Somatic mutations reveal complex metastatic seeding from multifocal primary prostate cancer.

Primary prostate cancer shows a striking intraorgan molecular heterogeneity, with multiple spatially separated malignant foci in the majority of patients. Metastatic prostate cancer, however, typically reveals more homogenous molecular profiles, suggesting a monoclonal origin of the metastatic lesions. Longitudinal mutational spectra, comparing multiple primary lesions with metastases from the same patients remain poorly defined. We have here analyzed somatic mutations in multisampled, spatio-temporal biobanked lesions (38 samples from primary foci and 1 sample from each of 8 metastases from seven prostate cancer patients) applying a custom-designed panel targeting 68 prostate cancer relevant genes. The metastatic samples were taken at time of primary surgery and up to 7 years later, and sampling included circulating tumor DNA in plasma or solid metastatic tissue samples. A total of 282 somatic mutations were detected, with a range of 0 to 25 mutations per sample. Although seven samples had solely private mutations, the remaining 39 samples had both private and shared mutations. Seventy-four percent of mutations in metastases were not found in any primary samples, and vice versa, 96% of mutations in primary cancers were not found in any metastatic samples. However, for three patients, shared mutations were found suggesting the focus of origin, including mutations in AKT1, FOXA1, HOXB13, RB1 and TP53. In conclusion, the spatio-temporal heterogeneous nature of multifocal disease is emphasized in our study, and underlines the importance of testing a recent sample in genomics-based precision medicine for metastatic prostate cancer.

The exon-level biomarker SLC39A14 has organ-confined cancer-specificity in colorectal cancer.

An alternative transcript variant of SLC39A14, caused by pre-mRNA splicing, was recently suggested as a biomarker for colorectal cancer (CRC). In our study, we have validated the cancer-specific splicing pattern of the mutually exclusive exons 4A and 4B in altogether 244 colorectal tissue samples. Exon-specific quantitative RT-PCR analyses across 136 Stage I-IV CRC samples and 44 normal colonic mucosa samples showed complete cancer-specificity, as well as 94% sensitivity of SLC39A14-exon4B relative to SLC39A14-exon4A expression. However, across 20 samples from a range of healthy tissues, 18 expressed the CRC variant. This was true also for ten benign lymph nodes, demonstrating that the cancer-specificity is mainly confined to the colon and rectum. Hence, clinical use of SLC39A14-exon4B as a detection marker for CRC other than in samples taken from the bowel wall is diminished. Prognostic value by detection of metastasis to lymph nodes is also abated, elucidating an important pitfall to biomarker discovery. However, analyses of ten nondysplastic biopsies from patients with active inflammatory bowel disease showed negative results in seven samples and only weakly positive results in three samples, suggesting value of SLC39A14-exon4B as a marker to distinguish CRC from other pathologic conditions of the colon. In conclusion, the SLC39A14-exon4B transcript variant is a CRC biomarker with high sensitivity and organ-confined specificity. Further use of the transcript and its encoded protein isoform should be explored in an organ-confined context.

Fusion gene microarray reveals cancer type-specificity among fusion genes.

Detection of fusion genes for diagnostic purposes and as a guide to treatment is well-established in hematological malignancies, and the prevalence of fusion genes in epithelial cancers is also increasingly appreciated. To study whether established fusion genes are present within additional cancer types, we have used an updated version of our fusion gene microarray in a systematic survey of reported fusion genes in multiple cancer types. We assembled a comprehensive database of published fusion genes, including those reported only in individual studies and samples, and fusion genes resulting from deep sequencing of cancer genomes and transcriptomes. From the total set of 548 fusion genes, we designed 599,839 oligonucleotides, targeting both chimeric transcript junctions as well as sequences internal to each of the fusion gene partners. We investigated the presence of fusion genes in a series of 67 cell lines representing 15 different cancer types. Data from ten leukemia cell lines with known fusion gene status were used to develop an automated scoring algorithm, and in five cell lines the correct fusion gene was the top scoring hit, and one came second. Two additional fusion genes, BCAS4-BCAS3 in the MCF-7 breast cancer cell line and CCDC6-RET in the TPC-1 thyroid cancer cell line were validated as true positive fusion transcripts. However, these fusion genes were not new to these cancer types, and none of 548 fusion genes were identified from a novel cancer type. We therefore find it unlikely that the assayed fusion genes are commonly present across multiple cancer types.

In situ expression of ERG protein in the context of tumor heterogeneity identifies prostate cancer patients with inferior prognosis.

Prognostic biomarkers for prostate cancer are needed to improve prediction of disease course and guide treatment decisions. However, biomarker development is complicated by the common multifocality and heterogeneity of the disease. We aimed to determine the prognostic value of candidate biomarkers transcriptional regulator ERG and related ETS family genes, while considering tumor heterogeneity. In a multisampled, prospective, and treatment-naïve radical prostatectomy cohort from one tertiary center (2010-2012, median follow-up 8.1 years), we analyzed ERG protein (480 patients; 2047 tissue cores), and RNA of several ETS genes in a subcohort (165 patients; 778 fresh-frozen tissue samples). Intra- and interfocal heterogeneity was identified in 29% and 33% (ERG protein) and 39% and 27% (ETS RNA) of patients, respectively. ERG protein and ETS RNA was identified exclusively in a nonindex tumor in 31% and 32% of patients, respectively. ERG protein demonstrated independent prognostic value in predicting biochemical (P = 0.04) and clinical recurrence (P = 0.004) and appeared to have greatest prognostic value for patients with Grade Groups 4-5. In conclusion, when heterogeneity is considered, ERG protein is a robust prognostic biomarker for prostate cancer.

High expression of SCHLAP1 in primary prostate cancer is an independent predictor of biochemical recurrence, despite substantial heterogeneity.

In primary prostate cancer, the common multifocality and heterogeneity are major obstacles in finding robust prognostic tissue biomarkers. The long noncoding RNA SCHLAP1 has been suggested, but its prognostic value has not been investigated in the context of tumor heterogeneity. In the present study, expression of SCHLAP1 was investigated using real-time RT-PCR in a multisampled series of 778 tissue samples from radical prostatectomies of 164 prostate cancer patients (median follow-up time 7.4 y). The prognostic value of SCHLAP1 was evaluated with biochemical recurrence as endpoint. In total, 29% of patients were classified as having high expression of SCHLAP1 in at least one malignant sample. Among these, inter- and intrafocal heterogeneity was detected in 72% and 56%, respectively. High expression of SCHLAP1 was shown to be a predictor of biochemical recurrence in both uni- and multivariable cox regression analyses (P < 0.001 and P = 0.02). High expression of SCHLAP1 was also significantly associated with adverse clinicopathological characteristics, including grade group, high pT stage, invasive cribriform growth/intraductal carcinoma of the prostate, and reactive stroma. In conclusion, high expression of SCHLAP1 in at least one malignant sample is a robust prognostic biomarker in primary prostate cancer. For the first time, high SCHLAP1 expression has been associated with the aggressive histopathologic feature reactive stroma. The expression of SCHLAP1 is highly heterogeneous, and analysis of multiple samples is therefore crucial in determination of the SCHLAP1 status of a patient.

Interfocal heterogeneity challenges the clinical usefulness of molecular classification of primary prostate cancer.

Prostate cancer is a highly heterogeneous disease and typically multiple distinct cancer foci are present at primary diagnosis. Molecular classification of prostate cancer can potentially aid the precision of diagnosis and treatment. A promising genomic classifier was published by The Cancer Genome Atlas (TCGA), successfully classifying 74% of primary prostate cancers into seven groups based on one cancer sample per patient. Here, we explore the clinical usefulness of this classification by testing the classifier's performance in a multifocal context. We analyzed 106 cancer samples from 85 distinct cancer foci within 39 patients. By somatic mutation data from whole-exome sequencing and targeted qualitative and quantitative gene expression assays, 31% of the patients were uniquely classified into one of the seven TCGA classes. Further, different samples from the same focus had conflicting classification in 12% of the foci. In conclusion, the level of both intra- and interfocal heterogeneity is extensive and must be taken into consideration in the development of clinically useful molecular classification of primary prostate cancer.

Multifocal Primary Prostate Cancer Exhibits High Degree of Genomic Heterogeneity.

BACKGROUND: Most primary prostate cancers are multifocal with individual tumors harboring different aggressiveness; however, the genomic heterogeneity among these tumors is poorly understood. OBJECTIVE: To better understand the biological basis for clinical variability among different lesions, we sought to comprehensively characterize the heterogeneity of somatic gene mutations in multifocal prostate cancer. DESIGN, SETTING, AND PARTICIPANTS: High-coverage whole-exome sequencing of 153 frozen tissue samples, taken from two to three distinct tumor foci and one non-cancerous area from each of 41 patients, covering a total of 89 tumor foci. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: State-of-the-art bioinformatics tools for mutation calling and copy number determination from whole-exome sequencing data. RESULTS AND LIMITATIONS: We found a very high degree of interfocal heterogeneity among tumors, that is, 76% of pairwise-compared tumor foci from the same prostatectomy specimen had no point mutations in common and DNA copy number changes were rarely shared across cancer foci. The few point mutations shared across tumor foci were seldom in cancer-critical genes. CONCLUSIONS: In this first large genomic heterogeneity study of primary prostate cancer, we observe that different tumor foci within the same patient are genetically distinct, only rarely sharing any somatic gene mutations, including those in cancer driver genes. This heterogeneity affects how genomics-based management of prostate cancer can be implemented, as information from all tumor foci is necessary to draw valid conclusions about the cancer's genomic alterations. PATIENT SUMMARY: Most primary prostate cancers consist of multiple tumors within the same organ, but little is known about their relationships. We have compared the sets of gene mutations among such tumors and found that they only exceptionally have any in common. This will influence treatment decisions in the future as each tumor's mutations will render it unique and have to be considered to gain the best treatment results.

Novel transcription-induced fusion RNAs in prostate cancer.

Prostate cancer is a clinically and pathologically heterogeneous disease with a broad spectrum of molecular abnormalities in the genome and transcriptome. One key feature is the involvement of chromosomal rearrangements creating fusion genes. Recent RNA-sequencing technology has uncovered that fusions which are not caused by chromosomal rearrangements, but rather meditated at transcription level, are common in both healthy and diseased cells. Such fusion transcripts have been proven highly associated with prostate cancer development and progression. To discover novel fusion transcripts, we analyzed RNA sequencing data from 44 primary prostate tumors and matched benign tissues from The Cancer Genome Atlas. Twenty-one high-confident candidates were significantly enriched in malignant vs. benign samples. Thirteen of the candidates have not previously been described in prostate cancer, and among them, five long intergenic non-coding RNAs are involved as fusion partners. Their expressions were validated in 50 additional prostate tumor samples and seven prostate cancer cell lines. For four fusion transcripts, we found a positive correlation between their expression and the expression of the 3' partner gene. Among these, differential exon usage and qRT-PCR analyses in particular support that SLC45A3-ELK4 is mediated by an RNA polymerase read-through mechanism.

High frequency of fusion transcripts involving TCF7L2 in colorectal cancer: novel fusion partner and splice variants.

VTI1A-TCF7L2 was reported as a recurrent fusion gene in colorectal cancer (CRC), found to be expressed in three out of 97 primary cancers, and one cell line, NCI-H508, where a genomic deletion joins the two genes [1]. To investigate this fusion further, we analyzed high-throughput DNA and RNA sequencing data from seven CRC cell lines, and identified the gene RP11-57H14.3 (ENSG00000225292) as a novel fusion partner for TCF7L2. The fusion was discovered from both genome and transcriptome data in the HCT116 cell line. By triplicate nested RT-PCR, we tested both the novel fusion transcript and VTI1A-TCF7L2 for expression in a series of 106 CRC tissues, 21 CRC cell lines, 14 normal colonic mucosa, and 20 normal tissues from miscellaneous anatomical sites. Altogether, 42% and 45% of the CRC samples expressed VTI1A-TCF7L2 and TCF7L2-RP11-57H14.3 fusion transcripts, respectively. The fusion transcripts were both seen in 29% of the normal colonic mucosa samples, and in 25% and 75% of the tested normal tissues from other organs, revealing that the TCF7L2 fusion transcripts are neither specific to cancer nor to the colon and rectum. Seven different splice variants were detected for the VTI1A-TCF7L2 fusion, of which three are novel. Four different splice variants were detected for the TCF7L2-RP11-57H14.3 fusion. In conclusion, we have identified novel variants of VTI1A-TCF7L2 fusion transcripts, including a novel fusion partner gene, RP11-57H14.3, and demonstrated detectable levels in a large fraction of CRC samples, as well as in normal colonic mucosa and other tissue types. We suggest that the fusion transcripts observed in a high frequency of samples are transcription induced chimeras that are expressed at low levels in most samples. The similar fusion transcripts induced by genomic rearrangements observed in individual cancer cell lines may yet have oncogenic potential as suggested in the original study by Bass et al.

Transforming pluripotency: an exon-level study of malignancy-specific transcripts in human embryonal carcinoma and embryonic stem cells.

To circumvent difficulties of isolating pure populations of cancer stem cells (CSCs) for the purpose of identifying malignancy-specific gene expression, we have compared exon-resolution transcriptomic profiles of 5 embryonal carcinoma (EC) cell lines, a histological subtype of germ cell tumor (GCT), to their nonmalignant caricature, specifically 6 human embryonic stem (ES) cell lines. Both cell types are readily accessible, and were purified for undifferentiated cells only. We identified a set of 28 differentially expressed genes, many of which had cancer and stemness roles. Overexpression of the recently discovered pluripotency gene NR5A2 in malignant EC cells revealed an intriguing indication of how WNT-mediated dysregulation of pluripotency is involved with malignancy. Expression of these 28 genes was further explored within 2 publically available data sets of primary EC tumors and normal testis. At the exon-level, alternative splicing events were detected in ZNF195, DNMT3B, and PMF1, and alternative promoters were detected for ASH2L and ETV5. These events were validated by reverse transcriptase-polymerase chain reaction-based methods in EC and ES lines, where the alternative splicing event in the de novo DNA methyltransferase DNMT3B may have functional consequences. In conclusion, we have identified malignancy-specific gene expression differences within a rigorous pluripotent stem cell context. These findings are of particular interest for both GCT and ES cell biology, and, in general, to the concept of CSCs.

Novel 5' fusion partners of ETV1 and ETV4 in prostate cancer.

Gene fusions involving the erythroblast transformation-specific (ETS) transcription factors ERG, ETV1, ETV4, ETV5, and FLI1 are a common feature of prostate carcinomas (PCas). The most common upstream fusion partner described is the androgen-regulated prostate-specific gene TMPRSS2, most frequently with ERG, but additional 5' fusion partners have been described. We performed 5' rapid amplification of cDNA ends in 18 PCas with ETV1, ETV4, or ETV5 outlier expression to identify the 5' fusion partners. We also evaluated the exon-level expression profile of these ETS genes in 14 cases. We identified and confirmed by fluorescent in situ hybridization (FISH) and reverse transcription-polymerase chain reaction the two novel chimeric genes OR51E2-ETV1 and UBTF-ETV4 in two PCas. OR51E2 encodes a G-protein-coupled receptor that is overexpressed in PCas, whereas UBTF is a ubiquitously expressed gene encoding an HMG-box DNA-binding protein involved in ribosome biogenesis. We additionally describe two novel gene fusion combinations of previously described genes, namely, SLC45A3-ETV4 and HERVK17-ETV4. Finally, we found one PCa with TMPRSS2-ETV1, one with C15orf21-ETV1, one with EST14-ETV1, and two with 14q133-q21.1-ETV1. In nine PCas (eight ETV1 and one ETV5), exhibiting ETS outlier expression and genomic rearrangement detected by FISH, no 5' fusion partner was found. Our findings contribute significantly to characterize the heterogeneous group of ETS gene fusions and indicate that all genes described as 5' fusion partners with one ETS gene can most likely be rearranged with any of the other ETS genes involved in prostate carcinogenesis.

A tumor-associated mutation of FYVE-CENT prevents its interaction with Beclin 1 and interferes with cytokinesis.

The tumor suppressor activity of Beclin 1 (BECN1), a subunit of class III phosphatidylinositol 3-kinase complex, has been attributed to its regulation of apoptosis and autophagy. Here, we identify FYVE-CENT (ZFYVE26), a phosphatidylinositol 3-phosphate binding protein important for cytokinesis, as a novel interacting protein of Beclin 1. A mutation in FYVE-CENT (R1945Q) associated with breast cancer abolished the interaction between FYVE-CENT and Beclin 1, and reduced the localization of these proteins at the intercellular bridge during cytokinesis. Breast cancer cells containing the FYVE-CENT R1945Q mutation displayed a significant increase in cytokinetic profiles and bi-multinuclear phenotype. Both Beclin 1 and FYVE-CENT were found to be downregulated in advanced breast cancers. These findings suggest a positive feedback loop for recruitment of FYVE-CENT and Beclin 1 to the intercellular bridge during cytokinesis, and reveal a novel potential tumor suppressor mechanism for Beclin 1.