Deficiency in Galectin-3, -8, and -9 impairs immunity to chronic Mycobacterium tuberculosis infection but not acute infection with multiple intracellular pathogens

Abstract

Macrophages employ an array of pattern recognition receptors to detect and eliminate intracellular pathogens that access the cytosol. The cytosolic carbohydrate sensors Galectin-3, -8, and -9 (Gal-3, Gal-8, and Gal-9) recognize damaged pathogen-containing phagosomes, and Gal-3 and Gal-8 are reported to restrict bacterial growth via autophagy in cultured cells. However, the contribution of these galectins to host resistance during bacterial infection in vivo remains unclear. We found that Gal-9 binds directly to Mycobacterium tuberculosis (Mtb) and Salmonella enterica serovar Typhimurium (Stm) and localizes to Mtb in macrophages. To determine the combined contribution of membrane damage-sensing galectins to immunity, we generated Gal-3, -8, and -9 triple knockout (TKO) mice. Mtb infection of primary macrophages from TKO mice resulted in defective autophagic flux but normal bacterial replication. Surprisingly, these mice had no discernable defect in resistance to acute infection with Mtb, Stm or Listeria monocytogenes, and had only modest impairments in bacterial growth restriction and CD4 T cell activation during chronic Mtb infection. Collectively, these findings indicate that while Gal-3, -8, and -9 respond to an array of intracellular pathogens, together these membrane damage-sensing galectins play a limited role in host resistance to bacterial infection.

Fig 1. Gal-9 binds and recruits to Mtb.

(A) Experimental design for Mtb pull-down mass spectrometry identification of Mtb-binding proteins. (B) Domain organization of Gal-9. CRD, carbohydrate recognition domain. (C) Immunoblot of in vitro binding reactions between indicated pathogens and Gal-9-FLAG THP-1 lysate, probed with anti-FLAG antibody; IN, input; Lm, Listeria monocytogenes; Stm, Salmonella enteria serovar Typhimurium; Cn, Cryptococcus neoformans; Mtb, Mycobacterium tuberculosis. (D) Confocal microscopy of WT BMMs infected with WT or ΔeccC Mtb-GFP (MOI = 2) 8 hours post-infection and immunostained for endogenous Gal-9 and Gal-3. (E) Quantification of Mtb-GFP colocalization with Gal-9 or Gal-3 at indicated time points. Figures represent two independent experiments (D, E). An average of 882 cells were analyzed per technical replicate (D, E). Error bars represent SD from 3 technical replicates. The schematic was created with BioRender.com.

Fig 2. TKO macrophages exhibit reduced autophagic flux.

(A) Immunoblots of bone marrow-derived macrophage lysates from WT and TKO mice (strain 1) probed for indicated proteins. (B) Confocal microscopy of WT or TKO BMMs infected with Mtb-GFP (MOI = 2) 8 hours post-infection and immunostained for polyubiquitin. (C) Quantification of (B) for WT or ΔeccC Mtb-GFP colocalization with polyubiquitin. (D) Quantification of Mtb-GFP colocalization with polyubiquitin in WT or TKO BMMs at indicated time points. (E) Same as in (B) but stained with Lysotracker. (F) Quantification of (E) for WT or ΔeccC Mtb colocalization with Lysotracker signal. (G) Confocal microscopy of WT, TKO, Atg5^fl/fl LysMcre^-/-, or Atg5^fl/fl LysMcre^+/- BMMs infected with Mtb-GFP (MOI = 2) 8 hours post-infection and immunostained for LC3. (H) Quantification of (G) for Mtb-GFP colocalization with LC3. (I) Quantification of Mtb-LC3 colocalization in WT or TKO BMMs infected with Mtb-GFP (MOI = 2) 8 hours post-infection and treated with indicated concentrations of bafilomycin A1 for 3 hours prior to fixation. Figures represent two (D), or are representative of two (I), or four (B, C, E-H) independent experiments. An average of 2115 (B, C), 933 (D), 925 (E, F), 1464 (G, H), and 1636 (I) cells were analyzed per technical replicate. Error bars represent SD from three (G-I) or four (B-F) technical replicates, and *p<0.05., **p<0.01, ***p<0.001, ****p<0.0001 by unpaired t-test.

Fig 3. Gal-3, -8, and -9 are not required for bacterial growth restriction in primary macrophages.

(A and B) Relative light units (RLU) fold change of WT or TKO BMMs infected with Mtb-LUX (MOI = 1, A; MOI = 5, B). (C) Colony-forming unit (CFU) fold change of WT or TKO BMMs infected with WT Mtb (MOI = 2). (D) CFU-fold change of WT or TKO BMMs infected with WT Stm (MOI = 10). (E) RLU of WT or TKO BMMs infected with WT or ΔflaA Lp-LUX (MOI = 0.01). (F) CFU per coverslip of WT or TKO BMMs infected with WT or ΔactA PlcA^H86A PlcB^H69G Lm (MOI = 0.25). Figures represent one (B, E), or two (F, A) independent experiments, or are representative of two (C, D) independent experiments. Error bars represent SD from three (D, F), four (A-C), or five (E) technical replicates, and *p<0.05., **p<0.01: compared to WT BMM + WT bacteria (A, B, E), ^#p<0.05, ^##p<0.01, ^####p<0.0001: compared to WT BMM + mutant bacteria (E, F). p-values determined by unpaired t-test.

Fig 4. TKO mice exhibit modestly impaired resistance to chronic Mtb infection.

(A) WT and TKO female mice were aerosol infected with ~75 CFUs of Mtb Erdman and bacterial loads in lungs were enumerated by plating for CFU at indicated time points. CFU-fold change (relative to day 0 inoculum) from two pooled experiments is shown. n = 10–12 mice per genotype. (B) Same as in (A) but bacterial loads in spleens. (C) Survival of aerosol-infected mice. n = 12–13 mice per genotype. Figures represent one experiment (C) or represent data from two independent experiments (A, B). Bars in (A) and (B) represent the mean, error bars in (A) and (B) represent SEM, and p-values were determined by unpaired t-test (A, B), or Mantel-Cox log (C).

Fig 5. TKO mice exhibit normal bacterial control during infection with Lm or Stm.

(A) WT and TKO mice were intravenously infected with 1 x 10⁵ CFUs of Lm and bacterial loads in spleens were enumerated by plating for CFU at 48 hours post-infection. n = 5 mice per genotype. (B) Same as in (A) but bacterial loads in livers. (C) WT and TKO mice were infected via the intraperitoneal route with 1000 CFUs of Stm and bacterial loads in spleens were enumerated by plating for CFU at 24 hours post-infection. n = 5 mice per genotype. (D) Same as in (C) but bacterial loads in livers. n = 4–5 mice per genotype. Figures represent data from one experiment (A, B), or are representative of three independent experiments (C, D). Bars represent the mean, error bars represent SEM, and p-values were determined by unpaired t-test.