Irsan E. Kooi

RNA-Seq of Tumor-Educated Platelets Enables Blood-Based Pan-Cancer, Multiclass, and Molecular Pathway Cancer Diagnostics

Summary Tumor-educated blood platelets (TEPs) are implicated as central players in the systemic and local responses to tumor growth, thereby altering their RNA profile. We determined the diagnostic potential of TEPs by mRNA sequencing of 283 platelet samples. We distinguished 228 patients with localized and metastasized tumors from 55 healthy individuals with 96% accuracy. Across six different tumor types, the location of the primary tumor was correctly identified with 71% accuracy. Also, MET or HER2-positive, and mutant KRAS, EGFR, or PIK3CA tumors were accurately distinguished using surrogate TEP mRNA profiles. Our results indicate that blood platelets provide a valuable platform for pan-cancer, multiclass cancer, and companion diagnostics, possibly enabling clinical advances in blood-based “liquid biopsies”.

Swarm Intelligence-Enhanced Detection of Non-Small-Cell Lung Cancer Using Tumor-Educated Platelets

Summary Blood-based liquid biopsies, including tumor-educated blood platelets (TEPs), have emerged as promising biomarker sources for non-invasive detection of cancer. Here we demonstrate that particle-swarm optimization (PSO)-enhanced algorithms enable efficient selection of RNA biomarker panels from platelet RNA-sequencing libraries (n = 779). This resulted in accurate TEP-based detection of early- and late-stage non-small-cell lung cancer (n = 518 late-stage validation cohort, accuracy, 88%; AUC, 0.94; 95% CI, 0.92–0.96; p < 0.001; n = 106 early-stage validation cohort, accuracy, 81%; AUC, 0.89; 95% CI, 0.83–0.95; p < 0.001), independent of age of the individuals, smoking habits, whole-blood storage time, and various inflammatory conditions. PSO enabled selection of gene panels to diagnose cancer from TEPs, suggesting that swarm intelligence may also benefit the optimization of diagnostics readout of other liquid biopsy biosources.

Proper genomic profiling of (BRCA1‐mutated) basal‐like breast carcinomas requires prior removal of tumor infiltrating lymphocytes

BRCA1‐mutated breast carcinomas may have distinct biological features, suggesting the involvement of specific oncogenic pathways in tumor development. The identification of genomic aberrations characteristic for BRCA1‐mutated breast carcinomas could lead to a better understanding of BRCA1‐associated oncogenic events and could prove valuable in clinical testing for BRCA1‐involvement in patients.

Genomic profiling of CHEK2*1100delC-mutated breast carcinomas

BackgroundCHEK2*1100delC is a moderate-risk breast cancer susceptibility allele with a high prevalence in the Netherlands. We performed copy number and gene expression profiling to investigate whether CHEK2*1100delC breast cancers harbor characteristic genomic aberrations, as seen for BRCA1 mutated breast cancers.MethodsWe performed high-resolution SNP array and gene expression profiling of 120 familial breast carcinomas selected from a larger cohort of 155 familial breast tumors, including BRCA1, BRCA2, and CHEK2 mutant tumors. Gene expression analyses based on a mRNA immune signature was used to identify samples with relative low amounts of tumor infiltrating lymphocytes (TILs), which were previously found to disturb tumor copy number and LOH (loss of heterozygosity) profiling. We specifically compared the genomic and gene expression profiles of CHEK2*1100delC breast cancers (n = 14) with BRCAX (familial non-BRCA1/BRCA2/CHEK2*1100delC mutated) breast cancers (n = 34) of the luminal intrinsic subtypes for which both SNP-array and gene expression data is available.ResultsHigh amounts of TILs were found in a relatively small number of luminal breast cancers as compared to breast cancers of the basal-like subtype. As expected, these samples mostly have very few copy number aberrations and no detectable regions of LOH. By unsupervised hierarchical clustering of copy number data we observed a great degree of heterogeneity amongst the CHEK2*1100delC breast cancers, comparable to the BRCAX breast cancers. Furthermore, copy number aberrations were mostly seen at low frequencies in both the CHEK2*1100delC and BRCAX group of breast cancers. However, supervised class comparison identified copy number loss of chromosomal arm 1p to be associated with CHEK2*1100delC status.ConclusionsIn conclusion, in contrast to basal-like BRCA1 mutated breast cancers, no apparent specific somatic copy number aberration (CNA) profile for CHEK2*1100delC breast cancers was found. With the possible exception of copy number loss of chromosomal arm 1p in a subset of tumors, which might be involved in CHEK2 tumorigenesis. This difference in CNAs profiles might be explained by the need for BRCA1-deficient tumor cells to acquire survival factors, by for example specific copy number aberrations, to expand. Such factors may not be needed for breast tumors with a defect in a non-essential gene such as CHEK2.

Loss of photoreceptorness and gain of genomic alterations in retinoblastoma reveal tumor progression.

Background Retinoblastoma is a pediatric eye cancer associated with RB1 loss or MYCN amplification (RB1+/+MYCNA). There are controversies concerning the existence of molecular subtypes within RB1−/− retinoblastoma. To test whether these molecular subtypes exist, we performed molecular profiling. Methods Genome-wide mRNA expression profiling was performed on 76 primary human retinoblastomas. Expression profiling was complemented by genome-wide DNA profiling and clinical, histopathological, and ex vivo drug sensitivity data. Findings RNA and DNA profiling identified major variability between retinoblastomas. While gene expression differences between RB1+/+MYCNA and RB1−/− tumors seemed more dichotomous, differences within the RB1−/− tumors were gradual. Tumors with high expression of a photoreceptor gene signature were highly differentiated, smaller in volume and diagnosed at younger age compared with tumors with low photoreceptor signature expression. Tumors with lower photoreceptor expression showed increased expression of genes involved in M-phase and mRNA and ribosome synthesis and increased frequencies of somatic copy number alterations. Interpretation Molecular, clinical and histopathological differences between RB1−/− tumors are best explained by tumor progression, reflected by a gradual loss of differentiation and photoreceptor expression signature. Since copy number alterations were more frequent in tumors with less photoreceptorness, genomic alterations might be drivers of tumor progression. Research in context Retinoblastoma is an ocular childhood cancer commonly caused by mutations in the RB1 gene. In order to determine optimal treatment, tumor subtyping is considered critically important. However, except for very rare retinoblastomas without an RB1 mutation, there are controversies as to whether subtypes of retinoblastoma do exist. Our study shows that retinoblastomas are highly diverse but rather than reflecting distinct tumor types with a different etiology, our data suggests that this diversity is a result of tumor progression driven by cumulative genetic alterations. Therefore, retinoblastomas should not be categorized in distinct subtypes, but be described according to their stage of progression.

Somatic genomic alterations in retinoblastoma beyond RB1 are rare and limited to copy number changes

Retinoblastoma is a rare childhood cancer initiated by RB1 mutation or MYCN amplification, while additional alterations may be required for tumor development. However, the view on single nucleotide variants is very limited. To better understand oncogenesis, we determined the genomic landscape of retinoblastoma. We performed exome sequencing of 71 retinoblastomas and matched blood DNA. Next, we determined the presence of single nucleotide variants, copy number alterations and viruses. Aside from RB1, recurrent gene mutations were very rare. Only a limited fraction of tumors showed BCOR (7/71, 10%) or CREBBP alterations (3/71, 4%). No evidence was found for the presence of viruses. Instead, specific somatic copy number alterations were more common, particularly in patients diagnosed at later age. Recurrent alterations of chromosomal arms often involved less than one copy, also in highly pure tumor samples, suggesting within-tumor heterogeneity. Our results show that retinoblastoma is among the least mutated cancers and signify the extreme sensitivity of the childhood retina for RB1 loss. We hypothesize that retinoblastomas arising later in retinal development benefit more from subclonal secondary alterations and therefore, these alterations are more selected for in these tumors. Targeted therapy based on these subclonal events might be insufficient for complete tumor control.

A Meta-Analysis of Retinoblastoma Copy Numbers Refines the List of Possible Driver Genes Involved in Tumor Progression.

Background While RB1 loss initiates retinoblastoma development, additional somatic copy number alterations (SCNAs) can drive tumor progression. Although SCNAs have been identified with good concordance between studies at a cytoband resolution, accurate identification of single genes for all recurrent SCNAs is still challenging. This study presents a comprehensive meta-analysis of genome-wide SCNAs integrated with gene expression profiling data, narrowing down the list of plausible retinoblastoma driver genes. Methods We performed SCNA profiling of 45 primary retinoblastoma samples and eight retinoblastoma cell lines by high-resolution microarrays. We combined our data with genomic, clinical and histopathological data of ten published genome-wide SCNA studies, which strongly enhanced the power of our analyses (N = 310). Results Comprehensive recurrence analysis of SCNAs in all studies integrated with gene expression data allowed us to reduce candidate gene lists for 1q, 2p, 6p, 7q and 13q to a limited gene set. Besides the well-established driver genes RB1 (13q-loss) and MYCN (2p-gain) we identified CRB1 and NEK7 (1q-gain), SOX4 (6p-gain) and NUP205 (7q-gain) as novel retinoblastoma driver candidates. Depending on the sample subset and algorithms used, alternative candidates were identified including MIR181 (1q-gain) and DEK (6p gain). Remarkably, our study showed that copy number gains rarely exceeded change of one copy, even in pure tumor samples with 100% homozygosity at the RB1 locus (N = 34), which is indicative for intra-tumor heterogeneity. In addition, profound between-tumor variability was observed that was associated with age at diagnosis and differentiation grades. Interpretation Since focal alterations at commonly altered chromosome regions were rare except for 2p24.3 (MYCN), further functional validation of the oncogenic potential of the described candidate genes is now required. For further investigations, our study provides a refined and revised set of candidate retinoblastoma driver genes.

Defects in the Fanconi Anemia Pathway and Chromatid Cohesion in Head and Neck Cancer

Failure to repair DNA damage or defective sister chromatid cohesion, a process essential for correct chromosome segregation, can be causative of chromosomal instability (CIN), which is a hallmark of many types of cancers. We investigated how frequent this occurs in head and neck squamous cell carcinoma (HNSCC) and whether specific mechanisms or genes could be linked to these phenotypes. The genomic instability syndrome Fanconi anemia is caused by mutations in any of at least 16 genes regulating DNA interstrand crosslink (ICL) repair. Since patients with Fanconi anemia have a high risk to develop HNSCC, we investigated whether and to which extent Fanconi anemia pathway inactivation underlies CIN in HNSCC of non-Fanconi anemia individuals. We observed ICL-induced chromosomal breakage in 9 of 17 (53%) HNSCC cell lines derived from patients without Fanconi anemia. In addition, defective sister chromatid cohesion was observed in five HNSCC cell lines. Inactivation of FANCM was responsible for chromosomal breakage in one cell line, whereas in two other cell lines, somatic mutations in PDS5A or STAG2 resulted in inadequate sister chromatid cohesion. In addition, FANCF methylation was found in one cell line by screening an additional panel of 39 HNSCC cell lines. Our data demonstrate that CIN in terms of ICL-induced chromosomal breakage and defective chromatid cohesion is frequently observed in HNSCC. Inactivation of known Fanconi anemia and chromatid cohesion genes does explain CIN in the minority of cases. These findings point to phenotypes that may be highly relevant in treatment response of HNSCC.

Coregulation of FANCA and BRCA1 in human cells

Fanconi anemia (FA) is a genetically heterogeneous syndrome associated with increased cancer predisposition. The underlying genes govern the FA pathway which functions to protect the genome during the S-phase of the cell cycle. While upregulation of FA genes has been linked to chemotherapy resistance, little is known about their regulation in response to proliferative stimuli. The purpose of this study was to examine how FA genes are regulated, especially in relation to the cell cycle, in order to reveal their possible participation in biochemical networks. Expression of 14 FA genes was monitored in two human cell-cycle models and in two RB1/E2F pathway-associated primary cancers, retinoblastoma and basal breast cancer. In silico studies were performed to further evaluate coregulation and identify connected networks and diseases. Only FANCA was consistently induced over 2-fold; FANCF failed to exhibit any regulatory fluctuations. Two tools exploiting public data sets indicated coregulation of FANCA with BRCA1. Upregulation of FANCA and BRCA1 correlated with upregulation of E2F3. Genes coregulated with both FANCA and BRCA1 were enriched for MeSH-Term id(s) genomic instability, microcephaly, and Bloom syndrome, and enriched for the cellular component centrosome. The regulation of FA genes appears highly divergent. In RB1-linked tumors, upregulation of FA network genes was associated with reduced expression of FANCF. FANCA and BRCA1 may jointly act in a subnetwork - supporting vital function(s) at the subcellular level (centrosome) as well as at the level of embryonic development (mechanisms controlling head circumference).

MR Imaging Features of Retinoblastoma: Association with Gene Expression Profiles

Purpose To identify associations between magnetic resonance (MR) imaging features and gene expression in retinoblastoma. Materials and Methods A retinoblastoma MR imaging atlas was validated by using anonymized MR images from referral centers in Essen, Germany, and Paris, France. Images were from 39 patients with retinoblastoma (16 male and 18 female patients [the sex in five patients was unknown]; age range, 5-90 months; inclusion criterion: pretreatment MR imaging). This atlas was used to compare MR imaging features with genome-wide messenger RNA (mRNA) expression data from 60 consecutive patients obtained from 1995 to 2012 (35 male patients [58%]; age range, 2-69 months; inclusion criteria: pretreatment MR imaging, genome-wide mRNA expression data available). Imaging pathway associations were analyzed by means of gene enrichment. In addition, imaging features were compared with a predefined gene expression signature of photoreceptorness. Statistical analysis was performed with generalized linear modeling of radiology traits on normalized log2-transformed expression values. P values were corrected for multiple hypothesis testing. Results Radiogenomic analysis revealed 1336 differentially expressed genes for qualitative imaging features (threshold P = .05 after multiple hypothesis correction). Loss of photoreceptorness gene expression correlated with advanced stage imaging features, including multiple lesions (P = .03) and greater eye size (P < .001). The number of lesions on MR images was associated with expression of MYCN (P = .04). A newly defined radiophenotype of diffuse-growing, plaque-shaped, multifocal tumors displayed overexpression of SERTAD3 (P = .003, P = .049, and P = .06, respectively), a protein that stimulates cell growth by activating the E2F network. Conclusion Radiogenomic biomarkers can potentially help predict molecular features, such as photoreceptorness loss, that indicate tumor progression. Results imply a possible role for radiogenomics in future staging and treatment decision making in retinoblastoma.