play a central role in the epigenetic silencing of target genes and functionally link stem cells, metastasis, and cancer survival.

Identifying significant associations using free text problem summary lists in a clinical data repository

Recent studies have been seeking common signatures in gene expression profiles in order to find novel associations. We chose to test the extensibility of a tool created to perform association analyses of gene expression signatures by applying the same technique to problem summary lists. Both commonly known as well as less known associations were found using this technique.

Molecular concepts analysis links tumors, pathways, mechanisms, and drugs

Global molecular profiling of cancers has shown broad utility in delineating pathways and processes underlying disease, in predicting prognosis and response to therapy, and in suggesting novel treatments. To gain further insights from such data, we have integrated and analyzed a comprehensive collection of "molecular concepts" representing > 2500 cancer-related gene expression signatures from Oncomine and manual curation of the literature, drug treatment signatures from the Connectivity Map, target gene sets from genome-scale regulatory motif analyses, and reference gene sets from several gene and protein annotation databases. We computed pairwise association analysis on all 13,364 molecular concepts and identified > 290,000 significant associations, generating hypotheses that link cancer types and subtypes, pathways, mechanisms, and drugs. To navigate a network of associations, we developed an analysis platform, the Molecular Concepts Map. We demonstrate the utility of the approach by highlighting molecular concepts analyses of Myc pathway activation, breast cancer relapse, and retinoic acid treatment.

Integrating biomedical knowledge to model pathways of prostate cancer progression

Due to pathologic, histologic, and biologic variation within prostate cancers, profiling the genetic changes associated with disease progression has been difficult. Although initial integration of data from profiling studies had been limited by platform variation, bioinformatic tools and analytic techniques have enabled integrative analysis of profiling studies and the identification of more robust and valid profiles. The identification of key transition points in the progression of prostate cancer relies on profiling precursor lesions and "pure" cell populations. Utilizing laser-capture microdissection to isolate 101 cell populations, a more specific genetic profile of progression from benign epithelium to metastatic disease was obtained. This laser-capture profile was analyzed in the context of the molecular concepts map (MCM), a compendium of over 20,000 molecular concepts including other expression profiles of prostate cancer, to obtain an integrative molecular model of progression. The conceptual connections associated with progression confirm that prostate cancer biology is largely driven by pathways related to androgen signaling and epithelial cell biology; however, further analysis of concepts associated with progression suggests stromal factors are highly associated with progression of prostate cancer. The effect of stromal signatures on the progression model suggests the impact of stromal signature downregulation may reflect both a change in the epithelia:stroma ratio within higher grade tumors and also a microenvironment influence on prostate epithelia. Analyzing complex gene expression signatures in the context of molecular concepts improves integrative models and may improve detection, prognostication, or targeted therapy.

Oncomine 3.0: genes, pathways, and networks in a collection of 18,000 cancer gene expression profiles

DNA microarrays have been widely applied to cancer transcriptome analysis; however, the majority of such data are not easily accessible or comparable. Furthermore, several important analytic approaches have been applied to microarray analysis; however, their application is often limited. To overcome these limitations, we have developed Oncomine, a bioinformatics initiative aimed at collecting, standardizing, analyzing, and delivering cancer transcriptome data to the biomedical research community. Our analysis has identified the genes, pathways, and networks deregulated across 18,000 cancer gene expression microarrays, spanning the majority of cancer types and subtypes. Here, we provide an update on the initiative, describe the database and analysis modules, and highlight several notable observations. Results from this comprehensive analysis are available at http://www.oncomine.org.

Bioinformatics approaches in the study of cancer

A revolution is underway in the approach to studying the genetic basis of cancer. Massive amounts of data are now being generated via high-throughput techniques such as DNA microarray technology and new computational algorithms have been developed to aid in analysis. At the same time, standards-based repositories, including the Stanford Microarray Database and the Gene Expression Omnibus have been developed to store and disseminate the results of microarray experiments. Bioinformatics, the convergence of biology, information science, and computation, has played a key role in these developments. Recently developed techniques include Module Maps, SLAMS (Stepwise Linkage Analysis of Microarray Signatures), and COPA (Cancer Outlier Profile Analysis). What these techniques have in common is the application of novel algorithms to find high-level gene expression patterns across heterogeneous microarray experiments. Large-scale initiatives are underway as well. The Cancer Genome Atlas (TCGA) project is a logical extension of the Human Genome Project and is meant to produce a comprehensive atlas of genetic changes associated with cancer. The Cancer Biomedical Informatics Grid (caBIG), led by the NCI, also represents a colossal initiative involving virtually all aspects of cancer research and may help to transform the way cancer research is conducted and data are shared.

Integrative molecular concept modeling of prostate cancer progression

Despite efforts to profile prostate cancer, the genetic alterations and biological processes that correlate with the observed histological progression are unclear. Using laser-capture microdissection to isolate 101 cell populations, we have profiled prostate cancer progression from benign epithelium to metastatic disease. By analyzing expression signatures in the context of over 14,000 'molecular concepts', or sets of biologically connected genes, we generated an integrative model of progression. Molecular concepts that demarcate critical transitions in progression include protein biosynthesis, E26 transformation-specific (ETS) family transcriptional targets, androgen signaling and cell proliferation. Of note, relative to low-grade prostate cancer (Gleason pattern 3), high-grade cancer (Gleason pattern 4) shows an attenuated androgen signaling signature, similar to metastatic prostate cancer, which may reflect dedifferentiation and explain the clinical association of grade with prognosis. Taken together, these data show that analyzing gene expression signatures in the context of a compendium of molecular concepts is useful in understanding cancer biology.

Whole transcriptome amplification for gene expression profiling and development of molecular archives

Expression profiling of clinically obtainable tumor specimens has been hindered by the need for microgram quantities of RNA. In vitro transcription (IVT)-based amplifications are most commonly used to amplify small quantities of RNA for microarray analysis. However, significant drawbacks exist with IVT-based amplification, and the need for alternative amplification methods remains. Herein, we validate whole transcriptome amplification (WTA), an exponential amplification technique that produces cDNA libraries and amplified target in 3 to 4 hours from nanogram quantities of total RNA using a combination of cDNA microarrays and quantitative polymerase chain reaction (PCR). We demonstrate that WTA material can serve as a "molecular archive" because a WTA cDNA library can be faithfully amplified through multiple rounds of PCR amplification, allowing it to serve as a bankable and distributable resource. To demonstrate applicability, WTA was combined with laser capture microdissection to profile frozen prostate tissues. Unlike most IVT-based and exponential amplification techniques, WTA does not depend on the presence of a poly-A tail. Thus, we demonstrate that WTA is compatible with artificially degraded RNA and RNA isolated from formalin-fixed paraffin-embedded tissues. Taken together, WTA represents a versatile approach to profile and archive cDNA from minute tumor samples and is compatible with partially degraded RNA.

Delineation, functional validation, and bioinformatic evaluation of gene expression in thyroid follicular carcinomas with the PAX8-PPARG translocation

A subset of follicular thyroid carcinomas contains a balanced translocation, t(2;3)(q13;p25), that results in fusion of the paired box gene 8 (PAX8) and peroxisome proliferator-activated receptor gamma (PPARG) genes with concomitant expression of a PAX8-PPARgamma fusion protein, PPFP. PPFP is thought to contribute to neoplasia through a mechanism in which it acts as a dominant-negative inhibitor of wild-type PPARgamma. To better understand this type of follicular carcinoma, we generated global gene expression profiles using DNA microarrays of a cohort of follicular carcinomas along with other common thyroid tumors and used the data to derive a gene expression profile characteristic of PPFP-positive tumors. Transient transfection assays using promoters of four genes whose expression was highly associated with the translocation showed that each can be activated by PPFP. PPFP had unique transcriptional activities when compared with PAX8 or PPARgamma, although it had the potential to function in ways qualitatively similar to PAX8 or PPARgamma depending on the promoter and cellular environment. Bioinformatics analyses revealed that genes with increased expression in PPFP-positive follicular carcinomas include known PPAR target genes; genes involved in fatty acid, amino acid, and carbohydrate metabolism; micro-RNA target genes; and genes on chromosome 3p. These results have implications for the neoplastic mechanism of these follicular carcinomas.

TMPRSS2:ETV4 gene fusions define a third molecular subtype of prostate cancer

Although common in hematologic and mesenchymal malignancies, recurrent gene fusions have not been well characterized in epithelial carcinomas. Recently, using a novel bioinformatic approach, we identified recurrent gene fusions between TMPRSS2 and the ETS family members ERG or ETV1 in the majority of prostate cancers. Here, we interrogated the expression of all ETS family members in prostate cancer profiling studies and identified marked overexpression of ETV4 in 2 of 98 cases. In one such case, we confirmed the overexpression of ETV4 using quantitative PCR, and by rapid amplification of cDNA ends, quantitative PCR, and fluorescence in situ hybridization, we show that the TMPRSS2 (21q22) and ETV4 (17q21) loci are fused in this case. This result defines a third molecular subtype of prostate cancer and supports the hypothesis that dysregulation of ETS family members through fusions with TMRPSS2 may be an initiating event in prostate cancer development.

Recurrent fusion of TMPRSS2 and ETS transcription factor genes in prostate cancer

Recurrent chromosomal rearrangements have not been well characterized in common carcinomas. We used a bioinformatics approach to discover candidate oncogenic chromosomal aberrations on the basis of outlier gene expression. Two ETS transcription factors, ERG and ETV1, were identified as outliers in prostate cancer. We identified recurrent gene fusions of the 5' untranslated region of TMPRSS2 to ERG or ETV1 in prostate cancer tissues with outlier expression. By using fluorescence in situ hybridization, we demonstrated that 23 of 29 prostate cancer samples harbor rearrangements in ERG or ETV1. Cell line experiments suggest that the androgen-responsive promoter elements of TMPRSS2 mediate the overexpression of ETS family members in prostate cancer. These results have implications in the development of carcinomas and the molecular diagnosis and treatment of prostate cancer.

Integrative genomic and proteomic analysis of prostate cancer reveals signatures of metastatic progression

Molecular profiling of cancer at the transcript level has become routine. Large-scale analysis of proteomic alterations during cancer progression has been a more daunting task. Here, we employed high-throughput immunoblotting in order to interrogate tissue extracts derived from prostate cancer. We identified 64 proteins that were altered in prostate cancer relative to benign prostate and 156 additional proteins that were altered in metastatic disease. An integrative analysis of this compendium of proteomic alterations and transcriptomic data was performed, revealing only 48%-64% concordance between protein and transcript levels. Importantly, differential proteomic alterations between metastatic and clinically localized prostate cancer that mapped concordantly to gene transcripts served as predictors of clinical outcome in prostate cancer as well as other solid tumors.

Probabilistic model of the human protein-protein interaction network

A catalog of all human protein-protein interactions would provide scientists with a framework to study protein deregulation in complex diseases such as cancer. Here we demonstrate that a probabilistic analysis integrating model organism interactome data, protein domain data, genome-wide gene expression data and functional annotation data predicts nearly 40,000 protein-protein interactions in humans-a result comparable to those obtained with experimental and computational approaches in model organisms. We validated the accuracy of the predictive model on an independent test set of known interactions and also experimentally confirmed two predicted interactions relevant to human cancer, implicating uncharacterized proteins into definitive pathways. We also applied the human interactome network to cancer genomics data and identified several interaction subnetworks activated in cancer. This integrative analysis provides a comprehensive framework for exploring the human protein interaction network.

Integrative analysis of the cancer transcriptome

DNA microarrays have been widely applied to the study of human cancer, delineating myriad molecular subtypes of cancer, many of which are associated with distinct biological underpinnings, disease progression and treatment response. These primary analyses have begun to decipher the molecular heterogeneity of cancer, but integrative analyses that evaluate cancer transcriptome data in the context of other data sources are often capable of extracting deeper biological insight from the data. Here we discuss several such integrative computational and analytical approaches, including meta-analysis, functional enrichment analysis, interactome analysis, transcriptional network analysis and integrative model system analysis.

Mining for regulatory programs in the cancer transcriptome

DNA microarrays have been widely applied to cancer transcriptome analysis. The Oncomine database contains a large collection of such data, as well as hundreds of derived gene-expression signatures. We studied the regulatory mechanisms responsible for gene deregulation in these cancer signatures by searching for the coordinate regulation of genes with common transcription factor binding sites. We found that genes with binding sites for the archetypal cancer transcription factor, E2F, were disproportionately overexpressed in a wide variety of cancers, whereas genes with binding sites for other transcription factors, such as Myc-Max, c-Rel and ATF, were disproportionately overexpressed in specific cancer types. These results suggest that alterations in pathways activating these transcription factors may be responsible for the observed gene deregulation and cancer pathogenesis.

ONCOMINE: a cancer microarray database and integrated data-mining platform

DNA microarray technology has led to an explosion of oncogenomic analyses, generating a wealth of data and uncovering the complex gene expression patterns of cancer. Unfortunately, due to the lack of a unifying bioinformatic resource, the majority of these data sit stagnant and disjointed following publication, massively underutilized by the cancer research community. Here, we present ONCOMINE, a cancer microarray database and web-based data-mining platform aimed at facilitating discovery from genome-wide expression analyses. To date, ONCOMINE contains 65 gene expression datasets comprising nearly 48 million gene expression measurements form over 4700 microarray experiments. Differential expression analyses comparing most major types of cancer with respective normal tissues as well as a variety of cancer subtypes and clinical-based and pathology-based analyses are available for exploration. Data can be queried and visualized for a selected gene across all analyses or for multiple genes in a selected analysis. Furthermore, gene sets can be limited to clinically important annotations including secreted, kinase, membrane, and known gene-drug target pairs to facilitate the discovery of novel biomarkers and therapeutic targets.

Overexpression, amplification, and androgen regulation of TPD52 in prostate cancer

Gains in the long arm of chromosome 8 (8q) are believed to be associated with poor outcome and the development of hormone-refractory prostate cancer. Based on a meta-analysis of gene expression microarray data from multiple prostate cancer studies (D. R. Rhodes et al., Cancer Res 2002;62:4427-33), a candidate oncogene, Tumor Protein D52 (TPD52), was identified in the 8q21 amplicon. TPD52 is a coiled-coil motif-bearing protein, potentially involved in vesicle trafficking. Both mRNA and protein levels of TPD52 were highly elevated in prostate cancer tissues. Array comparative genomic hybridization and amplification analysis using single nucleotide polymorphism arrays demonstrated increased DNA copy number in the region encompassing TPD52. Fluorescence in situ hybridization on tissue microarrays confirmed TPD52 amplification in prostate cancer epithelia. Furthermore, our studies suggest that TPD52 protein levels may be regulated by androgens, consistent with the presence of androgen response elements in the upstream promoter of TPD52. In summary, these findings suggest that dysregulation of TPD52 by genomic amplification and androgen induction may play a role in prostate cancer progression.

Bioinformatics strategies for translating genome-wide expression analyses into clinically useful cancer markers

The DNA microarray has revolutionized cancer research. Now, scientists can obtain a genome-wide perspective of cancer gene expression. One potential application of this technology is the discovery of novel cancer biomarkers for more accurate diagnosis and prognosis, and potentially for the earlier detection of disease or the monitoring of treatment effectiveness. Because microarray experiments generate a tremendous amount of data and because the number of laboratories generating microarray data is rapidly growing, new bioinformatics strategies that promote the maximum utilization of such data are necessary. Here, we describe a method to validate multiple microarray data sets, a Web-based cancer microarray database for biomarker discovery, and methods for integrating gene ontology annotations with microarray data to improve candidate biomarker selection.

Large-scale meta-analysis of cancer microarray data identifies common transcriptional profiles of neoplastic transformation and progression

Many studies have used DNA microarrays to identify the gene expression signatures of human cancer, yet the critical features of these often unmanageably large signatures remain elusive. To address this, we developed a statistical method, comparative metaprofiling, which identifies and assesses the intersection of multiple gene expression signatures from a diverse collection of microarray data sets. We collected and analyzed 40 published cancer microarray data sets, comprising 38 million gene expression measurements from >3,700 cancer samples. From this, we characterized a common transcriptional profile that is universally activated in most cancer types relative to the normal tissues from which they arose, likely reflecting essential transcriptional features of neoplastic transformation. In addition, we characterized a transcriptional profile that is commonly activated in various types of undifferentiated cancer, suggesting common molecular mechanisms by which cancer cells progress and avoid differentiation. Finally, we validated these transcriptional profiles on independent data sets.

Humoral Immune Response to -Methylacyl-CoA Racemase and Prostate Cancer

BACKGROUND

Although prostate-specific antigen (PSA) is a prototypic biomarker for prostate cancer, it has poor specificity. Expression of alpha-methylacyl-CoA racemase (AMACR), which is involved in the conversion of R-stereoisomers of branched-chain fatty acids to S-stereoisomers, has been shown to be specifically increased in prostate cancer epithelia. However, attempts to detect AMACR in circulation have not been successful. Hence, we determined whether an immune response to AMACR could be used as a serum biomarker for prostate cancer.

METHODS

Sera from patients with biopsy-proven prostate cancer and from control subjects were screened for a humoral immune response to selected tumor antigens, including AMACR, by using protein microarrays (46 patients, 28 control subjects). Humoral immune response to AMACR was then validated using high-throughput immunoblot analysis (151 patients, 259 control subjects) and enzyme-linked immunosorbent assay (ELISA) (54 patients, 55 control subjects). Receiver operating characteristic curves were used to determine the sensitivity and specificity of the immune response to AMACR to detect prostate cancer.

RESULTS

Immunoreactivity against AMACR was statistically significantly higher in sera from patients with prostate cancer than in control subjects by all three techniques (P(protein microarray) =.009, P(immunoblot)<.001, P(ELISA) =.011). High-throughput immunoblot analysis revealed that, in subjects with intermediate PSA levels (4-10 ng/mL), the immune response against AMACR was more sensitive and specific than was PSA in distinguishing sera from prostate cancer patients relative to control subjects (sensitivity and specificity of 77.8% and 80.6% versus 45.6% and 50%, respectively; area under the curve of 0.789 versus 0.492; P<.001).

CONCLUSION

Assays to detect a humoral immune response against AMACR may have the potential to supplement PSA screening in identifying patients with clinically significant prostate cancer, especially those with intermediate PSA levels.

Prostate cancer biomarkers: a current perspective

Genomic- and proteomic-based studies have led to the identification of a large number of candidate biomarkers, as well as signature patterns of multiple markers for prostate cancer diagnosis, disease progression and prediction of survival. While these candidates include the usual suspects, including oncogenes, proliferation markers and cytoskeletal proteins, there are many additional unexpected molecules such as those involved in processes such as transcriptional repression and fatty acid metabolism. Patterns of expression serving as useful biomarkers is a new and, as yet, clinically untested concept which promises to permit a consideration of the complex milieu of cancer. Exciting as these developments are, clinical application will have to await careful validation of these candidates by independent biochemical approaches over large and diverse samples.

Multiplex biomarker approach for determining risk of prostate-specific antigen-defined recurrence of prostate cancer

BACKGROUND

Molecular signatures in cancer tissue may be useful for diagnosis and are associated with survival. We used results from high-density tissue microarrays (TMAs) to define combinations of candidate biomarkers associated with the rate of prostate cancer progression after radical prostatectomy that could identify patients at high risk for recurrence.

METHODS

Fourteen candidate biomarkers for prostate cancer for which antibodies are available included hepsin, pim-1 kinase, E-cadherin (ECAD; cell adhesion molecule), alpha-methylacyl-coenzyme A racemase, and EZH2 (enhancer of zeste homolog 2, a transcriptional repressor). TMAs containing more than 2000 tumor samples from 259 patients who underwent radical prostatectomy for localized prostate cancer were studied with these antibodies. Immunohistochemistry results were evaluated in conjunction with clinical parameters associated with prostate cancer progression, including tumor stage, Gleason score, and prostate-specific antigen (PSA) level. Recurrence was defined as a postsurgery PSA level of more than 0.2 ng/mL. All statistical tests were two-sided.

RESULTS

Moderate or strong expression of EZH2 coupled with at most moderate expression of ECAD (i.e., a positive EZH2:ECAD status) was the biomarker combination that was most strongly associated with the recurrence of prostate cancer. EZH2:ECAD status was statistically significantly associated with prostate cancer recurrence in a training set of 103 patients (relative risk [RR] = 2.52, 95% confidence interval [CI] = 1.09 to 5.81; P =.021), in a validation set of 80 patients (RR = 3.72, 95% CI = 1.27 to 10.91; P =.009), and in the combined set of 183 patients (RR = 2.96, 95% CI = 1.56 to 5.61; P<.001). EZH2:ECAD status was statistically significantly associated with disease recurrence even after adjusting for clinical parameters, such as tumor stage, Gleason score, and PSA level (hazard ratio = 3.19, 95% CI = 1.50 to 6.77; P =.003).

CONCLUSION

EZH2:ECAD status was statistically significantly associated with prostate cancer recurrence after radical prostatectomy and may be useful in defining a cohort of high-risk patients.