Viral infection augments Nod1/2 signaling to potentiate lethality associated with secondary bacterial infections

Abstract

Secondary bacterial infection is a common sequela to viral infection and is associated with increased lethality and morbidity. However, the underlying mechanisms remain poorly understood. We show that the TLR3/MDA5 agonist poly I:C or viral infection dramatically augments signaling via the NLRs Nod1 and Nod2 and enhances the production of proinflammatory cytokines. Enhanced Nod1 and Nod2 signaling by poly I:C required the TLR3/MDA5 adaptors TRIF and IPS-1 and was mediated by type I IFNs. Mechanistically, poly I:C or IFN-β induced the expression of Nod1, Nod2, and the Nod-signaling adaptor Rip2. Systemic administration of poly I:C or IFN-β or infection with murine norovirus-1 promoted inflammation and lethality in mice superinfected with E. coli, which was independent of bacterial burden but attenuated in the absence of Nod1/Nod2 or Rip2. Thus, crosstalk between type I IFNs and Nod1/Nod2 signaling promotes bacterial recognition, but induces harmful effects in the virally infected host.

Figure 1. Poly I:C augments Nod2 activation via TRIF- and IPS-1-dependent signaling pathways in macrophages

(A–D) BMDMs from WT and indicated mutant mice were left untreated (−) or pretreated with LPS (A), pam3CSK4 (A), or poly I:C (A–D) for 24 h and then restimulated with MDP. Cell extracts were collected at the indicated times and assessed for MAPK and NF-κB activation using phosphospecific antibodies. (E) BMDMs from WT and Trif^−/−Ips1^−/− mice were treated with poly I:C or left untreated for 24 h. The macrophages were then re-stimulated with MDP. Cell-free supernatants were analyzed by ELISA for production of IL-6 and TNF-α. (*** p < 0.001, compared with untreated and poly I:C-treated or WT and mutant macrophage cultures). N.D. denotes not detected. Results are representative of 2 or 3 independent experiments. Data are expressed as mean ± S.D.

Figure 2. Enhancement of Nod2 signaling by poly I:C and the chemotherapeutic molecule DMXAA is mediated via type-I IFNs

(A and C) BMDMs from WT and Ifnar1^−/− mice were left untreated (control) or pretreated with poly I:C (A), DMXAA, or IFN-β (C) for 24 h and then restimulated with MDP. Cell extracts were collected at the indicated times and assessed for MAPK and NF-κB activation using phosphospecific antibodies. (B and D) BMDMs from WT, Ifnar1^−/− (B), or Nod2^−/− mice (D) were treated with poly I:C, DMXAA, or left untreated for 24 h and then re-stimulated with MDP. Cell-free supernatants were analyzed for production of IL-6 and TNF-α (*** p < 0.001, compared with WT and mutant macrophage cultures). N.D. denotes not detected. Results are representative of 3 independent experiments. Data are expressed as mean ± S.D.

Figure 3. Type-I IFN signaling induces expression of RIP2

(A) BMDMs from WT mice were treated with poly I:C, DMXAA, or IFN-β for the indicated times, and cell extracts were assessed for RIP2 and α-tubulin levels by immunoblotting. (B) BMDMs from WT, Trif^−/−Ips1^−/−, and Ifnar1^−/− mice were treated with poly I:C, DMXAA, or IFN-β for 24 h, and cell extracts were assessed for RIP2 and α-tubulin levels by immunoblotting. (C) BMDMs from WT mice were treated with poly I:C, DMXAA, or IFN-β for the indicated times, and total RNA was isolated to analyze for Nod1 and Nod2 expression. Gene expression was normalized to that of β-actin. (D) BMDMs from WT, Trif^−/−Ips1^−/−, and Ifnar1^−/− mice were treated with poly I:C, DMXAA, or IFN-β for 8 h, and total RNA was isolated to analyze expression of Nod1 and Nod2. * p < 0.05, *** p < 0.001, comparison between untreated and treated macrophage cultures. Results are representative of 2 or 3 independent experiments.

Figure 4. Viral infection augments Nod1 and Nod2 signaling

(A and B) BMDMs from WT mice were mock-infected (−) or infected with MNV-1 at MOIs of 0.1 and 0.3 (A) or VSV at MOIs of 0.002 and 0.01 (B) for 24 h and then restimulated with MDP. Cell extracts were collected at the indicated times and assessed for MAPK and NF-κB activation using phosphospecific antibodies. (C and D) BMDMs from WT mice were mock-infected (−) or infected with MNV-1 at MOI of 0.3 or VSV at MOI of 0.01 for 24 h. The macrophages were then restimulated with MDP. Cell-free supernatants were analyzed for IL-6 (C) and TNF-α (D) production. (*** p < 0.001, compared with mock-infected and infected cultures). (E and F) BMDMs from WT mice were left untreated (−) or treated with poly I:C, DMXAA, or IFN-β (E) or infected with MNV-1 (F) for 24 hr and then restimulated with the Nod1 ligand KF1B. Cell extracts were collected at the indicated times and assessed for MAPK and NF-κB activation using phosphospecific antibodies. (G) BMDMs from WT mice were infected with MNV-1 for the indicated times, and total RNA was isolated to analyze for the expression of Nod1 and Nod2. (H) WT mice were injected i.p. with thioglycollate for 3 days and then treated with PBS(−) or infected with MNV-1 (1 × 10⁶ PFU/mouse) i. p. Elicited peritoneal macrophages were isolated 24 h after MNV-1 infection and stimulated with MDP ex-vivo. Cell extracts were collected at the indicated time and assessed for MAPK and NF-κB activation using phosphospecific antibodies. Data are expressed as mean ± S.D. Results are representative of 3 independent experiments.

Figure 5. Nod1 and Nod2 contribute to bacteria-induced TNF-α production and NF-κB activation in macrophages infected with MNV-1

(A and B) BMDMs from WT and Nod1^−/−Nod2^−/− mice were left untreated, treated with poly I:C, or infected with MNV-1 for 24 h and then infected with live E. coli (A) or P. aeruginosa (B) at a bacteria/macrophage ratio (B/M) of 1, 5, and 10. Cell-free supernatants were analyzed for production of TNF-α (* p < 0.05, ** p < 0.01, *** p < 0.001, compared with WT and mutant macrophages cultures). (C) BMDMs from WT and Rip2^−/− mice were left untreated, treated with poly I:C, or infected with MNV-1 for 24 h and then infected with live E. coli or P. aeruginosa at a bacterial/macrophage ratio (B/M) of 1. Cell free supernatants were analyzed for production of TNF-α (* p < 0.05, ** p < 0.01, compared with WT and mutant macrophages cultures). (D) BMDMs from WT and Nod1/Nod2^−/− mice were left untreated or treated with poly I:C or infected with MNV-1 for 24 h and then infected with E. coli at bacteria/macrophage ratio of 10 for the indicated times. Cell extracts were assessed for MAPK, NF-κB activation and α-tubulin levels by immunoblotting. Data are expressed as mean ± S.D. Results are representative of 3 independent experiments

Figure 6. Murine norovirus infection promotes bacteria-induced TNF-α production and lethality in mice via Nod1 and Nod2

(A and B) WT, and Nod1^−/−Nod2^−/− mice were injected i.p. with PBS (n=5) or poly I:C (200 μg/mouse, n=10) and then infected with E. coli 3 × 10⁸ CFU/mouse i. p. 24 h after poly I:C administration. Lethality was monitored for 5 days after infection (A). At 3 h post-infection, the levels of TNF-α in serum were assessed by ELISA (n=5 or 10) (B). (C and D) WT, and Nod1^−/−Nod2^−/− mice were injected with PBS or poly I:C (200 μg/mouse) i.p. and then infected with P. aeruginosa 2.5 × 10⁷ CFU/mouse i. n. 24 h after poly I:C administration. Lethality was monitored for 5 days after infection (n=10–11) (C). At 3 hrpost-infection, the level of TNF-α in lung homogenates was assessed by ELISA (n=6) (D) (E–I) WT, Nod1^−/−Nod2^−/−, and Rip2^−/− mice were mock-infected or infected orally with MNV-1 (1 × 10⁷ PFU/mouse) and then infected with E. coli 3 × 10⁸ CFU/mouse i. p. 24 h after the MNV-1 infection. Lethality was monitored for 5 days after E. coli infection (n=9–10) (E). Bacterial loads in lung (F) and blood (G) samples were determined at 8 h post-infection by plating (n=7–8). The short black horizontal lines show the mean values for the groups. Each symbol represents one animal. N.S. denotes not significant. At 3 h post-infection, serum TNF-α levels were assessed by ELISA (n=9–10). N.S. denotes not significant (H). 1 h before E. coli infection, WT mice were injected i.p. with control rat IgG or anti-TNF-α rat IgG and lethality was monitored for 5 days after infection (n=10) (I). Data are expressed as mean ± S.D. Results are representative of 2 or 3 independent experiments.

Figure 7. IFN-β enhances MDP-induced cytokine production in vivo and promotes bacteria-induced lethality via Nod1 and Nod2

(A) WT (n=7), and Nod2^−/− (n=6) mice were injected with PBS or PEGylated murine IFN-β (5 × 10³ U/mouse) i.p. and 18 h later injected with MDP (300 μg/mouse). 3 h later, the levels of IL-6 and CCL2 in serum were assessed by ELISA. (B) WT mice (n=5/group) were injected with PBS or PEGylated murine IFN-β (5 × 10³ U/mouse) i. p. or with unmodified murine IFN-β (2.5 × 10⁴ U/mouse, twice separated by 3 h) s.c. and 18 h later infected with 3 × 10⁸ CFU of E. coli/mouse. 3 h post-injection with E. coli, the level of TNF-α in serum was assessed by ELISA. (C) WT, and Nod1^−/−Nod2^−/− mice were injected with PBS (n=15) or PEGylated murine IFN-β (5 × 10³ U/mouse, n=15) i.p. and 18 h later infected with 3 × 10⁸ CFU of E. coli/mouse i. p. Mouse survival was monitored for 5 days after infection. (D) 3 h post-infection, the level of TNF-α in serum was assessed by ELISA (n=9 or 10).