In this work we used single cell RNA sequencing to identify heterogeneity of the innate lymphoid cell population in the liver and the effects of microbiota. We collected data from 3320 innate lymphoid cells (ILCs) form three conditions, germ free, antibiotics treated and conventional specific pathogen free mice. To perform single cell RNAseq of innate lymphoid cells we cleaned intestines, surgically removed adipose tissue and Peyer’s patches. We performed tissue digestion and cell isolation using Lamina Propria Dissociation kit from Miltenyi Biotech. Cell suspension was stained with antibodies against CD45, CD3, CD19, Gr-1, NKp46, CD127, and KLRG-1. Single cells were sorted into wells of 384 well plates and sequencing libraries were prepared using MARSseq protocol. The libraries were sequenced using Ilumina Nextseq. To analyse the data, we used the R package “MetaCell” this allowed to infer sub-clusters of cells within the canonical ILC subsets, their functional annotation and the interactions between cells.
Furthermore, I used the acetaminophen and thioacetamide ALF models to characterise over 40.000 single cell transcriptomes and to define the ALF cellular atlas. Using hierarchical clustering I identified 49 distinct populations and described their molecular states. Then I performed analysis of secretome and cell-to-cell interactions. The analysis of activation patterns revealed a common signature of 82-genes and further examination of promoters of these showed that MYC is a transcriptional regulator of this signature. MYC is driving expression of Ccl2 in stellate, Kupffer and endothelial cells which in turn drives infiltration of monocytes. Then we validated this using MYC inhibitors and flow cytometry, activity of liver enzymes in serum, histology and single cell sequencing. Then we performed experiments in germ free and antibiotics treated mice to examine the role of microbiota on the transcriptomic responses using single cell RNA sequencing and we were able to show that in absence of microbiota the responses are ameliorated and additionally there was less infiltration of monocytes. I validated this with flow cytometry, activity of liver enzymes in serum and histology. We hypothesised that microbiota as well as damage signal through the same pathway, i.e. TLR signalling and we used MyD88 Trif double knock out mice which do not have TLR signalling, because MyD88 and Trif are adaptor proteins of TLRs. With single cell RNA sequencing we found that in disease MyD88 Trif double knock out mice phenocopy MYC inhibition, suggesting that indeed TLR signalling is upstream MYC. We also tested inhibitors of MAPK pathway proteins (IRAK4, p38, TAK1, TPL2, RIP1 and ERK) in mouse disease models to find out which of the proteins mediate the signalling cascade and used flow cytometry to measure the extent of monocyte infiltration and liver damage with activity of liver enzymes in serum and histology.
