Microbiota produced indole metabolites disrupt host cell mitochondrial energy production and inhibit Cryptosporidium parvum growth

Abstract

Cryptosporidiosis is a leading cause of life-threatening diarrhea in young children in resource-poor settings. Susceptibility rapidly declines with age, associated with changes in the microbiota. To explore microbial influences on susceptibility, we screened 85 microbiota- associated metabolites enriched in the adult gut for their effects on C. parvum growth in vitro. We identified eight inhibitory metabolites in three main classes: secondary bile salts/acids, a vitamin B ₆ precursor, and indoles. Growth restriction of C. parvum by indoles did not depend on the host aryl hydrocarbon receptor (AhR) pathway. Instead, treatment impaired host mitochondrial function and reduced total cellular ATP, as well as directly reduced the membrane potential in the parasite mitosome, a degenerate mitochondria. Oral administration of indoles, or reconstitution of the gut microbiota with indole producing bacteria, delayed life cycle progression of the parasite in vitro and reduced severity of C. parvum infection in mice. Collectively, these findings indicate that microbiota metabolites contribute to colonization resistance to Cryptosporidium infection.

Fig. 1:. Gut metabolites, specifically secondary bile acids and indoles, inhibit C. parvum infection in vitro.

a) Effects of 85 intestinal metabolites at 1 mM (circles) or 0.1 mM (squares) on C. parvum (Cp) infection in HCT-8 cells 24 hpi. Data plotted represents mean ± S.D. of Cp or mean host cell numbers (blue line) relative to PBS control for six independent experiments. Differences between Cp numbers for each metabolite and the PBS control were analyzed using a one-way ANOVA followed by a Dunnett’s test for multiple comparisons. Metabolites that significantly inhibited Cp growth are indicated in red. *P < 0.05, ***P < 0.001. b) Chemical structures of five inhibitory metabolites with their respective EC50 values for Cp and host cells and fold selectivity (host EC50 divided by Cp EC50). EC50 values were calculated using a nonlinear regression curve fit with six replicates (three technical replicates from two independent experiments) per concentration. c) Screen of indole analogs (1 mM) modified at the 3- (teal), 4- (pink), 5- (orange), 6- (green), or 7-carbon (purple) positions and their effects on Cp infection in HCT-8 cells. Data plotted represents mean ± S.D. of six replicates (three technical replicates from two independent experiments). Differences between mean Cp numbers for each metabolite and the DMSO control were analyzed using a one-way ANOVA followed by a Dunnett’s test for multiple comparisons. ***P < 0.001. ILA = indole-3-lactic acid; I3AM = indole-3-acetamide; I3S = indoxyl-3-sulfate; IAA = indole-3-acetic acid; IPA = indole-3-propionic acid; Trypt = tryptamine; I3ACN = indole-3-acetonitrile; MeI = methylindole; CNI = cyanoindole; AI = aminoindole; HI = hydroxyindole; MeOHI = methoxyindole.

Fig 2.. Indoles do not inhibit C. parvum through the host AhR pathway.

a) Ratio of C. parvum (Cp) relative to DMSO control in HCT-8 cells after 24 h treatment with serial dilutions of AhR agonists. Starting concentrations (1X) were 1 μM for VAF347 and FICZ and 1 mM for kynurenic acid, indole and 4-methylindole. Data plotted represents mean ± S.D. of nine replicates (three technical replicates from three independent experiments). Differences between mean Cp ratio in 1X or 0.5X treated cultures versus 0.25X-treated cultures for each compound were analyzed with a two-way ANOVA followed by a Dunnett’s test for multiple comparisons. ***P < 0.001. b) Fold change in gene expression of human AHRR or CYP1A1 genes (normalized to GAPDH) in uninfected HCT-8 cultures after 24 h treatment with VAF347 (250 nM), indole (1.5 mM) or 4-hydroxyindole (4HI, 2.5 mM) relative to 1% DMSO control. Data plotted represents mean ± S.D. of 3 – 4 technical replicates from a single experiment. For each gene, differences between fold change in gene expression for each treatment versus DMSO control were analyzed with a one-way ANOVA followed by a Dunnett’s test for multiple comparisons. ***P < 0.001. c) Fold change in gene expression of human CYP1A1 genes (normalized to GAPDH) in uninfected HCT-8 AhR WT (gray) or KO cell lines (blue) after 24 h treatment with VAF347 (500 nM) relative to 1% DMSO control. Data plotted represents mean ± S.D. of three technical replicates from a single experiment. Differences between fold change in gene expression for each AhR KO versus AhR WT cell line for each treatment were analyzed with a two-way ANOVA followed by a Dunnett’s test for multiple comparisons. ***P < 0.001. d) Ratio of Cp relative to DMSO control in HCT-8 AhR WT (gray) or KO cell lines (blue) after 24 h treatment with 0.5% DMSO, indole (1 mM), 4HI (1 mM) or VAF347 (500 nM). Data plotted represents mean ± S.D. of six replicates (three technical replicates from two independent experiments). Differences between Cp ratio in each AhR KO versus AhR WT cell line for each treatment were analyzed with a two-way ANOVA followed by a Dunnett’s test for multiple comparisons. **P < 0.01.

Fig. 3:. Indoles delay C. parvum life cycle progression.

a) Ratio of C. parvum (Cp) numbers relative to DMSO control in HCT-8 cells after treatment with 1% DMSO or EC90 concentrations of indole (880 μM) or 7-cyanoindole (7CNI, 500 μM) for the indicated hours post infection (hpi). Data plotted represents mean ± S.D. of six replicates (three technical replicates from two independent experiments). Differences between mean Cp ratio in indole or 7CNI-treated cultures vs in the DMSO control at each time point were analyzed with a two-way ANOVA followed by a Dunnett’s test for multiple comparisons. *P < 0.05, ***P < 0.001. b) Total number of Cp in HCT-8 cultures treated with 1% DMSO or EC90 concentrations of indole or 7CNI for the indicated hours post infection. Data plotted represents mean ± S.D. of three independent experiments (same experiments as in c). Differences between mean Cp numbers in indole or 7CNI-treated cultures vs in the DMSO control at each time point were analyzed with a two-way ANOVA followed by a Dunnett’s test for multiple comparisons. **P < 0.01, ***P < 0.001. c) Ratio of the number of Cp in the trophozoite, early meront, middle meront or late meront stages in infected HCT-8 cultures treated with 1% DMSO or EC90 concentrations of indole or 7CNI at the indicated hours post infection. N = the number of nuclei per parasite. Data plotted represents mean ± S.D. of three independent experiments. d) Immunofluorescence images of Cp in HCT-8 cultures treated with 1% DMSO or EC90 concentrations of indole or 7CNI 22 hpi. Parasites are labeled with membrane marker 1E12 (green) and a general Cp antibody Pan Cp (red). Nuclei are stained with Hoechst. Scale bar = 3 μm. e) Washout experiments in Cp-infected air-liquid interface (ALI) cultures treated with 1% DMSO or indole at EC50 (577 μM), EC90 (1894 μM) or 2 x EC90 (3788 μM); or 7CNI at EC50 (379 μM), EC90 (688 μM) or 2 x EC90 (1376 μM) for 48 h before washout. Cp genome eq. were normalized to the DMSO control at each time point. Data plotted represents mean ± S.D. of six replicates (three technical replicates from two independent experiments). Differences between mean % Cp after washout versus mean % Cp at time of washout (2 dpi) for each indole concentration were analyzed using a two-way ANOVA followed by a Dunnett’s test for multiple comparisons. **P < 0.01, ***P < 0.001.

Fig 4.. Indole induces ER stress and transporter upregulation in HCT-8 cells.

a) Volcano plot of fold-change vs P-value after GSA analysis of indole vs DMSO-treated HCT-8 cells highlighting genes significantly (P < 0.05) upregulated (red) or downregulated (blue) after indole treatment by >2-fold. b) Hierarchical clustering analysis of the 30 most differentially regulated genes (FDR-corrected P < 1×10 ) between indole and DMSO-treated HCT-8 cells. c,d) Gene Ontology (GO) pathway analysis performed in Enrichr using genes significantly upregulated after c) 4 h or d) 12 h of indole treatment as input. Upregulated genes associated with each pathway are listed to the right of the bar graph. e) Ratio of C. parvum (Cp) relative to DMSO control in HCT-8 cells after 24 h treatment with a serial dilution of indole in growth medium supplemented with 1 mM tryptophan (Trp) or 1 mM Trp plus 1 mM phenylalanine (PHE). EC50 values were calculated for each medium using a nonlinear regression curve fit with six replicates (three technical replicates from two independent experiments) per indole concentration.

Fig 5.. Indole impairs host mitochondrial ATP production.

a) Metabolic analysis using the Seahorse XF Cell Mito Stress Test kit on HCT-8 cells treated for 18 h with 1% DMSO or indole (0.5 mM, 1 mM or 2 mM). Data calculated as a percentage of the oxygen consumption rate (OCR) for each well relative to the mean basal OCR of DMSO control cells for that experiment. Spare respiratory capacity = maximal respiratory rate – basal respiratory rate for each well. Data plotted represents mean ± S.D. of 12 replicates (six technical replicates from two independent experiments). Differences between % OCR for each indole concentration vs the DMSO control for each measurement were analyzed with a two-way ANOVA followed by a Dunnett’s test for multiple comparisons. *P < 0.05, ***P < 0.001. b) Metabolic analysis using the Seahorse XF Real-time ATP Rate assay on HCT-8 cells treated for 18 h with 1% DMSO or indole (0.5 mM, 1 mM or 2 mM). Data plotted represents mean ± S.D. of ATP production rate (pmol/min) produced by glycolysis, the mitochondria, or total ATP (glycolysis + mitochondrial ATP rates) for 12 replicates (six technical replicates from two independent experiments). For each source of ATP, differences between ATP production rate for each indole concentration vs the DMSO control were analyzed with a two-way ANOVA followed by a Dunnett’s test for multiple comparisons. **P < 0.01, ***P < 0.001. c) Metabolic analysis using the Seahorse XF Cell Energy Phenotype Test kit on HCT-8 AhR WT cells (gray) or AhR KO cells (blue) treated for 18 h with 1% DMSO or indole (0.5 mM, 1 mM or 2 mM). Data calculated as a percentage of OCR for each well relative to the mean basal OCR of DMSO control cells for that experiment. Data plotted represents mean ± S.D. of 12 replicates (six technical replicates from two independent experiments). For each cell line, differences between % OCR for each indole concentration vs the DMSO control were analyzed with a one-way ANOVA followed by a Dunnett’s test for multiple comparisons. *P < 0.05, **P < 0.01, ***P < 0.001. d) Ratio of C. parvum (Cp) relative to DMSO control in HCT-8 cells after 24 h treatment with serial dilutions of mitochondrial Complex I and III inhibitors rotenone and antimycin A, respectively (Rot/AA); ATP synthase inhibitor oligomycin; or proton gradient uncoupler carbonyl cyanide m-chlorophenyl hydrazone (CCCP). Inhibition curves were calculated for each compound using a nonlinear regression curve fit with six replicates (three technical replicates from two independent experiments) per concentration.

Fig 6.. Exogenous indole treatment, or reconstitution with indole-producing bacteria, suppresses C. parvum infection in GKO mice.

a) GKO mice were treated twice daily by gavage with vehicle (10% DMSO in water), or 50 mg/kg indole or 7-cyanoindole (7CNI) for 7 days. Cp oocysts numbers were quantified from a single fecal pellet per mouse collected 3, 5, 7 and 9 dpi. All data plotted represents 7 mice per treatment group sampled over time from two independent experiments. b) Cp oocysts per mg feces for each mouse at the indicated dpi. Statistical analyses comparing each treatment group to vehicle control on individual days were performed using two-tailed Mann-Whitney U tests. *P < 0.05. c) Percent of original body weight plotted as mean ± S.D. Statistical analysis performed using a mixed-effects model with a Geisser-Greenhouse correction for matched values, followed by a Dunnett’s test for multiple comparisons. *P < 0.05. d) Combined survival curves of all 7 mice for the first 10 days then for the 4 mice from the second experiment for days 11 – 30. e) GKO mice were treated with antibiotic to suppress endogenous flora and then reconstituted with WT B. theta or ∆tnaA B. theta followed by challenge with Cp. Oocysts numbers were quantified from fecal pellets collected 3, 5, 7 and 9 dpi. All data plotted represents 4 mice per treatment group sampled over time. f) Estimation of bacterial burdens in the gut by 16S rRNA qPCR. Means ± S.D. Statistical analyses comparing vehicle to WT B. theta group (shown in blue asterisk) or ∆tnaA B. theta group (shown in red asterisk) on individual days were analyzed with a two-way ANOVA followed by a Dunnett’s test for multiple comparisons. *P < 0.05, **P < 0.01. g) Relative abundance of Bacteroides at the genus level from each mouse at the indicated dpi. h) Cp oocysts per mg feces for each mouse at the indicated dpi. Statistical analyses comparing WT B. theta group to ∆tnaA B. theta group on individual days were performed using two-tailed Mann-Whitney U tests. *P < 0.05.