The Cancer Genome Atlas RNA-seq Research Articles

Abstract 3034: Krüppel-like factors in regulation of gastric cancer progression

Abstract Introduction: Gastric cancer (GC) is the third leading cause of cancer-associated mortality and the fifth most frequently diagnosed cancer worldwide. The development and progression of GC is a complicated multistage process subjected to dynamic changes in gene regulation. Investigating the effects of these changes is therefore critical for improving our knowledge about GC initiation and progression. Krüppel-like factors (KLFs) are evolutionarily conserved zinc finger-containing transcription factors with altered expressions and diverse regulatory functions in GC. We previously reported that there is loss of KLF4 in GC. Here we report that KLF4 is a crucial regulator of KLF5 expression and vice versa and together they regulate GC cell behavior. Methods: The Cancer Genome Atlas (TCGA) and pan cancer data of differential expressions of KLFs in GC was analyzed. TCGA RNAseq and clinical data for the differential expression status of KLF genes in normal and cancerous tissues and the clinicopathological association was accessed from the UALCAN web portal. The expressions of KLF4 and KLF5 in human normal stomach and GC and orthotopic xenografts were detected by immunohistochemistry (IHC). KLF4, KLF5 cross regulation was evaluated by gene manipulation of each gene individually and checking the expression of the other factor in GC cells. The effect of KLFs on GC cell proliferation, migration, and invasion was determined by Ki67 staining of tissues, PrestoBlue cell viability assay, scratch wound and transwell cell invasion assays. Results and Conclusion: Our results indicated that not only KLF4 and KLF5 have contrasting expressions and effects in GC but also these two factors cross regulate each other. While lower expressions of KLF4 was found in human GC and orthotopic xenografts, increased expression of KLF5 was found in both human GC tissue and orthotopic xenografts. KLF4 silencing and CRISPR/Cas9 knockdown in human GC cells resulted in increased expression of KLF5 which was associated with increased GC cell proliferation and invasiveness. Interestingly, KLF5 knockdown by CRISPR/Cas9 in human GC cells on the other hand resulted in increased expression of KLF4, which was associated with decreased GC cell growth. Our data therefore suggests that the differential expressions of KLF4 and KLF5 might serve as valuable prognostic markers in GC. Citation Format: Sooraj Kakkat, Prabhat Suman, Sandeep Goswami, Sujit Basu, Martin J. Heslin, Elba A. Turbat-Herrera, Veronica Ramirez Alcantara, Chandrani Sarkar, Debanjan Chakroborty. Krüppel-like factors in regulation of gastric cancer progression [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3034.

Abstract 6548: Leveraging compact feature sets for TCGA-based molecular subtype classification on new samples

Abstract The NCI's The Cancer Genome Atlas (TCGA) project profiled over 10,000 tumor samples over the course of 10 years. As different tissue-specific working groups reviewed all of the available data, these patient samples were separated into distinct molecular subtypes, and these clusters were reported in various marker papers. While these assignments provided invaluable information about the common patterns of molecular characteristics in different types of cancer there was no consistent methodology for assigning new samples to these defined molecular subtypes.The NCI's Tumor Molecular Pathology group was formulated to create machine learning-based models that could be applied to non-TCGA samples and determine their TCGA mapped subtypes. Five modeling systems, JADBio, SKGrid by the Oregon Health and Science University, CloudForest by the Institute of Systems Biology, AKLIMATE by University of California Santa Cruz and subSCOPE by BC Cancer’s Genome Sciences Centre, were trained to recognize TCGA subtypes using multi-omic measurements from gene expression, DNA methylation, miRNA expression, copy number, and somatic mutation calls. While the TCGA samples were profiled using multi-omic technologies, single platform and/or compact feature set models also were assessed for their ability to assign these classifications. Each machine learning system created predictive models for 106 subtypes from 26 cancer types using as few features as possible, with a maximum of 100 features allowed for scored models. A set of 411,706 models was developed, composed of results of each of the learning methods across the various omic platforms. Top models, both multi-omic and single platform, were selected for each cancer type. On average, models were able to achieve an overall weighted F1 score of 0.895 with 42 features. While the top models for each cancer type had an overall weighted F1 mean performance of 0.936 with a mean of 29 features, in 20 of the 26 cancer types models using only gene expression provided the best performance. Analysis of features selected by the models showed some known onco-drivers were selected by many models, but many times different models would utilize features of different genes with similar levels of performance. Network-level analysis revealed that many genes of these selected features operated within the same pathways.Transferability of these models to external datasets was tested, taking TCGA breast cancer trained models and applying them to AURORA and METABRIC datasets. Interestingly, despite the data platform difference between TCGA (RNAseq) and METABRIC (microarray), model performance saw only minimal degradation of F1 values in transfer. This set of models and the training dataset will provide new opportunities for researchers and translational scientists to connect new tumors to the subtypes seen in the TCGA cohorts. Citation Format: Kyle Ellrott, Chris K. Wong, Christina Yau, Mauro A. Castro, Jordan Lee, Brian Karlberg, Jasleen K. Grewal, Vincenzo Lagani, Bahar Tercan, Verena Friedl, Toshinori Hinoue, Vladislav Uzunangelov, Lindsay Westlake, Xavier Loinaz, Ina Felau, Peggy Wang, Anab Kemal, Samantha J. Caesar-Johnson, Ilya Shmulevich, Alexander J. Lazar, Ioannis Tsamardinos, Katherine A. Hoadley, The Cancer Genome Atlas Analysis Network, Gordon A. Robertson, Theo A. Knijnenburg, Christopher C. Benz, Joshua M. Stuart, Jean C. Zenklusen, Andrew D. Cherniack, Peter W. Laird. Leveraging compact feature sets for TCGA-based molecular subtype classification on new samples [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6548.

LRRC59 promotes the progression of oral squamous cell carcinoma by interacting with SRP pathway components and enhancing the secretion of CKAP4-containing exosomes

BackgroundAs a ribosome receptor, LRRC59 was thought to regulate mRNA translation on the ER membrane. Evidence suggests that LRRC59 is overexpressed in a number of human malignancies and is associated with poor prognoses, but its primary biological function in the development of oral squamous cell carcinoma (OSCC) remains obscure. ObjectiveThe purpose of this study is to investigate at the expression changes and functional role of LRRC59 in OSCC. MethodsLRRC59 gene expression and correlation with prognosis of OSCC patients were first examined using the data from The Cancer Genome Atlas (TCGA) databases. Following that, a series of functional experiments, including cell counting kit-8, cell cycle analysis, wound healing assays, and transwell assays, were carried out to analyze the biological roles of LRRC59 in tumor cells. Mechanistically, we employed Tandem Affinity Purification-Mass Spectrometry (TAP-MS) approach to isolate and identify protein complexes of LRRC59. Downstream regulatory proteins of LRRC59 were verified through immunoprecipitation and immunofluorescence experiments. Furthermore, we isolated exosomes from OSCC cell supernatant and conducted co-culture experiments to examine the effect of LRRC59 knockdown on OSCC cells. ResultsIn samples from OSCC patients, LRRC59 was highly expressed and correlated with poor prognoses. Moreover, the gene sets analysis based on TCGA RNA-seq data indicated that LRRC59 seemed to be strongly related with protein secretory and OSCC migration. Upregulated levels of LRRC59 are more prone to lymph node metastasis in OSCC patients. LRRC59 knockdown impaired the ability of OSCC cell proliferation, migration, and invasion invitro. Mechanistically, our TAP-MS data situate LRRC59 in a functional nexus for mRNA translation regulation via interactions with SRP pathway components, translational initiation factors, CRD-mediated mRNA stabilization factors. More importantly, we found that LRRC59 interacted with cytoskeleton-associated protein 4 (CKAP4) and promoted the formation of CKAP4-containing exosomes. We also revealed that the LRRC59-CKAP4 axis was a crucial regulator of CKAP4-containing exosome secretion in OSCC cells for migration and invasion. ConclusionsTherefore, based on our findings, LRRC59 may serve as a potential biomarker for OSCC patients, and LRRC59-induced exosome secretion via the CKAP4 axis may serve as a potential therapeutic target for OSCC.

Abstract A102: Secreted frizzled related-protein 2 is prognostic for survival in pancreatic cancer

Abstract INTRODUCTION: Secreted frizzled-related protein 2 (SFRP2) is overexpressed in several tumor types and blocks tumor apoptosis, increases angiogenesis, and promotes T-cell exhaustion. It has also been shown to be associated with fibrosis in non-neoplastic pathologies. We hypothesize SFRP2 is prognostic for survival and associated with PDAC fibrosis and mesenchymal transformation. METHODS: SFRP2 gene expression and protein levels in patients with PDAC were analyzed using NCI’s The Cancer Genome Atlas (TCGA) data and data from resected tumors analyzed for protein levels by western blot. Disease free survival (DFS) and overall survival (OS) were analyzed using Kaplan Meier test. The association of stroma content and SFPR2 RNA expression in human PDAC was studied using Spearman correlation between the ESTIMATE algorithm stromal score and TCGA data. The GEPIA tool was used to compare SFRP2 mRNA levels between PDAC and normal pancreas tissue. Immunohistochemistry (IHC) for SFRP2 in KPC PDAC tumors were correlated with fibrosis stains Picrosirus red and Massone trichome. The role of SFRP2 in mesenchymal transformation was assessed by western blot in fibroblasts. RESULTS: TCGA RNAseq analysis of SFRP2 expression showed highest SFRP2 mRNA levels in pancreatic cancer out of 32 cancer types. GEPIA analysis of PDAC (n=179) and normal pancreas (n=177) mRNA data showed higher levels of SFRP2 mRNA in PDAC compared to normal pancreas (p<.01). High SFRP2 mRNA levels were associated with decreased disease-free survival hazard ratio = 2.6; p=0.0097). PDAC tissue showed higher SFRP2 protein compared to matched normal-adjacent tissue (n=24, p=0.0003). High SFRP2 expression was associated with decreased OS (p=0.0158. The ESTIMATE stromal score correlated with SFRP2 expression in human PDAC (Spearman correlation coefficient 0.81). Fibrosis levels in KPC PDAC tumors correlated with SFRP2 IHC (Spearman correlation coefficient 0.75). Fibroblasts treated with SFRP2 and TGFβ showed higher mesenchymal markers compared to untreated control. CONCLUSIONS: SFRP2 is increased in PDAC, associated with tumor fibrosis, and prognostic for survival. SFRP2 is a novel therapeutic target in PDAC. Citation Format: Lillian Hsu, Julie Siegel, Patrick Nasarre, Rupak Mukherjee, Meghan Gormley, Bailey Richardson, Jordan Morningstar, Imran Khan, Eleanor Hilliard, John O’Bryan, Krist Helke, Laura Spruill, Nathan Dolloff, Nancy Klauber-Demore. Secreted frizzled related-protein 2 is prognostic for survival in pancreatic cancer [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Pancreatic Cancer; 2023 Sep 27-30; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(2 Suppl):Abstract nr A102.

Abstract PO-011: Analysis of nuclear receptor expression in head and neck cancer

Abstract Background: The nuclear receptor (NR) superfamily contains a wide array of transcription factors responsible for cellular metabolism, inflammation, differentiation, and hormonal functions. For 45 years these receptors have been of interest in head and neck squamous cell carcinoma (HNSCC) treatment and prevention, especially retinoic acid receptor (RAR), peroxisome proliferator-activated receptor (PPAR), and vitamin D receptor (VDR) families. In the current study we examined expression patterns of these receptors in The Cancer Genome Atlas (TCGA). Methods: TCGA RNA-Seq data within the cBioportal platform comparing HNSCC samples (n=515 patients/samples) to normal tissue (n=82 patients/samples) was interrogated for significant differences in nuclear receptor expression. Genes with a z-score relative to normal samples (log RNA-Seq V2 RSEM normalized) greater than +2 or less than -2 in ≥50% of cases were considered significant. Unpaired, two-tailed t-tests were used to analyze comparisons in nuclear receptor genes’ z-scores across multiple clinical and pathologic parameters, p<0.05 was deemed significant. Statistical stringency for differential gene expression conformed to cBioportal and the University of Alabama at Birmingham’s Cancer Data Analysis Portal (UALCAN). Results: Of the 46 NR genes and 19 NR cofactors examined, ninety nine percent of tumor samples in the TCGA had some form of NR gene ‘alteration’ compared to normal tissue. These alterations predominantly encompass expression up- or down-regulation. NR genes Peroxisome Proliferator Activated Receptor Gamma (PPARγ) and Retinoic Acid Receptor Related Orphan Receptor C (RORC) (both in the thyroid hormone receptor-like family), and the NR cofactor, Nuclear Receptor Coactivator 1 (NCOA1), were differentially expressed and downregulated in tumors compared to normal tissue. Differences in specific NR expression varied significantly with HPV status (30 NRs), organ site (oral cavity vs hypopharynx & larynx; 20 NRs), tumor size (T1/2 vs. T3/4; 21 NRs), gender (male vs. female; 7 NRs); tumor staging (stage I/II vs. stage III/IV; 6 NRs); lymph node status (N0 vs. N+; 5 NRs), smoking history (smokers vs. non-smokers; 3 NRs), and patient race (White vs. Black/African American; 2 NRs). Conclusions: In prior studies of nuclear receptors in aerodigestive cancer, the retinoid receptor axis was downregulated, and a subsequent goal was stimulation and restoration of activity with natural or synthetic retinoids as a rationale for treatment. Presently, we have discovered significant decreases in PPARγ expression and are targeting this receptor with thiazolidinediones in a series of preclinical and human trials in a similar rationale to promote restoration of normal signaling and events like maturation. Here, we observed significant differences in a variety of parameters of several receptors associated with risk factors, stage, race, and gender that are observed and may become the focus of interest in future targeting efforts. Citation Format: Lindsey Mortensen, Beverly R. Wuertz, Frank G. Ondrey. Analysis of nuclear receptor expression in head and neck cancer [abstract]. In: Proceedings of the AACR-AHNS Head and Neck Cancer Conference: Innovating through Basic, Clinical, and Translational Research; 2023 Jul 7-8; Montreal, QC, Canada. Philadelphia (PA): AACR; Clin Cancer Res 2023;29(18_Suppl):Abstract nr PO-011.

Y-Chromosome-Linked Genes Are Associated With Sex-Related Head-Neck Squamous Cell Carcinoma Survival.

To define novel gene biomarkers for prognosis of head and neck squamous cell carcinoma (HNSCC) patients' survival. Retrospective study. The Cancer Genome Atlas (TCGA) HNSCC RNA-Seq dataset. Coexpressed gene clusters were extracted from TCGA RNA-seq data using our previously published method (EPIG). Kaplan-Meier estimator was then used for overall survival-relevant analysis, with patients partitioned into 3 groups based on gene expression levels: female, male_low, and male_high. Male had better overall survival than female and male with higher expression level of Y-chromosome-linked (Y-linked) genes had significantly better survival than those with lower expression levels. In addition, male with a higher expression level of Y-linked genes showed even better survival when they have a higher level of coexpressed cluster of genes related to B or T cell immune response. Other clinical conditions related to immune responses also consistently showed favorable effects on the Y-linked genes for survival estimation. Male patients with higher expression level of Y-linked genes also have significantly higher tumor/normal tissue (T/N) ratio of those genes and higher level of several immune responses related clinical measurements (eg, lymphocyte and TCR related). Male patients with lower expression level of Y-linked genes benefited from radiation-only treatment. The favorable role of a cluster of coexpressed Y-linked genes in HNSCC patients' survival is potentially associated with elevated level of immune responses. These Y-linked genes could serve as useful prognostic biomarkers for HNSCC patients' survival estimation and treatment.

VirMAP for cancer: Characterization of the intratumoral virome in virally associated cancers and a resource for investigators.

2568 Background: A major barrier to understanding the role of viruses in cancer progression has been the lack of sensitivity of current tools to assemble viral genomes in an untargeted way. Here, we utilize our novel VirMAP algorithm to identify eukaryotic and prokaryotic viruses associated with cancers, using RNA sequencing data from The Cancer Genome Atlas (TCGA). Methods: VirMAP was run on 16 TCGA RNASeq datasets of primary tumor tissues, with a focus on cancers with known viral associations (including cervical, head and neck, liver, stomach, esophagus and ovarian). Previously published survival data were used for exploratory analysis. Overall survival (OS) was assessed using the Kaplan-Meier method, and p-values were calculated using log-rank test. Multivariate analysis (MVA) was done using Cox proportional hazards model to control for age, sex, stage, and histology. Results: Viral sequence recovery (among 537 billion sequencing reads from 6185 patients) was at least equal to previously published algorithms such as Kraken, but with higher taxonomic resolution; overall recovery was 0.0041% from 16 tumor types. Cancers with high viral loads such as cervical cancer demonstrated high recovery rates (0.012% of all reads mapped to viral taxa, followed by 0.004% for liver, 0.0018% head and neck, 0.0016% stomach, 0.0009% esophageal, and 0.0002% ovarian). There was a wide range of viral signature richness, with cervical cancer showing the highest viral read count predominated by 18 different HPV serotypes; in contrast, stomach cancer had lower abundance but contained over 300 unique viral species. Exploratory analysis revealed novel correlations between viral signatures and OS. In cervical cancer, presence of HPV45 predicted worse survival (median OS 837 days vs 4086, p=0.0043); MVA demonstrated hazard ratio (HR) of 5.7 [2.6-11.4, p<0.0001]. In stomach cancer, patients often had more than one viral taxon present (mean 3.7, range 0-18); presence of seven or more unique taxa predicted worse OS (median OS 403 days vs 1153, p=0.0001). MVA showed HR of 1.1 for every additional taxon (p=0.0004). Across all 16 tumor types, the detection of phage virus was associated with worse OS (1073 days vs. 2094 days for non-phage virus and 2131 days for no virus, p<0.0001). Conclusions: VirMAP performs deep characterization of the tumor-associated virome in both high and low abundance settings, with generation of partial genome reconstructions and strain level taxonomic classification including phages. Further work is ongoing to optimize viral read normalization based on genome length, as well as subset analysis with disease-specific modifiers such as histology, race, and exposures. Utilizing publicly available platforms such as TCGA, clinical and correlative published data can be leveraged to better understand the nuanced pleotropic effects of the tumor virome on cancer progression.

TCGA RNA-Seq and Tumor-Infiltrating Lymphocyte Imaging Data Reveal Cold Tumor Signatures of Invasive Ductal Carcinomas and Estrogen Receptor-Positive Human Breast Tumors.

Comparative studies of immune-active hot and immune-deserted cold tumors are critical for identifying therapeutic targets and strategies to improve immunotherapy outcomes in cancer patients. Tumors with high tumor-infiltrating lymphocytes (TILs) are likely to respond to immunotherapy. We used the human breast cancer RNA-seq data from the cancer genome atlas (TCGA) and classified them into hot and cold tumors based on their lymphocyte infiltration scores. We compared the immune profiles of hot and cold tumors, their corresponding normal tissue adjacent to the tumor (NAT), and normal breast tissues from healthy individuals from the Genotype-Tissue Expression (GTEx) database. Cold tumors showed a significantly lower effector T cells, lower levels of antigen presentation, higher pro-tumorigenic M2 macrophages, and higher expression of extracellular matrix (ECM) stiffness-associated genes. Hot/cold dichotomy was further tested using TIL maps and H&E whole-slide pathology images from the cancer imaging archive (TCIA). Analysis of both datasets revealed that infiltrating ductal carcinoma and estrogen receptor ER-positive tumors were significantly associated with cold features. However, only TIL map analysis indicated lobular carcinomas as cold tumors and triple-negative breast cancers (TNBC) as hot tumors. Thus, RNA-seq data may be clinically relevant to tumor immune signatures when the results are supported by pathological evidence.

Excavation of diagnostic biomarkers and construction of prognostic model for clear cell renal cell carcinoma based on urine proteomics.

Clear cell renal cell carcinoma (ccRCC) is the most common pathology type in kidney cancer. However, the prognosis of advanced ccRCC is unsatisfactory. Thus, early diagnosis becomes one of the most important research priorities of ccRCC. However, currently available studies about ccRCC lack urine-related further studies. In this study, we applied proteomics to search urinary biomarkers to assist early diagnosis of ccRCC. In addition, we constructed a prognostic model to assist judge patients' prognosis. Urine which was used to perform 4D label-free quantitative proteomics was collected from 12 ccRCC patients and 11 non-tumor patients with no urinary system diseases. The urine of 12 patients with ccRCC confirmed by pathological examination after surgery was collected before operatoin. Bioinformatics analysis was used to describe the urinary proteomics landscape of these patients with ccRCC. The top ten proteins with the highest expression content were selected as the basis for subsequent validation. Urine from 46 ccRCC patients and 45 control patients were collected to use for verification by enzyme linked immunosorbent assay (ELISA). In order to assess the prognostic value of urine proteomics, a prognostic model was constructed by COX regression analysis on the intersection of RNA-sequencing data in The Cancer Genome Atlas (TCGA) database and our urine proteomic data. 133 proteins differentially expressed in the urinary samples were found and 85 proteins (Fold Change, FC>1.5) were identified up-regulated while 48 down-regulated (FC<0.5). Top 10 proteins including S100A14, PKHD1L1, FABP4, ITIH2, C3, C8G, C2, ATF6, ANGPTL6, F13B were performed ELISA to verify. The results showed that PKHD1L1, ANGPTL6, FABP4 and C3 were statistically significant (P<0.05). We performed multivariate logistic regression analysis and plotted a nomogram. Receiver operating characteristic (ROC) curve indicted that the diagnostic efficiency of combined indicators is satisfactory (Aare under curve, AUC=0.835). Furthermore, the prognostic value of the urine proteomics was explored through the intersection between urine proteomics and TCGA RNA-seq data. Thus, COX regression analysis showed that VSIG4, HLA-DRA, SERPINF1, and IGLV2-23 were statistically significant (P<0.05). Our study indicated that the application of urine proteomics to explore diagnostic biomarkers and to construct prognostic models of renal clear cell carcinoma is of certain clinical value. PKHD1L1, ANGPTL6, FABP4 and C3 can assist to diagnose ccRCC. The prognostic model constituted of VSIG4, HLA-DRA, SERPINF1, and IGLV2-23 can significantly predict the prognosis of ccRCC patients, but this still needs more clinical trials to verify.

Abstract 641: VirMAP for cancer: Characterization of the intratumoral virome in virally-associated cancers and a resource for investigators

Abstract Background: A major barrier to understanding the role of viruses in cancer progression has been the lack of sensitivity of current tools to identify and assemble viral genomes in an untargeted way. Here, we utilize our novel VirMAP algorithm to identify eukaryotic and prokaryotic viruses associated with cancers, using RNA sequencing data from The Cancer Genome Atlas (TCGA). Methods: VirMAP was used to analyze TCGA RNA-Seq data derived from primary tumor tissues, with a focus on cancers with known viral associations (cervical, ovarian, head and neck, liver, stomach and esophagus). Previously published survival data were matched with viral signatures for exploratory analysis. Ancom-BC, a differential abundance pipeline, was utilized to screen for viruses which drive outcome. Overall survival (OS) was assessed using the Kaplan-Meier method, and p-values were calculated using log-rank test. Multivariate analysis (MVA) was done using the Cox proportional hazards model to control for age, sex, stage, and histology. Results: Viral sequence recovery (among 723 billion sequencing reads from 2090 patients) was at least equal to previously published algorithms such as Kraken, but with higher taxonomic resolution. Cancers with high viral loads such as cervical cancer demonstrated high recovery rates (0.012% of all sequencing reads mapped to viral taxa, followed by 0.004% for liver, 0.0018% for head and neck squamous cell, 0.0016% for stomach, 0.0009% for esophageal, and 0.0002% for ovarian). There was a wide range of viral signature richness, with cervical cancer showing the highest viral read count predominated by HPV serotypes; in contrast, stomach cancer had lower abundance but contained over 300 unique viral taxa. Exploratory clinical analysis revealed novel correlations between viral signatures and OS. In cervical cancer, presence of HPV45 predicted worse survival (median OS 837 days vs 4086, p=0.0043); MVA demonstrated hazard ratio (HR) of 5.7 [2.6-11.4, p&amp;lt;0.0001]. In stomach cancer, patients often had more than one viral taxon present (mean 3.7, range 0-18); presence of 7 or more unique taxa predicted worse OS (median OS 403 days vs 1153, p=0.0001). MVA showed HR 1.1 for every additional taxon (p=0.0004). Conclusions: VirMAP performs deep characterization of the tumor-associated virome in both high and low abundance settings, with generation of partial genome reconstructions and strain level taxonomic classification. Utilizing publicly available platforms such as TCGA, clinical and correlative published data can be leveraged to better understand the nuanced pleotropic effects of the tumor virome on cancer progression. Citation Format: Yongwoo D. Seo, Neal Bhutiani, Matthew C. Wong, Ashish V. Damania, Golnaz Morad, Matthew Lastrapes, Alexander J. Lazar, Jennifer A. Wargo, Nadim J. Ajami. VirMAP for cancer: Characterization of the intratumoral virome in virally-associated cancers and a resource for investigators [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 641.

Abstract 2107: Development a customized panel with clinical application potential for transcriptome analyses in selected cancer types using FFPE specimens

Abstract Purpose: HTG EdgeSeq is a Next Generation Sequencing (NGS) based RNA Sequencing (RNA Seq) platform using a quantitative nuclease protection chemistry that enables extraction-free quantitation of mRNA/miRNAs with significantly reduces FFPE sample requirements compared to standard RNA-Seq, which makes it a feasible option for potential clinical applications. A customized gene panel was designed and developed for detection gene signatures. Procedure: The customized gene panel consists of 1327 genes, including 900 genes in signaling pathways in oncology &amp; immuno-oncology, which including Nature Kill (NK) cells, neutrophils, Plasmacytoid dendritic cells (pDCs), DCs, monocytes, Tregs, B cells, tumor associated macrophages (TAMs), TGFβ, EMT, interferon-γ signature, IL-17, β-catenin etc. and 15 house-keeping genes. The validation included in silico probe design, feasibility testing, data comparison with The Cancer Genome Atlas (TCGA) and precision studies in six cancer indications including: Non-Small-Cell Lung Cancer (NSCLC), Head and Neck Squamous Cell Carcinoma (HNSCC), Hepatocellular Carcinoma (HCC), Colorectal Cancer (CRC), gastric cancer (OC), ovarian cancer (GC). Data Summary: After the In-silico probe design for all genes the custom probe pool for quality control test in biological samples was performed by evaluating directional and expected expression of 1327 selected genes in 59 FFPE samples from six indications. All probes performed as expected according to specification; gene expression signatures data were consistent with the different cancer indication tested. Expected differential expression of the selected genes was based on RNA-Seq data as reported in TCGA. In addition, principal component analysis (PCA) performed on the entire data et, showed a distinct clustering of samples by indication, demonstrating that the custom panel was able to reliably identify each indication. For the validation, precision studies including inter-well, inter-day, inter-processor, inter-sequencer and site-to-site comparisons were performed using FFPE tissue lysate pools prepared from each of the 6 indications. Predefined acceptance criteria are ≥ 90% of all Lin`s Concordance Correlation Coefficients (LCCC) comparisons to be greater than 0.90. The results showed that the predefined acceptance criteria were met, which are 95.9% of LCCC &amp;gt; 0.90 for inter-well repeatability, and 100% of LCCC &amp;gt;0.90 for inter-day, inter-processor, inter-sequencer and inter-site comparisons. Conclusions: The designed panel met all the predefined quality and analytical acceptance criteria, also, the reported results is corrected with TCGA RNA-Seq data, and the precision studies confirmed its performance. Customized gene panel can be possibly utilized for potential predictive and prognostic biomarkers identification to inform new cancer therapies development. Citation Format: Vittorio D'Urso, Dorothee Foernzler, Dennis O’Rourke, Danyi Wang, Jorge F Sanchez-Garcia, Heidi Giese, Thomas Clarke IV, XiaoZhe (Janet) Wang, Andreas Machl, Ti Cai, Juergen Scheuenpflug, Zheng Feng. Development a customized panel with clinical application potential for transcriptome analyses in selected cancer types using FFPE specimens [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 2107.

Stromal Interaction Molecule 1 (STIM1) is a Potential Prognostic Biomarker and Correlates with Immune Infiltrates in Solid Tumors.

Increasing evidence has shown that stromal interaction molecule 1 (STIM1), a key subunit of store-operated Ca2+ entry (SOCE), is closely associated with tumor growth, development, and metastasis. However, there is no report of a comprehensive assessment of STIM1 in pan-cancer. This study aimed to perform a general analysis of STIM1 in human tumors, including its molecular characteristics, functional mechanisms, clinical significance, and immune infiltrates correlation based on pan-cancer data from The Cancer Genome Atlas (TCGA). Gene expression analysis was investigated using TCGA RNA-seq data, the Tumor Immune Estimation Resource (TIMER). Phosphorylation analysis was undertaken using the Clinical Proteomic Tumor Analysis Consortium (CP-TAC) and the PhosphoNET database. Genetic alterations of STIM1 were analyzed using cBioPortal. Prognostic analysis was via the R package "survival" function and the Kaplan-Meier plotter. Functional enrichment analysis was via by the R package "cluster Profiler" function. The association between STIM1 and tumor-infiltrating immune cells and immune markers was by the R package "GSVA" function and TIMER. STIM1 was differentially expressed and associated with distinct clinical stages in multiple tumors. The phosphorylation of STIM1 at S673 is highly expressed in clear cell renal carcinoma and lung adenocarcinoma tumors compared to normal tissues. STIM1 genetic alterations correlate with poor prognosis in several tumors, including ovarian cancer and lung squamous cell carcinomas. High STIM1 expression is associated with good or poor prognosis across diverse tumors. Overall survival (OS) analysis indicated that STIM1 is a favorable prognostic factor for patients with BRCA, KIRC, LIHC, LUAD, OV, SARC, and UCEC, and is a risk prognostic factor for BLCA, KIRP, STAD, and UVM. There is a close correlation between STIM1 expression and immune cell infiltration, immune-regulated genes, chemokines, and immune checkpoints in a variety of tumors. STIM1 functions differently in diverse tumors, playing an oncogenic or antitumor role. Moreover, It may serve as a prognostic biomarker and an immunotherapy target across multiple tumors.

Abstract P003: Computational identification of ovarian cancer candidate genes with mutated protein structures caused by missense variants

Abstract Ovarian cancer is one of the leading killers in the US for women. The American Cancer Society (ACR) ranks ovarian cancer as the fifth leading cause of cancer deaths among women. This cancer is hereditary, with higher risks if the patient has more relatives who have had ovarian cancer before. The Cancer Genome Atlas (TCGA) is a joint-effort project by NCI and NHGRI that generates genomic data for the cancer patients and matches normal data samples with these patients. Missense mutations are single nucleotide polymorphisms (SNPs) in which the amino acids are changed. A missense variant could disrupt protein structures; thus the function of the disrupted protein is altered. In this study, we investigated genes with high expression levels in the TCGA RNA-seq dataset by characterizing protein structure change patterns of candidate genes due to missense mutations in association with ovarian cancer. We also analyzed the association of these genes with ovarian cancer prognosis. We aim to select genes that have pathogenic consequences. We selected 44 genes with the highest expression values based on TCGA ovarian cancer RNA-seq data. We then retrieved protein sequences from the Ensembl BioMart database and missense mutation annotation information from cBioPortal for 44 input genes. Specifically, the missense mutation dataset, which is projected by the Pan-Cancer Analysis of Whole Genomes, was retrieved from cBioPortal. We then used the Phyre2 database to predict the 3-D structures of the wild and mutated type genes to identify structural changes. However, some genes were not suitable for Phyre2 analysis. For example, RPL41 was not long enough for analysis with Phyre2 as it only had 25 amino acids whereas Phyre2 required 30. TMSL3 (TMSB4XP8) is a pseudogene not found in both NCBI and AlphaFold. We then searched DrugBank for 42 genes with the predicted protein structure changes by Phyre2 to identify drug targets. We also performed Kaplan-Meier survival analysis to further evaluate the clinical consequences of the candidate genes. Our preliminary results show 26 genes that have drug targets and several of them have been identified to either decrease or increase the survival rate when the expression is high in ovarian cancer patients. This indicates that the highly expressed genes with protein structure changes due to missense mutations could be potential biomarkers for treating ovarian cancer. The functional annotation study could further validate ovarian cancer association for our prioritized gene list. In the future, we also plan to compare prediction results between genders and race groups. Our results could guide medical researchers in prioritizing drug targets and developing better treatment strategies for ovarian cancer. Our approach in prioritizing candidate genes is applicable to other cancers as well, and we plan to develop bioinformatic pipelines for automatically predicting candidate genes with significant effect on patient survival for any cancer type. Citation Format: Ian W. Hou, Yongsheng Bai. Computational identification of ovarian cancer candidate genes with mutated protein structures caused by missense variants. [abstract]. In: Proceedings of the AACR Special Conference: Precision Prevention, Early Detection, and Interception of Cancer; 2022 Nov 17-19; Austin, TX. Philadelphia (PA): AACR; Can Prev Res 2023;16(1 Suppl): Abstract nr P003.

Characterization of a lipid droplet and endoplasmic reticulum stress related gene risk signature to evaluate the clinical and biological value in hepatocellular carcinoma

IntroductionLipid metabolism and endoplasmic reticulum (ER) stress play an important role in the progression and metastasis of hepatocellular carcinoma (HCC). We aimed to establish lipid droplet (LD)-associated and ER stress-related gene risk signature as prognostic indicators.Materials and methodsLiterature searches for LD-associated proteins was screened and validated in The Cancer Genome Atlas (TCGA) and International Cancer Genome Collaboratory (ICGC) databases. A total of 371 samples were enrolled from the TCGA RNA-seq dataset (training cohort) and 240 samples from IGGC RNA-seq dataset (validation cohort). A 10-gene risk signature was established by the last absolute shrinkage and selection operator (LASSO) regression analysis. The prognostic value of the risk signature was evaluated by Cox regression, Kaplan–Meier and ROC Curve analyses. Biological features associated with LD and ER stress-related factors were explored by functional analysis and in vitro experiment.ResultsBased on the medical literatures, 124 lipid droplet-associated proteins were retrieved, and three genes failed to establish a valid prognostic model. ER stress was considered as an important component by functional analysis. A 10-gene risk signature compared the clinicopathology characteristics, immunosuppressive events and a nomogram in HCC patients.ConclusionLD-associated and ER stress-related gene risk signatures highlighted poor prognosis for clinicopathological features, positively correlate with macrophages and T cell immunoglobulin and mucin-3 (TIM-3) expression in the tumor microenvironment, and might act as independent prognostic factors.

Analysis of aberrant miRNA-mRNA interaction networks in prostate cancer to conjecture its molecular mechanisms.

MicroRNAs (miRNAs) capable of post-transcriptionally regulating mRNA expression are essential to tumor occurrence and progression. This study aims to find negatively regulatory miRNA-mRNA pairs in prostate adenocarcinoma (PRAD). Combining The Cancer Genome Atlas (TCGA) RNA-Seq data with Gene Expression Omnibus (GEO) mRNA/miRNA expression profiles, differently expressed miRNA/mRNA (DE-miRNAs/DE-mRNAs) were identified. MiRNA-mRNA pairs were screened by miRTarBase and TarBase, databases collecting experimentally confirmed miRNA-mRNA pairs, and verified in 30 paired prostate specimens by real-time reverse transcription polymerase chain reaction (RT-qPCR). The diagnostic values of miRNA-mRNA pairs were measured by receiver operation characteristic (ROC) curve and Decision Curve Analysis (DCA). DAVID-mirPath database and Connectivity Map were employed in GO/KEGG analysis and compounds research. Interactions between miRNA-mRNA pairs and phenotypic features were analyzed with correlation heatmap in hiplot. Based on TCGA RNA-Seq data, 22 miRNA and 14 mRNA GEO datasets, 67 (20 down and 47 up) miRNAs and 351 (139 up and 212 down) mRNAs were selected. After screening from 2 databases, 8 miRNA (up)-mRNA (down) and 7 miRNA (down)-mRNA (up) pairs were identified with Pearson's correlation in TCGA. By external validation, miR-221-3p (down)/GALNT3 (up) and miR-20a-5p (up)/FRMD6 (down) were chosen. The model combing 4 signatures possessed better diagnostic value. These two miRNA-mRNA pairs were significantly connected with immune cells fraction and tumor immune microenvironment. The diagnostic model containing 2 negatively regulatory miRNA-mRNA pairs was established to distinguish PRADs from normal controls.

Cancer cell intrinsic TIM-3 induces glioblastoma progression

Glioblastoma (GBM) is identified to share common signal pathways between glioma and immune cells. Here, we find that Tcell immunoglobulin domain and mucin domain protein 3 (TIM-3) is one of the most common co-inhibitory immune checkpoints in GBM shared by tumor and non-tumor cells. Glioma cell-intrinsic TIM-3 is involved in not only regulating malignant behaviors of glioma cells but also inducing macrophage migration and transition to anti-inflammatory/pro-tumorigenic phenotype by a TIM-3/interleukin 6 (IL6) signal. In mechanism, as one of the major regulators of IL6, TIM-3 regulates its expression through activating NF-κB. Blocking this feedback loop by Tocilizumab, an IL6R inhibitor, inhibited the above effects and repressed the tumorigenicity of GBM invivo. Our work identifies glioma cell-intrinsic functions of TIM-3/IL6 signal mediating the crosstalk feedback loop between glioma cells and tumor-associated macrophages (TAMs). Blocking this feedback loop may provide a novel therapeutic strategy for GBM.

Establishment and Validation of a Tumor Microenvironment Prognostic Model for Predicting Bladder Cancer Survival Status Based on Integrated Bioinformatics Analyses.

This study was designed to analyze the characteristics of bladder cancer-related genes and establish a prognostic model of bladder cancer. The model passed an independent external validation set test. Differentially expressed genes (DEGs) related to bladder cancer were obtained from the Gene Expression Omnibus (GEO), The Cancer Genome Atlas (TCGA), and Genotype-Tissue Expression (GTEx) databases. WGCNA was used to fit the GSE188715, TCGA, and GTEx RNA-Seq data. Fusing the module genes with the high significance in tumor development extracted from WGCNA and DEGs screened from multiple databases. 709 common prognostic-related genes were obtained. The 709 genes were enriched in the Gene Ontology database. Univariate Cox and LASSO regression analyses were used to screen out 21 prognostic-related genes and further multivariate Cox regression established a bladder cancer prognostic model consisting of 8 genes. After the eight-gene prognostic model was established, the Human Protein Atlas (HPA) database, GEPIA 2, and quantitative real-time PCR (qRT-PCR) verified the differential expression of these genes. Gene Set Enrichment Analysis and immune infiltration analysis found biologically enrichment pathways and cellular immune infiltration related to this bladder cancer prognostic model. Then, we selected bladder cancer patients in the TCGA database to evaluate the predictive ability of the model on the training set and validation set. The overall survival status of the two TCGA patient groups in the training and the test sets was obtained by Kaplan–Meier survival analysis. Three-year survival rates in the training and test sets were 37.163% and 25.009% for the low-risk groups and 70.000% and 62.235% for the high-risk groups, respectively. Receiver operating characteristic curve (ROC) analysis showed that the areas under the curve (AUCs) for the training and test sets were above 0.7. In an external independent validation database GSE13507, Kaplan–Meier survival analysis showed that the three-year survival rates of the high-risk and the low-risk groups in this database were 56.719% and 76.734%, respectively. The AUCs of the ROC drawn in the external validation set were both above 0.65. Here, we constructed a prognostic model of bladder cancer based on data from the GEO, TCGA, and GTEx databases. This model has potential prognostic and clinical auxiliary diagnostic value.

Removing unwanted variation from large-scale RNA sequencing data with PRPS

Accurate identification and effective removal of unwanted variation is essential to derive meaningful biological results from RNA sequencing (RNA-seq) data, especially when the data come from large and complex studies. Using RNA-seq data from The Cancer Genome Atlas (TCGA), we examined several sources of unwanted variation and demonstrate here how these can significantly compromise various downstream analyses, including cancer subtype identification, association between gene expression and survival outcomes and gene co-expression analysis. We propose a strategy, called pseudo-replicates of pseudo-samples (PRPS), for deploying our recently developed normalization method, called removing unwanted variation III (RUV-III), to remove the variation caused by library size, tumor purity and batch effects in TCGA RNA-seq data. We illustrate the value of our approach by comparing it to the standard TCGA normalizations on several TCGA RNA-seq datasets. RUV-III with PRPS can be used to integrate and normalize other large transcriptomic datasets coming from multiple laboratories or platforms.

PTPRN Serves as a Prognostic Biomarker and Correlated with Immune Infiltrates in Low Grade Glioma.

Background: Glioma is one of the most common malignant tumors of the central nervous system. Immune infiltration of tumor microenvironment was associated with overall survival in low grade glioma (LGG). However, effects of Tyrosine phosphatase receptor type N (PTPRN) on the progress of LGG and its correlation with tumor infiltration are unclear. Methods: Here, datasets of LGG were from The Cancer Genome Atlas (TCGA) and normal samples were from GTEx dataset. Gepia website and Human Protein Atlas (HPA) Database were used to analyze the mRNA and protein expression of PTPRN. We evaluated the influence of PTPRN on survival of LGG patients. MethSurv was used to explore the expression and prognostic patterns of single CpG methylation of PTPRN gene in LGG. The correlations between the clinical information and PTPRN expression were analyzed using logistic regression and Multivariate Cox regression. We also explored the correlation between PTPRN expression and cancer immune infiltration by TIMER. Gene set enrichment analysis (GSEA) was formed using TCGA RNA-seq datasets. Results: PTPRN mRNA and protein expression decreased in LGG compared to normal brain tissue in TCGA and HPA database. Kaplan-Meier analysis showed that the high expression level of PTPRN correlated with a good overall survival (OS) of patients with LGG. The Multivariate Cox analysis demonstrated that PTPRN expression and other clinical-pathological factors (age, WHO grade, IDH status, and primary therapy outcome) significantly correlated with OS of LGG patients. The DNA methylation pattern of PTPRN with significant prognostic value were confirmed, including cg00672332, cg06971096, cg01382864, cg03970036, cg10140638, cg16166796, cg03545227, and cg25569248. Interestingly, PTPRN expression level significantly negatively correlated with infiltrating level of B cell, CD4+ T cells, Macrophages, Neutrophils, and DCs in LGG. Finally, GSEA showed that signaling pathways, mainly associated with tumor microenvironment and immune cells, were significantly enriched in PTPRN high expression. Conclusion: PTPRN is a potential biomarker and correlates with tumor immune infiltration in LGG.

Expression Profiles of HOXC6 Predict the Survival of Glioblastoma Patients and Correlate with Cell Cycle.

The goal of this study was to investigate the homeobox (HOX) gene expression status and its prognostic value in glioblastoma multiforme (GBM) and to uncover the biological processes related to its expression. The prognostic value of HOX genes in GBM was systematically investigated by a genome-wide analysis of HOX gene expression profiles in GBM patient samples in The Cancer Genome Atlas (TCGA) project (microarray dataset) and validation datasets. Using the differentially expressed gene (DEG) analysis and a Cox regression model, we discovered that the HOXC6 could stratify patients into significantly different survival (p = 0.0012, log-rank test) groups in the training cohort. TCGA RNA-seq and GSE16011 datasets were used for validation. Multivariate Cox and stratification analysis indicated that HOXC6 was an independent prognostic factor after adjusting for other clinical covariates. Bioinformatic analysis suggested that the HOXC6 might be involved in the cell cycle-related biological processes and pathways that are well established in the context of glioblastoma tumorigenesis. We further explored the bioinformatic implications by gene set enrichment analysis (GSEA). Tumor cell biology experiments verified the role of HOXC6 in proliferation and cell cycle progression. In conclusion, HOXC6 might be a candidate biomarker gene for individual treatment optimization of glioblastoma. HOXC6 expression has a significant prognostic value and is related to the cell cycle process in glioblastoma.