Restricting datasets to classifiable samples augments discovery of immune disease biomarkers.

Immunological diseases are typically heterogeneous in clinical presentation, severity and response to therapy. Biomarkers of immune diseases often reflect this variability, especially compared to their regulated behaviour in health. This leads to a common difficulty that frustrates biomarker discovery and interpretation - namely, unequal dispersion of immune disease biomarker expression between patient classes necessarily limits a biomarker's informative range. To solve this problem, we introduce dataset restriction, a procedure that splits datasets into classifiable and unclassifiable samples. Applied to synthetic flow cytometry data, restriction identifies biomarkers that are otherwise disregarded. In advanced melanoma, restriction finds biomarkers of immune-related adverse event risk after immunotherapy and enables us to build multivariate models that accurately predict immunotherapy-related hepatitis. Hence, dataset restriction augments discovery of immune disease biomarkers, increases predictive certainty for classifiable samples and improves multivariate models incorporating biomarkers with a limited informative range. This principle can be directly extended to any classification task.

Modeling metastatic progression from cross-sectional cancer genomics data.

MOTIVATION: Metastasis formation is a hallmark of cancer lethality. Yet, metastases are generally unobservable during their early stages of dissemination and spread to distant organs. Genomic datasets of matched primary tumors and metastases may offer insights into the underpinnings and the dynamics of metastasis formation. RESULTS: We present metMHN, a cancer progression model designed to deduce the joint progression of primary tumors and metastases using cross-sectional cancer genomics data. The model elucidates the statistical dependencies among genomic events, the formation of metastasis, and the clinical emergence of both primary tumors and their metastatic counterparts. metMHN enables the chronological reconstruction of mutational sequences and facilitates estimation of the timing of metastatic seeding. In a study of nearly 5000 lung adenocarcinomas, metMHN pinpointed TP53 and EGFR as mediators of metastasis formation. Furthermore, the study revealed that post-seeding adaptation is predominantly influenced by frequent copy number alterations. AVAILABILITY AND IMPLEMENTATION: All datasets and code are available on GitHub at https://github.com/cbg-ethz/metMHN.

Anomaly detection in mixed high-dimensional molecular data.

MOTIVATION: Mixed molecular data combines continuous and categorical features of the same samples, such as OMICS profiles with genotypes, diagnoses, or patient sex. Like all high-dimensional molecular data, it is prone to incorrect values that can stem from various sources for example the technical limitations of the measurement devices, errors in the sample preparation, or contamination. Most anomaly detection algorithms identify complete samples as outliers or anomalies. However, in most cases, not all measurements of those samples are erroneous but only a few one-dimensional features within the samples are incorrect. These one-dimensional data errors are continuous measurements that are either located outside or inside the normal ranges of their features but in both cases show atypical values given all other continuous and categorical features in the sample. Additionally, categorical anomalies can occur for example when the genotype or diagnosis was submitted wrongly. RESULTS: We introduce ADMIRE (Anomaly Detection using MIxed gRaphical modEls), a novel approach for the detection and correction of anomalies in mixed high-dimensional data. Hereby, we focus on the detection of single (one-dimensional) data errors in the categorical and continuous features of a sample. For that the joint distribution of continuous and categorical features is learned by mixed graphical models, anomalies are detected by the difference between measured and model-based estimations and are corrected using imputation. We evaluated ADMIRE in simulation and by screening for anomalies in one of our own metabolic datasets. In simulation experiments, ADMIRE outperformed the state-of-the-art methods of Local Outlier Factor, stray, and Isolation Forest. AVAILABILITY AND IMPLEMENTATION: All data and code is available at https://github.com/spang-lab/adadmire. ADMIRE is implemented in a Python package called adadmire which can be found at https://pypi.org/project/adadmire.

An overview and a roadmap for artificial intelligence in hematology and oncology.

BACKGROUND: Artificial intelligence (AI) is influencing our society on many levels and has broad implications for the future practice of hematology and oncology. However, for many medical professionals and researchers, it often remains unclear what AI can and cannot do, and what are promising areas for a sensible application of AI in hematology and oncology. Finally, the limits and perils of using AI in oncology are not obvious to many healthcare professionals. METHODS: In this article, we provide an expert-based consensus statement by the joint Working Group on "Artificial Intelligence in Hematology and Oncology" by the German Society of Hematology and Oncology (DGHO), the German Association for Medical Informatics, Biometry and Epidemiology (GMDS), and the Special Interest Group Digital Health of the German Informatics Society (GI). We provide a conceptual framework for AI in hematology and oncology. RESULTS: First, we propose a technological definition, which we deliberately set in a narrow frame to mainly include the technical developments of the last ten years. Second, we present a taxonomy of clinically relevant AI systems, structured according to the type of clinical data they are used to analyze. Third, we show an overview of potential applications, including clinical, research, and educational environments with a focus on hematology and oncology. CONCLUSION: Thus, this article provides a point of reference for hematologists and oncologists, and at the same time sets forth a framework for the further development and clinical deployment of AI in hematology and oncology in the future.

Low-rank tensor methods for Markov chains with applications to tumor progression models.

Cancer progression can be described by continuous-time Markov chains whose state space grows exponentially in the number of somatic mutations. The age of a tumor at diagnosis is typically unknown. Therefore, the quantity of interest is the time-marginal distribution over all possible genotypes of tumors, defined as the transient distribution integrated over an exponentially distributed observation time. It can be obtained as the solution of a large linear system. However, the sheer size of this system renders classical solvers infeasible. We consider Markov chains whose transition rates are separable functions, allowing for an efficient low-rank tensor representation of the linear system's operator. Thus we can reduce the computational complexity from exponential to linear. We derive a convergent iterative method using low-rank formats whose result satisfies the normalization constraint of a distribution. We also perform numerical experiments illustrating that the marginal distribution is well approximated with low rank.

Salt-inducible kinase 3 protects tumor cells from cytotoxic T-cell attack by promoting TNF-induced NF-κB activation.

BACKGROUND: Cancer immunotherapeutic strategies showed unprecedented results in the clinic. However, many patients do not respond to immuno-oncological treatments due to the occurrence of a plethora of immunological obstacles, including tumor intrinsic mechanisms of resistance to cytotoxic T-cell (TC) attack. Thus, a deeper understanding of these mechanisms is needed to develop successful immunotherapies. METHODS: To identify novel genes that protect tumor cells from effective TC-mediated cytotoxicity, we performed a genetic screening in pancreatic cancer cells challenged with tumor-infiltrating lymphocytes and antigen-specific TCs. RESULTS: The screening revealed 108 potential genes that protected tumor cells from TC attack. Among them, salt-inducible kinase 3 (SIK3) was one of the strongest hits identified in the screening. Both genetic and pharmacological inhibitions of SIK3 in tumor cells dramatically increased TC-mediated cytotoxicity in several in vitro coculture models, using different sources of tumor and TCs. Consistently, adoptive TC transfer of TILs led to tumor growth inhibition of SIK3-depleted cancer cells in vivo. Mechanistic analysis revealed that SIK3 rendered tumor cells susceptible to tumor necrosis factor (TNF) secreted by tumor-activated TCs. SIK3 promoted nuclear factor kappa B (NF-κB) nuclear translocation and inhibited caspase-8 and caspase-9 after TNF stimulation. Chromatin accessibility and transcriptome analyses showed that SIK3 knockdown profoundly impaired the expression of prosurvival genes under the TNF-NF-κB axis. TNF stimulation led to SIK3-dependent phosphorylation of the NF-κB upstream regulators inhibitory-κB kinase and NF-kappa-B inhibitor alpha on the one side, and to inhibition of histone deacetylase 4 on the other side, thus sustaining NF-κB activation and nuclear stabilization. A SIK3-dependent gene signature of TNF-mediated NF-κB activation was found in a majority of pancreatic cancers where it correlated with increased cytotoxic TC activity and poor prognosis. CONCLUSION: Our data reveal an abundant molecular mechanism that protects tumor cells from cytotoxic TC attack and demonstrate that pharmacological inhibition of this pathway is feasible.

Evaluation of radio-immunotherapy sequence on immunological responses and clinical outcomes in patients with melanoma brain metastases (ELEKTRA).

In patients with melanoma brain metastases (MBM), a combination of radiotherapy (RT) with immune checkpoint inhibitors (ICI) is routinely used. However, the best sequence of radio-immunotherapy (RIT) remains unclear. In an exploratory phase 2 trial, MBM patients received RT (stereotactic or whole-brain radiotherapy depending on the number of MBM) combined with ipilimumab (ipi) ± nivolumab (nivo) in different sequencing (Rad-ICI or ICI-Rad). Comparators arms included patients treated with ipi-free systemic treatment or without RT (in MBM-free patients). The primary endpoints were radiological and immunological responses in the peripheral blood. Secondary endpoints were progression-free survival (PFS) and overall survival (OS). Of 106 screened, 92 patients were included in the study. Multivariate analysis revealed an advantage for patients starting with RT (Rad-ICI) for overall response rate (RR: p = .007; HR: 7.88 (95%CI: 1.76-35.27)) and disease control rate (DCR: p = .036; HR: 6.26 (95%CI: 1.13-34.71)) with a trend for a better PFS (p = .162; HR: 1.64 (95%CI: 0.8-3.3)). After RT plus two cycles of ipi-based ICI in both RIT sequences, increased frequencies of activated CD4, CD8 T cells and an increase in melanoma-specific T cell responses were observed in the peripheral blood. Lasso regression analysis revealed a significant clinical benefit for patients treated with Rad-ICI sequence and immunological features, including high frequencies of memory T cells and activated CD8 T cells in the blood. This study supports increasing evidence that sequencing RT followed by ICI treatment may have better effects on the immunological responses and clinical outcomes in MBM patients.

Epstein-Barr virus status of sporadic Burkitt lymphoma is associated with patient age and mutational features.

Sporadic Burkitt lymphoma (BL) is the most frequent tumour of children and adolescents but a rare subtype of lymphomas in adults. To date most molecular data have been obtained from lymphomas arising in the young. Recently, Epstein-Barr virus (EBV) positive and negative BL in young patients was shown to differ in molecular features. In the present study, we present a large age-overarching cohort of sporadic BL (n = 162) analysed by immunohistochemistry, translocations of MYC proto-oncogene, basic helix-loop-helix transcription factor (MYC), B-cell leukaemia/lymphoma 2 (BCL2) and B-cell leukaemia/lymphoma 6 (BCL6) and by targeted sequencing. We illustrate an age-associated inter-tumoral molecular heterogeneity in this disease. Mutations affecting inhibitor of DNA binding 3, HLH protein (ID3), transcription factor 3 (TCF3) and cyclin D3 (CCND3), which are highly recurrent in paediatric BL, and expression of sex determining region Y-box transcription factor 11 (SOX11) declined with patient age at diagnosis (P = 0·0204 and P = 0·0197 respectively). In contrast, EBV was more frequently detected in adult patients (P = 0·0262). Irrespective of age, EBV-positive sporadic BL showed significantly less frequent mutations in ID3/TCF3/CCND3 (P = 0·0088) but more often mutations of G protein subunit alpha 13 (GNA13; P = 0·0368) and forkhead box O1 (FOXO1; P = 0·0044) compared to EBV-negative tumours. Our findings suggest that among sporadic BL an EBV-positive subgroup of lymphomas increases with patient age that shows distinct pathogenic features reminiscent of EBV-positive endemic BL.

Metabolic imbalance of T cells in COVID-19 is hallmarked by basigin and mitigated by dexamethasone.

Metabolic pathways regulate immune responses and disrupted metabolism leads to immune dysfunction and disease. Coronavirus disease 2019 (COVID-19) is driven by imbalanced immune responses, yet the role of immunometabolism in COVID-19 pathogenesis remains unclear. By investigating 87 patients with confirmed SARS-CoV-2 infection, 6 critically ill non-COVID-19 patients, and 47 uninfected controls, we found an immunometabolic dysregulation in patients with progressed COVID-19. Specifically, T cells, monocytes, and granulocytes exhibited increased mitochondrial mass, yet only T cells accumulated intracellular reactive oxygen species (ROS), were metabolically quiescent, and showed a disrupted mitochondrial architecture. During recovery, T cell ROS decreased to match the uninfected controls. Transcriptionally, T cells from severe/critical COVID-19 patients showed an induction of ROS-responsive genes as well as genes related to mitochondrial function and the basigin network. Basigin (CD147) ligands cyclophilin A and the SARS-CoV-2 spike protein triggered ROS production in T cells in vitro. In line with this, only PCR-positive patients showed increased ROS levels. Dexamethasone treatment resulted in a downregulation of ROS in vitro and T cells from dexamethasone-treated patients exhibited low ROS and basigin levels. This was reflected by changes in the transcriptional landscape. Our findings provide evidence of an immunometabolic dysregulation in COVID-19 that can be mitigated by dexamethasone treatment.

Learning from Embryogenesis-A Comparative Expression Analysis in Melanoblast Differentiation and Tumorigenesis Reveals miRNAs Driving Melanoma Development.

Malignant melanoma is one of the most dangerous tumor types due to its high metastasis rates and a steadily increasing incidence. During tumorigenesis, the molecular processes of embryonic development, exemplified by epithelial-mesenchymal transition (EMT), are often reactivated. For melanoma development, the exact molecular differences between melanoblasts, melanocytes, and melanoma cells are not completely understood. In this study, we aimed to identify microRNAs (miRNAs) that promote melanoma tumorigenesis and progression, based on an in vitro model of normal human epidermal melanocyte (NHEM) de-differentiation into melanoblast-like cells (MBrCs). Using miRNA-sequencing and differential expression analysis, we demonstrated in this study that a majority of miRNAs have an almost equal expression level in NHEMs and MBrCs but are significantly differentially regulated in primary tumor- and metastasis-derived melanoma cell lines. Further, a target gene analysis of strongly regulated but functionally unknown miRNAs yielded the implication of those miRNAs in many important cellular pathways driving malignancy. We hypothesize that many of the miRNAs discovered in our study are key drivers of melanoma development as they account for the tumorigenic potential that differentiates melanoma cells from proliferating or migrating embryonic cells.

Virus-specific memory T cell responses unmasked by immune checkpoint blockade cause hepatitis.

Treatment of advanced melanoma with combined PD-1/CTLA-4 blockade commonly causes serious immune-mediated complications. Here, we identify a subset of patients predisposed to immune checkpoint blockade-related hepatitis who are distinguished by chronic expansion of effector memory CD4+ T cells (TEM cells). Pre-therapy CD4+ TEM cell expansion occurs primarily during autumn or winter in patients with metastatic disease and high cytomegalovirus (CMV)-specific serum antibody titres. These clinical features implicate metastasis-dependent, compartmentalised CMV reactivation as the cause of CD4+ TEM expansion. Pre-therapy CD4+ TEM expansion predicts hepatitis in CMV-seropositive patients, opening possibilities for avoidance or prevention. 3 of 4 patients with pre-treatment CD4+ TEM expansion who received αPD-1 monotherapy instead of αPD-1/αCTLA-4 therapy remained hepatitis-free. 4 of 4 patients with baseline CD4+ TEM expansion given prophylactic valganciclovir and αPD-1/αCTLA-4 therapy remained hepatitis-free. Our findings exemplify how pathogen exposure can shape clinical reactions after cancer therapy and how this insight leads to therapeutic innovations.

Identification of a miRNA based model to detect prognostic subgroups in patients with aggressive B-cell lymphoma.

In order to differentiate prognostic subgroups of patients with aggressive B-cell lymphoma, we analyzed the expression of 800 miRNAs with the NanoString nCounter human miRNA assay on a cohort of 228 FFPE samples of patients enrolled in the RICOVER-60 and MegaCHOEP trials. We identified significant miRNA signatures for overall survival (OS) and progression-free survival (PFS) by LASSO-penalized linear Cox-regression. High expression levels of miR-130a-3p and miR-423-5p indicate a better prognosis, whereas high levels of miR-374b-5p, miR-590-5p, miR-186-5p, and miR-106b-5p increase patients' risk levels for OS. Regarding PFS high expression of miR-365a-5p in addition to the other two miRNAs improves the prognosis and high levels of miR374a-5p, miR-106b-5p, and miR-590-5p, connects with increased risk and poor prognosis. We identified miRNA signatures to subdivide patients into two different risk groups. These prognostic models may be used in risk stratification in future clinical trials and help making personalized therapy decisions.

Platform independent protein-based cell-of-origin subtyping of diffuse large B-cell lymphoma in formalin-fixed paraffin-embedded tissue.

Diffuse large B-cell lymphoma (DLBCL) is commonly classified by gene expression profiling according to its cell of origin (COO) into activated B-cell (ABC)-like and germinal center B-cell (GCB)-like subgroups. Here we report the application of label-free nano-liquid chromatography - Sequential Window Acquisition of all THeoretical fragment-ion spectra - mass spectrometry (nanoLC-SWATH-MS) to the COO classification of DLBCL in formalin-fixed paraffin-embedded (FFPE) tissue. To generate a protein signature capable of predicting Affymetrix-based GCB scores, the summed log2-transformed fragment ion intensities of 780 proteins quantified in a training set of 42 DLBCL cases were used as independent variables in a penalized zero-sum elastic net regression model with variable selection. The eight-protein signature obtained showed an excellent correlation (r = 0.873) between predicted and true GCB scores and yielded only 9 (21.4%) minor discrepancies between the three classifications: ABC, GCB, and unclassified. The robustness of the model was validated successfully in two independent cohorts of 42 and 31 DLBCL cases, the latter cohort comprising only patients aged >75 years, with Pearson correlation coefficients of 0.846 and 0.815, respectively, between predicted and NanoString nCounter based GCB scores. We further show that the 8-protein signature is directly transferable to both a triple quadrupole and a Q Exactive quadrupole-Orbitrap mass spectrometer, thus obviating the need for proprietary instrumentation and reagents. This method may therefore be used for robust and competitive classification of DLBCLs on the protein level.

Author Correction: Mechanisms governing the pioneering and redistribution capabilities of the non-classical pioneer PU.1.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Mechanisms governing the pioneering and redistribution capabilities of the non-classical pioneer PU.1.

Establishing gene regulatory networks during differentiation or reprogramming requires master or pioneer transcription factors (TFs) such as PU.1, a prototype master TF of hematopoietic lineage differentiation. To systematically determine molecular features that control its activity, here we analyze DNA-binding in vitro and genome-wide in vivo across different cell types with native or ectopic PU.1 expression. Although PU.1, in contrast to classical pioneer factors, is unable to access nucleosomal target sites in vitro, ectopic induction of PU.1 leads to the extensive remodeling of chromatin and redistribution of partner TFs. De novo chromatin access, stable binding, and redistribution of partner TFs both require PU.1's N-terminal acidic activation domain and its ability to recruit SWI/SNF remodeling complexes, suggesting that the latter may collect and distribute co-associated TFs in conjunction with the non-classical pioneer TF PU.1.

Loss-Function Learning for Digital Tissue Deconvolution.

The gene expression profile of a tissue averages the expression profiles of all cells in this tissue. Digital tissue deconvolution addresses the following inverse problem: given the expression profile y of a tissue, what is the cellular composition c of that tissue? If X is a matrix whose columns are reference profiles of individual cell types, the composition c can be computed by minimizing ℒ ( y - X c ) for a given loss function ℒ . Current methods use predefined all-purpose loss functions. They successfully quantify the dominating cells of a tissue, while often falling short in detecting small cell populations. In this study we use training data to learn the loss function ℒ along with the composition c . This allows us to adapt to application-specific requirements such as focusing on small cell populations or distinguishing phenotypically similar cell populations. Our method quantifies large cell fractions as accurately as existing methods and significantly improves the detection of small cell populations and the distinction of similar cell types.

DTD: An R Package for Digital Tissue Deconvolution.

Digital tissue deconvolution (DTD) estimates the cellular composition of a tissue from its bulk gene-expression profile. For this, DTD approximates the bulk as a mixture of cell-specific expression profiles. Different tissues have different cellular compositions, with cells in different activation states, and embedded in different environments. Consequently, DTD can profit from tailoring the deconvolution model to a specific tissue context. Loss-function learning adapts DTD to a specific tissue context, such as the deconvolution of blood, or a specific type of tumor tissue. We provide software for loss-function learning, for its validation and visualization, and for applying the DTD models to new data.

A novel lymphoma-associated macrophage interaction signature (LAMIS) provides robust risk prognostication in diffuse large B-cell lymphoma clinical trial cohorts of the DSHNHL.

Diffuse large B-cell lymphoma (DLBCL) is a disease with heterogeneous outcome. Stromal signatures have been correlated to survival in DLBCL. Their use, however, is hampered by the lack of assays for formalin-fixed paraffin-embedded material (FFPE). We constructed a lymphoma-associated macrophage interaction signature (LAMIS) interrogating features of the microenvironment using a NanoString assay applicable to FFPE. The clinical impact of the signature could be validated in a cohort of 466 patients enrolled in prospective clinical trials of the German High-Grade Non-Hodgkin Lymphoma Study Group (DSHNHL). Patients with high expression of the signature (LAMIShigh) had shorter EFS, PFS, and OS. Multivariate analyses revealed independence from IPI factors in EFS (HR 1.7, 95% CI 1.2-2.4, p-value = 0.001), PFS (HR 1.8, 95% CI 1.2-2.5, p-value = 0.001) and OS (HR 1.8, 95% CI 1.3-2.7, p-value = 0.001). Multivariate analyses adjusted for the IPI factors showed the signature to be independent from COO, MYC rearrangements and double expresser status (DE). LAMIShigh and simultaneous DE status characterized a patient subgroup with dismal prognosis and early relapse. Our data underline the importance of the microenvironment in prognosis. Combined analysis of stromal features, the IPI and DE may provide a new rationale for targeted therapy.

Modelling cancer progression using Mutual Hazard Networks.

MOTIVATION: Cancer progresses by accumulating genomic events, such as mutations and copy number alterations, whose chronological order is key to understanding the disease but difficult to observe. Instead, cancer progression models use co-occurrence patterns in cross-sectional data to infer epistatic interactions between events and thereby uncover their most likely order of occurrence. State-of-the-art progression models, however, are limited by mathematical tractability and only allow events to interact in directed acyclic graphs, to promote but not inhibit subsequent events, or to be mutually exclusive in distinct groups that cannot overlap. RESULTS: Here we propose Mutual Hazard Networks (MHN), a new Machine Learning algorithm to infer cyclic progression models from cross-sectional data. MHN model events by their spontaneous rate of fixation and by multiplicative effects they exert on the rates of successive events. MHN compared favourably to acyclic models in cross-validated model fit on four datasets tested. In application to the glioblastoma dataset from The Cancer Genome Atlas, MHN proposed a novel interaction in line with consecutive biopsies: IDH1 mutations are early events that promote subsequent fixation of TP53 mutations. AVAILABILITY AND IMPLEMENTATION: Implementation and data are available at https://github.com/RudiSchill/MHN. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.