Big Picture

• identifying biomarkers of response to immune-checkpoint blockers is major challenge in immuno-oncology
• most studies look at specific mechanisms that activate tumor's suppression of immune response → tumors are complex in terms of their microenvironments, and understanding immune response in this microenvironment (TME) requires better hollistic understanding
  • paper uses RNA-sequencing data with different types of prior knowledge, using machine learnignt o extract from data mechanistic biomarkers of antitumor immune responses that predicted pateint's response to ICBs
  • approach only requires patient's tumor RNA-eq data to quantify biomarkers and predict patient's likelihood of response to ICB

Intro:

• immunotherapy revolutionized cancer treatment, based on antibodies targeting immune checkpoints such as the cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), programmed cell death protein (PD-1), or ligand (PD-L1)

  • Immune checkpoints:

• immune-checkpoint blockers boost patient's immune system to effectively recognize and attack cancer cells

  • pateitns treated with immune-based therapies have shown promising results especially in long-term patient survival

• only minor percent of patients get a perfect esponse → high immunological toxicity and considerable costs are callenges for ICB therapy

  • finding biomarkers to measure prior to providing treatment can save money and identify ideal candidates

• different mechanisms in tumor microenvironment involved in mediating immune response and affect efficacy of ICB therapy

• cell-type composition of this TME can reveal more info about treatment → different types of TME cells and immune cells in the TME can have a pro- or anti- tumor role in treatment

  • key role in anticancer response is effector T cells → phenotype, abundance, and localization within TME are major determinants of immunotherapy success

• also helps learn more about cellular function regulation based on the receptors and inhibitors → signals from outside of cells are processed by intracellular signalling pathways leading to changes in transcription factor activity in gene expression (binding factor)

  • for tumor cells, these regulatory networks are involved in growth and mutations + stimuli-responding mechanisms that tumor cells exploit to resist immune attacks
    • accomplished by upregulation of immune checkpoints, reduced release of inflammatory cytokines, and impaired antigen presentation by MHCs
  • these mechanisms used by cancer cells are primary ways they communicate with surrounding cells like T-cells
    • ligand-receptor interactions regulate cell-cell communication between cells in microenvironment, including tumor cells, immune cells, fibroblasts, and antitimour responses

  The Approach:

  • use RNA-seq data combined with prior knoweldge on intracellular signaling and immune-checkpoint inhibotors to learn more about said TME
  • quantify tumor infiltraing immune cells, activity and tfs, and extent of intercellular communicatoin from bulk-tumor RNA-seq data
  • use multi-task machine learning algorithms to assess how system-based signatures f TME are associated with 14 different transcriptome-based predictors of anticancer immune responses (models the response to ICB therapy, act as classes)
  • trained on 7550 pateints across 18 solid cancers from TCGA
  • helsp show how derived biomarkers that are known to be associated with immune response and response to ICBs for independent datasets and patients
    • tool built is Estimate Systems Immune Response

The Results:

• found 5 key signatures of TME of 7550 cancer patients across 18 solid cancers
• type 1: immune-cell fractions obtained with quanTIseq which quantifies the ten different immune cell types and fraction of other uncharacterized cells present in heterogeneous samples → pipeline is as shown
  • cell fractions estimated based on deconvolution approach which leverages cell-type-specific expression signatures for B cells and activated macrophages, monocytes, neutrophils, NK cells, non-regulatory CD4 T cells, CD* T cells, regulatory T cells, and myeloid dendritic cells