Checkpoint receptor TIGIT expressed on Tim-1+ B cells regulates tissue inflammation

Tim-1, a phosphatidylserine receptor expressed on B cells induces IL-10 production by sensing apoptotic cells. Here we show that mice with B cell-specific Tim-1 deletion developed tissue inflammation in multiple organs including spontaneous paralysis with inflammation in the central nervous system (CNS). Transcriptomic analysis demonstrated that besides IL-10, Tim-1+ B cells also differentially expressed a number of co-inhibitory “checkpoint” receptors including TIGIT. Mice with B cell-specific TIGIT deletion developed spontaneous paralysis with CNS inflammation but with limited inflammation in other organs. Our findings suggest that Tim-1+ B cells are essential for maintaining self-tolerance and restraining tissue inflammation, and that Tim-1 dependent TIGIT expression on B cells is essential for maintaining CNS-specific tolerance. A possible critical role of AhR in regulating the B cell function is discussed, as we found that AhR is a top-ranked transcription factor expressed in Tim-1+ B cells and regulates their TIGIT and IL-10 expression. Overall design: Samples from isolated Tim-1- and Tim-1+ B cells were processed with the SMART-Seq2 protocol (Picelli et al., 2013), and sequenced on Illumina Hi-Seq 2500. The paired-end 38bp reads sequenced (26.53M ± 1.62M pairs of reads per sample) for each of the 8 samples (4 Tim-1+ and 4 Tim-1- B cell samples derived from naïve WT mice) were aligned to the mouse mm10 UCSC reference genome using Tophat version 2.0.10 (Kim et al., 2013) with default settings (read alignment rate of 82.81% ± 0.44% properly mapped pairs per sample). Gene expression levels were quantified for 22,815 genes in the mouse mm10 UCSC reference gene annotations using Cuffquant in the Cufflinks suite version 2.2.1 (Trapnell et al., 2012). These levels were subsequently normalized across all 8 samples using Cuffnorm with default settings based on “geometric” normalization, which normalizes samples based on the median expression level in each sample, thus reporting normalized expression levels (FPKM: fragments per kilo bases of exons for per million mapped reads). Out of the 22,815 quantified reference genes, 15,519 genes were detected in at least one of the 8 samples at a normalized FPKM level > 0. Differential gene expression analysis comparing the 4 Tim-1+ and 4 Tim-1- B cell samples was performed by estimating the asymptotic t-test (Student's t-distribution) p-values, False Discovery Rate (FDR) (Benjamini and Hochberg) values and fold-changes using the ComparativeMarkerSelection module in GenePattern (Reich et al., 2006). Significantly differentially expressed genes were identified using the asymptotic t-test p-value = 0.05 and absolute fold change = 2 selection criteria. Heatmaps visualizing the normalized gene expression levels (zero mean centering and unit standard deviation scaling of the expression levels for each gene, followed by saturating these normalized gene expression levels at +1 and -1) were generated using GENE-E/Morpheus.