A conserved role for a GATA transcription factor in regulating epithelial innate immune responses

Abstract

Innate immunity is an ancient and conserved defense mechanism. Although host responses toward various pathogens have been delineated, how these responses are orchestrated in a whole animal is less understood. Through an unbiased genome-wide study performed in Caenorhabditis elegans, we identified a conserved function for endodermal GATA transcription factors in regulating local epithelial innate immune responses. Gene expression and functional RNAi-based analyses identified the tissue-specific GATA transcription factor ELT-2 as a major regulator of an early intestinal protective response to infection with the human bacterial pathogen Pseudomonas aeruginosa. In the adult worm, ELT-2 is required specifically for infection responses and survival on pathogen but makes no significant contribution to gene expression associated with intestinal maintenance or to resistance to cadmium, heat, and oxidative stress. We further demonstrate that this function is conserved, because the human endodermal transcription factor GATA6 has a protective function in lung epithelial cells exposed to P. aeruginosa. These findings expand the repertoire of innate immunity mechanisms and illuminate a yet-unknown function of endodermal GATA proteins.

Fig. 1.

P. aeruginosa accumulation in the worm gut is accompanied by robust gene expression changes in the host. (A) Representative images (×400) of worms (10–20 per group) exposed to PA14-GFP for 4, 12, and 24 h. Asterisks mark the posterior pharyngeal bulb; wedges mark the intestinal lumen boundaries. Yellow signal is autofluorescence of intestinal granules; in addition, a yellow pharyngeal signal of an unknown source appeared consistently at the 12-h time point. (B) Genes differentially expressed during P. aeruginosa infection. Hierarchically clustered expression profiles (rows) for 248 PCR products, corresponding to 232 genes, which are differentially expressed under exposure to PA14 compared with OP50. Data from three independent experiments (columns) are shown for each time point, separated by dotted lines. Vertical bars mark clusters of genes repressed (top) or induced (bottom).

Fig. 2.

Functional consequences of elt-2 expression knockdown. (A) Gene expression changes for elt-2 and C18G1.2 during infection. Means ± SEM for two to three microarray measurements are shown. (B) Survival assays for spe-26 mutants fed with E. coli either expressing elt-2 RNAi (open circles; n = 102) or containing the control empty RNAi vector (filled circles; n = 99) and subsequently transferred (at time 0) to PA14 lawns. Shown for each curve are means ± SD of the fraction of live animals on each of three plates. (C) Lifespan assays for the same experimental groups as above [n = 85 for elt-2(RNAi) animals; n = 92 for controls]. Worms were exposed, after RNAi treatment, to kanamycin-killed OP50-1. (D and E) Faster intestinal accumulation of P. aeruginosa. Representative images (×200) of glp-4;rrf-3 mutants grown on control (D) or elt-2 -RNAi-expressing bacteria (E). Asterisks mark the pharynx.

Fig. 3.

elt-2 knockdown selectively affects the expression of infection-response genes. RNA levels were measured by using qRT-PCR. Excluding D, columns represent means ± SD for three separate experiments. (A and B) Effects on constitutive expression. Shown are normalized fractions of specific RNA levels in elt-2(RNAi) animals relative to control-treated animals. Double asterisks mark significant decreases (P < 0.001; t test). (A) Infection-response genes. (B) Intestine functional and structural markers. (C and D) Effects on induced expression. (C) Infection responses. Normalized fold changes in animals exposed for 24 h to PA14 compared with those exposed to OP50-1. Open columns represent control-treated animals; filled columns represent elt-2(RNAi) animals. Asterisks mark significant changes (P < 0.02; t test). (D) Cadmium responses. Normalized fold changes in animals exposed to 100 μM cadmium for 24 h relative to no-cadmium controls. Columns are as described above. Shown are results of one experiment (means ± SEM). The results presented were obtained in spe-26 worms. Wild-type worms showed similar trends.

Fig. 4.

Consequences of elt-2 knockdown for protection from infection vs. general stress conditions. Survival assays for cdc-25.1-RNAi-sterilized wild-type worms, fed with elt-2-RNAi-expressing bacteria (filled) or control bacteria (open) and subsequently subjected to P. aeruginosa infection (A), oxidative stress (B), cadmium stress (C), or heat shock (D). (A) Killing by P. aeruginosa. Means ± SD of fraction of live animals on each of three plates. Control animals (n = 90); elt-2(RNAi) animals (n = 110). (B) Sensitivity to oxidative stress. Fraction survival as a function of paraquat concentration (6-h exposure) for control and elt-2(RNAi) animals. Shown are means ± SD for experiments performed in duplicate (18–45 animals per point). (C) Sensitivity to cadmium. Means ± SD of the fraction of live animals on each of three plates [control, n = 75; elt-2(RNAi), n = 85] are shown. Cadmium concentration was 100 μM, in which RNAi directed against cdr-1 resulted in a measurable increase in sensitivity (results not shown). (D) Sensitivity to heat stress. Means ± SD of the fraction of live animals on each of three plates subjected to 37°C for 17 h [control, n = 69; elt-2(RNAi), n = 66] are shown.

Fig. 5.

GATA6 is important for protecting human epithelial cells from P. aeruginosa infection. (A) GATA6 is induced upon infection of A549 cells. Fold changes of GATA6 RNA levels in A549 cells, treated with GATA6 RNAi (filled bars), or transfection reagent alone (open bars), and after exposure to PA14 or to PBS alone. Fold changes are over control-treated noninfected cells (set to 1). RNA levels were measured by qRT-PCR. Shown are means ± SD of two experiments. (B) GATA6 knockdown increases cells' susceptibility to infection. Percentage of damaged membrane-permeabilized cells, marked by Trypan blue accumulation, of the total number of cells. Counts were performed on randomly captured images (109–187 cells per image), each of a different well in a six-well plate. Shown are means ± SD for three to five wells, in one of two experiments with similar results. Statistically significant values (t test; P < 0.01) are marked with asterisks.