Unveiling the immune infiltrate modulation in cancer and response to immunotherapy by MIXTURE-an enhanced deconvolution method

Fourteen years of cellular deconvolution: methodology, applications, technical evaluation and outstanding challenges

Single-cell RNA sequencing (scRNA-Seq) is a recent technology that allows for the measurement of the expression of all genes in each individual cell contained in a sample. Information at the single-cell level has been shown to be extremely useful in many areas. However, performing single-cell experiments is expensive. Although cellular deconvolution cannot provide the same comprehensive information as single-cell experiments, it can extract cell-type information from bulk RNA data, and therefore it allows researchers to conduct studies at cell-type resolution from existing bulk datasets. For these reasons, a great effort has been made to develop such methods for cellular deconvolution. The large number of methods available, the requirement of coding skills, inadequate documentation, and lack of performance assessment all make it extremely difficult for life scientists to choose a suitable method for their experiment. This paper aims to fill this gap by providing a comprehensive review of 53 deconvolution methods regarding their methodology, applications, performance, and outstanding challenges. More importantly, the article presents a benchmarking of all these 53 methods using 283 cell types from 30 tissues of 63 individuals. We also provide an R package named DeconBenchmark that allows readers to execute and benchmark the reviewed methods (https://github.com/tinnlab/DeconBenchmark).

From complexity to clarity: unravelling tumor heterogeneity through the lens of tumor microenvironment for innovative cancer therapy

Despite the tremendous clinical successes recorded in the landscape of cancer therapy, tumor heterogeneity remains a formidable challenge to successful cancer treatment. In recent years, the emergence of high-throughput technologies has advanced our understanding of the variables influencing tumor heterogeneity beyond intrinsic tumor characteristics. Emerging knowledge shows that drivers of tumor heterogeneity are not only intrinsic to cancer cells but can also emanate from their microenvironment, which significantly favors tumor progression and impairs therapeutic response. Although much has been explored to understand the fundamentals of the influence of innate tumor factors on cancer diversity, the roles of the tumor microenvironment (TME) are often undervalued. It is therefore imperative that a clear understanding of the interactions between the TME and other tumor intrinsic factors underlying the plastic molecular behaviors of cancers be identified to develop patient-specific treatment strategies. This review highlights the roles of the TME as an emerging factor in tumor heterogeneity. More particularly, we discuss the role of the TME in the context of tumor heterogeneity and explore the cutting-edge diagnostic and therapeutic approaches that could be used to resolve this recurring clinical conundrum. We conclude by speculating on exciting research questions that can advance our understanding of tumor heterogeneity with the goal of developing customized therapeutic solutions.

Informing immunotherapy with multi-omics driven machine learning

Progress in sequencing technologies and clinical experiments has revolutionized immunotherapy on solid and hematologic malignancies. However, the benefits of immunotherapy are limited to specific patient subsets, posing challenges for broader application. To improve its effectiveness, identifying biomarkers that can predict patient response is crucial. Machine learning (ML) play a pivotal role in harnessing multi-omic cancer datasets and unlocking new insights into immunotherapy. This review provides an overview of cutting-edge ML models applied in omics data for immunotherapy analysis, including immunotherapy response prediction and immunotherapy-relevant tumor microenvironment identification. We elucidate how ML leverages diverse data types to identify significant biomarkers, enhance our understanding of immunotherapy mechanisms, and optimize decision-making process. Additionally, we discuss current limitations and challenges of ML in this rapidly evolving field. Finally, we outline future directions aimed at overcoming these barriers and improving the efficiency of ML in immunotherapy research.

Cyclooxygenase-2 Blockade Is Crucial to Restore Natural Killer Cell Activity before Anti-CTLA-4 Therapy against High-Grade Serous Ovarian Cancer

Chronic inflammation influences the tumor immune microenvironment (TIME) in high-grade serous ovarian cancer (HGSOC). Specifically, cyclooxygenase-2 (COX-2) overexpression promotes cytotoxic T-lymphocyte-associated protein-4 (CTLA-4) expression. Notably, elevated COX-2 levels in the TIME have been associated with reduced response to anti-CTLA-4 immunotherapy. However, the precise impact of COX-2, encoded by PTGS2, on the immune profile remains unknown. To address this, we performed an integrated bioinformatics analysis using data from the HGSOC cohorts (TCGA-OV, n = 368; Australian cohort AOCS, n = 80; GSE26193, n = 62; and GSE30161, n = 45). Employing Gene Set Variation Analysis (GSVA), MIXTURE and Ecotyper cell deconvolution algorithms, we concluded that COX-2 was linked to immune cell ecosystems associated with shorter survival, cell dysfunction and lower NK cell effector cytotoxicity capacity. Next, we validated these results by characterizing circulating NK cells from HGSOC patients through flow cytometry and cytotoxic assays while undergoing COX-2 and CTLA-4 blockade. The blockade of COX-2 improved the cytotoxic capacity of NK cells against HGSOC cell lines. Our findings underscore the relevance of COX-2 in shaping the TIME and suggest its potential as a prognostic indicator and therapeutic target. Increased COX-2 expression may hamper the effectivity of immunotherapies that require NK cell effector function. These results provide a foundation for experimental validation and clinical trials investigating combined therapies targeting COX-2 and CTLA-4 in HGSOC.

Novel evaluation approach for molecular signature-based deconvolution methods

The tumoral immune microenvironment (TIME) plays a key role in prognosis, therapeutic approach and pathophysiological understanding over oncological processes. Several computational immune cell-type deconvolution methods (DM), supported by diverse molecular signatures (MS), have been developed to uncover such TIME interplay from RNA-seq tumor biopsies. MS-DM pairs were benchmarked against each other by means of different metrics, such as Pearson's correlation, R2 and RMSE, but these only evaluate the linear association of the estimated proportion related to the expected one, missing the analysis of prediction-dependent bias trends and cell identification accuracy. We present a novel protocol composed of four tests allowing appropriate evaluation of the cell type identification performance and proportion prediction accuracy of molecular signature-deconvolution method pair by means of certainty and confidence cell-type identification scores (F1-score, distance to the optimal point and error rates) as well the Bland-Altman method for error-trend analysis. Our protocol was used to benchmark six state-of-the-art DMs (CIBERSORTx, DCQ, DeconRNASeq, EPIC, MIXTURE and quanTIseq) paired to five murine tissue-specific MSs, revealing a systematic overestimation of the number of different cell types across almost all methods.

Blocking soluble TNFα sensitizes HER2-positive breast cancer to trastuzumab through MUC4 downregulation and subverts immunosuppression

BACKGROUND

The success of HER2-positive (HER2+) breast cancer treatment with trastuzumab, an antibody that targets HER2, relies on immune response. We demonstrated that TNFα induces mucin 4 (MUC4) expression, which shields the trastuzumab epitope on the HER2 molecule decreasing its therapeutic effect. Here, we used mouse models and samples from HER2+ breast cancer patients to unravel MUC4 participation in hindering trastuzumab effect by fostering immune evasion.

METHODS

We used a dominant negative TNFα inhibitor (DN) selective for soluble TNFα (sTNFα) together with trastuzumab. Preclinical experiments were performed using two models of conditionally MUC4-silenced tumors to characterize the immune cell infiltration. A cohort of 91 patients treated with trastuzumab was used to correlate tumor MUC4 with tumor-infiltrating lymphocytes.

RESULTS

In mice bearing de novo trastuzumab-resistant HER2+ breast tumors, neutralizing sTNFα with DN induced MUC4 downregulation. Using the conditionally MUC4-silenced tumor models, the antitumor effect of trastuzumab was reinstated and the addition of TNFα-blocking agents did not further decrease tumor burden. DN administration with trastuzumab modifies the immunosuppressive tumor milieu through M1-like phenotype macrophage polarization and NK cells degranulation. Depletion experiments revealed a cross-talk between macrophages and NK cells necessary for trastuzumab antitumor effect. In addition, tumor cells treated with DN are more susceptible to trastuzumab-dependent cellular phagocytosis. Finally, MUC4 expression in HER2+ breast cancer is associated with immune desert tumors.

CONCLUSIONS

These findings provide rationale to pursue sTNFα blockade combined with trastuzumab or trastuzumab drug conjugates for MUC4+ and HER2+ breast cancer patients to overcome trastuzumab resistance.

Identification of circadian clock genes as regulators of immune infiltration in Hepatocellular Carcinoma

Background: Multiple studies have reported that the immune system is under the control of a circadian clock, especially in cancers, but how circadian clock genes shape tumor immune cell infiltration in hepatocellular carcinoma (HCC) remains unclear. Methods: The rhythmicity of circadian clock genes was investigated using the GETx database. The expression and methylation level of circadian clock genes in HCC and paracancerous was evaluated using the GETx and TCGA databases. The differential expression of circadian clock genes in HCC was analyzed using the "limma" package of the R 4.0.4 software. The prognosis of each circadian clock gene was accessed by Kaplan-Meier survival analysis and Cox proportional hazards regression analysis. Quantitative real-time PCR and immunohistochemistry (IHC) was carried out to confirm the results. The relationship between circadian rhythm and immune infiltration in HCC was evaluated using the TIMER database and the CIBERSORT algorithm. Results: In addition to RORA, RORB, and ARNTL2, there was a rhythmic expression of other circadian clock genes in liver tissue. The correlation between the expression of circadian clock genes differed when comparing HCC and liver tissue. HCC patients who express low levels of PER-1and CRY2 had a poor overall survival (OS). In contrast, patients with higher expression of NPAS2 had a poor prognosis. In HCC, the expression of the PER-1, CRY2, and NPAS2 genes was closely related to immune infiltration. Conclusion: Our study indicated the disruption of the expression of circadian clock-regulated genes in HCC and identified PER-1, CRY2, and NPAS2 as independent predictors of survival. These genes may be applied as candidate molecular targets for diagnosis and therapy of HCC.

Oral squamous cell carcinoma (OSCC) tumors from heavy alcohol consumers are associated with higher levels of TLR9 and a particular immunophenotype: Impact on patient survival

Oral squamous cell carcinoma (OSCC) is one of the most frequent types of oral cancer in developing countries and its burden correlates with exposure to tobacco and excessive alcohol consumption. Toll like receptors (TLRs) are major sensors of inflammatory stimuli, from both microbial and sterile causes and as such, they have been related to tumor progression and metastasis. Here, we evaluated the expression of TLR2, 4 and 9 as well as CD3+, CD8+ and Granzyme B+ cell infiltration by immunohistochemistry in oral samples of 30 patients with OSCC, classified according to their consumption of alcohol. Our findings indicate that there is a significant association between heavy alcohol consumption and tumors with higher expression levels of TLR9. Moreover, patients with TLR9high tumors, as well as those who indicated high consumption of alcohol exhibited a diminished overall survival. TCGA data analysis indicated that TLR9high tumors express a significant increase in some genes related with the oral cavity itself, inflammation and tumor promotion. Our analysis of tumor infiltrating leukocytes demonstrated that the major differences perceived in heavy alcohol consumers was the location of CD8+ T cells infiltrating the tumor, which showed lower numbers intratumorally. Our data suggest the existence of a pathogenic loop that involves alcohol consumption, high TLR9 expression and the immunophenotype, which might have a profound impact on the progression of the disease.

Comprehensive analysis of tumor immune microenvironment and prognosis of m6A-related lncRNAs in gastric cancer

Background

N6-methyladenosine (m6A) modification and long non-coding RNAs (lncRNAs) play pivotal roles in gastric cancer (GC) progression. The emergence of immunotherapy in GC has created a paradigm shift in the approaches of treatment, whereas there is significant heterogeneity with regard to degree of treatment responses, which results from the variability of tumor immune microenvironment (TIME). How the interplay between m6A and lncRNAs enrolling in the shaping of TIME remains unclear.

Methods

The RNA sequencing and clinical data of GC patients were collected from TCGA database. Pearson correlation test and univariate Cox analysis were used to screen out m6A-related lncRNAs. Consensus clustering method was implemented to classify GC patients into two clusters. Survival analysis, the infiltration level of immune cells, Gene set enrichment analysis (GSEA) and the mutation profiles were analyzed and compared between two clusters. A competing endogenous RNA (ceRNA) network and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were applied for the identification of pathways in which m6A-related lncRNAs enriched. Then least absolute shrinkage and selection operator (LASSO) COX regression was implemented to select pivotal lncRNAs, and risk model was constructed accordingly. The prognosis value of the risk model was explored. In addition, the response to immune checkpoint inhibitors (ICIs) therapy were compared between different risk groups. Finally, we performed qRT-PCR to detect expression patterns of the selected lncRNAs in the 35 tumor tissues and their paired adjacent normal tissues, and validated the prognostic value of risk model in our cohort (N = 35).

Results

The expression profiles of 15 lncRNAs were included to cluster patients into 2 subtypes. Cluster1 with worse prognosis harbored higher immune score, stromal score, ESTIMATE score and lower mutation rates of the genes. Different immune cell infiltration patterns were also displayed between the two clusters. GSEA showed that cluster1 preferentially enriched in tumor hallmarks and tumor-related biological pathways. KEGG pathway analysis found that the target mRNAs which m6A-related lncRNAs regulated by sponging miRNAs mainly enriched in vascular smooth muscle contraction, cAMP signaling pathway and cGMP-PKG signaling pathway. Next, eight lncRNAs were selected by LASSO regression algorithm to construct risk model. Patients in the high-risk group had poor prognoses, which were consistent in our cohort. As for predicting responses to ICIs therapy, patients from high-risk group were found to have lower tumor mutation burden (TMB) scores and account for large proportion in the Microsatellite Instability-Low (MSI-L) subtype. Moreover, patients had distinct immunophenoscores in different risk groups.

Conclusion

Our study revealed that the interplay between m6A modification and lncRNAs might have critical role in predicting GC prognosis, sculpting TIME landscape and predicting the responses to ICIs therapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-022-09377-8.

Multimetric feature selection for analyzing multicategory outcomes of colorectal cancer: random forest and multinomial logistic regression models

Colorectal cancer (CRC) is one of the most common cancers worldwide, and a leading cause of cancer deaths. Better classifying multicategory outcomes of CRC with clinical and omic data may help adjust treatment regimens based on individual's risk. Here, we selected the features that were useful for classifying four-category survival outcome of CRC using the clinical and transcriptomic data, or clinical, transcriptomic, microsatellite instability and selected oncogenic-driver data (all data) of TCGA. We also optimized multimetric feature selection to develop the best multinomial logistic regression (MLR) and random forest (RF) models that had the highest accuracy, precision, recall and F1 score, respectively. We identified 2073 differentially expressed genes of the TCGA RNASeq dataset. MLR overall outperformed RF in the multimetric feature selection. In both RF and MLR models, precision, recall and F1 score increased as the feature number increased and peaked at the feature number of 600-1000, while the models' accuracy remained stable. The best model was the MLR one with 825 features based on sum of squared coefficients using all data, and attained the best accuracy of 0.855, F1 of 0.738 and precision of 0.832, which were higher than those using clinical and transcriptomic data. The top-ranked features in the MLR model of the best performance using clinical and transcriptomic data were different from those using all data. However, pathologic staging, HBS1L, TSPYL4, and TP53TG3B were the overlapping top-20 ranked features in the best models using clinical and transcriptomic, or all data. Thus, we developed a multimetric feature-selection based MLR model that outperformed RF models in classifying four-category outcome of CRC patients. Interestingly, adding microsatellite instability and oncogenic-driver data to clinical and transcriptomic data improved models' performances. Precision and recall of tuned algorithms may change significantly as the feature number changes, but accuracy appears not sensitive to these changes.

Evaluating the impact of age on immune checkpoint therapy biomarkers

Both tumors and aging alter the immune landscape of tissues. These interactions may play an important role in tumor progression among elderly patients and may suggest considerations for patient care. We leverage large-scale genomic and clinical databases to perform comprehensive comparative analysis of molecular and cellular markers of immune checkpoint blockade (ICB) response with patient age. These analyses demonstrate that aging is associated with increased tumor mutational burden, increased expression and decreased promoter methylation of immune checkpoint genes, and increased interferon gamma signaling in older patients in many cancer types studied, all of which are expected to promote ICB efficacy. Concurrently, we observe age-related alterations that might be expected to reduce ICB efficacy, such as decreases in T cell receptor diversity. Altogether, these changes suggest the capacity for robust ICB response in many older patients, which may warrant large-scale prospective study on ICB therapies among patients of advanced age.