A genome-wide CRISPR screen identifies regulation factors of the TLR3 signalling pathway

Abstract

A subset of TLRs is specialised in the detection of incoming pathogens by sampling endosomes for nucleic acid contents. Among them, TLR3 senses the abnormal presence of double-stranded RNA in the endosomes and initiates a potent innate immune response via activation of NF-κB and IRF3. Nevertheless, mechanisms governing TLR3 regulation remain poorly defined. To identify new molecular players involved in the TLR3 pathway, we performed a genome-wide screen using CRISPR/Cas9 technology. We generated TLR3⁺ reporter cells carrying a NF-κB-responsive promoter that controls GFP expression. Cells were next transduced with a single-guide RNA (sgRNA) library, subjected to sequential rounds of stimulation with poly(I:C) and sorting of the GFP-negative cells. Enrichments in sgRNA estimated by deep sequencing identified genes required for TLR3-induced activation of NF-κB. Among the hits, five genes known to be critically involved in the TLR3 pathway, including TLR3 itself and the chaperone UNC93B1, were identified by the screen, thus validating our strategy. We further studied the top 40 hits and focused on the transcription factor aryl hydrocarbon receptor (AhR). Depletion of AhR had a dual effect on the TLR3 response, abrogating IL-8 production and enhancing IP-10 release. Moreover, in primary human macrophages exposed to poly(I:C), AhR activation enhanced IL-8 and diminished IP-10 release. Overall, these results reveal AhR plays a role in the TLR3 cellular innate immune response.

Keywords: AhR; TLR; TLR3; genetic screen; innate immunity.

Figure 1.

Outline of the genome-wide CRISPR/Cas9 forward genetic screen to identify genes required for TLR3 signalling. (a) Transduction of KBM7 cells with TLR3 cDNA leads to the expression of both full-length (*) and cleaved/mature (**) TLR3. Immunoblot of cell lysates of KBM7 cells complemented by the indicated cDNA was revealed by an anti-TLR3 mAb. As a loading control, tubulin was revealed with a specific mAb. (b) Design of the forward genetic screen in human near-haploid KBM7 cells. Creation of the NF-κB reporter KBM7 cell line by transduction of three independent expression cassettes encoding for dscGFP, TLR3 and Cas9. DscGFP expression was controlled by a NF-κB dependent promoter. Cells were transduced with the lentiviral GeCKO v2 sgRNA library (122,411 sgRNAs). Cells that were successfully transduced were selected with puromycin. After selection, the population was split into two; half was not sorted to represent the entire library, while the remaining population was sorted by flow cytometry after poly(I:C) stimulation to enrich for dscGFP negative cells. After three rounds of poly(I:C) stimulation/enrichment, the DNA from the enriched populations (rounds 1, 2 and 3) was harvested, and enriched sgRNAs were identified by sequencing and compared to an unsorted library. FACS plots obtained after each sort on dscGFP-negative cells show progressive enrichment rates of dscGFP-negative cells. (c) and (d) Proportion of cells responding to poly(I:C) as judged by dscGFP expression before and after sorting during the sequential enrichment process. Polyclonal or clonal cell populations were stimulated with poly(I:C; twice at 35 µg/ml, 4 h apart) or TNF-α (10 ng/ml) for 16 h before sorting at each round. In both cases, the proportion of dscGFP negative cells increased at each round of enrichment. sgRNA: single-guide RNA; NF-κB TRE: NF-κB transcription responsive element; min CMV: minimal CMV; dscGFP: destabilized copepod GFP; SFFV: spleen focus-forming virus promoter; 2A: peptide bond skipping sequence; BFP: blue fluorescent protein: PGK: phosphoglycerate kinase promoter; Hygro: hygromycin resistance gene.

Figure 2.

Genome-wide screen of the TLR3 pathway identifies expected and new genes. (a) –log P value of the results of sgRNA enrichments identified by sequencing, corresponding to 42 genes. After one, two or three rounds of enrichment, the DNA from the enriched cell population was harvested, and enriched sgRNAs were identified by sequencing and comparison to an unsorted library. This determination was carried out for the two independent screens performed in parallel, and sequencing was performed twice generating a total of 12 enriched cell populations. Genes are presented when –log(P value) > 4 for at least 2/12 comparisons. Each dot represents the gene enrichments calculated from one condition. The RSA algorithm was used to identify the significantly enriched genes targeted in the selected cells. Red: known key members of TLR3 pathway; blue: further validated genes. (b) Plot illustrating the hits from the genetic screen. Mean log(P values) determined with the clonal population as a function of those found with polyclonal population. Each dot is a gene. Red squares: known key members of the TLR3 pathway; light blue triangles: genes further validated; grey dots: genes further tested; density colours: remaining genes out of the 42 found in panel (a).

Figure 3.

DLX1, SOD1 and AhR are required for a proper TLR3 response. (a) Heat map of the GFP expression results obtained with KBM7Rep cells transduced with the various sgRNAs (n = 3 independent experiments). KBM7Rep cells transduced with the indicated sgRNAs were exposed to increasing doses of poly(I:C) or TNF-α (10 ng/ml) for 16 h before measurement of the GFP expression by FACS. The cluster of genes containing two sgRNAs targeting UNC93B1 and one targeting TLR3 is magnified. red: positive controls; blue: genes further studied; grey: negative controls. (b) NF-κB activity or (c) IL-8 and (d) IP-10 were measured by FACS or cytometric bead assay (CBA), respectively, in response to poly(I:C) exposure of the indicated cells (two-way ANOVA between one sgRNA condition and ctrl Cas9 ctrl). (e), (f) and (g) The indicated parameters were measured in the absence or presence of TNF-α (n = 3 independent experiments). *, **, *** or ****: significantly different from ctrl Cas9 (Friedman test, with Dunn’s multiple comparison test between sgRNA conditions and Cas9 ctrl). UNC: UNC93B1. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001.

Figure 4.

Cells edited for AhR exhibit a decreased IL-8 and an increased IP-10 response to poly(I:C). (a) Immunoblots of cell lysates of KBM7Rep cells KO for AhR or ctrl Cas9 cells revealed with an anti-AhR mAb or with an anti-actin mAb. (b) NF-κB activity or (c) IL-8 and (d) IP-10 were measured by FACS or CBA, respectively, under the various conditions and cells as indicated, in response to poly(I:C; 25 μg/ml) or TNF-α (10 ng/ml; n = 4 independent experiments). *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001 (two-way ANOVA followed by Sidak’s multiple comparisons test for each panel).

Figure 5.

AhR agonist and antagonist affect IL-8 response to poly(I:C). (a) GFP expression, (b) IL-8 and (c) IP-10 productions were determined in control or AHR KO cells in response to poly(I:C; 25 μg/ml) or TNF-α (10 ng/ml) in the presence of increasing doses of AhR agonist (FICZ) and antagonist (SR-1; n = 4 independent experiments). *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001 (for each curve vs. the appropriate DMSO control, two-way ANOVA).

Figure 6.

Primary myeloid cells response to poly(I:C) is affected by AhR agonist and antagonist. (a) and (b) AhR modulators impact human monocyte derived macrophages (MDMs) response to poly(I:C) and TNF-α. MDMs were treated with FICZ (5 µM), SR-1 (5 µM) or vehicle (DMSO) for 1 h before stimulation with increasing doses of poly(I:C) or TNF-α. (a) IL-8 and (b) IP-10 were measured by CBA (n = 5 independent donors). *P ≤ 0.05; **P ≤ 0.01; ****P ≤ 0.0001 (for each curve vs. the appropriate DMSO control, two-way ANOVA). (c) AhR activity was enhanced by the agonist and inhibited by the antagonist. CYP1A1 mRNA expression of MDM treated with FICZ (5 μM) or SR-1 (5 μM) for 17 h was quantified by quantitative RT-PCR (n = 5 independent donors). *P ≤ 0.05; ****P ≤ 0.0001 (Kruskal–Wallis test with Dunn’s multiple comparisons test).