The Cancer Genome Atlas Pan-Cancer analysis project

Knowledge-guided learning methods for integrative analysis of multi-omics data

Integrative analysis of multi-omics data has the potential to yield valuable and comprehensive insights into the molecular mechanisms underlying complex diseases such as cancer and Alzheimer's disease. However, a number of analytical challenges complicate multi-omics data integration. For instance, -omics data are usually high-dimensional, and sample sizes in multi-omics studies tend to be modest. Furthermore, when genes in an important pathway have relatively weak signal, it can be difficult to detect them individually. There is a growing body of literature on knowledge-guided learning methods that can address these challenges by incorporating biological knowledge such as functional genomics and functional proteomics into multi-omics data analysis. These methods have been shown to outperform their counterparts that do not utilize biological knowledge in tasks including prediction, feature selection, clustering, and dimension reduction. In this review, we survey recently developed methods and applications of knowledge-guided multi-omics data integration methods and discuss future research directions.

Novel benzoylurea derivative decreases TRPM7 channel function and inhibits cancer cells migration

The transient receptor potential melastatin 7 channel (TRPM7) is a nonselective cation channel highly expressed in some human cancer tissues. TRPM7 is involved in the proliferation, migration, invasion, and epithelial-mesenchymal transition (EMT) of cancer cells. Modulation of TRPM7 could be a promising therapeutic strategy for treating cancer; however, efficient and selective pharmacological TRPM7 modulators are lacking. In this study we investigated N- [4- (4, 6-dimethyl- 2-pyrimidinyloxy) - 3- methylphenyl] -N' - [2 -(dimethylamino)] benzoylurea (SUD), a newly synthesized benzoylurea derivative, for its effects on cancer cell migration and EMT and on functional expression of TRPM7. Our previous studies showed that SUD induces cell cycle arrest and apoptosis of MCF-7 and BGC-823 cells (human breast cancer and gastric cancer cell lines, respectively). Here, we show that SUD significantly decreased the migration of both types of cancer cells. Moreover, SUD decreased vimentin expression and increased E-cadherin expression in both cell types, indicating that EMT is also decreased by SUD. Importantly, SUD potentially reduced the TRPM7-like current in a concentration-dependent manner and decreased TRPM7 expression through the PI3K/Akt signaling pathway. Finally, molecular docking simulations were used to investigate potential SUD binding sites on TRPM7. In summary, our research demonstrated that SUD is an effective TRPM7 inhibitor and a potential agent to suppress the metastasis of breast and gastric cancer by inhibiting TRPM7 expression and function.

Benchmarking feature selection and feature extraction methods to improve the performances of machine-learning algorithms for patient classification using metabolomics biomedical data

Objective

Classification tasks are an open challenge in the field of biomedicine. While several machine-learning techniques exist to accomplish this objective, several peculiarities associated with biomedical data, especially when it comes to omics measurements, prevent their use or good performance achievements. Omics approaches aim to understand a complex biological system through systematic analysis of its content at the molecular level. On the other hand, omics data are heterogeneous, sparse and affected by the classical "curse of dimensionality" problem, i.e. having much fewer observation, samples (n) than omics features (p). Furthermore, a major problem with multi-omics data is the imbalance either at the class or feature level. The objective of this work is to study whether feature extraction and/or feature selection techniques can improve the performances of classification machine-learning algorithms on omics measurements.

Methods

Among all omics, metabolomics has emerged as a powerful tool in cancer research, facilitating a deeper understanding of the complex metabolic landscape associated with tumorigenesis and tumor progression. Thus, we selected three publicly available metabolomics datasets, and we applied several feature extraction techniques both linear and non-linear, coupled or not with feature selection methods, and evaluated the performances regarding patient classification in the different configurations for the three datasets.

Results

We provide general workflow and guidelines on when to use those techniques depending on the characteristics of the data available. To further test the extension of our approach to other omics data, we have included a transcriptomics and a proteomics data. Overall, for all datasets, we showed that applying supervised feature selection improves the performances of feature extraction methods for classification purposes. Scripts used to perform all analyses are available at: https://github.com/Plant-Net/Metabolomic_project/.

Inflammatory response signature score model for predicting immunotherapy response and pan-cancer prognosis

Background

Inflammatory responses influence the outcome of immunotherapy and tumorigenesis by modulating host immunity. However, systematic inflammatory response assessment models for predicting cancer immunotherapy (CIT) responses and survival across human cancers remain unexplored. Here, we investigated an inflammatory response score model to predict CIT responses and patient survival in a pan-cancer analysis.

Methods

We retrieved 12 CIT response gene expression datasets from the Gene Expression Omnibus database (GSE78220, GSE19423, GSE100797, GSE126044, GSE35640, GSE67501, GSE115821 and GSE168204), Tumor Immune Dysfunction and Exclusion database (PRJEB23709, PRJEB25780 and phs000452.v2.p1), European Genome-phenome Archive database (EGAD00001005738), and IMvigor210 cohort. The tumor samples from six cancers types: metastatic urothelial cancer, metastatic melanoma, gastric cancer, primary bladder cancer, renal cell carcinoma, and non-small cell lung cancer.We further established a binary classification model to predict CIT responses using the least absolute shrinkage and selection operator (LASSO) computational algorithm.

Findings

The model had high predictive accuracy in both the training and validation cohorts. During sub-group analysis, area under the curve (AUC) values of 0.82, 0.80, 0.71, 0.7, 0.67, and 0.64 were obtained for the non-small cell lung cancer, gastric cancer, metastatic urothelial cancer, primary bladder cancer, metastatic melanoma, and renal cell carcinoma cohorts, respectively. CIT response rates were higher in the high-scoring training cohort subjects (51%) than the low-scoring subjects (27%). The five-year survival rates in the high- and low score groups of the training cohorts were 62% and 21%, respectively, while those of the validation cohorts were 54% and 22%, respectively (P < 0·001 in all cases). Inflammatory response signature score derived from on-treatment tumor specimens are highly predictive of response to CIT in patients with metastatic melanoma. A significant correlation was observed between the inflammatory response scores and tumor purity. Regardless of the tumor purity, patients in the low score group had a significantly poorer prognosis than those in the high score group. Immune cell infiltration analysis indicated that in the high score cohort, tumor-infiltrating lymphocytes were significantly enriched, particularly effector and natural killer cells. Inflammatory response scores were positively correlated with immune checkpoint genes, suggesting that immune checkpoint inhibitors may have benefited patients with high scores. Analysis of signature scores across different cancer types from The Cancer Genome Atlas revealed that the prognostic performance of inflammatory response scores for survival in patients who have not undergone immunotherapy can be affected by tumor purity. Interleukin 21 (IL21) had the highest weight in the inflammatory response model, suggesting its vital role in the prediction mode. Since the number of metastatic melanoma patients (n = 429) was relatively large among CIT cohorts, we further performed a co-culture experiment using a melanoma cell line and CD8 + T cell populations generated from peripheral blood monocytes. The results showed that IL21 therapy combined with anti-PD1 (programmed cell death 1) antibodies (trepril monoclonal antibodies) significantly enhanced the cytotoxic activity of CD8 + T cells against the melanoma cell line.

Conclusion

In this study, we developed an inflammatory response gene signature model that predicts patient survival and immunotherapy response in multiple malignancies. We further found that the predictive performance in the non-small cell lung cancer and gastric cancer group had the highest value among the six different malignancy subgroups. When compared with existing signatures, the inflammatory response gene signature scores for on-treatment samples were more robust predictors of the response to CIT in metastatic melanoma.

A review on trends in development and translation of omics signatures in cancer

The field of cancer genomics and transcriptomics has evolved from targeted profiling to swift sequencing of individual tumor genome and transcriptome. The steady growth in genome, epigenome, and transcriptome datasets on a genome-wide scale has significantly increased our capability in capturing signatures that represent both the intrinsic and extrinsic biological features of tumors. These biological differences can help in precise molecular subtyping of cancer, predicting tumor progression, metastatic potential, and resistance to therapeutic agents. In this review, we summarized the current development of genomic, methylomic, transcriptomic, proteomic and metabolic signatures in the field of cancer research and highlighted their potentials in clinical applications to improve diagnosis, prognosis, and treatment decision in cancer patients.

Identification of feature genes and molecular mechanisms involved in cell communication in uveal melanoma through analysis of single‑cell sequencing data

Uveal melanoma (UM) is a highly metastatic cancer with resistance to immunotherapy. The present study aimed to identify novel feature genes and molecular mechanisms in UM through analysis of single-cell sequencing data. For this purpose, data were downloaded from The Cancer Genome Atlas and National Center for Biotechnology Information Gene Expression Omnibus public databases. The statistical analysis function of the CellPhoneDB software package was used to analyze the ligand-receptor relationships of the feature genes. The Metascape database was used to perform the functional annotation of notable gene sets. The randomForestSRC package and random survival forest algorithm were applied to screen feature genes. The CIBERSORT algorithm was used to analyze the RNA-sequencing data and infer the relative proportions of the 22 immune-infiltrating cell types. In vitro, small interfering RNAs were used to knockdown the expression of target genes in C918 cells. The migration capability and viability of these cells were then assessed by gap closure and Cell Counting Kit-8 assays. In total, 13 single-cell sample subtypes were clustered by t-distributed Stochastic Neighbor Embedding and annotated by the R package, SingleR, into 7 cell categories: Tissue stem cells, epithelial cells, fibroblasts, macrophages, natural killer cells, neurons and endothelial cells. The interactions in NK cells|Endothelial cells, Neurons|Endothelial cells, CD74_APP, and SPP1_PTGER4 were more significant than those in the other subsets. T-Box transcription factor 2, tropomyosin 4, plexin D1 (PLXND1), G protein subunit α I2 (GNAI2) and SEC14-like lipid binding 1 were identified as the feature genes in UM. These marker genes were found to be significantly enriched in pathways such as vasculature development, focal adhesion and cell adhesion molecule binding. Significant correlations were observed between key genes and immune cells as well as immune factors. Relationships were also observed between the expression levels of the key genes and multiple disease-related genes. Knockdown of PLXND1 and GNAI2 expression led to significantly lower viability and gap closure rates of C918 cells. Therefore, the results of the present study uncovered cell communication between endothelial cells and other cell types, identified innovative key genes and provided potential targets of gene therapy in UM.

Comprehensive multi-omics profiling identifies novel molecular subtypes of pancreatic ductal adenocarcinoma

Pancreatic cancer, a highly fatal malignancy, is predicted to rank as the second leading cause of cancer-related death in the next decade. This highlights the urgent need for new insights into personalized diagnosis and treatment. Although molecular subtypes of pancreatic cancer were well established in genomics and transcriptomics, few known molecular classifications are translated to guide clinical strategies and require a paradigm shift. Notably, chronically developing and continuously improving high-throughput technologies and systems serve as an important driving force to further portray the molecular landscape of pancreatic cancer in terms of epigenomics, proteomics, metabonomics, and metagenomics. Therefore, a more comprehensive understanding of molecular classifications at multiple levels using an integrated multi-omics approach holds great promise to exploit more potential therapeutic options. In this review, we recapitulated the molecular spectrum from different omics levels, discussed various subtypes on multi-omics means to move one step forward towards bench-to-beside translation of pancreatic cancer with clinical impact, and proposed some methodological and scientific challenges in store.

m6A modification of VEGFA mRNA by RBM15/YTHDF2/IGF2BP3 contributes to angiogenesis of hepatocellular carcinoma

Vascular endothelial growth factor A (VEGFA) plays a critical role as a potent angiogenesis factor and is highly expressed in hepatocellular carcinoma (HCC). Although the expression of VEGFA has been strongly linked to the aggressive nature of HCC, the specific posttranscriptional modifications that might contribute to VEGFA expression and HCC angiogenesis are not yet well understood. In this study, we aimed to investigate the epitranscriptome regulation of VEGFA in HCC. A comprehensive analysis integrating MeRIP-seq, RNA-seq, and crosslinking-immunprecipitation-seq data revealed that VEGFA was hypermethylated in HCC and identified the potential m6A regulators of VEGFA including a m6A methyltransferase complex component RBM15 and the two readers, YTHDF2 and IGF2BP3. Through rigorous cell and molecular biology experiments, RBM15 was validated as a key component of methyltransferase complex responsible for m6A methylation of VEGFA, which was subsequently recognized and stabilized by IGF2BP3 and YTHDF2, leading to enhanced VEGFA expression and VEGFA-related functions such as human umbilical vascular endothelial cells (HUVEC) migration and tube formation. In the HCC xenograft model, knockdown of RBM15, IGF2BP3, or YTHDF2 resulted in reduced expression of VEGFA, accompanied by significant inhibition of tumor growth closely associated with VEGFA expression and angiogenesis. Furthermore, our analysis of HCC clinical samples identified positive correlations between the expression levels of VEGFA and the regulators RBM15, IGF2BP3, and YTHDF2. Collectively, these findings offer novel insights into the posttranscriptional modulation of VEGFA and provide potential avenues for alternative approaches to antiangiogenesis therapy targeting VEGFA.

The pan-cancer landscape presented ITGA7 as a prognostic determinant, tumor suppressor, and oncogene in multiple tumor types

Integrin α7 (ITGA7) is an extracellular matrix-binding protein. Integrins are the main type of cell adhesive molecules in mammals, playing a role in many biological pathways. Although various studies have shown correlations between ITGA7 and various types of cancer, a comprehensive study at a pan-cancer level has not yet been conducted. In this study, we investigated the function of ITGA7 in distinct tumor types using the multi-omics relevant information, then two CeRNA regulatory network was drawn to identify the ITGA7 hub regulatory RNAs. The results indicated that the expression of ITGA7 varies in different tumors. Overexpression of ITGA7 was correlated with a worse OS in BLCA, LGG, and UVM, and the downregulation of ITGA7 was related to a worse OS in PAAD. In addition, BLCA, and UVM showed poor PFS in association with ITGA7 overexpression, and PAAD, SARC, and THCA indicated poor PFS in correlation with ITGA7 under expression. Further analyses of ITGA7 gene alteration data showed that ITGA7 amplifications may have an impact on Kidney Chromophobe prognosis. In 20 types of tumors, ITGA7 expression was linked to cancer-associated fibroblast infiltration. ITGA7 expression was linked to cancer-associated fibroblast infiltration. ITGA7-Related Gene Enrichment Analysis indicated that ITGA7 expression-correlated and functional binding genes were enriched in homotypic cell-cell adhesion, focal adhesion, and ECM-receptor interaction. This pan-cancer study found that abnormal expression of ITGA7 was correlated with poor prognosis and metastasis in different types of tumors. Thus, the ITGA7 gene may prove to be a promising biomarker for the prognosis and complication prevention of different cancers.

Survival prediction and analysis of drug-resistance genes in HER2-positive breast cancer

Despite the approval of several therapeutic agents for HER2-positive breast cancer, drug resistance remains a significant challenge, hindering the patient's prognosis. Thus, our study aimed to establish a risk model to predict the prognosis of patients and identify key genes regulating drug resistance in HER2-positive breast cancer. Utilizing data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO), a predictive model was constructed based on 5 drug resistance-related genes, which demonstrated a notable capacity to indicate the survival rates of patients. Besides, through eccDNA and transcriptome sequencing of drug-sensitive and resistant cancer cells, 3 significant DEGs were identified: MED1, MED24, and NMD3. Among them, MED1 showed the most significant elevation in drug-resistance cells, highlighting its crucial role in mediating drug resistance. MED1 may serve as a valuable target for alleviating drug resistance in HER2-positive breast cancer.

Multi-omics analysis identifies BCAT2 as a potential pan-cancer biomarker for tumor progression and immune microenvironment modulation

Branched-chain amino acid transaminase 2 (BCAT2) encodes a crucial protein involved in the initial catalysis of branched-chain amino acid (BCAA) catabolism, with emerging evidence suggesting its association with tumor progression. This study explores BCAT2 in a pan-cancer multi-omics context and evaluates its prognostic significance. We utilized a multi-database approach, analyzing cBioPortal for genetic alterations, RNA-Seq data from TCGA and GTEx for expression patterns, and RSEM for transcript analysis. Protein expression and interaction networks were assessed using the Human Protein Atlas, UniProt, and STRING. Prognostic value was determined through Cox regression analysis of TCGA clinical survival data, while immune cell infiltration across various cancers was examined using TCGA data and the TIMER2 platform. Our results revealed that BCAT2 alterations are primarily amplifications and is upregulated in various tumors, correlating with poor survival rates in several tumor types, including GBMLGG, LGG, and UVM. Elevated BCAT2 protein levels were common in pan-cancer, interacting with a range of metabolic enzymes. Additionally, BCAT2 expression significantly influenced CD4+ T cells, CD8+ T cells, and Treg cells infiltration, with varied correlations across cancer types. These findings indicate BCAT2 as a potential biomarker for cancer diagnosis and therapy, potentially regulating key metabolic and immune factors to mediate tumor progression and the microenvironment.

Scalable, compressed phenotypic screening using pooled perturbations

High-throughput phenotypic screens using biochemical perturbations and high-content readouts are constrained by limitations of scale. To address this, we establish a method of pooling exogenous perturbations followed by computational deconvolution to reduce required sample size, labor and cost. We demonstrate the increased efficiency of compressed experimental designs compared to conventional approaches through benchmarking with a bioactive small-molecule library and a high-content imaging readout. We then apply compressed screening in two biological discovery campaigns. In the first, we use early-passage pancreatic cancer organoids to map transcriptional responses to a library of recombinant tumor microenvironment protein ligands, uncovering reproducible phenotypic shifts induced by specific ligands distinct from canonical reference signatures and correlated with clinical outcome. In the second, we identify the pleotropic modulatory effects of a chemical compound library with known mechanisms of action on primary human peripheral blood mononuclear cell immune responses. In sum, our approach empowers phenotypic screens with information-rich readouts to advance drug discovery efforts and basic biological inquiry.

Metastasis-free survival of uveal melanoma by tumour size category based on The Cancer Genome Atlas (TCGA) classification in 1001 cases

BACKGROUND

Uveal melanoma (UM) can be classified by tumour size category and by The Cancer Genome Atlas (TCGA) groups (cytogenetic-based, 4-category prognostic classification into Groups A-D). This study was conducted to assess impact on metastasis-free survival (MFS) in UM by tumour size category based on correlation with TCGA classification.

METHODS

Retrospective analysis of 1001 cases categorised as small (0.0-3.0 mm), medium (3.1-8.0 mm) and large (≥8.1 mm), grouped by TCGA classification.

RESULTS

Of 1001 cases, TCGA Groups (A/B/C/D) included small (n = 270, 75%/11%/13%/1%), medium (n = 503, 46%/14%/27%/13%) and large (n = 228, 23%/19%/38%/20%) UM. The 5-and 10-year Kaplan-Meier MFS for small UM revealed Group A (98%, 98%), Group B (100%, 100%), Group C (86%, NA) and Group D (100%, NA). For medium UM, the values dropped with Group A (95%, 93%), Group B (90%, 90%), Group C (68%, 38%), and Group D (44%, NA). For large UM, the values dropped further with Group A (94%, 86%), Group B (85%, NA), Group C (40%, 28%), and Group D (23%, NA). Additionally, a comparison (small vs. medium vs. large tumour size category) revealed TCGA low-risk grouping (Groups A or B) in 86% vs. 60% vs. 58% cases with UM.

CONCLUSION

By tumour size category, favourable cytogenetics (Groups A or B) is found in 86% of small tumours, 60% of medium tumours, and 58% of large tumours. The MFS at 10 years for favourable cytogenetics was 98% for small tumours, 92% for medium tumours, and 54% for large tumours. Tumour size category can serve as a surrogate for TCGA.

Melanoma-specific mutation hotspots in distal, non-coding, promoter-interacting regions implicate novel candidate driver genes

BACKGROUND

To develop targeted treatments, it is crucial to identify the full spectrum of genetic drivers in melanoma, including those in non-coding regions. However, recent efforts to explore non-coding regions have primarily focused on gene-adjacent elements such as promoters and non-coding RNAs, leaving intergenic distal regulatory elements largely unexplored.

METHODS

We used Hi-C chromatin contact data from melanoma cells to map distal, non-coding, promoter-interacting regulatory elements genome-wide in melanoma. Using this "promoter-interaction network", alongside whole-genome sequence and gene expression data from the Pan Cancer Analysis of Whole Genomes, we developed multivariate linear regression models to identify distal somatic mutation hotspots that affect promoter activity.

RESULTS

We identified eight recurrently mutated hotspots that are novel, melanoma-specific, located in promoter-interacting distal regulatory elements, alter transcription factor binding motifs, and affect the expression of genes (e.g., HSPB7, CLDN1, ADCY9 and FDXR) previously implicated as tumour suppressors/oncogenes in various cancers.

CONCLUSIONS

Our study suggests additional non-coding drivers beyond the well-characterised TERT promoter in melanoma, offering new insights into the disruption of complex regulatory networks by non-coding mutations that may contribute to melanoma development. Furthermore, our study provides a framework for integrating multiple levels of biological data to uncover cancer-specific non-coding drivers.

Analysis of the relationship between resistin with prognosis, cell migration, and p38 and ERK1/2 activation in breast cancer

Obesity increases the risk and mortality of breast cancer through dysregulated secretion of proinflammatory cytokines and tumor adipokines that induce an inflammatory breast microenvironment. Resistin is an adipokine secreted by adipocytes, immune cells, and predominantly macrophages, which contributes to cancer progression, but its molecular mechanism in cancer is not completely described. In this study, we analyzed the relationship of resistin on breast cancer prognosis and tumor progression and the effect in vitro of resistin on p38 and ERK1/2 activation in breast cancer cell lines. By bioinformatic analysis, we found that resistin is overexpressed in the basal subtype triple-negative breast cancer and is related to poor prognosis. In addition, we demonstrated a positive correlation between RETN and MAPK3 expression in basal triple-negative breast cancer. Importantly, we found amplifications of the RETN gene in at least 20 % of metastatic samples from patients with breast cancer. Most samples with RETN amplifications metastasized to bone and showed high expression of IL-8 (CXCL8) and IL-6 (IL6). Finally, resistin could be considered a prognostic marker for basal triple-negative breast cancer, and we also proposed the possibility that resistin-induced cell migration involves the activation of MAPK in breast cancer cells.

Aneuploidy as a driver of human cancer

Aneuploidy, an abnormal chromosome composition, is a major contributor to cancer development and progression and an important determinant of cancer therapeutic responses and clinical outcomes. Despite being recognized as a hallmark of human cancer, the exact role of aneuploidy as a 'driver' of cancer is still largely unknown. Identifying the specific genetic elements that underlie the recurrence of common aneuploidies remains a major challenge of cancer genetics. In this Review, we discuss recurrent aneuploidies and their function as drivers of tumor development. We then delve into the context-dependent identification and functional characterization of the driver genes underlying driver aneuploidies and examine emerging strategies to uncover these driver genes using cancer genomics data and cancer models. Lastly, we explore opportunities for targeting driver aneuploidies in cancer by leveraging the functional consequences of these common genetic alterations.

Deep learning-based approaches for multi-omics data integration and analysis

BACKGROUND

The rapid growth of deep learning, as well as the vast and ever-growing amount of available data, have provided ample opportunity for advances in fusion and analysis of complex and heterogeneous data types. Different data modalities provide complementary information that can be leveraged to gain a more complete understanding of each subject. In the biomedical domain, multi-omics data includes molecular (genomics, transcriptomics, proteomics, epigenomics, metabolomics, etc.) and imaging (radiomics, pathomics) modalities which, when combined, have the potential to improve performance on prediction, classification, clustering and other tasks. Deep learning encompasses a wide variety of methods, each of which have certain strengths and weaknesses for multi-omics integration.

METHOD

In this review, we categorize recent deep learning-based approaches by their basic architectures and discuss their unique capabilities in relation to one another. We also discuss some emerging themes advancing the field of multi-omics integration.

RESULTS

Deep learning-based multi-omics integration methods were categorized broadly into non-generative (feedforward neural networks, graph convolutional neural networks, and autoencoders) and generative (variational methods, generative adversarial models, and a generative pretrained model). Generative methods have the advantage of being able to impose constraints on the shared representations to enforce certain properties or incorporate prior knowledge. They can also be used to generate or impute missing modalities. Recent advances achieved by these methods include the ability to handle incomplete data as well as going beyond the traditional molecular omics data types to integrate other modalities such as imaging data.

CONCLUSION

We expect to see further growth in methods that can handle missingness, as this is a common challenge in working with complex and heterogeneous data. Additionally, methods that integrate more data types are expected to improve performance on downstream tasks by capturing a comprehensive view of each sample.

DeepFusionCDR: Employing Multi-Omics Integration and Molecule-Specific Transformers for Enhanced Prediction of Cancer Drug Responses

Deep learning approaches have demonstrated remarkable potential in predicting cancer drug responses (CDRs), using cell line and drug features. However, existing methods predominantly rely on single-omics data of cell lines, potentially overlooking the complex biological mechanisms governing cell line responses. This paper introduces DeepFusionCDR, a novel approach employing unsupervised contrastive learning to amalgamate multi-omics features, including mutation, transcriptome, methylome, and copy number variation data, from cell lines. Furthermore, we incorporate molecular SMILES-specific transformers to derive drug features from their chemical structures. The unified multi-omics and drug signatures are combined, and a multi-layer perceptron (MLP) is applied to predict IC50 values for cell line-drug pairs. Moreover, this MLP can discern whether a cell line is resistant or sensitive to a particular drug. We assessed DeepFusionCDR's performance on the GDSC dataset and juxtaposed it against cutting-edge methods, demonstrating its superior performance in regression and classification tasks. We also conducted ablation studies and case analyses to exhibit the effectiveness and versatility of our proposed approach. Our results underscore the potential of DeepFusionCDR to enhance CDR predictions by harnessing the power of multi-omics fusion and molecular-specific transformers. The prediction of DeepFusionCDR on TCGA patient data and case study highlight the practical application scenarios of DeepFusionCDR in real-world environments.

Biological roles of THRAP3, STMN1 and GNA13 in human blood cancer cells

Blood cancers, such as diffuse large B-cell lymphoma (DLBCL), Burkitt's lymphoma (BL) and acute myeloid leukemia (AML), are aggressive neoplasms that are characterized by undesired clinical courses with dismal survival rates. The objective of the current work is to study the expression THRAP3, STMN1 and GNA13 in DLBCL, BL and AML, and to investigate if these proteins are implicated in the prognosis and progression of the blood cancers. Isolation of normal blood cells was performed using lymphoprep coupled with gradient centrifugation and magnetic beads. Flow-cytometric analysis showed high quality of the isolated cells. Western blotting identified THRAP3, STMN1 and GNA13 to be overexpressed in the blood cancer cells but hardly detected in normal blood cells from healthy donors. Consistently, investigations performed using genotype-tissue expression (GTEx) and gene expression profiling interactive analysis (GEPIA) showed that the three proteins had higher mRNA expression in various cancers compared with matched normal tissues (p ≤ 0.01). Furthermore, the up-regulated transcript expression of these proteins was a feature of short overall survival (OS; p ≤ 0.02) in patients with the blood cancers. Interestingly, functional profiling using gProfiler and protein-protein interaction network analysis using STRING with cytoscape reported THRAP3 to be associated with cancer-dependent proliferation and survival pathways (corrected p ≤ 0.05) and to interact with proteins (p = 1 × 10-16) implicated in tumourigenesis and chemotherapy resistance. Taken together, these findings indicated a possible implication of THRAP3, STMN1 and GNA13 in the progression and prognosis of the blood cancers. Additional work using clinical samples of the blood cancers is required to further investigate and validate the results reported here.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13205-024-04093-5.

PPIC-labeled CAFs: Key players in neoadjuvant chemotherapy resistance for gastric cancer

BACKGROUND

Gastric cancer (GC) is the fourth leading cause of cancer deaths, with advanced cases having a median survival of less than one year. Neoadjuvant chemotherapy (NCT) is vital but faces drug resistance issues, partly due to cancer-associated fibroblasts (CAFs). Yet, specific CAF subpopulations contributing to resistance are poorly understood.

METHODS

Differentially expressed genes (DEGs) between chemosensitive and resistant GC patients were identified using GEO2R. Single-cell sequencing (scRNA-seq) identified CAF-related genes. Immunohistochemistry verified key genes in NCT-treated GC samples, analyzing their correlation with tumor regression grade (TRG) and clinicopathological characteristics.

RESULTS

PPIC as a gene highly expressed in CAFs was closely associated with NCT resistance in gastric cancer. Immunohistochemistry results revealed positivity for the expression of cyclophilin C (CypC), encoded by PPIC, in the 5-fluorouracil and cisplatin NCT resistant and -sensitive groups of gastric cancer patients at rates of 69.7 % (76/109) and 43.6 % (24/55), respectively (p < 0.001). The high expression of CypC in CAFs was positively correlated to tumor size (p = 0.025), T stage (p = 0.004), TNM stage (p = 0.004), and vascular invasion (p = 0.027). In cancer cells the expression of CypC was associated with OS (p = 0.026). However, in CAFs, CypC expression was not related to OS (p = 0.671).

CONCLUSIONS

PPIC-labeled CAF subgroups are related to NCT resistance and poor prognosis in GC and they may cause drug resistance through signaling pathways such as glucose metabolism and extracellular matrix remodeling. However, the exact mechanism behind the involvement of PPIC-labeled CAF in drug resistance of GC requires further study.

Interferon regulatory factor 4 modulates epigenetic silencing and cancer-critical pathways in melanoma cells

Interferon regulatory factor 4 (IRF4) was initially identified as a key controller in lymphocyte differentiation and function, and subsequently as a dependency factor and therapy target in lymphocyte-derived cancers. In melanocytes, IRF4 takes part in pigmentation. Although genetic studies have implicated IRF4 in melanoma, how IRF4 functions in melanoma cells has remained largely elusive. Here, we confirmed prevalent IRF4 expression in melanoma and showed that high expression is linked to dependency in cells and mortality in patients. Analysis of genes activated by IRF4 uncovered, as a novel target category, epigenetic silencing factors involved in DNA methylation (DNMT1, DNMT3B, UHRF1) and histone H3K27 methylation (EZH2). Consequently, we show that IRF4 controls the expression of tumour suppressor genes known to be silenced by these epigenetic modifications, for instance cyclin-dependent kinase inhibitors CDKN1A and CDKN1B, the PI3-AKT pathway regulator PTEN, and primary cilium components. Furthermore, IRF4 modulates activity of key downstream oncogenic pathways, such as WNT/β-catenin and AKT, impacting cell proliferation and survival. Accordingly, IRF4 modifies the effectiveness of pertinent epigenetic drugs on melanoma cells, a finding that encourages further studies towards therapeutic targeting of IRF4 in melanoma.

ASM variants in the spotlight: A structure-based atlas for unraveling pathogenic mechanisms in lysosomal acid sphingomyelinase

Lysosomal acid sphingomyelinase (ASM), a critical enzyme in lipid metabolism encoded by the SMPD1 gene, plays a crucial role in sphingomyelin hydrolysis in lysosomes. ASM deficiency leads to acid sphingomyelinase deficiency, a rare genetic disorder with diverse clinical manifestations, and the protein can be found mutated in other diseases. We employed a structure-based framework to comprehensively understand the functional implications of ASM variants, integrating pathogenicity predictions with molecular insights derived from a molecular dynamics simulation in a lysosomal membrane environment. Our analysis, encompassing over 400 variants, establishes a structural atlas of missense variants of lysosomal ASM, associating mechanistic indicators with pathogenic potential. Our study highlights variants that influence structural stability or exert local and long-range effects at functional sites. To validate our predictions, we compared them to available experimental data on residual catalytic activity in 135 ASM variants. Notably, our findings also suggest applications of the resulting data for identifying cases suited for enzyme replacement therapy. This comprehensive approach enhances the understanding of ASM variants and provides valuable insights for potential therapeutic interventions.

Updates in Molecular Profiling of Pancreatic Ductal Adenocarcinoma

Outcomes from pancreatic ductal adenocarcinoma (PDAC) remain poor and better methods of prognostication and therapeutic approaches are needed. Recent advances in cancer genomics have led to the development of molecular subtypes of PDAC associated with clinical outcomes. Current evidence also suggests that the subtypes have differential response to first-line chemotherapy regimens. PDAC is also characterized by different stroma and immune environments. Further work is needed to confirm the utility of these subtypes to predicting response to different systemic therapies.

Gene expression-phenotype association study reveals the dual role of TNF-α/TNFR1 signaling axis in confined breast cancer cell migration

AIMS

While enhanced tumor cell migration is a key process in the tumor dissemination, mechanistic insights into causal relationships between tumor cells and mechanical confinement are still limited. Here we combine the use of microfluidic platforms to characterize confined cell migration with genomic tools to systematically unravel the global signaling landscape associated with the migratory phenotype of breast cancer (BC) cells.

METERIALS AND METHODS

The spontaneous migration capacity of seven BC cell lines was evaluated in 3D microfluidic devices and their migration capacity was correlated with publicly available molecular signatures. The role of identified signaling pathways on regulating BC migration capacity was determined by receptor stimulation through ligand binding or inhibition through siRNA silencing. Downstream effects on cell migration were evaluated in microfluidic devices, while the molecular changes were monitored by RT-qPCR.

KEY FINDINGS

Expression of 715 genes was correlated with BC cells migratory phenotype, revealing TNF-α as one of the top upstream regulators. Signal transduction experiments revealed that TNF-α stimulates the confined migration of triple negative, mesenchymal-like BC cells that are also characterized by high TNFR1 expression, but inhibits the migration of epithelial-like cells with low TNFR1 expression. TNFR1 was strongly associated with the migration capacity and triple-negative, mesenchymal phenotype. Downstream of TNF/TNFR1 signaling, transcriptional regulation of NFKB seems to be important in driving cell migration in confined spaces.

SIGNIFICANCE

TNF-α/TNFR1 signaling axis reveals as a key player in driving BC cells confined migration, emerging as a promising therapeutic strategy in targeting dissemination and metastasis of triple negative, mesenchymal BC cells.

Defining precancer: a grand challenge for the cancer community

The term 'precancer' typically refers to an early stage of neoplastic development that is distinguishable from normal tissue owing to molecular and phenotypic alterations, resulting in abnormal cells that are at least partially self-sustaining and function outside of normal cellular cues that constrain cell proliferation and survival. Although such cells are often histologically distinct from both the corresponding normal and invasive cancer cells of the same tissue origin, defining precancer remains a challenge for both the research and clinical communities. Once sufficient molecular and phenotypic changes have occurred in the precancer, the tissue is identified as a 'cancer' by a histopathologist. While even diagnosing cancer can at times be challenging, the determination of invasive cancer is generally less ambiguous and suggests a high likelihood of and potential for metastatic disease. The 'hallmarks of cancer' set out the fundamental organizing principles of malignant transformation but exactly how many of these hallmarks and in what configuration they define precancer has not been clearly and consistently determined. In this Expert Recommendation, we provide a starting point for a conceptual framework for defining precancer, which is based on molecular, pathological, clinical and epidemiological criteria, with the goal of advancing our understanding of the initial changes that occur and opportunities to intervene at the earliest possible time point.

A live single-cell reporter system reveals drug-induced plasticity of a cancer stem cell-like population in cholangiocarcinoma

Cancer stem cells (CSC) play an important role in carcinogenesis and are acknowledged to be responsible for chemoresistance in cholangiocarcinoma (CCA). Studying CCA CSC has been challenging, due to lack of consensus CSC markers, and to their plastic nature. Since dual expression of the core pluripotent factors SOX2/OCT4 has been shown to correlate with poor outcome in CCA patients, we selected the SOX2/OCT4 activating short half-life GFP-based live reporter (SORE6-dsCopGFP) to study CSC dynamics at the single-cell level. Transduction of five human CCA cell lines resulted in the expression of 1.8-13.1% GFP-positive (SORE6POS) cells. By live imaging, we found that SORE6POS CCA cells possess self-renewal capacity and that they can be induced to differentiate. Significantly, the SORE6POS cells were highly tumorigenic, both in vitro and in vivo, thus implicating the characteristics of primary CSCs. When we then analyzed for selected CSC-related markers, we found that the majority of both CD133+/CD44+, and CD133+/LGR5+ CCA cells were SORE6POS cells. Exposing transduced cells to standard CCA chemotherapy revealed higher growth rate inhibition at 50% (GR50s) for SORE6POS cells compared to GFP-negative (SORE6NEG) ones indicating that these CSC-like cells were more resistant to the treatment. Moreover, the chemotherapy induced SORE6POS from SORE6NEG cells, while retaining the existing SORE6POS population. Finally, treatment of transduced cells with CDK4/6 inhibitors in vitro for 3 days resulted in a lowered CSC number in the culture. Thus, applying a live reporter system allowed us to elucidate the stem cell diversity and drug-induced plasticity of CCA CSCs. These findings have clear implications for future management of such patients.

Facilitating integrative and personalized oncology omics analysis with UCSCXenaShiny

The continuous generation of multi-omics and phenotype data is propelling advancements in precision oncology. UCSCXenaShiny was developed as an interactive tool for exploring thousands of cancer datasets available on UCSC Xena. However, its capacity for comprehensive and personalized pan-cancer data analysis is being challenged by the growing demands. Here, we introduce UCSCXenaShiny v2, a milestone update through a variety of improvements. Firstly, by integrating multidimensional data and implementing adaptable sample settings, we create a suite of robust TPC (TCGA, PCAWG, CCLE) analysis pipelines. These pipelines empower users to conduct in-depth analyses of correlation, comparison, and survival in three modes: Individual, Pan-cancer and Batch screen. Additionally, the tool includes download interfaces that enable users to access diverse data and outcomes, several features also facilitate the joint analysis of drug sensitivity and multi-omics of cancer cell lines. UCSCXenaShiny v2 is an open-source R package and a web application, freely accessible at https://github.com/openbiox/UCSCXenaShiny .

Scanning sample-specific miRNA regulation from bulk and single-cell RNA-sequencing data

BACKGROUND

RNA-sequencing technology provides an effective tool for understanding miRNA regulation in complex human diseases, including cancers. A large number of computational methods have been developed to make use of bulk and single-cell RNA-sequencing data to identify miRNA regulations at the resolution of multiple samples (i.e. group of cells or tissues). However, due to the heterogeneity of individual samples, there is a strong need to infer miRNA regulation specific to individual samples to uncover miRNA regulation at the single-sample resolution level.

RESULTS

Here, we develop a framework, Scan, for scanning sample-specific miRNA regulation. Since a single network inference method or strategy cannot perform well for all types of new data, Scan incorporates 27 network inference methods and two strategies to infer tissue-specific or cell-specific miRNA regulation from bulk or single-cell RNA-sequencing data. Results on bulk and single-cell RNA-sequencing data demonstrate the effectiveness of Scan in inferring sample-specific miRNA regulation. Moreover, we have found that incorporating the prior information of miRNA targets can generally improve the accuracy of miRNA target prediction. In addition, Scan can contribute to construct cell/tissue correlation networks and recover aggregate miRNA regulatory networks. Finally, the comparison results have shown that the performance of network inference methods is likely to be data-specific, and selecting optimal network inference methods is required for more accurate prediction of miRNA targets.

CONCLUSIONS

Scan provides a useful method to help infer sample-specific miRNA regulation for new data, benchmark new network inference methods and deepen the understanding of miRNA regulation at the resolution of individual samples.

ENO2 in progression and treatment of colon adenocarcinoma: integrative bioinformatics analysis on non-apoptotic cell death

Colon adenocarcinoma (COAD) is one of the most common types of cancer. The interconnection between non-apoptotic cell death and COAD has not been adequately addressed. In our study, an integrative bioinformatics analysis was performed to explore non-apoptotic cell death-related biomarkers in COAD. ENO2 was determined as a potent biomarker for prognosis, drug response, immunity, and immunotherapy prediction. We used EdU and RT-qPCR assays to test our hypothesis and investigate how the ENO2 gene may influence or regulate cancer-related processes. ENO2 was expected to be a potential target in COAD.

Age- and ethnic-driven molecular and clinical disparity of East Asian breast cancers

BACKGROUND

Breast cancer (BC) is a complex disease with profound genomic aberrations. However, the underlying molecular disparity influenced by age and ethnicity remains elusive.

METHODS

In this study, we aimed to investigate the molecular properties of 843 primary and metastatic BC patients enrolled in the K-MASTER program. By categorizing patients into two distinct age subgroups, we explored their unique molecular properties. Additionally, we leveraged large-scale genomic data from the TCGA and MSK-IMPACT studies to examine the ethnic-driven molecular and clinical disparities.

RESULTS

We observed a high prevalence of PI3KCA mutations in K-MASTER HER2 + tumors, particularly in older patients. Moreover, we identified increased mutation rates in DNA damage response molecules, including ARID1A, MSH6, and MLH1. The K-MASTER patients were mainly comprised of triple-negative breast cancer (TNBC) and HER2-positive tumors, while the TCGA and MSK-IMPACT cohorts exhibited a predominance of hormone receptor-positive (HR +) subtype tumors. Importantly, GATA3 mutations were less frequently observed in East Asian patients, which correlated with poor clinical outcomes. In addition to characterizing the molecular disparities, we developed a gradient-boosting multivariable model to identify a new molecular signature that could predict the therapeutic response to platinum-based chemotherapy.

CONCLUSIONS

Our findings collectively provide unprecedented insights into the significance of age and ethnicity on the molecular and clinical characteristics of BC patients.

Potential crosstalk between Naïve CD4+ T cells and SPP1+ Macrophages is associated with clinical outcome and therapeutic response in hepatocellular carcinoma

BACKGROUND

The highly heterogeneity of the tumor microenvironment (TME) in hepatocellular carcinoma (HCC) results in diverse clinical outcomes and therapeutic responses. This study aimed to investigate potential intercellular crosstalk and its impact on clinical outcomes and therapeutic responses.

METHODS

Single-cell RNA sequencing (scRNA-seq), spatial transcriptomics (ST) and bulk RNA sequencing (RNA-seq) datasets were integrated to comprehensively analyze the intercellular interactions within the TME. Multiplex immunohistochemistry was conducted to validate the intercellular interactions. A machine learning-based integrative procedure was used in bulk RNA-seq datasets to generate a risk model to predict prognosis and therapeutic responses.

RESULTS

Survival analyses based on the bulk RNA-seq datasets revealed the negative impact of the naïve Cluster of Differentiation 4+ (CD4) T cells and Secreted Phosphoprotein 1+ (SPP1) macrophages on prognosis. Furthermore, their intricate intercellular crosstalk and spatial colocalization were also observed by scRNA-seq and ST analyses. Based on this crosstalk, a machine learning model, termed the naïve CD4+ T cell and SPP1+ macrophage prognostic score (TMPS), was established in the bulk-RNA seq datasets for prognostic prediction. The TMPS achieved C-index values of 0.785, 0.715, 0.692 and 0.857, respectively, across 4 independent cohorts. A low TMPS was associated with a significantly increased survival rates, improved response to immunotherapy and reduced infiltration of immunosuppressive cells, such as. regulatory T cells. Finally, 8 potential sensitive drugs and 6 potential targets were predicted for patients based on their TMPS.

CONCLUSION

The crosstalk between naïve CD4+ T cells and SPP1+ macrophages play a crucial role in the TME. TMPS can reflect this crosstalk and serve as a valuable tool for prognostic stratification and guiding clinical decision-making.

A Graph-Informed Modeling Framework Empowering Gene Pathway Discovery

This study introduces a novel graph-informed modeling framework for improving the statistical analysis of gene expression data, particularly in the context of identifying differentially expressed gene pathways and gene expression-assisted disease classification in a high-dimensional data setting. By integrating gene regulatory network information into hypothesis testing for the difference between mean vectors and linear discriminant analysis, we aim to effectively capture and utilize previously validated external gene interaction information. Our method leverages a block-coordinate descent approach which enables us to incorporate mixed graph information into linear structural equation modeling, accommodating directed/undirected edges and potential cycles in gene regulatory networks. Extensive simulations under various data scenarios have demonstrated the effectiveness of our approach with improved power for gene pathway tests and disease classification over existing methods. An application to a lung cancer dataset from the Cancer Genome Atlas Program (TCGA) further exemplifies the potential of our graph-informed approach in empowering the detection of differentially expressed gene pathways and gene expression-based classification of different lung cancer stages. Our findings underscore the potential utility of incorporating gene regulatory network information in gene pathway analysis, setting the stage for future advancements in gene pathway discovery, disease diagnosis, and treatment strategies.

Characterisation of APOBEC3B-Mediated RNA editing in breast cancer cells reveals regulatory roles of NEAT1 and MALAT1 lncRNAs

RNA editing is a crucial post-transcriptional process that influences gene expression and increases the diversity of the proteome as a result of amino acid substitution. Recently, the APOBEC3 family has emerged as a significant player in this mechanism, with APOBEC3A (A3A) having prominent roles in base editing during immune and stress responses. APOBEC3B (A3B), another family member, has gained attention for its potential role in generating genomic DNA mutations in breast cancer. In this study, we coupled an inducible expression cell model with a novel methodology for identifying differential variants in RNA (DVRs) to map A3B-mediated RNA editing sites in a breast cancer cell model. Our findings indicate that A3B engages in selective RNA editing including targeting NEAT1 and MALAT1 long non-coding RNAs that are often highly expressed in tumour cells. Notably, the binding of these RNAs sequesters A3B and suppresses global A3B activity against RNA and DNA. Release of A3B from NEAT1/MALAT1 resulted in increased A3B activity at the expense of A3A activity suggesting a regulatory feedback loop between the two family members. This research substantially advances our understanding of A3B's role in RNA editing, its mechanistic underpinnings, and its potential relevance in the pathogenesis of breast cancer.

Clinician’s Guide to Epitranscriptomics: An Example of N1-Methyladenosine (m1A) RNA Modification and Cancer

Epitranscriptomics is the study of modifications of RNA molecules by small molecular residues, such as the methyl (-CH3) group. These modifications are inheritable and reversible. A specific group of enzymes called “writers” introduces the change to the RNA; “erasers” delete it, while “readers” stimulate a downstream effect. Epitranscriptomic changes are present in every type of organism from single-celled ones to plants and animals and are a key to normal development as well as pathologic processes. Oncology is a fast-paced field, where a better understanding of tumor biology and (epi)genetics is necessary to provide new therapeutic targets and better clinical outcomes. Recently, changes to the epitranscriptome have been shown to be drivers of tumorigenesis, biomarkers, and means of predicting outcomes, as well as potential therapeutic targets. In this review, we aimed to give a concise overview of epitranscriptomics in the context of neoplastic disease with a focus on N1-methyladenosine (m1A) modification, in layman’s terms, to bring closer this omics to clinicians and their future clinical practice.

PPARγ antagonism as a new tool for preventing or overcoming endocrine resistance in luminal A breast cancers

PURPOSE

This research investigates the role of PPARγ in the complex molecular events underlying the acquisition of resistance to tamoxifen (Tam) in luminal A breast cancer (BC) cells. Furthermore, it focuses on evaluating the possibility of repurposing Imatinib mesylate, an FDA-approved anticancer agent recently recognized also as a PPARγ antagonist, for the personalized therapy of endocrine-resistant BC with increased PPARγ expression.

METHODS

Differential gene expression between parental and Tam-resistant MCF7 cells was assessed by RNA-seq followed by bioinformatics analysis and validation by RT-qPCR. PPARγ was downregulated by esiRNAs or inhibited by the antagonist GW9662. Cell viability and proliferation were measured by MTT and colony formation assays. Spheroids were prepared from parental and Tam-resistant MCF7 cells. Other luminal A BC cell lines resistant to Tam were generated.

RESULTS

In MCF7-TamR cells, PPARγ and several of its target genes were significantly upregulated. Increased PPARγ expression was due to the modulation of its positive/negative transcriptional regulators. Downregulating PPARγ with esiRNAs or GW9662 effectively killed parental and Tam-resistant cells and spheroids. Imatinib revealed to be as effective as GW9662 in restoring Tam susceptibility of these cells. PPARγ overexpression was also observed in the newly-selected Tam-resistant luminal A BC cells, in which GW9662 and Imatinib restored their susceptibility to Tam.

CONCLUSION

Our findings demonstrate that the overexpression of PPARγ is a frequent occurrence during acquisition of Tam resistance in luminal A BC cells, and that PPARγ antagonism represents an alternative therapeutic approach for the personalized treatment of BC showing dysregulation of this nuclear receptor.

A Machine Learning-Based Investigation of Integrin Expression Patterns in Cancer and Metastasis

Background

Integrins, a family of transmembrane receptor proteins, play complex roles in cancer development and metastasis. These roles could be better delineated through machine learning of transcriptomic data to reveal relationships between integrin expression patterns and cancer.

Methods

We collected publicly available RNA-Seq integrin expression from 8 healthy tissues and their corresponding tumors, along with data from metastatic breast cancer. We then used machine learning methods, including t-SNE visualization and Random Forest classification, to investigate changes in integrin expression patterns.

Results

Integrin expression varied across tissues and cancers, and between healthy and cancer samples from the same tissue, enabling the creation of models that classify samples by tissue or disease status. The integrins whose expression was important to these classifiers were identified. For example, ITGA7 was key to classification of breast samples by disease status. Analysis in breast tissue revealed that cancer rewires co-expression for most integrins, but the co-expression relationships of some integrins remain unchanged in healthy and cancer samples. Integrin expression in primary breast tumors differed from their metastases, with liver metastasis notably having reduced expression.

Conclusions

Integrin expression patterns vary widely across tissues and are greatly impacted by cancer. Machine learning of these patterns can effectively distinguish samples by tissue or disease status.

Spatial multi-omics reveal intratumoral humoral immunity niches associated with tertiary lymphoid structures in pancreatic cancer immunotherapy pathologic responders

Pancreatic adenocarcinoma (PDAC) is a rapidly progressing cancer that responds poorly to immunotherapies. Intratumoral tertiary lymphoid structures (TLS) have been associated with rare long-term PDAC survivors, but the role of TLS in PDAC and their spatial relationships within the context of the broader tumor microenvironment remain unknown. We generated a spatial multi-omics atlas encompassing 26 PDAC tumors from patients treated with combination immunotherapies. Using machine learning-enabled H&E image classification models and unsupervised gene expression matrix factorization methods for spatial transcriptomics, we characterized cellular states within TLS niches spanning across distinct morphologies and immunotherapies. Unsupervised learning generated a TLS-specific spatial gene expression signature that significantly associates with improved survival in PDAC patients. These analyses demonstrate TLS-associated intratumoral B cell maturation in pathological responders, confirmed with spatial proteomics and BCR profiling. Our study also identifies spatial features of pathologic immune responses, revealing TLS maturation colocalizing with IgG/IgA distribution and extracellular matrix remodeling.

GRAPHICAL ABSTRACT

HIGHLIGHTS

Integrated multi-modal spatial profiling of human PDAC tumors from neoadjuvant immunotherapy clinical trials reveal diverse spatial niches enriched in TLS.TLS maturity is influenced by tumor location and the cellular neighborhoods in which TLS immune cells are recruited.Unsupervised machine learning of genome-wide signatures on spatial transcriptomics data characterizes the TLS-enriched TME and associates TLS transcriptomes with survival outcomes in PDAC.Interactions of spatially variable gene expression patterns showed TLS maturation is coupled with immunoglobulin distribution and ECM remodeling in pathologic responders.Intratumoral plasma cell and immunoglobin gene expression spatial dynamics demonstrate trafficking of TLS-driven humoral immunity in the PDAC TME.

Significance

We report a spatial multi-omics atlas of PDAC tumors from a series of immunotherapy neoadjuvant clinical trials. Intratumorally, pathologic responders exhibit mature TLS that propagate plasma cells into malignant niches. Our findings offer insights on the role of TLS-associated humoral immunity and stromal remodeling during immunotherapy treatment.

Global RNA Interaction and Transcriptome Profiles Demonstrate the Potential Anti-Oncogenic Targets and Pathways of RBM6 in HeLa Cells

BACKGROUND

The fate and functions of RNAs are coordinately regulated by RNA-binding proteins (RBPs), which are often dysregulated in various cancers. Known as a splicing regulator, RNA-binding motif protein 6 (RBM6) harbors tumor-suppressor activity in many cancers; however, there is a lack of research on the molecular targets and regulatory mechanisms of RBM6.

METHODS

In this study, we constructed an RBM6 knock-down (shRBM6) model in the HeLa cell line to investigate its functions and molecular targets. Then we applied improved RNA immunoprecipitation coupled with sequencing (iRIP-seq) and whole transcriptome sequencing approaches to investigate the potential role and RNA targets of RBM6.

RESULTS

Using The Cancer Genome Atlas dataset, we found that higher expression of RBM6 is associated with a better prognosis in many cancer types. In addition, we found that RBM6 knockdown promoted cell proliferation and inhibited apoptosis, demonstrating that RBM6 may act as an anti-oncogenic protein in cancer cells. RBM6 can regulate the alternative splicing (AS) of genes involved in DNA damage response, proliferation, and apoptosis-associated pathways. Meanwhile, RBM6 knockdown activated type I interferon signaling pathways and inhibited the expression of genes involved in the cell cycle, cellular responses to DNA damage, and DNA repair pathways. The differentially expressed genes (DEGs) by shRBM6 and their involved pathways were likely regulated by the transcription factors undergoing aberrant AS by RBM6 knockdown. For iRIP-seq analysis, we found that RBM6 could interact with a large number of mRNAs, with a tendency for binding motifs GGCGAUG and CUCU. RBM6 bound to the mRNA of cell proliferation- and apoptosis-associated genes with dysregulated AS after RBM6 knockdown.

CONCLUSIONS

In summary, our study highlights the important role of RBM6, as well as the downstream targets and regulated pathways, suggesting the potential regulatory mechanisms of RBM6 in the development of cancer.

Mitochondria-targeted cancer analysis using survival and expression: Prioritizing mitochondrial targets that alleviate pancreatic cancer cell phenotypes

Substantial changes in energy metabolism are a hallmark of pancreatic cancer. To adapt to hypoxic and nutrient-deprived microenvironments, pancreatic cancer cells remodel their bioenergetics from oxidative phosphorylation to glycolysis. This bioenergetic shift makes mitochondria an Achilles' heel. Since mitochondrial function remains essential for pancreatic cancer cells, further depleting mitochondrial energy production is an appealing treatment target. However, identifying effective mitochondrial targets for treatment is challenging. Here, we developed an approach, mitochondria-targeted cancer analysis using survival and expression (mCAUSE), to prioritize target proteins from the entire mitochondrial proteome. Selected proteins were further tested for their impact on pancreatic cancer cell phenotypes. We discovered that targeting a dynamin-related GTPase, OPA1, which controls mitochondrial fusion and cristae, effectively suppresses pancreatic cancer activities. Remarkably, when combined with a mutation-specific KRAS inhibitor, OPA1 inhibition showed a synergistic effect. Our findings offer a therapeutic strategy against pancreatic cancer by simultaneously targeting mitochondria dynamics and KRAS signaling.

Multivariate analysis of metabolic state vulnerabilities across diverse cancer contexts reveals synthetically lethal associations

Targeting the distinct metabolic needs of tumor cells has recently emerged as a promising strategy for cancer therapy. The heterogeneous, context-dependent nature of cancer cell metabolism, however, poses challenges to identifying effective therapeutic interventions. Here, we utilize various unsupervised and supervised multivariate modeling approaches to systematically pinpoint recurrent metabolic states within hundreds of cancer cell lines, elucidate their association with tumor lineage and growth environments, and uncover vulnerabilities linked to their metabolic states across diverse genetic and tissue contexts. We validate key findings via analysis of data from patient-derived tumors and pharmacological screens and by performing genetic and pharmacological experiments. Our analysis uncovers synthetically lethal associations between the tumor metabolic state (e.g., oxidative phosphorylation), driver mutations (e.g., loss of tumor suppressor PTEN), and actionable biological targets (e.g., mitochondrial electron transport chain). Investigating the mechanisms underlying these relationships can inform the development of more precise and context-specific, metabolism-targeted cancer therapies.

Dynamic changes of the immune microenvironment in the development of gastric cancer caused by inflammation

Precancerous lesions typically precede gastric cancer (GC), but the molecular mechanisms underlying the transition from these lesions to GC remain unclear. Therefore, it is urgent to understand this transition from precancerous lesions to GC, which is crucial for the early diagnosis and treatment of GC. In this study, we merged multiple single-cell RNA sequencing datasets to investigate the molecular changes in distinct cell types associated with the progression of GC. First, we observed an increasing abundance of immune cells and a decrease in non-immune cells from non-atrophic gastritis to GC. Five immune cell types were significantly enriched in GC compared to precancerous lesions. Moreover, we found that the interleukin (IL)-17 signaling pathway and Th17 cell differentiation were significantly up-regulated in immune cell subsets during GC progression. Some genes in these processes were predominantly expressed at the GC stage, highlighting their potential as diagnostic markers. Furthermore, we validated our findings using bulk RNA sequencing data from The Cancer Genome Atlas and confirmed consistent immune cell changes during GC progression. Our study provides insights into the immune infiltration and signaling pathways involved in the development of GC, contributing to the development of early diagnosis and targeted treatment strategies for this malignancy.

Macrophage-mediated myelin recycling fuels brain cancer malignancy

Tumors growing in metabolically challenged environments, such as glioblastoma in the brain, are particularly reliant on crosstalk with their tumor microenvironment (TME) to satisfy their high energetic needs. To study the intricacies of this metabolic interplay, we interrogated the heterogeneity of the glioblastoma TME using single-cell and multi-omics analyses and identified metabolically rewired tumor-associated macrophage (TAM) subpopulations with pro-tumorigenic properties. These TAM subsets, termed lipid-laden macrophages (LLMs) to reflect their cholesterol accumulation, are epigenetically rewired, display immunosuppressive features, and are enriched in the aggressive mesenchymal glioblastoma subtype. Engulfment of cholesterol-rich myelin debris endows subsets of TAMs to acquire an LLM phenotype. Subsequently, LLMs directly transfer myelin-derived lipids to cancer cells in an LXR/Abca1-dependent manner, thereby fueling the heightened metabolic demands of mesenchymal glioblastoma. Our work provides an in-depth understanding of the immune-metabolic interplay during glioblastoma progression, thereby laying a framework to unveil targetable metabolic vulnerabilities in glioblastoma.

TYRP1 directed CAR T cells control tumor progression in preclinical melanoma models

Despite therapeutic efficacy observed with immune checkpoint blockade in advanced melanoma, many tumors do not respond to treatment, representing a need for new therapies. Here, we have generated chimeric antigen receptor (CAR) T cells targeting TYRP1, a melanoma differentiation antigen expressed on the surface of melanomas, including rare acral and uveal melanomas. TYRP1-targeted CAR T cells demonstrate antigen-specific activation and cytotoxic activity in vitro and in vivo against human melanomas independent of the MHC alleles and expression. In addition, the toxicity to pigmented normal tissues observed with T lymphocytes expressing TYRP1-targeted TCRs was not observed with TYRP1-targeted CAR T cells. Anti-TYRP1 CAR T cells provide a novel means to target advanced melanomas, serving as a platform for the development of similar novel therapeutic agents and as a tool to interrogate the immunobiology of melanomas.

GATDE: A graph attention network with diffusion-enhanced protein-protein interaction for cancer classification

Cancer classification is crucial for effective patient treatment, and recent years have seen various methods emerge based on protein expression levels. However, existing methods oversimplify by assuming uniform interaction strengths and neglecting intermediate influences among proteins. Addressing these limitations, GATDE employs a graph attention network enhanced with diffusion on protein-protein interactions. By constructing a weighted protein-protein interaction network, GATDE captures the diversity of these interactions and uses a diffusion process to assess multi-hop influences between proteins. This information is subsequently incorporated into the graph attention network, resulting in precise cancer classification. Experimental results on breast cancer and pan-cancer datasets demonstrate that GATDE surpasses current leading methods. Additionally, in-depth case studies further validate the effectiveness of the diffusion process and the attention mechanism, highlighting GATDE's robustness and potential for real-world applications.

Spatial transcriptomic analysis of primary and metastatic pancreatic cancers highlights tumor microenvironmental heterogeneity

Although the spatial, cellular and molecular landscapes of resected pancreatic ductal adenocarcinoma (PDAC) are well documented, the characteristics of its metastatic ecology remain elusive. By applying spatially resolved transcriptomics to matched primary and metastatic PDAC samples, we discovered a conserved continuum of fibrotic, metabolic and immunosuppressive spatial ecotypes across anatomical regions. We observed spatial tumor microenvironment heterogeneity spanning beyond that previously appreciated in PDAC. Through comparative analysis, we show that the spatial ecotypes exhibit distinct enrichment between primary and metastatic sites, implying adaptability to the local environment for survival and progression. The invasive border ecotype exhibits both pro-tumorigenic and anti-tumorigenic cell-type enrichment, suggesting a potential immunotherapy target. The ecotype heterogeneity across patients emphasizes the need to map individual patient landscapes to develop personalized treatment strategies. Collectively, our findings provide critical insights into metastatic PDAC biology and serve as a valuable resource for future therapeutic exploration and molecular investigations.

Deconvolution from bulk gene expression by leveraging sample-wise and gene-wise similarities and single-cell RNA-Seq data

BACKGROUND

The widely adopted bulk RNA-seq measures the gene expression average of cells, masking cell type heterogeneity, which confounds downstream analyses. Therefore, identifying the cellular composition and cell type-specific gene expression profiles (GEPs) facilitates the study of the underlying mechanisms of various biological processes. Although single-cell RNA-seq focuses on cell type heterogeneity in gene expression, it requires specialized and expensive resources and currently is not practical for a large number of samples or a routine clinical setting. Recently, computational deconvolution methodologies have been developed, while many of them only estimate cell type composition or cell type-specific GEPs by requiring the other as input. The development of more accurate deconvolution methods to infer cell type abundance and cell type-specific GEPs is still essential.

RESULTS

We propose a new deconvolution algorithm, DSSC, which infers cell type-specific gene expression and cell type proportions of heterogeneous samples simultaneously by leveraging gene-gene and sample-sample similarities in bulk expression and single-cell RNA-seq data. Through comparisons with the other existing methods, we demonstrate that DSSC is effective in inferring both cell type proportions and cell type-specific GEPs across simulated pseudo-bulk data (including intra-dataset and inter-dataset simulations) and experimental bulk data (including mixture data and real experimental data). DSSC shows robustness to the change of marker gene number and sample size and also has cost and time efficiencies.

CONCLUSIONS

DSSC provides a practical and promising alternative to the experimental techniques to characterize cellular composition and heterogeneity in the gene expression of heterogeneous samples.

HMGA1 stimulates cancer stem-like features and sensitivity to monensin in gastric cancer

Gastric cancer represents a serious health problem worldwide, with insufficient molecular biomarkers and therapeutic options. Consequently, several efforts have been directed towards finding specific disease markers in order to develop new therapies capable of defeating gastric cancer. Attention has been pointed to cancer stem cells (CSCs) as they are primarily responsible for tumor initiation and recurrence, making them essential therapeutic targets. Using the SORE6-GFP reporter system, based on the expression of SOX2 and/or OCT4 to drive GFP expression, we isolated gastric cancer stem-like cells (SORE6+ cells) enriched in several molecules, including SOX2, C-MYC, KLF4, HIF-1α, NOTCH1 and HMGA1. Here, we explored the previously undisclosed link of HMGA1 with gastric CSCs. Our results indicated that HMGA1 can activate a transcriptional program that includes SOX2, C-MYC, and KLF4 and endows cells with CSC features. We further showed that chemical induction of gastric CSCs using ciclopirox (CPX) can be mediated by HMGA1. Finally, we showed that HMGA1 GFP+ cells were sensitive to monensin confirming the selective activity of this drug over CSCs. Thus, HMGA1 is a key player in the cellular reprogramming of gastric non-CSCs to cancer stem-like cells.

A novel prognostic model based on pyroptosis signature in AML

Acute myeloid leukemia (AML), a highly heterogeneous myeloid malignancy, remains a challenge in terms of proper risk stratification. In this study, we developed a novel pyroptosis prognostic model based on pyroptosis-related gene pairs, which exhibited excellent prognostic performance across multiple cohorts (N = 1506) and accurately predicted both adult and pediatric AML prognosis. Additionally, we integrated the pyroptosis risk model with other clinical risk factors to construct a highly operational nomogram. Moreover, our findings indicate a significant correlation between elevated pyroptosis risk scores and increased stemness of AML. Using CIBERSORT immune analysis, we found a decreased proportion of resting NK cells and activated mast cells in the high-risk group. Through analyzing the correlation between chemotherapy drug response sensitivity and risk scores, we found that AZD1332 and BPD-0008900 were extremely sensitive in the high-risk group.

An integrated pan-cancer assessment of prognosis, immune infiltration, and immunotherapy response for B7 family using multi-omics data

AIMS

B7 molecules (B7s) are crucial synergistic signals for effective immune surveillance against tumor cells. While previous studies have explored the association between the B7 family and cancer, most have been limited to specific genes or cancer subtypes.

MAIN METHODS

Our study utilized multi-omics data to investigate potential correlations between B7s expression (B7s exp.) and prognosis, clinicopathological features, somatic mutations (SMs), copy number variations (CNVs), immune characteristics, tumor microenvironment (TME), microsatellite instability, tumor mutation burden, immune checkpoint gene (ICG), and drug responsiveness in TCGA tumors. Furthermore, the connection between B7s exp. and immunotherapy (IT) performance assessed in various validated datasets. Following this, immune infiltration analysis (IIA) was conducted based on B7s exp., CNVs, or SMs in bladder cancer (BLCA), complemented by real-time PCR (RT-PCR) and protein confirmation of B7-H3.

KEY FINDINGS

Across most cancer types, B7s exp. was related to prognosis, clinicopathological characteristics, mutations, CNVs, ICG, TMB, TME. The examination of sensitivity to anticancer drugs unveiled correlations between B7 molecules and different drug sensitivities. Specific B7s exp. patterns were linked to the clinical effectiveness of IT. Using GSEA, several enriched immune-related functions and pathways were identified. Particularly in BLCA, IIA revealed significant connections between B7 CNVs, mutation status, and various immune cell infiltrates. RT-PCR confirmed elevated B7-H3 gene levels in BLCA tumor tissues.

SIGNIFICANCE

This study confirmed the significance of B7s exp. and genomic changes in predicting outcomes and treatment across different cancer types. Moreover, they indicate a critical function of B7s in BLCA and their potential as IT biomarkers.