Exploring inflammatory signatures in arthritic joint biopsies with Spatial Transcriptomics

Abstract

Lately it has become possible to analyze transcriptomic profiles in tissue sections with retained cellular context. We aimed to explore synovial biopsies from rheumatoid arthritis (RA) and spondyloarthritis (SpA) patients, using Spatial Transcriptomics (ST) as a proof of principle approach for unbiased mRNA studies at the site of inflammation in these chronic inflammatory diseases. Synovial tissue biopsies from affected joints were studied with ST. The transcriptome data was subjected to differential gene expression analysis (DEA), pathway analysis, immune cell type identification using Xcell analysis and validation with immunohistochemistry (IHC). The ST technology allows selective analyses on areas of interest, thus we analyzed morphologically distinct areas of mononuclear cell infiltrates. The top differentially expressed genes revealed an adaptive immune response profile and T-B cell interactions in RA, while in SpA, the profiles implicate functions associated with tissue repair. With spatially resolved gene expression data, overlaid on high-resolution histological images, we digitally portrayed pre-selected cell types in silico. The RA displayed an overrepresentation of central memory T cells, while in SpA effector memory T cells were most prominent. Consequently, ST allows for deeper understanding of cellular mechanisms and diversity in tissues from chronic inflammatory diseases.

Figure 1

Comparison of the top differentially expressed genes in RA vs. SpA when analysing the entire tissue section as a bulk compared to pre-selection of morphologically interesting areas. (A) An example of an H&E stained RA tissue section and the spots were the local transcriptome was captured in black. (B) A heatmap with the top 20 differentially expressed genes (DEGs) in the bulk data (without pre-selecting cluster) and ordered by lowest p-values (from top). The colour labeling represent the log2 normalized expression values and only genes with greater than twofold the difference in normalized expression value were selected. Each column represents a section, and three sections per patient were used. All sections were treated as a bulk with red columns representing RA and blue representing SpA. The tissue sections from the same biopsies cluster together and the SpA samples are clearly different to the RA samples. (C) The same tissue section with the annotated lymphocyte infiltrate areas in red and the unbiased analysis showing which spots fall into different clusters (depicted by different colours) as generated with an hierarchical clustering approach. For this tissue section, the black cluster (cluster 5) represents the cluster with strongest association to the histological annotation (red areas). (D) A heatmap with the top 20 differentially expressed genes (DEGs) in the pre-selected clusters overlapping the annotated infiltrates and ordered by the lowest p-values (from top). The colour labeling represent the log2 normalized expression values and only genes with greater than twofold the difference in normalized expression value were selected. Each column represents a selected cluster from each tissue section, and three sections per patient were used with red columns representing RA sections and blue representing SpA. The arrowheads show differentially expressed genes seen in both the bulk analysis and the infiltrates analysis.

Figure 2

Shared and unique pathways prominent in affected joint tissue from RA versus SpA. (A) Circos plot showing overlapping genes from the in silico bulk analysis based on differentially expressed genes. Each segment of the inner circle represents a gene list, whereas each line is originating from a single gene. Segments colored dark orange cover genes appearing in both the RA and SpA gene lists while light orange represents unique genes. Purple lines connect the genes being shared in the RA and SpA gene lists. Blue lines link the different genes where they fall into the same ontology term. A high number of connecting purple lines as well as a long dark orange segment thus implies an extensive overlap in the RA and SpA gene lists. (B) Heatmap showing the top 20 pathways from metascape analysis based on differentially expressed genes, sorted by p-value. (C) Circos plot of selected clusters associated with infiltrates shows that the higher resolution results in a stronger similarity between the infiltrate genes in diseases as well as higher heterogeneity within the diseases. (D) Heatmap showing the top 20 pathways for the analysis of selected clusters. Cluster selection was based on the association with infiltrates. The higher resolution obtained by selecting clusters reveals different pathways than the bulk analysis. The arrowheads show pathways overlapping both the bulk analysis and the infiltrates analysis.

Figure 3

The RA lymphocyte infiltrates are enriched in central memory CD4 and CD8 T cells as well as CD4 effector memory T cells. (A) Examples from one of the RA tissue sections depicting both the H&E stained section with the annotated mononuclear infiltrates as well as visualizations of transcriptomic data to show the localizations of different lymphocyte subsets as assigned by xCell (v1.1.0) method and viewed as spatial heatmaps to the right. The value for the respective cell subsets is an enrichment score showing how much enriched each cell type is compared to other regions within each individual tissue section. N.B. it is not possible to compare the amount of cell types for the sections. For that reason significance numbers were not considered for this figure. (B) Validation of xCell data with IHC analysis of T-cells (CD3-positive) from a nearby section in an area corresponding to the grey box on the CD4+ Tcm image in (A). (C) Validation of xCell data with IHC analysis of plasma cells (CD138-positive) from a nearby section in an area corresponding to the grey box on the plasma cell image in (A) with expected reciprocal T cell and Plasma cell distributions.

Figure 4

The SpA lymphocyte infiltrates are enriched in CD4 effector memory T cells and lack signals for the central memory T cells. Examples from one of the SpA tissue sections depicting both the H&E stained section with the annotated lymphocyte infiltrates as well as visualizations of transcriptomic data to show the localization of different lymphocyte subsets as assigned by xCell (v1.1.0) method and viewed as spatial heatmaps to the right. The value for the respective cell subsets is an enrichment score showing how much enriched each cell type is compared to other regions within each individual tissue section. N.B. it is not possible to compare the amount of cell types for the sections. For that reason significance numbers were not considered for this figure.