High-throughput screening using patient-derived tumor xenografts to predict clinical trial drug response.

Profiling candidate therapeutics with limited cancer models during preclinical development hinders predictions of clinical efficacy and identifying factors that underlie heterogeneous patient responses for patient-selection strategies. We established ∼1,000 patient-derived tumor xenograft models (PDXs) with a diverse set of driver mutations. With these PDXs, we performed in vivo compound screens using a 1 × 1 × 1 experimental design (PDX clinical trial or PCT) to assess the population responses to 62 treatments across six indications. We demonstrate both the reproducibility and the clinical translatability of this approach by identifying associations between a genotype and drug response, and established mechanisms of resistance. In addition, our results suggest that PCTs may represent a more accurate approach than cell line models for assessing the clinical potential of some therapeutic modalities. We therefore propose that this experimental paradigm could potentially improve preclinical evaluation of treatment modalities and enhance our ability to predict clinical trial responses.

Pharmacological and genomic profiling identifies NF-κB-targeted treatment strategies for mantle cell lymphoma.

Mantle cell lymphoma (MCL) is an aggressive malignancy that is characterized by poor prognosis. Large-scale pharmacological profiling across more than 100 hematological cell line models identified a subset of MCL cell lines that are highly sensitive to the B cell receptor (BCR) signaling inhibitors ibrutinib and sotrastaurin. Sensitive MCL models exhibited chronic activation of the BCR-driven classical nuclear factor-κB (NF-κB) pathway, whereas insensitive cell lines displayed activation of the alternative NF-κB pathway. Transcriptome sequencing revealed genetic lesions in alternative NF-κB pathway signaling components in ibrutinib-insensitive cell lines, and sequencing of 165 samples from patients with MCL identified recurrent mutations in TRAF2 or BIRC3 in 15% of these individuals. Although they are associated with insensitivity to ibrutinib, lesions in the alternative NF-κB pathway conferred dependence on the protein kinase NIK (also called mitogen-activated protein 3 kinase 14 or MAP3K14) both in vitro and in vivo. Thus, NIK is a new therapeutic target for MCL treatment, particularly for lymphomas that are refractory to BCR pathway inhibitors. Our findings reveal a pattern of mutually exclusive activation of the BCR-NF-κB or NIK-NF-κB pathways in MCL and provide critical insights into patient stratification strategies for NF-κB pathway-targeted agents.

Protein kinase C inhibitor sotrastaurin selectively inhibits the growth of CD79 mutant diffuse large B-cell lymphomas.

The activated B-cell-like (ABC) subtype of diffuse large B-cell lymphoma (DLBCL) correlates with poor prognosis. The ABC subtype of DLBCL is associated with constitutive activation of the NF-κB pathway, and oncogenic lesions have been identified in its regulators, including CARD11/CARMA1 (caspase recruitment domain-containing protein 11), A20/TNFAIP3, and CD79A/B. In this study, we offer evidence of therapeutic potential for the selective PKC (protein kinase C) inhibitor sotrastaurin (STN) in preclinical models of DLBCL. A significant fraction of ABC DLBCL cell lines exhibited strong sensitivity to STN, and we found that the molecular nature of NF-κB pathway lesions predicted responsiveness. CD79A/B mutations correlated with STN sensitivity, whereas CARD11 mutations rendered ABC DLBCL cell lines insensitive. Growth inhibitory effects of PKC inhibition correlated with NF-κB pathway inhibition and were mediated by induction of G1-phase cell-cycle arrest and/or cell death. We found that STN produced significant antitumor effects in a mouse xenograft model of CD79A/B-mutated DLBCL. Collectively, our findings offer a strong rationale for the clinical evaluation of STN in ABC DLBCL patients who harbor CD79 mutations also illustrating the necessity to stratify DLBCL patients according to their genetic abnormalities.

Characterization CSMD1 in a large set of primary lung, head and neck, breast and skin cancer tissues.

The Cub and Sushi Multiple Domains-1 (CSMD1) is a tumor suppressor gene on 8p23.2, where allelic loss is both frequent and associated with poor prognosis in head and neck squamous cell carcinoma (HNSCC). To understand the extent of CSMD1 aberrations in vivo, we characterized 184 primary tumors from the head and neck, lung, breast and skin for gene copy number and analyzed expression in our HNSCCs and lung squamous cell carcinomas (SCCs). We detected loss of CSMD1 in a large proportion of HNSCCs (50%), lung (46%) and breast cancers (55%), and to a lesser extent in cutaneous SCCs (29%) and basal cell carcinomas (BCCs, 17%) using array-based comparative genomic hybridization (aCGH). Studying the region more closely with quantitative real-time PCR (qPCR), the loss of CSMD1 increased to 80% in HNSCCs and 93% in lung SCCs. CSMD1 expression was decreased in tumors compared to adjacent benign tissue (65%, 13/20) and was likely due to gene loss in 45% of cases (9/20). We also identified truncated transcripts lacking exons due to DNA copy number loss (30%, 5/17) or aberrant splicing (24%, 4/17). We show loss of CSMD1 in primary HNSCC tissues, and document for the first time that CSMD1 is lost in breast, lung and cutaneous SCCs. We also show that deletions of CSMD1 and aberrant splicing contribute to altered CSMD1 function in vivo.

Estrogen receptor status could modulate the genomic pattern in familial and sporadic breast cancer.

PURPOSE: Familial breast cancer represents 5% to 10% of all breast tumors. Mutations in the two known major breast cancer susceptibility genes, BRCA1 and BRCA2, account for a minority of familial breast cancer, whereas families without mutations in these genes (BRCAX group) account for 70% of familial breast cancer cases. EXPERIMENTAL DESIGN: To better characterize and define the genomic differences between the three classes of familial tumors and sporadic malignancies, we have analyzed 19 BRCA1, 24 BRCA2, and 31 BRCAX samples from familial breast cancer patients and 19 sporadic breast tumors using a 1-Mb resolution bacterial artificial chromosome array-based comparative genomic hybridization. RESULTS: We found that BRCA1/2 tumors showed a higher genomic instability than BRCAX and sporadic cancers. There were common genomic alterations present in all breast cancer groups, such as gains of 1q and 16p or losses of 8ptel-p12 and 16q. We found that the presence/absence of the estrogen receptor (ER) may play a crucial role in driving tumor development through distinct genomic pathways independently of the tumor type (sporadic or familial) and mutation status (BRCA1 or BRCA2). ER(-) tumors presented higher genomic instability and different altered regions than ER+ ones. CONCLUSIONS: According to our results, the BRCA gene mutation status (mainly BRCA1) would contribute to the genomic profile of abnormalities by increasing or modulating the genome instability.

STAC: A method for testing the significance of DNA copy number aberrations across multiple array-CGH experiments.

Regions of gain and loss of genomic DNA occur in many cancers and can drive the genesis and progression of disease. These copy number aberrations (CNAs) can be detected at high resolution by using microarray-based techniques. However, robust statistical approaches are needed to identify nonrandom gains and losses across multiple experiments/samples. We have developed a method called Significance Testing for Aberrant Copy number (STAC) to address this need. STAC utilizes two complementary statistics in combination with a novel search strategy. The significance of both statistics is assessed, and P-values are assigned to each location on the genome by using a multiple testing corrected permutation approach. We validate our method by using two published cancer data sets. STAC identifies genomic alterations known to be of clinical and biological significance and provides statistical support for 85% of previously reported regions. Moreover, STAC identifies numerous additional regions of significant gain/loss in these data that warrant further investigation. The P-values provided by STAC can be used to prioritize regions for follow-up study in an unbiased fashion. We conclude that STAC is a powerful tool for identifying nonrandom genomic amplifications and deletions across multiple experiments. A Java version of STAC is freely available for download at http://cbil.upenn.edu/STAC.

microRNAs exhibit high frequency genomic alterations in human cancer.

MicroRNAs (miRNAs) are endogenous noncoding RNAs, which negatively regulate gene expression. To determine genomewide miRNA DNA copy number abnormalities in cancer, 283 known human miRNA genes were analyzed by high-resolution array-based comparative genomic hybridization in 227 human ovarian cancer, breast cancer, and melanoma specimens. A high proportion of genomic loci containing miRNA genes exhibited DNA copy number alterations in ovarian cancer (37.1%), breast cancer (72.8%), and melanoma (85.9%), where copy number alterations observed in >15% tumors were considered significant for each miRNA gene. We identified 41 miRNA genes with gene copy number changes that were shared among the three cancer types (26 with gains and 15 with losses) as well as miRNA genes with copy number changes that were unique to each tumor type. Importantly, we show that miRNA copy changes correlate with miRNA expression. Finally, we identified high frequency copy number abnormalities of Dicer1, Argonaute2, and other miRNA-associated genes in breast and ovarian cancer as well as melanoma. These findings support the notion that copy number alterations of miRNAs and their regulatory genes are highly prevalent in cancer and may account partly for the frequent miRNA gene deregulation reported in several tumor types.

Distinct genomic profiles in hereditary breast tumors identified by array-based comparative genomic hybridization.

Mutations in BRCA1 and BRCA2 account for a significant proportion of hereditary breast cancers. Earlier studies have shown that inherited and sporadic tumors progress along different somatic genetic pathways and that global gene expression profiles distinguish between these groups. To determine whether genomic profiles similarly discriminate among BRCA1, BRCA2, and sporadic tumors, we established DNA copy number profiles using comparative genomic hybridization to BAC-clone microarrays providing <1 Mb resolution. Tumor DNA was obtained from BRCA1 (n = 14) and BRCA2 (n = 12) mutation carriers, as well as sporadic cases (n = 26). Overall, BRCA1 tumors had a higher frequency of copy number alterations than sporadic breast cancers (P = 0.00078). In particular, frequent losses on 4p, 4q, and 5q in BRCA1 tumors and frequent gains on 7p and 17q24 in BRCA2 tumors distinguish these from sporadic tumors. Distinct amplicons at 3q27.1-q27.3 were identified in BRCA1 tumors and at 17q23.3-q24.2 in BRCA2 tumors. A homozygous deletion on 5q12.1 was found in a BRCA1 tumor. Using a set of 169 BAC clones that detect significantly (P < 0.001) different frequencies of copy number changes in inherited and sporadic tumors, these could be discriminated into separate groups using hierarchical clustering. By comparing DNA copy number and RNA expression for genes in these regions, several candidate genes affected by up- or down-regulation were identified. Moreover, using support vector machines, we correctly classified BRCA1 and BRCA2 tumors (P < 0.0000004 and 0.00005, respectively). Further validation may prove this tumor classifier to be useful for selecting familial breast cancer cases for further mutation screening, particularly, as these data can be obtained using archival tissue.

High-resolution detection and mapping of genomic DNA alterations in neuroblastoma.

We used array-based comparative genomic hybridization (aCGH) to measure genomic copy number alterations (CNAs) in 42 neuroblastoma cell lines with known 1p36.3, 2p24 (MYCN), 11q23, and 17q23 allelic status. All cell lines showed CNAs, with an average of 22.0% of the genome of each sample showing evidence of gain (11.6%) or loss (10.4%). MYCN amplification was detected in 81% of cell lines, but other regions with high-level genomic amplification were observed only rarely. Gain of 17q material was present in 75% of the samples, and four discrete genomic regions at 17q23.2-17q25.3 were defined. Novel regions of gain were identified, including a 2.6-Mb subtelomeric region at 5p that includes the telomerase reverse transcriptase gene (TERT), which was found in 45% of the cell lines. Hemizygous deletions were noted at 1p36.23-1p36.32 and 11q23.3-11q25 in 60% and 36%, respectively, of the samples, with other frequent (>25%) regions of deletion localized to 1p32.1, 3p21.31-3p22.1, 5q35.2-5q35.3, 7q31.2, 7q34, 9q22.3-9q24.1, 10q26.11-10q26.12, 16q23.1-16q24.3, 18q21.32-18q23, and 20p11.21-20p11.23. A smallest region of overlap (SRO) for CNAs was mapped across all experiments and in each case was consistent with or refined the published data. A single cell line showed a homozygous deletion at 3p22.3, which was verified, and this location was refined by FISH and PCR. There was outstanding concordance of aCGH with PCR-based CNA detection methods. Several potential cooperating loci were identified, including deletion of 11q23-25, which was highly associated with both regional gain and loss at multiple chromosomal loci but was inversely correlated with the deletion of 1p36. Taking all of this together indicates that aCGH can accurately measure CNAs in the neuroblastoma genome and facilitate gene discovery efforts by high-throughput refinement of candidate loci.

CGHAnalyzer: a stand-alone software package for cancer genome analysis using array-based DNA copy number data.

SUMMARY: This synopsis provides an overview of array-based comparative genomic hybridization data display, abstraction and analysis using CGHAnalyzer, a software suite, designed specifically for this purpose. CGHAnalyzer can be used to simultaneously load copy number data from multiple platforms, query and describe large, heterogeneous datasets and export results. Additionally, CGHAnalyzer employs a host of algorithms for microarray analysis that include hierarchical clustering and class differentiation. AVAILABILITY: CGHAnalyzer, the accompanying manual, documentation and sample data are available for download at http://acgh.afcri.upenn.edu. This is a Java-based application built in the framework of the TIGR MeV that can run on Microsoft Windows, Macintosh OSX and a variety of Unix-based platforms. It requires the installation of the free Java Runtime Environment 1.4.1 (or more recent) (http://www.java.sun.com).

High resolution genomic analysis of sporadic breast cancer using array-based comparative genomic hybridization.

INTRODUCTION: Genomic aberrations in the form of subchromosomal DNA copy number changes are a hallmark of epithelial cancers, including breast cancer. The goal of the present study was to analyze such aberrations in breast cancer at high resolution. METHODS: We employed high-resolution array comparative genomic hybridization with 4,134 bacterial artificial chromosomes that cover the genome at 0.9 megabase resolution to analyze 47 primary breast tumors and 18 breast cancer cell lines. RESULTS: Common amplicons included 8q24.3 (amplified in 79% of tumors, with 5/47 exhibiting high level amplification), 1q32.1 and 16p13.3 (amplified in 66% and 57% of tumors, respectively). Moreover, we found several positive correlations between specific amplicons from different chromosomes, suggesting the existence of cooperating genetic loci. Queried by gene, the most frequently amplified kinase was PTK2 (79% of tumors), whereas the most frequently lost kinase was PTK2B (hemizygous loss in 34% of tumors). Amplification of ERBB2 as measured by comparative genomic hybridization (CGH) correlated closely with ERBB2 DNA and RNA levels measured by quantitative PCR as well as with ERBB2 protein levels. The overall frequency of recurrent losses was lower, with no region lost in more than 50% of tumors; the most frequently lost tumor suppressor gene was RB1 (hemizygous loss in 26% of tumors). Finally, we find that specific copy number changes in cell lines closely mimicked those in primary tumors, with an overall Pearson correlation coefficient of 0.843 for gains and 0.734 for losses. CONCLUSION: High resolution CGH analysis of breast cancer reveals several regions where DNA copy number is commonly gained or lost, that non-random correlations between specific amplicons exist, and that specific genetic alterations are maintained in breast cancer cell lines despite repeat passage in tissue culture. These observations suggest that genes within these regions are critical to the malignant phenotype and may thus serve as future therapeutic targets.

1-Mb resolution array-based comparative genomic hybridization using a BAC clone set optimized for cancer gene analysis.

Array-based comparative genomic hybridization (aCGH) is a recently developed tool for genome-wide determination of DNA copy number alterations. This technology has tremendous potential for disease-gene discovery in cancer and developmental disorders as well as numerous other applications. However, widespread utilization of a CGH has been limited by the lack of well characterized, high-resolution clone sets optimized for consistent performance in aCGH assays and specifically designed analytic software. We have assembled a set of approximately 4100 publicly available human bacterial artificial chromosome (BAC) clones evenly spaced at approximately 1-Mb resolution across the genome, which includes direct coverage of approximately 400 known cancer genes. This aCGH-optimized clone set was compiled from five existing sets, experimentally refined, and supplemented for higher resolution and enhancing mapping capabilities. This clone set is associated with a public online resource containing detailed clone mapping data, protocols for the construction and use of arrays, and a suite of analytical software tools designed specifically for aCGH analysis. These resources should greatly facilitate the use of aCGH in gene discovery.