Immune cell specific GBP5 expression in the colon of CD patients
To better understand the tissue distribution of GBP5, IHC staining of GBP5 was examined at the entire depth of the inflamed and non-inflamed colonic tissues of CD patients underwent colon resection. In the inflamed colon, GBP5 positive cells were densely populated in mucosa, mostly lamina propria (Fig. 2A, I). Besides, GBP5 positive cells were frequently observed in other layers of the colonic tissue including submucosa, circular muscle, longitudinal muscle, serosa, and mesentery (Fig. 2A, II-VI). Less GBP5 positive cells were observed in colonic layers of the non-inflamed tissue from the same patient (Fig. 2B). It is noteworthy that GBP5 was not detected in muscle cell (Fig. 2A, III, IV) or endothelial cell, but detected in the blood cells (Fig. 2A, VI). Another outstanding observation was the enrichment of GBP5 positive cells in Peyer’s patch (mucosa, Fig. 2A, VII) and lymph node (submucosa, Fig. 2A, VIII). The elevated expression of GBP5 across the colonic layers of CD patients is in line with the fact that CD is usually inflicted by transmural inflammation.
To determine the cellular distribution of GBP5, inflamed colon tissues from CD patients were subjected to immunofluorescence staining for GBP5 and immune cell marker proteins. Confocal microscopy showed that, most of the CD3 positive cells, most of the CD40 positive cells, and the majority of CD68 positive cells expressed GBP5 (Fig. 2C). CD3, CD40 and CD68 are marker proteins for T lymphocyte, antigen presenting cells (dendritic cells, macrophages and B cells)(22) and macrophages, respectively. Thus the above results demonstrated immune cell specific expression of GBP5.
Transcriptome analysis reveals association between GBP5 and inflammatory reaction pathways
To identify potential links between GBP5 and IBD pathogenesis, we performed hierarchical clustering of GBP5 with 102 available cytokine and chemokine genes with the transcriptome dataset generated from colonic mucosa of IBD patients (GSE16879)(21). The clustering result showed that GBP5 shared a similar expression pattern with genes coding for proinflammatory cytokines and chemokines including IL1B and IL6, and they were highly elevated in most of the CD and UC patients compared to the healthy controls (Fig. 3A). Interestingly, the anti-inflammatory cytokine IL10 exhibited a similar expression pattern as GBP5. The elevated IL10 expression in IBD was observed previously and was thought to reflect a futile effort of patient immune system to control the excessive inflammatory reaction(23). Next, we performed Pearson correlation analyses between every two genes based on the transcriptome data (Fig. 3B). With a threshold of correlation coefficient greater than 0.6, 486 genes were correlated with GBP5. In comparison, 226, 123, and 112 genes were correlated with GBP1, GBP2 and GBP4, respectively. Apparently, at the transcription level, GBP5 had the largest impact on IBD among all GBP family genes. We then performed Gene Ontology (GO) enrichment analysis with all 486 genes correlated with GBP5. Top “Biological Process”, “cellular compartment” and “molecular function” are listed in Fig. 3C. The top “Biological Process”, including “leukocyte migration”, “response to molecule of bacterial origin”, “response to lipopolysaccharide” and “cellular response to molecule of bacterial origin” are related to inflammatory responses to bacterial infection. In addition, the top “cellular compartment” and the top “molecular function” such as “collagen-containing extracellular matrix” and “extracellular matrix structural constituent”, are closely related to lymphocyte migration and infiltration. These results are consistent with previous reports that GBP5 plays a critical role in host defense against bacterial pathogens(24, 25).
GBP5 deficiency down-regulates pro-inflammatory chemokines and cytokines in cultured cells
For further understanding of the association between GBP5 and inflammatory processes, we performed GBP5 siRNA knockdown in Jurkat cells to determine the impact of GBP5 on chemokine and cytokine secretion. Diminished GBP5 expression in Western blot analysis indicated efficient knockdown of GBP5 gene in Jurkat cells (Supplementary Fig. S1A). The cell culture supernatants were then subjected to Luminex chemokine and cytokine assay. Compared to cells treated with control RNA, upon stimulation with IFNγ and lipopolysaccharide (LPS), GBP5 siRNA treated cells exhibited decreased levels of CCL2, CCL8, CCL13, CCL25, CXCL10, CXCL11, CXCL12, CXCL16, CX3CL1, IL-1β, IL-10 and MIF in the cell culture supernatant (Supplementary Fig. S1B). In addition, CCL19, CCL20, CCL22, CXCL2 and CXCL13 were not detected in the supernatant after GBP5 knockdown, but detected in the supernatant of the cells treated with control RNA (Supplementary Fig. S1B). GBP5 knockdown also caused a trend of decreased levels in some proinflammatory chemokines and cytokines, including CCL1, CCL11, CCL15, CCL21, CCL26, CCL27, CXCL1, CXCL5, IL-16 and TNF-α, but statistical significance was not achieved (Supplementary Fig. S1B). CCL7, CCL17, CCL23, CXCL6, CXCL9, GM-CSF, IL-2, IL-4 and IL-6 were not detected in any of the samples (Supplementary Fig. S1B).
For a precise evaluation of GBP5 impact on chemokine and cytokine secretion, GBP5 gene was removed from THP-1 cell by CRISPR/CAS9 method and confirmed by Western blot showing no GBP5 expression in GBP5−/− THP-1 cells (Fig. 4A). Global mRNA expression of the GBP5−/− THP-1 cells were assessed by RNA sequencing. As expected, GBP5 mRNA was reduced in GBP5−/− THP-1 cells compared to the wildtype controls, with or without induction by IFNγ plus LPS (Fig. 4B, C). Surprisingly, GBP5 deficiency greatly reduced the mRNA expressions of many inflammation and immune related genes, including (1) IFNγ response genes such as AIM2, CASP1, GSDMD, and IL1B, and (2) IFNγ non-response genes such as NLRP3, CASP11 and IL18 (Fig. 4B).
Comparing the GBP5−/− with wildtype THP-1 cells, many differentially expressed genes, including 2298 genes up-regulated and 2813 genes down-regulated in GBP5−/− cells, were identified (Fig. 4C). The mRNA expression for most proinflammatory cytokines and chemokines were reduced or undetected in GBP5−/− cells (Supplementary Fig. S2). We then performed GO enrichment analysis with the list of genes down-regulated in GBP5−/− cells. The identified top “Biological Process”, “cellular compartment” and “molecular function” are related to inflammatory signaling, immune cell migration, neutrophil activation, secretion of cytokines and chemokines, and other immune and inflammatory events (Fig. 4D).
Using a different approach for functional analysis, gene set enrichment analysis (GSEA) identified a similarly broad suppression of immune and inflammatory functions in GBP5−/− THP-1 cells, including antigen processing, chemokine signaling, cytokine signaling, leukocyte migration and NK cell cytotoxicity (Fig. 4E). Compared to GO enrichment analysis that uses partial information of the biological pathways (in our case, the down-regulated genes), GSEA considers all available information of relevant genes for functional analysis. Thus both methods identified immune and inflammatory pathways as down-regulated pathways in GBP5−/− THP-1 cells.
Next, we examined the altered immune and inflammatory pathways at proteins level. Compared to the wildtype controls, THP-1 cells with GBP5 deficiency exhibited decreased levels of CCL2, CCL8, CCL13, CCL21, CCL25, CCL26, CXCL5, CXCL11, CXCL12, CX3CL1, IL1β, IL-10 and IL16 in the cell culture supernatant (Fig. 5). In addition, CCL1, CCL3, CCL7, CCL11, CCL15, CCL19, CCL20, CCL22, CCL23, CCL24, CCL27, CXCL1, CXCL2, CXCL6, CXCL9, CXCL10, CXCL13, CXCL16, IL-2, IL-4, IL-6, CXCL8, TNF-α and GM-CSF were not detected in the supernatant of GBP5 knockout cells, but detected in the WT controls (Fig. 5). CCL17 was again not detected in any of the samples (Fig. 5). Therefore, GBP5 deficiency led to decreased secretion for most of the inflammatory mediators analyzed.
Given the close relation among GBP5, inflammasomes and IL1β, we examined the intracellular protein expression of some related molecules. Western blots showed that GBP5 deficiency greatly decreased the expression of gasdermin D, caspase 1 and pro-IL1β (Fig. 6). Since the active form for gasdermin D or caspase 1 was not observed in any of the samples, the decreased production of IL1β is not likely due to impaired proteolytic activation by inflammasomes in GBP5 deficiency. Rather, the transcription and expression analysis indicate that the decreased IL1β production is part of the consequence of broad inhibition of inflammatory gene expression in GBP5 deficiency.