Stromal score is a promising index in tumor patients' outcome determination.

Background: Immune status is widely acknowledged as a valuable marker for predicting cancer prognosis and therapy response. However, there has been a limited understanding of the stromal landscape in cancer. Methods: By employing ESTIMATE, stromal- and immune-scores were inferred for 6193 tumor samples spanning 12 cancer types sourced from The Cancer Genome Atlas (TCGA). Subsequently, the samples were categorized into seven groups based on their stromal and immune scores. A comparison of prognosis, lymphocyte and stromal cell infiltration, and the response to programmed death ligand 1 (PD-L1) therapy was conducted among these subtypes. Results: It was unveiled by the analysis that, in the majority of cancer types, stromal score exhibited a more potent predictive capability for outcomes compared to the immune score. Furthermore, it was observed that in four cancer types, intermediate immune infiltration coupled with low stromal infiltration correlated with the most favorable overall survival, whereas an unfavorable outcome was predicted in colorectal cancer (CRC) and stomach adenocarcinoma (STAD) when high immune infiltration coexisted with intermediate or high stromal infiltration. Conclusion: In summary, while high immune scores frequently correlate with a positive prognosis, such correlation is not universal. A potential strategy to address the current limitations of the immune score in specific circumstances could involve a focus on stromal scores or a subtle integration of stromal and immune status.

miR-433 Inhibits Glioblastoma Progression by Suppressing the PI3K/Akt Signaling Pathway Through Direct Targeting of ERBB4.

Glioblastoma multiforme (GBM) is a highly malignant brain tumor where new biomarkers and drug targets are much needed in the oncology clinic. miR-433 was identified as a tumor-suppressing miRNA in several different types of human cancer. However, the integrative biology of miR-433 in GBM is still largely unknown. By analyzing the expression profiles of miR-433 in 198 patients with glioma at The Cancer Genome Atlas, we found that the miR-433 expression was decreased in glioma whereas the low expression of miR-433 was significantly associated with shorter overall survival. We then conducted in vitro studies and demonstrated that increased expression of miR-433 suppressed the proliferation, migration, and invasion of LN229 and T98G cells, two representative glioma cell lines. Further, using in vivo mouse model, we found that upregulation of miR-433 inhibited the tumor growth of glioma cells. To situate the integrative biology understanding of the action of miR-433 in glioma, we identified ERBB4 as a gene targeted directly by miR-433 in LN229 and T98G cells. Overexpressed ERBB4 rescued the phenotype caused by overexpression of miR-433. Finally, we showed that miR-433 suppressed the PI3K/Akt pathway in glioma cells. In conclusion, our study demonstrated that miR-433 could potentially act as a tumor suppressor for GBM and may serve as a potential therapeutic target for GBM. Further integrative biology and clinical translational research are warranted to evaluate miR-433 in GBM.

Hypoxia Confers Tumor with a Higher Immune Infiltration but Lower Mutation Burden in Gastrointestinal Cancer.

Background: Hypoxia is one of the driving forces of cancer progression, recurrence, and metastasis. However, the association between the tumor hypoxic tumor microenvironment and the tumor mutation burden (TMB) is poorly understood in gastrointestinal cancer. Methods: Approximately 2,000 samples from colorectal cancer (CRC) and stomach adenocarcinoma (STAD) patients were obtained from the gene expression omnibus database and the cancer genome Atlas databases and were clustered and subtyped by nonnegative matrix factorization. Significant differentially expressed genes that were possibly related to survival differences between the hypoxic and normoxic groups were subjected to multivariate Cox regression. Results: Gastrointestinal cancer patients with CRC and STAD were further divided into two subgroups, namely, the hypoxia group and the normoxia group, and hypoxia was correlated with unfavorable outcomes. Notably, hypoxic tumors had lower TMB but significantly higher levels of immune and stromal infiltration. A signature of HEYL and NRP1 selected by LASSO classified gastrointestinal cancer patients into either a low or high-risk group, allowing for the combination of TMB status with markers of hypoxia in future clinical applications. Conclusions: Hypoxia is an independent prognostic factor and a strong immune infiltration indicator in gastrointestinal tumors of different organs, especially for cancers with low TMB.

Association Mining Identifies MAL2 as a Novel Tumor Suppressor in Colorectal Cancer.

Introduction: Colorectal cancer (CRC) is a leading cause of cancer-related deaths worldwide. However, the driver genes that promote CRC metastasis remain poorly understood. Association mining mines and extracts the repeated correlations and relevance in a dataset to predict the appearance of other data items according to the appearance of one item. Methods: Here, the Apriori algorithm was used to find the frequent mutational gene sets (FMGSs) and hidden association rules (ARs) within these FMGSs from 383 CRCs with whole exome sequencing datasets. The weighted correlation network analysis (WGCNA) was used to identify the hub genes in CRC. CCK8, colony formation, cell migration and invasion assays were adopted to detect the roles of hub genes in CRC. Results: Intriguingly, we found that MAL2 (myelin and lymphocyte protein 2) was associated with TP53 and APC in stage IV of CRC, and further subnetwork exploration based on WGCNA identified MAL2 as a potent hub gene. To validate the metastasis-related role of MAL2 in CRC, a lentivirus-based overexpression system was utilized to construct MAL2-overexpressing human CRC LOVO cells. Overexpression of MAL2 remarkably inhibited CRC cell proliferation and invasion. Conclusion: Our results highlighted that MAL2 acts as a tumor suppressor in CRC and could serve as a potential therapeutic target.

A pretreatment transcriptomic signature that predicts outcomes of immunotherapy in melanoma.

Identifying indicators of immunotherapy response are key to clinical treatment decisions. To date, immunotherapy is most widely used in melanoma because of its higher tumor mutation burden compared to other cancer types. However, less than half of melanoma patients can benefit from immune checkpoint inhibitor (ICI) therapy. For this reason, we deciphered pretreatment transcriptomes across a cohort of melanoma patients receiving anti-PD-1 or CTLA-4 alone (sICI) or in combination (cICI). We developed a two-gene signature that could predict the curative effect of ICI in melanoma by using the LASSO method. The pre-ICI signature displayed an equally competitive predictive power as the post-ICI irRECIST assessment that could offer clues regarding long-term ICI therapy response and facilitate risk stratification and treatment strategies.

The Effects of Differentially-Expressed Homeobox Family Genes on the Prognosis and HOXC6 on Immune Microenvironment Orchestration in Colorectal Cancer.

Background: The homeobox (HOX) gene family encodes highly conserved transcription factors, that play important roles in the morphogenesis and embryonic development of vertebrates. Mammals have four similar HOX gene clusters, HOXA, HOXB, HOXC, and HOXD, which are located on chromosomes 7, 17,12 and 2 and consist of 38 genes. Some of these genes were found to be significantly related to a variety of tumors; however, it remains unknown whether abnormal expression of the HOX gene family affects prognosis and the tumor microenvironment (TME) reshaping in colorectal cancer (CRC). Therefore, we conducted this systematic exploration to provide additional information for the above questions. Methods: RNA sequencing data from The Cancer Genome Atlas (TCGA) and mRNA expression data from Gene Expression Omnibus (GEO) combined with online tumor analysis databases (UALCAN, TIMER, PrognoScan) were utilized to explore the relationship among abnormal expression of HOX family genes, prognosis and the tumor immune microenvironment in CRC. Results: 1. Differential expression and prognosis analysis: 24 genes were significantly differentially expressed in CRC compared to adjacent normal tissues, and seven upregulated genes were significantly associated with poor survival. Among these seven genes, univariate and multivariate Cox regression analysis revealed that only high expression of HOXC6 significantly contributed to poor prognosis; 2. The influence of overexpressed HOXC6 on the pathway and TME: High HOXC6 expression was significantly related to the cytokine pathway and expression of T cell attraction chemokines, the infiltration ratio of immune cells, expression of immune checkpoint markers, tumor mutation burden (TMB) scores and microsatellite instability-high (MSI-H) scores; 3. Stratified analysis based on stages: In stage IV, HOXC6 overexpression had no significant impact on TMB, MSI-H, infiltration ratio of immune cells and response prediction of immune checkpoint blockers (ICBs), which contributed to significantly poor overall survival (OS). Conclusion: Seven differentially expressed HOX family genes had significantly worse prognoses. Among them, overexpressed HOXC6 contributed the most to poor OS. High expression of HOXC6 was significantly associated with high immunogenicity in nonmetastatic CRC. Further research on HOXC6 is therefore worthwhile to provide potential alternatives in CRC immunotherapy.

Characterization and discrimination of human colorectal cancer cells using terahertz spectroscopy.

Terahertz technology has been widely used in biomedical research. Herein, terahertz time-domain attenuated total reflection (THz TD-ATR) spectroscopy was employed to characterize and discriminate human cancer cell lines (DLD-1 and HT-29). Terahertz responses of the cell lines were measured and Savitzky-Golay algorithm was applied to smooth the spectra of refractive index, absorption coefficient and dielectric loss tangent in terahertz regime. Principal component analysis (PCA) was then adopted for feature extraction and cell characterization. Based on the processed data, cancer cell lines were discriminated by applying random forests (RF) method to analyze three characteristic parameters separately and the results from them were compared. Results indicate that absorption coefficient was the most sensitive parameter for cancer cell discrimination. Our study suggests great potential for human cancer cell recognition and provides experimental basis for liquid biopsy.

HomeoboxC6 promotes metastasis by orchestrating the DKK1/Wnt/ß-catenin axis in right-sided colon cancer.

Patients with right-sided colon cancer (RCC) generally have a poorer prognosis than those with left-sided colon cancer (LCC). We previously found that homeobox C6 (HOXC6) was the most significantly upregulated gene in RCC compared to LCC. However, it remains unclear whether HOXC6 plays a role in tumor proliferation and metastasis. Our study aimed to explore the potential oncogenic role and the detailed molecular mechanism of HOXC6 in RCC. In this study, HOXC6 was validated to be overexpressed in RCC and associated with poor prognosis. Furthermore, overexpression of HOXC6 promoted the migration and invasion of colon cancer cells through inducing EMT by activating the Wnt/ß-catenin signaling pathway and inhibition of DKK1 secretion. Lastly, we preliminary explored the translational effect of HOXC6 and found that silencing of HOXC6 made HCT116 and HT29 cells more sensitive to irinotecan.

Hypermutated tumours across 11 cancer types show three distinct immune subtypes.

BACKGROUND: Complete remission is observed in less than half of hypermutated (HM) tumours after immune checkpoint blockade therapy, indicating that HM tumours are very heterogeneous. Thus, there is an urgent requirement to decipher the unknown intrinsic HM tumour subtypes. METHODS: Statistical analysis was performed on somatic mutation data from 5519 tumours across 11 cancer types obtained from The Cancer Genome Atlas and 338 colorectal cancer (CRC) samples obtained from an Asian cohort. Samples with a tumour mutation burden >10 mut/Mb were classified as HM. A total of 1040 HM samples harbouring corresponding transcriptomes were used for non-negative matrix factorisation clustering. Tumour mutational burden, neoantigens, T cell receptor (TCR) diversity, stromal score and immune score were compared between the subtypes. RESULTS: HM tumours fell into three distinct immune subtypes: HM1, HM2 and HM3. HM3 tumours were correlated with increased CD8 T cell infiltration, high TCR diversity, a high immune score and prolonged survival. HM2 tumours were correlated with an abundant stromal component, epithelial-mesenchymal transition, TGFß, angiogenesis hallmarks and poor outcomes. The infiltration of more CD8 T cells and increased chemokine expression in HM3 were validated in CRC by immunofluorescence. CONCLUSIONS: These findings will facilitate the development of a subtype-oriented therapy strategy to enhance the treatment effect in the near future.

Pan-cancer analyses demonstrate that ANKRD6 is associated with a poor prognosis and correlates with M2 macrophage infiltration in colon cancer.

OBJECTIVE: Ankyrin repeat domain-containing protein 6 (ANKRD6) is an ankyrin repeat-containing protein which is structurally related to vertebrate inversin and Drosophila Diego. However, the correlations between ANKRD6 and tumor-infiltrating immune cells in cancers is not clear. METHODS: ANKRD6 expression was analyzed by Oncomine, Tumor Immune Estimation Resource (TIMER) and Gene Expression Profiling Interactive Analysis (GEPIA). PrognoScan and GEPIA were used to evaluate the influence of ANKRD6 on clinical prognosis. TIMER and CIBERSORT were used to analyze correlations between ANKRD6 expression levels and tumor immune cell infiltrates. Immunohistochemical analysis of the relationship between ANKRD6 expression and overall survival, as well as the relationship between ANKRD6 expression and M2 macrophage infiltration, was performed. RESULTS: High level of ANKRD6 expression was associated with poor prognosis of colon cancer. ANKRD6 expression level was positively correlated with infiltrating levels of CD8+ T cells, CD4+ T cells, macrophages, neutrophils and dendritic cells in colon cancer by using TIMER. Using CIBERSORT, we found that in plasma cells, CD8+ T cells, CD4+ memory resting T cells, follicular helper T cells and activated natural killer cells were significantly lower in the ANKRD6-high group than in the ANKRD6-low group. M0 and M2 macrophages were significantly higher in the ANKRD6-high group than in the ANKRD6-low group. Immunohistochemistry confirmed that M2 macrophage infiltration in the ANKRD6-high group significantly increased. CONCLUSIONS: The high ANKRD6 expression is associated with poor prognosis of colon cancer. ANKRD6 expression is positively correlated with M2 macrophage infiltration in colon cancer.

Survival outcome and prognostic factors for colorectal cancer with synchronous bone metastasis: a population-based study.

Prognostic factors of synchronous bone metastatic colorectal cancer (CRC) are still undetermined. We aimed to investigate survival outcome and prognostic factors of patients with synchronous bone metastatic CRC. Information of patients with synchronous bone metastatic CRC were obtained from the Surveillance, Epidemiology, and End Results (SEER) and West China Hospital (WCH) databases. Cases from SEER database composed construction cohort, while cases from WCH database were used as validation cohort. A novel nomogram was constructed to predict individual survival probability based on Cox regression model. The performance of the nomogram was internally and externally validated using calibration curves and concordance index (C-index). Three hundred and eighty-one patients from SEER database were eligible. The median disease specific OS was 9.0 months (95% confidence interval [CI]: 7.3-10.7 months). Multivariate Cox analysis identified seven independent prognostic factors including histological type, differentiation grade, T stage of primary tumor, CEA level, systemic chemotherapy, combined with liver metastasis and combined with lung metastasis. A novel nomogram was established based on these variables. In the internal validation, the C-index (0.72, 95% CI 0.69-0.75) and calibration curve indicated well performance of this nomogram at predicting survival outcome in bone metastatic CRC. In the external validation, the C-index was 0.57 (95% CI 0.46-0.68). The prognosis of synchronous bone metastatic CRC is very poor. Histological type, differentiation grade, T stage of primary tumor, CEA level, systemic chemotherapy, combined with liver metastasis and combined with lung metastasis are independent prognostic factors. Further study is warranted to confirm the practicality of the prognostic nomogram.

Heterogeneity of MSI-H gastric cancer identifies a subtype with worse survival.

BACKGROUND: Microsatellite instability-high (MSI-H) tumour patients generally have a better prognosis than microsatellite-stable (MSS) ones due to the large number of non-synonymous mutations. However, an increasing number of studies have revealed that less than half of MSI-H patients gain survival benefits or symptom alleviation from immune checkpoint-blockade treatment. Thus, an in-depth inspection of heterogeneous MSI-H tumours is urgently required. METHODS: Here, we used non-negative matrix factorisation (non-NMF)-based consensus clustering to define stomach adenocarcinoma (STAD) MSI-H subtypes in samples from The Cancer Genome Atlas and an Asian cohort, GSE62254. RESULTS: MSI-H STAD samples are basically clustered into two subgroups (MSI-H1 and MSI-H2). Further examination of the immune landscape showed that immune suppression factors were enriched in the MSI-H1 subgroup, which may be associated with the poor prognosis in this subgroup. CONCLUSIONS: Our results illustrate the genetic heterogeneity within MSI-H STADs, with important implications for cancer patient risk stratification, prognosis and treatment.

Ultra-mutated colorectal cancer patients with POLE driver mutations exhibit distinct clinical patterns.

POLE mutations, which lead to an ultramutated phenotype in colorectal cancer (CRC), have been reported as a promising marker in immunotherapy. We performed sequencing of CRC cases in Zhejiang University (ZJU) and extracted obtainable data from recently published results, including The Cancer Genome Atlas (TCGA), Japanese studies and clinical trials, to present clinical patterns of POLE driver-mutated CRC and reveal its heterogeneity. The rate of somatic POLE driver mutations has been reported as 2.60% (ZJU cohort), 1.50% (TCGA cohort), 1.00% (Japan cohort), and 1.00% (Lancet cohort). POLE driver mutations show a clearly increased mutation burden (mean TMB: 217.98 mut/Mb in ZJU; 203.13 mut/Mb in TCGA). Based on pooled data, more than 70.00% of patients with POLE driver mutations were diagnosed before they were 55 years old and at an early disease stage (Stage 0-II >70.00%), and more than 70.00% were male. Among Asian patients, 68.40% developed POLE driver mutations in the left-side colon, whereas 64.00% of non-Asian patients developed them in the right-side colon (p < 0.01). The top three amino acid changes due to POLE driver mutations are P286R, V411L, and S459F. Investigators and physicians should ascertain the heterogeneity and clinical patterns of POLE driver mutations to be better equipped to design clinical trials and analyze the data.

Hypoxia Correlates With Poor Survival and M2 Macrophage Infiltration in Colorectal Cancer.

BACKGROUND: It is widely accepted that the oxygen level in tumor tissue is significantly lower than the adjacent normal tissue, thus termed hypoxia. Intratumoral hypoxia represents a major driving force in cancer progression, recurrence, metastasis, and decreased survival. Though multiple gene signatures reflect the complex cellular response to hypoxia have been established in several cancer types such as head and neck, breast, and lung cancers, the hypoxic panorama in colorectal cancer (CRC) remains poorly understood. METHODS: A hypoxic signature constituted by a total of 356 genes, including canonical hypoxia-responsive ADM, ANGPTL4, CA9, and VEGFA, was established based on systemic literature search. A total of 1,730 CRC samples across four independent cohorts were used for nonnegative matrix factorization clustering and subtyping. Prognosis, molecular signatures, pathways, and tumor-infiltrating lymphocytes were compared between the subtypes. RESULTS: CRCs mainly fell into two subgroups, one indicated as hypoxia and the other one designated as normoxia. Hypoxia was correlated with poor outcomes in CRC and will increase the risk of a subset of stage II patients to the level of normoxic stage III. Additionally, hypoxia was closely associated with activation of RAS signaling pathway independent of KRAS mutation. More M2 macrophage infiltration was another hypoxic marker indicated that subsets of patients with high M2 macrophages may benefit from macrophage-targeting therapy. CONCLUSIONS: These findings will facilitate the development of a hypoxia-oriented therapy strategy to enhance the treatment effect in the near future.

Hypermethylation of DRD5 Promoter Is a Biomarker Across 12 Cancer Types.

Aberrant DNA methylation is thought to be an early event in cancer development. Thus, identification of DNA methylation-based markers may provide valuable evidence in clinical decision-making. In this study, a DNA methylation dataset from 514 normal-tumor pairs is used to explore possible shared differentially methylated regions (DMRs) across 12 cancer types. Results showed that DMR in Dopamine receptor D5 (DRD5) promoter may be serviced as a good candidate biomarker across different cancer types. We further validated the extended DMR (292bp) in DRD5 promoter using SEQUENOM MassARRAY platform. Detection of DRD5 promoter dynamic methylation will allow rapid risk assessment at diagnosis, for suspicious tumor with the tissue biopsies.

High-risk Stage III colon cancer patients identified by a novel five-gene mutational signature are characterized by upregulation of IL-23A and gut bacterial translocation of the tumor microenvironment.

The heterogeneities of colorectal cancer (CRC) lead to staging inadequately of patients' prognosis. Here, we performed a prognostic analysis based on the tumor mutational profile and explored the characteristics of the high-risk tumors. We sequenced 338 colorectal carcinomas as the training dataset, constructed a novel five-gene (SMAD4, MUC16, COL6A3, FLG and LRP1B) prognostic signature, and validated it in an independent dataset from The Cancer Genome Atlas (TCGA). Kaplan-Meier and Cox regression analyses confirmed that the five-gene signature is an independent predictor of recurrence and prognosis in patients with Stage III colon cancer. The mutant signature translated to an increased risk of death (hazard ratio = 2.45, 95% confidence interval = 1.15-5.22, p = 0.016 in our dataset; hazard ratio = 4.78, 95% confidence interval = 1.33-17.16, p = 0.008 in TCGA dataset). RNA and bacterial 16S rRNA sequencing of high-risk tumors indicated that mutations of the five-gene signature may lead to intestinal barrier integrity, translocation of gut bacteria and deregulation of immune response and extracellular related genes. The high-risk tumors overexpressed IL23A and IL1RN genes and enriched with cancer-related bacteria (Bacteroides fragilis,Peptostreptococcus, Parvimonas, Alloprevotella and Gemella) compared to the low-risk tumors. The signature identified the high-risk group characterized by gut bacterial translocation and upregulation of interleukins of the tumor microenvironment, which was worth further researching.

Significant prognostic values of differentially expressed-aberrantly methylated hub genes in breast cancer.

Introduction: Abnormal status of gene expression plays an important role in tumorigenesis, progression and metastasis of breast cancer. Mechanisms of gene silence or activation were varied. Methylation of genes may contribute to alteration of gene expression. This study aimed to identify differentially expressed hub genes which may be regulated by DNA methylation and evaluate their prognostic value in breast cancer by bioinformatic analysis. Methods: GEO2R was used to obtain expression microarray data from GSE54002, GSE65194 and methylation microarray data from GSE20713, GSE32393. Differentially expressed-aberrantly methylated genes were identified by FunRich. Biological function and pathway enrichment analysis were conducted by DAVID. PPI network was constructed by STRING and hub genes was sorted by Cytoscape. Expression and DNA methylation of hub genes was validated by UALCAN and MethHC. Clinical outcome analysis of hub genes was performed by Kaplan Meier-plotter database for breast cancer. IHC was performed to analyze protein levels of EXO1 and Kaplan-Meier was used for survival analysis. Results: 677 upregulated-hypomethylated and 361 downregulated-hypermethylated genes were obtained from GSE54002, GSE65194, GSE20713 and GSE32393 by GEO2R and FunRich. The most significant biological process, cellular component, molecular function enriched and pathway for upregulated-hypomethylated genes were viral process, cytoplasm, protein binding and cell cycle respectively. For downregulated-hypermethylated genes, the result was peptidyl-tyrosine phosphorylation, plasma membrane, transmembrane receptor protein tyrosine kinase activity and Rap1 signaling pathway (All p< 0.05). 12 hub genes (TOP2A, MAD2L1, FEN1, EPRS, EXO1, MCM4, PTTG1, RRM2, PSMD14, CDKN3, H2AFZ, CCNE2) were sorted from 677 upregulated-hypomethylated genes. 4 hub genes (EGFR, FGF2, BCL2, PIK3R1) were sorted from 361 downregulated-hypermethylated genes. Differential expression of 16 hub genes was validated in UALCAN database (p<0.05). 7 in 12 upregulated-hypomethylated and 2 in 4 downregulated-hypermethylated hub genes were confirmed to be significantly hypomethylated or hypermethylated in breast cancer using MethHC database (p<0.05). Finally, 12 upregulated hub genes (TOP2A, MAD2L1, FEN1, EPRS, EXO1, MCM4, PTTG1, RRM2, PSMD14, CDKN3, H2AFZ, CCNE2) and 3 downregulated genes (FGF2, BCL2, PIK3R1) contributed to significant unfavorable clinical outcome in breast cancer (p<0.05). High expression level of EXO1 protein was significantly associated with poor OS in breast cancer patients (p=0.03). Conclusion: Overexpression of TOP2A, MAD2L1, FEN1, EPRS, EXO1, MCM4, PTTG1, RRM2, PSMD14, CDKN3, H2AFZ, CCNE2 and downregulation of FGF2, BCL2, PIK3R1 might serve as diagnosis and poor prognosis biomarkers in breast cancer by more research validation. EXO1 was identified as an individual unfavorable prognostic factor. Methylation might be one of the major causes leading to abnormal expression of those genes. Functional analysis and pathway enrichment analysis of those genes would provide novel ideas for breast cancer research.

Subtyping of microsatellite instability-high colorectal cancer.

BACKGROUND: Patients with microsatellite instability-high (MSI-H) colorectal cancer (CRC) generally have a better prognosis than patients with microsatellite stable (MSS) CRC. However, some MSI-H CRC patients do not gain overall survival benefits from immune checkpoint-blockade treatment. In other words, heterogeneity within the subgroup of MSI-H tumors remains poorly understood. Thus, an in-depth molecular characterization of MSI-H tumors is urgently required. METHODS: Here, we use nonnegative matrix factorization (NMF)-based consensus clustering to define CRC MSI-H subtypes in The Cancer Genome Atlas and a French multicenter cohort GSE39582. CIBERSORT was used to calculate the proportions of 22 lymphocytes in tumor tissue in MSI-H subtypes. RESULTS: MSI-H CRC samples basically clustered into two subgroups (MSI-H1 and MSI-H2). MSI-H1 was characterized by a lower BRAF mutational status, higher frequency of chromosomal instability, global hypomethylation, and worse survival than MSI-H2. Further examination of the immune landscape showed that macrophages of the M2 phenotype were enriched in MSI-H1, which may be associated with poor prognosis in this subgroup. CONCLUSIONS: Our results illustrate the genetic heterogeneity in MSI-H CRCs and macrophages may serve as good targets for anticancer therapy in MSI-H1.

PTPRB promotes metastasis of colorectal carcinoma via inducing epithelial-mesenchymal transition.

Dysregulation of protein tyrosine phosphatase, receptor type B (PTPRB) correlates with the development of a variety of tumors. Here we show that PTPRB promotes metastasis of colorectal cancer (CRC) cells via inducing epithelial-mesenchymal transition (EMT). We find that PTPRB is expressed at significantly higher levels in CRC tissues compared to adjacent nontumor tissues and in CRC cell lines with high invasion. PTPRB knockdown decreased the number of invasive CRC cells in an in vitro wound healing model, and also reduced tumor metastasis in vivo. Conversely, PTPRB overexpression promoted CRC cell invasion in vitro and metastasis in vivo. PTPRB overexpression decreased vimentin expression and promoted E-cadherin expression, consistent with promotion of EMT, while PTPRB knockdown had the opposite effect. Hypoxic conditions induced EMT and promoted invasion in CRC cells, but these effects were eliminated by PTPRB knockdown. EMT blockade via TWIST1 knockdown inhibited the migration and invasiveness of CRC cells, and even increased PTPRB expression could not reverse this effect. Altogether, these data support the conclusion that PTPRB promotes invasion and metastasis of CRC cells via inducing EMT, and that PTPRB would be a novel therapeutic target for the treatment of CRC.