Mycobacterium abscessus-Induced Granuloma Formation Is Strictly Dependent on TNF Signaling and Neutrophil Trafficking

Abstract

Mycobacterium abscessus is considered the most common respiratory pathogen among the rapidly growing non-tuberculous mycobacteria. Infections with M. abscessus are increasingly found in patients with chronic lung diseases, especially cystic fibrosis, and are often refractory to antibiotic therapy. M. abscessus has two morphotypes with distinct effects on host cells and biological responses. The smooth (S) variant is recognized as the initial airway colonizer while the rough (R) is known to be a potent inflammatory inducer associated with invasive disease, but the underlying immunopathological mechanisms of the infection remain unsolved. We conducted a comparative stepwise dissection of the inflammatory response in S and R pathogenesis by monitoring infected transparent zebrafish embryos. Loss of TNFR1 function resulted in increased mortality with both variants, and was associated with unrestricted intramacrophage bacterial growth and decreased bactericidal activity. The use of transgenic zebrafish lines harboring fluorescent macrophages and neutrophils revealed that neutrophils, like macrophages, interact with M. abscessus at the initial infection sites. Impaired TNF signaling disrupted the IL8-dependent neutrophil mobilization, and the defect in neutrophil trafficking led to the formation of aberrant granulomas, extensive mycobacterial cording, unrestricted bacterial growth and subsequent larval death. Our findings emphasize the central role of neutrophils for the establishment and maintenance of the protective M. abscessus granulomas. These results also suggest that the TNF/IL8 inflammatory axis is necessary for protective immunity against M. abscessus and may be of clinical relevance to explain why immunosuppressive TNF therapy leads to the exacerbation of M. abscessus infections.

Fig 1. Expression of TNF by macrophages during Mabs infection.

(A) Measurement of relative tnf-α expression by qRT-PCR using ef1α as a reference gene in whole embryo over PBS. Around 150 Mabs R or S variants were iv injected and assessed at 2 and 5 dpi. Mean log₁₀ value of three independent experiments with error bars representing the standard error of the mean (SEM). (B-D) Tg(tnf-α:eGFP-F) larvae were injected in the otic cavity with ≈100 Mabs R or S variants (tdTomato) or with PBS. (B) Bright-field and fluorescence overlay microscopy showing the representative expression of tnf close to the injection site at 2 hpi. Scale bars, 100 μm. (C-D) Mean proportion of infected larvae with tnf ⁺ cells (n = 10) (C) and quantification of tnf ⁺ cells per infected larvae (each symbol represents individual embryos and horizontal lines indicate the mean values) (D) in (C) after 2 hpi. The data are representative of two experiments. (E) Distribution of tnfα-expressing cells revealed by the tnf-α:eGFP-F reporter transgene (arrows) in whole Tg(tnf-α:eGFP-F) larvae imaged live at the indicated time points post iv injection of PBS, Mabs R or S (tdTomato, ≈150 colony forming units (CFU)). The yolk (*) is auto-fluorescent. Scale bars, 200 μm. (F-H) Confocal images showing the representative tnf expression in a 3 dpi-granuloma (F) (scale bar, 20 μm), in a 5 dpi-brain abscess (G) (scale bar, 50 μm) or close to a 3 dpi-cord (H) (scale bars, 20 μm) in Tg(tnf-α:eGFP-F) embryos iv infected with Mabs R (tdTomato). (I-J) Confocal microscopy of a 3 dpi-granuloma showing the tnf expression in Tg(tnf-α:eGFP-F/mpeg1:mCherryF) (I) or Tg(tnf-α:eGFP-F/LysC:DsRed) (J) double transgenic embryos iv infected with Mabs R (E2-Crimson). Scale bars, 50 μm. Statistical significance was determined by Kruskal-Wallis test with Dunns post-test (A), Fisher’s exact test of a contingency table (C) or one-tailed unpaired Student’s t test (D).

Fig 2. tnfr morphants are highly susceptible to Mabs infections.

(A-H) WT or tnfr morphants were iv infected with either R (A-H) or S (A-D) variants of Mabs (tdTomato, ≈150 CFU). (A) Survival of infected embryos versus PBS-injected embryos (n = 90, average of three independent experiments). (B-C) Representative fluorescence images of: (B) Bacterial loads (FPC, two independent experiments, horizontal lines indicate the median values); (C) 3 dpi embryos infected by either R or S variants. Scale bars, 200 μm. (D) Proportion of larvae with abscesses after 13 dpi expressed as mean ± SEM from two independent experiments (n = 40–50). (E) Kinetic of R-cord formation in whole infected embryos. Mean ± SEM from three independent experiments (n = 30). (F) Fluorescence microscopy of tnfr morphants exhibiting widespread R-cording in the vasculature and in the CNS at 3 dpi. Scale bar, 200μm. (G-H) Proportion of embryos containing <5 or >5 cords (G) and localization of cords (H) in (E). Error bars represent the SEM. Statistical significance was determined by log-rank test (A), one-tailed Mann-Whitney’s t test (C) or Fisher’s exact test of a contingency table (D-E and G-H).

Fig 3. Inhibition of intramacrophage Mabs growth correlates with TNF-mediated ROS production.

(A-E) WT or tnfr morphant Tg(mpeg1:mCherry-F) embryos were iv infected with ≈150 R or S Mabs expressing tdTomato (A and B) or E2 crimson (C-E). (A) Number of infected macrophages in the caudal hematopoietic tissue (CHT) at 4 hpi. Horizontal lines indicate the mean values. (B) Average proportions of macrophages containing <5, 5 to 10, or >10 bacteria in the CHT at 1 dpi (n = 20 embryos for each group). Top panel shows confocal images of each representative class of infected macrophages. Scale bar, 5 μm. (C-E) CellRox green (green) staining of ROS production in infected embryos. (C) Confocal microscopy of ROS-positive infected macrophages. Scale bars, 10 μm. (D) Average proportions ROS-producing macrophages within the CHT in 1 dpi embryos (n = 30 embryos for each group). Error bars represent the SEM. (E) Representative fluorescence microscopy of 3 dpi-granuloma. Scale bars, 10 μm. Statistical significance was determined by one-tailed unpaired Student’s t test (A) or Fisher’s exact test of a contingency table (B and D). Data are representative of two independent experiments.

Fig 4. M. abscessus infections are associated with a massive neutrophil mobilization.

(A-D) Confocal live imaging of Tg(mpx:eGFP) larvae after iv (A and D) or intramuscular (B and C) infection with ≈100 Mabs R (tdTomato). (A-B) Time-lapse of neutrophils recruitment in caudal vein (A), monitored from 20 mpi to 110 mpi, or into muscle (B) monitored from 10 mpi to 150 mpi. Scale bars, 20 μm. (C) Recruited neutrophils phagocytizing individual mycobacteria. Scale bar, 5 μm. (D) Neutrophils (arrows) and presumptive macrophages (*), containing mycobacteria. Scale bar, 20 μm. (E) Number of infected neutrophils and macrophages counted in CHT at 4 hrs post intravenous infection with both Mabs variants (≈150 CFU, three independent experiments, horizontal lines indicate mean values). M, macrophage; N, neutrophils. Statistical significance was determined by one-tailed unpaired Student’s t test. (F) Confocal imaging of neutrophil-bacteria interactions in the context of R-abscess. Scale bars, 20 μm. See also S1 Movie.

Fig 5. IL8-dependent recruited neutrophils phagocytize Mabs at initial sites of infection.

(A) qRT-PCR measurement of il8 transcripts during Mabs R- or S-infections (≈150 CFU) in whole embryos at 2 and 5dpi. Mean log₁₀ ± SEM of three independent experiments. (B) WT or il8 morphants Tg(mpx:eGFP) embryos were infected into the muscle (left), the hindbrain (middle) or the otic vesicle (right) with ≈100 R or S Mabs. Number of recruited neutrophils in injection sites at 3 hpi (two independent experiments, horizontal lines indicate the median values). (C) Microscopy showing representative neutrophil recruitment to R-abscesses in the brains of WT versus il8 morphants Tg(mpx:eGFP) larvae at 5 dpi. Scale bar, 100 μm. (D and E) WT or il8 morphants were iv infected with either R or S (tdTomato, ≈150 CFU). (D) Survival of R-(left) or S-infected (right) embryos (n = 50–60, average of two independent experiments). (E) Representative images of infected larvae at 3 dpi. Scale bars, 200 μm. (F and G) WT or csf3r morphants were iv infected with either R or S expressing tdTomato (≈150 CFU). (G) Survival of R-(left) or S-infected (right) embryos (n = 50–60, two independent experiments). (H) Representative images of infected larvae at 3 dpi. Scale bars, 200 μm. Statistical significance was determined by Kruskal-Wallis test with Dunns post-test (A), one-tailed unpaired Student’s t test (B) or by log-rank test (D and F).

Fig 6. TNF signaling is required for neutrophil recruitment.

(A-D) WT or tnfr morphant Tg(mpx:eGFP) larvae were injected in the otic cavity (A-B) or hindbrain (C-D), with both variants of Mabs (tdTomato, ≈100 CFU, two or three experiments) and imaged at 3 hpi. Representative images (A and C) and number (B and D) (horizontal lines indicate the mean values) of neutrophils at the infection site. Scale bars, 100 μm. (E and F) Confocal image of neutrophils surrounding a 3 dpi R-cord (E) (Scale bars, 50 μm) or a 5 dpi R-abscess (F) (Scale bars, 100 μm) in iv infected WT or tnfr morphant Tg(mpx:eGFP) embryos. Statistical significance was determined by one-tailed Student’s t test (B and D). See also S2–S5 Movies.

Fig 7. Recruitment of neutrophils is crucial for granuloma development.

(A and B) WT or tnfr morphants were iv infected with R Mabs (tdTomato, ≈150 CFU). Kinetics of granuloma formation (A) and mean ± SEM number of granuloma per infected embryos (B) in (A); (n = 30–45, three independent experiments). The top panel shows confocal images of representative granuloma. Scale bars, 100 μm. (C) Confocal images showing a representative 3 dpi granulomas in WT versus tnfr1 morphant Tg(mpeg:mcherry-F) embryos iv infected with S expressing Wasabi. Arrows indicate extracellular aggregates. Scale bars, 15 μm. (D) Confocal microscopy showing representative granulomas in WT and tnfr1 morphant Tg(mpx:eGFP) embryos iv infected with R expressing tdTomato at 4 dpi. Arrows indicate extracellular cords. Scale bars, 30 μm. (E-F) Confocal microscopy monitored WT or tnfr morphant Tg(mpx:eGFP) embryos that were iv infected with R (tdTomato, ≈150 CFU). (E) Representative kinetics of neutrophil mobilization during granuloma formation. Scale bars, 20 μm. (F) Number of neutrophils recruited to WT (up) or TNFR1-depleted (down) granuloma as a function of granuloma volume. Statistical significance was determined by Fisher’s exact test of a contingency table (A), one-tailed unpaired Student’s t test (B) or Pearson correlation (F). See also S6 and S7 Movies.

Fig 8. IL8-dependent mobilization of neutrophils is crucial for controlling M. abscessus infection and elaborating granulomas.

(A) qRT-PCR of il8 (normalized to ef1α) upon Mabs infection in R Mabs iv infected WT, tnfr morphants or LipoC embryos (≈150 CFU). Fold induction to PBS- injected animals at 3 dpi. Data are mean ± SEM of two independent experiments. (B and C) WT, LipoC embryos (B) or tnfr morphants (C) Tg(mpx:eGFP) embryos were infected into the otic vesicle with ≈150 R Mabs and treated with IL8 injection. Mean number of recruited neutrophils into the otic cavity in response to Mabs/IL8 injection at 3 hpi. Each symbol represents individual embryos and horizontal lines indicate the mean values. (D-F) WT or tnfr morphants were injected into the otic cavity with either ≈150 R or S variants expressing tdTomato and treated with IL8 injection. (D) Survival of R-(up) or S-infected (down) embryos (n = 50–60, two independent experiments). (E-F) Bacterial loads (FPC, two independent experiments, horizontal lines indicate the mean values) (E) and representative confocal images (F) of 3 dpi embryos. Scale bars, 200 μm. (F and G) WT, il8 or csf3r morphants were iv infected with R Mabs (tdTomato, ≈150 CFU). Kinetics of granuloma formation (G) and number of granuloma per infected embryos (H) assessed at 3 dpi. Mean ± SEM from two independent experiments (n = 40). (I) Number of neutrophils recruited to WT or IL8-depleted granulomas as a function of the granuloma volume at 2 dpi. Statistical significance was determined by Kruskal-Wallis test with Dunns post-test (A and H), ANOVA with Tukey post-test (B, C and E), log-rank test (D), Fisher’s exact test of a contingency table (G) or Spearman correlation (I).

Fig 9. IL8- and TNF-dependent neutrophil mobilization in Mabs granuloma formation and immunoprotection.

The upper panel summarizes the contribution of the TNF/IL8 axis in the early and late response to Mabs infection leading to a controlled and chronic infection. The lower panel illustrates the detrimental effects of impaired TNF signaling that, through disruption of the IL8 production and associated defect in neutrophil mobilization, leads to impaired granuloma formation, acute infections and larval death.