ABSTRACT
Regulatory T (Treg) cells are a barrier for tumor immunity and a target for immunotherapy. Using single-cell transcriptomics, we found that CD4+ T cells infiltrating primary and metastatic colorectal cancer and non-small-cell lung cancer are highly enriched for two subsets of comparable size and suppressor function comprising forkhead box protein P3+ Treg and eomesodermin homolog (EOMES)+ type 1 regulatory T (Tr1)-like cells also expressing granzyme K and chitinase-3-like protein 2. EOMES+ Tr1-like cells, but not Treg cells, were clonally related to effector T cells and were clonally expanded in primary and metastatic tumors, which is consistent with their proliferation and differentiation in situ. Using chitinase-3-like protein 2 as a subset signature, we found that the EOMES+ Tr1-like subset correlates with disease progression but is also associated with response to programmed cell death protein 1-targeted immunotherapy. Collectively, these findings highlight the heterogeneity of Treg cells that accumulate in primary tumors and metastases and identify a new prospective target for cancer immunotherapy.
Subject(s)
Carcinoma, Non-Small-Cell Lung/immunology , Clonal Hematopoiesis/immunology , Colorectal Neoplasms/immunology , Lung Neoplasms/immunology , T-Lymphocytes, Regulatory/immunology , Adult , Aged , Aged, 80 and over , Carcinoma, Non-Small-Cell Lung/genetics , Carcinoma, Non-Small-Cell Lung/secondary , Carcinoma, Non-Small-Cell Lung/therapy , Cell Differentiation/genetics , Cell Differentiation/immunology , Cell Proliferation/genetics , Chemotherapy, Adjuvant/methods , Chitinases/metabolism , Colectomy , Colon/pathology , Colon/surgery , Colorectal Neoplasms/genetics , Colorectal Neoplasms/pathology , Colorectal Neoplasms/therapy , Datasets as Topic , Disease Progression , Drug Resistance, Neoplasm/immunology , Female , Flow Cytometry , Forkhead Transcription Factors/metabolism , Gene Expression Regulation, Neoplastic/immunology , Granzymes/metabolism , Humans , Immune Checkpoint Inhibitors/pharmacology , Immune Checkpoint Inhibitors/therapeutic use , Kaplan-Meier Estimate , Lung Neoplasms/genetics , Lung Neoplasms/pathology , Lung Neoplasms/therapy , Male , Middle Aged , Primary Cell Culture , Programmed Cell Death 1 Receptor/antagonists & inhibitors , RNA-Seq , Single-Cell Analysis , T-Box Domain Proteins/metabolism , T-Lymphocytes, Regulatory/metabolismABSTRACT
Fluorescence in situ hybridization (FISH) is a powerful single-cell technique that harnesses nucleic acid base pairing to detect the abundance and positioning of cellular RNA and DNA molecules in fixed samples. Recent technology development has paved the way to the construction of FISH probes entirely from synthetic oligonucleotides (oligos), allowing the optimization of thermodynamic properties together with the opportunity to design probes against any sequenced genome. However, comparatively little progress has been made in the development of computational tools to facilitate the oligos design, and even less has been done to extend their accessibility. OligoMiner is an open-source and modular pipeline written in Python that introduces a novel method of assessing probe specificity that employs supervised machine learning to predict probe binding specificity from genome-scale sequence alignment information. However, its use is restricted to only those people who are confident with command line interfaces because it lacks a Graphical User Interface (GUI), potentially cutting out many researchers from this technology. Here, we present OligoMinerApp (http://oligominerapp.org), a web-based application that aims to extend the OligoMiner framework through the implementation of a smart and easy-to-use GUI and the introduction of new functionalities specially designed to make effective probe mining available to everyone.
