Protective Role of MAVS Signaling for Murine Lipopolysaccharide-Induced Acute Kidney Injury

Abstract

Despite treatment advances, acute kidney injury (AKI)–related mortality rates are still high in hospitalized adults, often due to sepsis. Sepsis and AKI could synergistically worsen the outcomes of critically ill patients. TLR4 signaling and mitochondrial antiviral signaling protein (MAVS) signaling are innate immune responses essential in kidney diseases, but their involvement in sepsis-associated AKI (SA-AKI) remains unclear. We studied the role of MAVS in kidney injury related to the TLR4 signaling pathway using a murine LPS-induced AKI model in wild-type and MAVS-knockout mice. We confirmed the importance of M1 macrophage in SA-AKI through in vivo assessment of inflammatory responses. The TLR4 signaling pathway was upregulated in activated bone marrow–derived macrophages, in which MAVS helped maintain the LPS-suppressed TLR4 mRNA level. MAVS regulated redox homeostasis via NADPH oxidase Nox2 and mitochondrial reverse electron transport in macrophages to alleviate the TLR4 signaling response to LPS. Hypoxia-inducible factor 1α (HIF-1α) and AP-1 were key regulators of TLR4 transcription and connected MAVS-dependent reactive oxygen species signaling with the TLR4 pathway. Inhibition of succinate dehydrogenase could partly reduce inflammation in LPS-treated bone marrow–derived macrophages without MAVS. These findings highlight the renoprotective role of MAVS in LPS-induced AKI by regulating reactive oxygen species generation-related genes and maintaining redox balance. Controlling redox homeostasis through MAVS signaling may be a promising therapy for SA-AKI.

Introduction

Acute kidney injury (AKI) is a life-threatening disease, and the prevalence of AKI is increasing worldwide (1). Sepsis is a major underlying cause, responsible for 40–70% of all AKI cases in critically ill patients (2), and sepsis-associated AKI (SA-AKI) carries a high risk of mortality and morbidity. The adjusted odds ratio for in-hospital mortality is reportedly 6.88 in stage 3 AKI patients compared with those without AKI (3). Therefore, it is important to understand the detailed pathophysiology and develop new treatments to improve prognosis.

Various factors contribute to the development of SA-AKI, including altered hemodynamics, hypercytokinemia, and bacterial toxins (4). LPS, a component of Gram-negative bacteria, is particularly important in SA-AKI, as it induces cellular inflammation and damage through TLR4 signaling (5, 6). In the kidney, TLR4 is expressed on tubular epithelial cells, endothelial cells, and immune cells such as macrophages (7). Consistent with these findings, LPS-induced AKI shows tubular injury, endothelial injury, and macrophage activation (8–10).

The retinoic acid–inducible gene-I (RIG-I)–mitochondrial antiviral signaling protein (MAVS) pathway contributes to antiviral innate immunity in response to RNA virus (11, 12), as well as various inflammatory diseases such as systemic lupus erythematosus, hepatic steatosis, and diabetic kidney disease (13–15). Recent studies demonstrated that LPS stimulation also regulates the MAVS pathway in endothelial cells and macrophages (16, 17). Although MAVS signaling contributes to the pathophysiology of kidney diseases (15, 18), its association with LPS stimulation remains to be investigated. Therefore, we explored the hypothesis that MAVS signaling is involved in LPS-induced AKI. To test this hypothesis, we administered LPS to MAVS-knockout (KO) mice and analyzed kidney injury, focusing on the role of macrophages.

Materials and Methods

Mice

C57BL/6J (B6) mice and MAVS-KO mice were purchased from Charles River Japan (Tokyo) and The Jackson Laboratory (Bar Harbor, ME), respectively. As described previously, MAVS-KO mice were generated and maintained on a B6 background (15). These mice were all bred and housed at Kanazawa University, where all animal experiments were conducted in accordance with Kanazawa University’s animal care regulations and authorized by the Institute for Experimental Animals at the Kanazawa University Advanced Science Research Center (registration no. AP-194072).

LPS-induced kidney injury model

Male mice 6–12 wk old were i.p. injected with LPS (Escherichia coli O111:B4 L2630, Sigma-Aldrich) at a lethal dose (20 mg/kg), and an equal volume of PBS was i.p. injected into the sham group.

Mice were kept separately in metabolic cages to collect their urine. Urine samples were obtained from metabolic cages 16–24 h after LPS injection. Mice were anesthetized to collect kidneys and blood for further analysis.

Renal functional assays

We measured urinary albumin and creatinine levels, then calculated proteinuria as the ratio of urinary albumin concentration (in milligrams) to creatinine concentration (in grams) as previously described (15).

Blood was collected via the inferior vena cava or retro-orbital sinus and separated into serum by centrifugation (1300 × g, 15 min). Subsequently, serum blood urea nitrogen (BUN) was determined by Oriental Yeast (Tokyo, Japan).

ELISAs

The mouse IL-6 DuoSet ELISA kit (R&D Systems, DY206-05) was used to determine IL-6 levels, following the manufacturer’s instructions. For measurements in mice, blood samples were collected from the inferior vena cava of mice and centrifuged at 10,000 × g at 4°C for 10 min. The pellet was then discarded, and diluted supernatants (serum) were used for IL-6 quantitation. To evaluate secreted IL-6 after LPS treatment, ELISAs were conducted using cell culture supernatants of bone marrow (BM)–derived macrophages (BMDMs) 24 h after stimulation.

Immunohistochemical analyses

After removing the renal capsule, half-kidneys were immediately snap-frozen at −80°C in Tissue-Tek OCT compound (Sakura Finetek) for cryosectioning, whereas the other half-kidneys were immediately fixed in 10% neutral buffered formalin, then embedded in paraffin.

The paraffin-embedded tissue or frozen blocks were cut into 2- to 3-μm sections and blocked by protein block (serum free) (Dako, X0909) or 1% BSA/PBS blocking solution for 15 min at room temperature (RT) before staining with primary Ab at 4°C overnight. Prior to primary Ab incubation, the paraffin-embedded samples required deparaffinization, Ag retrieval with proteinase K (Dako, S3020), and peroxidase blocking with peroxidase blocking solution (Dako, S2023).

Specific primary Abs used for staining were goat polyclonal anti-mouse kidney injury marker 1 (KIM-1) (1:500) (R&D Systems, AF1817) and purified rat anti-mouse Ly6G and Ly6C (1:100) (BD Pharmingen, 550291). Primary Ab incubation was followed by a 1-h incubation of an appropriate secondary Ab incubation at RT. A Histofine simple stain mouse MAX-PO (rat) kit (Nichirei Bioscience, 414311F) was used for Gr-1 staining; Ag was then revealed by Histofine diaminobenzidine substrate kit (Nichirei Biosciences, 425011). Rabbit anti-goat IgG H&L (Alexa Fluor 594) (Abcam, ab150144) was used for KIM-1 staining. Fluorescent staining required light avoidance, and nuclei counterstaining was achieved with Vectashield antifade mounting medium with DAPI (Vector Laboratories, H-1200-10). Finally, the number of KIM-1⁺ and Gr-1⁺ cells were then counted in 10–15 different fields of each specimen and expressed as the median number with 95% confidence intervals (95% CIs) per field.

Flow cytometry analyses of the leukocyte population

Single-cell suspensions were isolated from kidneys, peripheral blood (retro-orbital sinus), and BM of B6 and MAVS-KO mice, then stained and analyzed by flow cytometry as described previously (15, 19). Abs for surface staining, including anti-mouse CD45-PE (103106), CD11b-FITC (101206), Ly6G allophycocyanin-Cy7 (127624) (BD Biosciences), as well as F4/80 (BM8) PE-Cy7 (25-4801-82), Ly6C (HK1.4), and PerCP-Cy5.5 (45-5932-82) (eBioscience), were incubated at 4°C for 30 min. After washing, stained cells were resuspended in PBS, treated with propidium iodide (Immunostep Biotech), and run on FACSymphony A5 (BD Biosciences). The results were analyzed using FlowJo analysis software (Tree Star).

Primary cell culture

BMDMs were isolated from femurs of B6 or MAVS-KO mice and cultured in RPMI 1640 (Nacalai Tissue) containing 20–25% L929 (CCL-1) (American Type Culture Collection) conditioned medium, 10% FBS (Sigma-Aldrich), and 1% penicillin/streptomycin (Fujifilm) as described (15, 20). After 5–6 d, purified BMDMs were collected and seeded in 6- or 12-well plates for various assays.

Mouse primary tubular epithelial cells (pTECs) were isolated and cultured as described previously (21).

Gene expression analyses

We isolated total cellular RNA from mouse kidney tissue, primary macrophage cells, or pTECs using the ISOSPIN cell and tissue RNA kit (Nippon Gene, 314-08211). Total extracted RNA was then converted into single-stranded cDNA using high-capacity cDNA reverse transcription kits (Applied Biosystems, 4368813), following the manufacturer’s instructions. To analyze the expression of genes of interest, we conducted real-time quantitative PCR (qPCR) analysis based on iQ SYBR Green supermix (Bio-Rad, 1708885) and the ViiA 7 real-time PCR system (Thermo Fisher Scientific) using a previously described protocol (22). The sequences of primers synthesized by FASMAC are listed in Table I. ACTB was used as a control. Data were analyzed using the ΔΔCt method to calculate the relative gene expression of target genes with ACTB expression.

Table I. Primer sequences used for qPCR.

Primer	Sequence (5′→3′)
ACTB F	AGCCATGTACGTAGCCATCC
ACTB R	CTCTCAGCTGTGGTGGTGAA
Cat F	GGACGCTCAGCTTTTCATTC
Cat R	TTGTCCAGAAGAGCCTGGAT
c-fos F	CGGGTTTCAACGCCGACTA
c-fos R	TTGGCACTAGAGACGGACAGA
c-jun F	TGGGCACATCACCACTACAC
c-jun R	TCTGGCTATGCAGTTCAGCC
Ccl5 F	AGATCTCTGCAGCTGCCCTCA
Ccl5 R	GGAGCACTTGCTGCTGGTGTAG
Cxcl3 F	TGAGACCATCCAGAGCTTGACG
Cxcl3 R	CCTTGGGGGTTGAGGCAAACTT
GPx1 F	GGTTCGAGCCCAATTTTACA
GPx1 R	CCCACCAGGAACTTCTCAAA
HIF-1α F	TGCAGCAAGATCTCGGCGAA
HIF-1α R	AGTGGCAACTGATGAGCAAGC
IFNb F	CCCTATGGAGATGACGGAGA
IFNb R	TCCCACGTCAATCTTTCCTC
IL-1β F	GGATGAGGACATGAGCACCT
IL-1β R	AGGCCACAGGTATTTTGTCG
IL-6 F	GAGGATACCACTCCCAACAGACC
IL-6 R	AAGTGCATCATCGTTGTTCATACA
KC (Cxcl1) F	CTGGGATTCACCTCAAGAACATC
KC (Cxcl1) R	CAGGGTCAAGGCAAGCCTC
MCP-1 F	CTTCCTCCACCACCATGCA
MCP-1 R	CCAGCCGGCAACTGTGA
MIP-2 (Cxcl2) F	CCAACCACCAGGCTACAGG
MIP-2 (Cxcl2) R	GCGTCACACTCAAGCTCTG
mt-ATP6 F	CCATAAATCTAAGTATAGCCATTCCAC
mt-ATP6 R	AGCTTTTTAGTTTGTGTCGGAAG
mt-Co1 F	CAGACCGCAACCTAAACACA
mt-Co1 R	TTCTGGGTGCCCAAAGAAT
mt-Co2 F	GCCGACTAAATCAAGCAACA
mt-Co2 R	ATGGCCTACCCATTCCAACT
mt-Cytb F	CATTCTGAGGTGCCACAGTT
mt-Cytb R	GATGAAGTGGAAAGCGAAGA
mt-ND1 F	ACACTTATTACAACCCAAGAACACAT
mt-ND1 R	TCATATTATGGCTATGGGTCAGG
mt-ND2 F	CCATCAACTCAATCTCACTTCTATG
mt-ND2 R	GAATCCTGTTAGTGGTGGAAGG
mt-ND4 F	GCTTACGCCAAACAGAT
mt-ND4 R	TAGGCAGAATAGGAGTGAT
MyD88 F	AGAGCTGCTGGCCTTGTTAG
MyD88 R	CGAAAAGTTCCGGCGTTTGT
NGAL F	GGACCAGGGCTGTCGCTACT
NGAL R	GGTGGCCACTTGCACATTGT
Nox2 F	CGCCCTTTGCCTCCATTCTC
Nox2 R	CCTTTCCTGCATCTGGGTCTCC
Rab10 F	TGGAACTACAAGGAAAGAAGAT
Rab10 R	AGTAGGAGGTTGTGATGGT
Rab11a F	TGGGAAAACAATAAAGGCACAGA
Rab11a R	ATGTGAGATGCTTAGCAATGTCA
Rab5 F	GCTAATCGAGGAGCAACAAGAC
Rab5 R	CCAGGCTTGATTTGCCAACAG
Rab7b F	CGAGGAATACCAGACCACACT
Rab7b R	GGCTGGCCAGAACCTCAAAGG
RelA F	TGTGGAGATCATCGAACAGCCG
RelA R	TGCTTCTCTCGCCAGGAATAC
Sdha F	GGAACACTCCAAAAACAGACCT
Sdha R	CCACCACTGGGTATTGAGTAGAA
Sdhb F	AATTTGCCATTTACCGATGGGA
Sdhb R	AGCATCCAACACCATAGGTCC
Sod2 F	TGGACAAACCTGAGCCCTAAG
Sod2 R	CCCAAAGTCACGCTTGATAGC
TLR4 F	TTCAGAACTTCAGTGGCTGGAT
TLR4 R	GTCTCCACAGCCACCAGATT
TNF-α F	ACCGCCTGGACTTCTGGAA
TNF-α R	ATCCGCGACGTGGAACTG
TRAF3 F	GACTCTTCTAAGGAGTGAGG
TRAF3 R	TGGATGCTCTTGTTTTTCTC
TRAF6 F	TGCTTTGCGTCCGTGCGATG
TRAF6 R	GGGTCCGAATGGTCCGTTTG
TRIF F	CAGCTCAAGACCCCTACAGC
TRIF R	CTCCCACACAGCCTCGTC

F, forward; R, reverse.

BMDM and pTEC treatment

Prepared B6 and MAVS-KO BMDMs, as already described, were seeded in 6- (8 × 10⁵ cells/well) or 12-well plates (4 × 10⁵ cells/well). pTECs were seeded with a split ratio of 4.5 × 10⁵ and 1 × 10⁵ cells/well, respectively. Following a 24-h incubation period, cells were serum starved (1% FBS) for 10–12 h before treatment with or without 100 ng/ml LPS (E. coli O111:B4 L2630, Sigma-Aldrich). After 24 h, cells were collected to perform qPCR assays or immunoblotting. For Western blot analysis of phosphorylated proteins, cells were stimulated for only 0.5–1 h.

To verify the crucial role of mitochondrial complex II (CII) in mediating LPS-induced TLR4 signaling pathway in correlation with MAVS activity, cells were treated with LPS or a dose of 1 μM atpenin A5 (AA5) (Cayman Chemical, 119509-24-9) alone, or a combination of both LPS and AA5 for a duration of 24 h. Control group cells were incubated in the reagent-free medium.

Western blot analyses and MAVS aggregation detection

Total protein was extracted from BMDMs by 1× RIPA buffer (Merck Millipore) with cOmplete Mini, EDTA-free protease inhibitor cocktail (Roche, 04693159001) and phosphatase inhibitor PhosSTOP (Roche, 490683700). A total of 25 μg of whole-cell protein lysates was loaded per well and separated by SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membranes (Thermo Fisher Scientific). For MAVS aggregation, lysates were run on 1.5% agarose, 0.1% SDS gel, then transferred to PVDF membranes per the semidenaturing detergent agarose gel electrophoresis (SDD-AGE) protocol described previously (15). After a 1-h incubation in PVDF blocking reagent for Can Get Signal (Toyobo, NYPBR01) (RT), the primary Abs were added and incubated overnight at 4°C. After washing with TBST, appropriate HRP secondary Abs were added, incubated for 1 h in the dark at RT, and membrane-derived protein bands were detected using Clarity Western ECL substrate (Bio-Rad, 1705061). The blotted membrane was imaged by the ChemiDoc imaging system (Bio-Rad), and quantification of interested proteins was performed with ImageJ software (National Institutes of Health, Rockville, MD). Abs used for immunoblotting are listed in Supplemental Table I.

Mitochondrial reactive oxygen species generation assay

A mitochondrial reactive oxygen species (ROS) assay was performed using MitoSOX red (Invitrogen, M36008), a fluorogenic marker for the highly selective detection of superoxide (O₂⁻) in the mitochondria of live cells. At 24 h after LPS stimulation, macrophages were incubated with 5 μM MitoSOX+DMSO for 20 min at 37°C, then washed with PBS. At 2 h prior to staining, the positive control group was treated with 200 nM hydrogen peroxide (Fujifilm Wako Pure Chemical, 081-04215). Subsequently, cells were harvested for flow cytometric detection and analysis of MitoSOX intensity, as well as the percentage of positively stained cells, using FACSymphony A5 (BD Biosciences) and FlowJo software (Tree Star).

Statistics analyses

Data are expressed as the median with 95% CI as determined with GraphPad Prism 9 software (version 9.5.1). The nonparametric Mann–Whitney U test was used for two-group comparisons, and the nonparametric Kruskal–Wallis test followed by the a Mann–Whitney U test was used for multiple-group comparisons. A p value <0.05 indicated statistical significance.

Results

LPS induces more severe AKI in MAVS-KO mice

MAVS-KO mice showed that kidney function worsened and more kidney injuries than for B6 mice after 16–24 h of LPS treatment. In the renal cortex of treated B6 and MAVS-KO mice, KIM-1, an index of tubule injury, was intensely stained in tubular segments (Fig. 1A), and levels were significantly higher in MAVS-KO mice (Fig. 1B). Additional biomarkers were used to assess kidney function after treatment, including the spot urine albumin-to-creatinine ratio and BUN. Consistent with KIM-1 staining results, treated MAVS-KO mice had significantly higher levels of these biomarkers compared with age-matched treated B6 mice (p < 0.05) (Fig. 1C).

FIGURE 1.

MAVS deficiency promotes more severe kidney injury compared with B6 mice after 16–24 h of LPS administration.

(A) Fluorescent immunostaining of KIM-1 (red). DAPI was used for nuclear staining (blue). Original magnification, ×200. (B) Staining score was determined by counting the number of positive tubules per field using a ×20 magnification lens. Ten to 15 fields were examined in each kidney section. (C) Renal function assessed by urinary albumin/urinary creatinine ratio (ACR) and blood urea nitrogen (BUN) in B6 and MAVS-KO mice with/without LPS treatment. (D) mRNA expression of neutrophil gelatinase-associated lipocalin (NGAL) in kidney tissues. (E) Serum IL-6 levels in wild-type and MAVS-KO mice were measured by a mouse ELISA kit after 16 hourly injections of PBS or LPS. (F) Effects of LPS on mRNA expression levels of proinflammatory cytokines (TNF-α, IL-6, and IL-1β) in renal tissues of the endotoxin-induced sepsis model. The expression levels of the target genes were normalized against ACTB mRNA expression. Data represent median with 95% CI and are representative of three separate experiments. *p < 0.05, **p < 0.01, ***p < 0.001.

The study found that the mRNA expression of renal neutrophil gelatinase-associated lipocalin (NGAL) was also significantly increased in LPS-treated MAVS-KO mice after 16 h of treatment when compared with control mice (p < 0.05) (Fig. 1D). Furthermore, the expression of renal representative inflammatory cytokines was remarkably upregulated at both transcriptional and protein levels in the LPS-treated MAVS-KO mice group (p < 0.05) (Fig. 1E, 1F). These findings suggest that MAVS-KO mice are more susceptible to an LPS-induced kidney injury than are B6 mice.

LPS-stimulated chemokine and AKI-associated hypoxia-inducible factor-1α production is enhanced in the kidneys of MAVS-KO mice

The administration of LPS (20 mg/kg) increased MCP-1 transcription in kidney tissues. After 16 h of LPS stimulation, there was a remarkable rise in MCP-1 mRNA expression in kidney tissues of MAVS-KO mice compared with the B6 group (p < 0.05) (196-fold change [95% CI 13.71–266.69] and 124.7-fold change [95% CI 48.20–90.61], respectively) (Fig. 2A).

FIGURE 2.

Analysis of representative AKI-related inflammatory factors in murine kidney tissues after LPS treatment.

(A and B) Reverse transcription–qPCR of whole kidney tissues for mRNA expression of leukocyte recruitment and activation chemokines (A) MCP-1 and (B) IL-8-analog chemokines (MIP-2, Cxcl3, KC) were normalized to the ACTB mRNA level. (C) Expression levels of inflammatory protein MCP-1 and HIF-1α in the renal cortex were detected by Western blot analysis and quantified (n = 5). Data represent median with 95% CI and are representative of three separate experiments. *p < 0.05, **p < 0.01, ***p < 0.001.

The murine chemokines KC/Cxcl1, macrophage inflammatory protein-2 (MIP-2/Cxcl2), and MIP-2β (Cxcl3/DCIP-1) are the major chemoattractants responsible for recruiting neutrophils, which are IL-8 analogs in mice. Following LPS stimulation, there was a noticeable increase in the mRNA level of these mentioned neutrophil chemokines (p < 0.05). However, only renal MIP-2 mRNA demonstrated a significant upregulation in MAV-KO mice compared with B6 mice (Fig. 2B).

Consistently, the analysis of renal MCP-1 and hypoxia-inducible factor-1α (HIF-1α) protein expression by immunoblotting, along with the quantification of serum IL-6 through ELISA testing, showed a trend for upregulation in the MAVS-KO group (Figs. 1E, 2C). These findings confirmed a more severe inflammatory response in the MAVS-KO group after LPS administration. It is noteworthy that these proteins are closely associated with the renal inflammatory state, particularly HIF-1α, a master transcription factor strongly involved in the pathogenesis of sepsis-induced AKI due to renal hypoxia during endotoxemia.

Inflammatory cell infiltration was highly stimulated by LPS in the kidneys of MAVS-KO mice

Sepsis-related kidney damage is frequently triggered by activated immune cells. We propose that MAVS deficiency might lead to an upsurge in the number of inflammatory macrophages and neutrophils in response to LPS treatment due to increased chemokine expression as MCP-1, MIP-2, CXCL3, and KC.

To verify this hypothesis, CD45⁺ hematopoietic cells and CD11b⁺F4/80⁺ cells were quantified in the kidneys, BM, and blood with FACS analysis after 16 h of LPS treatment. Following LPS stimulation, the number and frequency of activated macrophage populations recruited to kidneys, including the M1 phenotype (F4/80^lo) and M2 phenotype (F4/80^hi), were compared between B6 and MAVS-KO mice, and the M2 phenotype was compared between B6 and MAVS-KO mice.

Interestingly, the total number and percentage of F4/80⁺ cells in the CD45⁺CD11b⁺ leukocyte population of B6 mice were higher in both kidneys, BM, and peripheral blood (Fig. 3B), However, in the subpopulation of renal F4/80^lo macrophages, although the cell numbers were not significantly different, MAVS-KO mice kidneys had a remarkably higher percentage of proinflammatory M1 macrophages expressing high levels of Ly6C (CD45⁺CD11b⁺F4/80^loLy6C^hi) (Fig. 3C). In contrast, after 16 h of stimulation in B6 mice, most renal macrophages were anti-inflammatory M2 phenotype F4/80^hi.

FIGURE 3.

LPS induced the elevated infiltration of proinflammatory macrophages and neutrophils in kidney tissues of MAVS-KO mice.

Wild-type and MAVS-KO mice were subjected to sham or LPS treatment. At 16 h after treatment, peripheral blood, bone marrow (BM), and kidneys were collected, and LPS-induced immune cell activation and kidney infiltration were evaluated using flow cytometry and immunohistostaining. (A) Diagram analysis using flow cytometry. (B) Total number and percentage of F4/80⁺ macrophages were evaluated using flow cytometry (within blood, myeloid, and renal CD45⁺CD11b⁺ leukocytes population). (C) Cell numbers and proportion of F4/80^loLy6C^hi, F4/80^lo, and F4/80^hi subsets within the total macrophage population isolated from the kidneys of wild-type and MAVS-KO mice after LPS administration. (D) Kidney sections were stained with Gr-1 (Ly6G) Ab for neutrophil detection. Gr-1–positive cells appeared brown after diaminobenzidine staining. Representative images are shown (original magnification, ×200; scale bars, 100 μm). For infiltrated neutrophil quantification, we counted the Gr-1–positive cells in 10 different fields (n = 3–9). (E) Identification and quantitative data analysis of Ly6G⁺ neutrophils (pregated on live CD45⁺CD11b⁺ population) using flow cytometry (peripheral blood, myeloid, and kidney). Data represent median with 95% CI. *p < 0.05, **p < 0.01.

Furthermore, LPS treatment of MAVS-KO mice for 16 h resulted in notably increased CD11b⁺Ly6G⁺ cell populations (Fig. 3F, 3G), which aligns with renal Gr-1⁺ immunohistochemistry staining results (Fig. 3D, 3E).

Taken together, MAVS deficiency might cause an increase in immune cells within the kidney, particularly inflammatory Ly6C^hi cells, while leading to a decrease in F4/80^hi cells in response to inflammation. This is likely due to an increase in the production of leukocyte-recruiting chemokines and activation of the LPS-activated inflammatory signaling pathway. These findings suggest a correlation between MAVS and LPS-activated inflammatory signaling pathways in proinflammatory and anti-inflammatory macrophages, ultimately contributing to the induction of kidney damage.

LPS induced a more severe inflammation response in MAVS-KO BMDMs

Gene expression of the representative LPS-induced proinflammatory cytokines/chemokines was evaluated using qPCR analysis in both unstimulated and stimulated samples with the primers listed in Table I (Fig. 4A–C). Afterward, protein expression was confirmed by Western blot analysis (MCP-1) and an ELISA kit test (IL-6) (Fig. 4D, Supplemental Fig. 2).

FIGURE 4.

LPS-triggered production of proinflammatory cytokines/chemokines in MAVS-KO mice’s BMDMs.

(A–C) qPCR-based analysis of (A) proinflammatory cytokines (TNF-α, IL-1β, IL-6), (B) monocyte chemoattractant protein-1 (MCP-1), and (C) neutrophil chemoattractant chemokines (KC, MIP-2, Cxcl2) in murine BMDMs cells after 24 of LPS treatment (100 ng/ml). Relative mRNA expression (fold change) is shown of cytokines/chemokines induced by BMDMs following endotoxin stimulation as analyzed through qPCR (n = 4–6). β-Actin was used as reference gene. (D) ELISA quantification of IL-6 level in the supernatants of LPS-stimulated BMDMs of B6 and MAVS-KO mice. Data represent median with 95% CI. *p < 0.05, **p < 0.01.

Twenty-four hours of LPS stimulation markedly increased the mRNA levels of inflammatory cytokines in both B6 and MAVS-KO cells when compared with the unstimulated control group. The mRNA levels of IL-1β, IL-6, and TNF-α were significantly more elevated in MAVS-KO BMDMs (p < 0.05) (Fig. 4A). Of note, the mRNA level of IL-6 was the most dramatically enhanced in MAVS-KO cells compared with B6 cells after LPS exposure (1008.4-fold change [95% CI 574.6–877.4] and 262.6-fold change [95% CI 112.5–144.8], respectively).

Reverse transcription–qPCR results also revealed that LPS significantly increased the transcription of leukocyte-recruiting chemokines in BMDMs of KO mice compared with the wild-type group (p < 0.05; Fig. 4B, 4C). Although the relative signal intensity of MCP-1 only has a trend of upregulation in the whole-cell lysate of MAVS-KO BMDMs (Supplemental Fig. 1C), secreted IL-6 of MAVS-KO cells measured by an ELISA test was much more elevated compared with that of B6 cells (Fig. 4D). These data seem to be consistent with the above in vivo analyzed data and confirmed the crucial role of macrophages in sepsis-induced kidney injury regulation in correlation with MAVS.

LPS-activated TLR4 signaling pathways have a key role in inducing severe inflammatory responses in MAVS-KO BMDMs

The results of this study indicated that TLR4/NF-κB signaling triggered more severe inflammatory responses in MAVS-KO BMDMs. There was an increase in expression of major components not only at gene levels such as MyD88, RelA, and c-fos (p < 0.05) (Fig. 5B) but also at protein levels of TRAF6, p–IκB kinase (IKK)α/IKKβ, p-IκBα, and p-c-jun (Fig. 5D–G). Furthermore, a trend of upregulated expression of p-p38-MAPK protein was observed in MAVS-KO cells (Supplemental Fig. 2), along with significantly elevated mRNA levels of TLR4/TRIF pathway components, which resulted in higher levels of downstream cytokine gene transcription (IFN-β1/Ifnb1, RANTES/CCL5) (Fig. 5C). These data suggest that MAVS protects macrophages from LPS stimulation by regulating TLR4-related signaling at the most upstream position, which is TLR4. Additionally, LPS induced the formation of endogenous MAVS high–molecular mass aggregates, which were detected via immunoblotting following resolution by SDD-AGE/immunoblotting (Fig. 5I). Taken together, these findings supported the hypothesis that MAVS was activated by LPS and was subsequently involved in regulating the LPS/TLR4 signaling pathway.

FIGURE 5.

LPS not only more strongly activated TLR4/MyD88-dependent and TLR4/MyD88-independent signaling pathways in MAVS-KO macrophages than in B6 cells, but it also induced MAVS aggregation.

(A–C) Gene expression levels of (A) TLR4 and (B) representative TLR4/MyD88-dependent inflammatory pathway components (MyD88, RelA, c-fos) or (C) TLR4/MyD88-independent inflammatory pathway components (TRIF, Ifnb1, Ccl5) were determined in 24-h LPS-treated BMDMs using a qPCR assay. β-Actin was used as reference gene. (D–H) Protein expression of some other crucial components in TLR4 signaling pathways including (D) TRAF6, (E) p-IKKαβ/total IKKαβ, (F) p-IκBα/IκBα, and (G) p-c-jun/total c-jun (AP-1) in LPS-stimulated BMDMs was detected by Western blot and quantitative data are shown (H). (I) Detection of LPS-induced endogenous MAVS aggregates in BMDMs by semidenaturing detergent agarose gel electrophoresis (SDD-AGE) and analyzed by immunoblot using anti-MAVS Ab. Data represent median with 95% CI. *p < 0.05, **p < 0.01.

Interestingly, although the mRNA level of TLR4 was reduced in murine primary macrophages after LPS treatment, it remained higher in MAVS-KO cells than in wild-type cells (p < 0.05) (Fig. 5A). Therefore, MAVS might be involved in TLR4 pathway regulation through transcription, posttranscriptional stabilization, or degradation. The upregulated gene expression of main components in the lysosomal TLR4 pathway, including Rab5, Rab7b, Rab10, and Rab11a in LPS-treated MAVS-KO BMDMs (p < 0.05) (Supplemental Fig. 3A), indicated that the mechanism by which MAVS modulated the TLR4 mRNA level was not related to degradation regulation.

Moreover, AP-1, a pivotal transcription factor composed of c-jun and c-fos, could increase mRNA synthesis by directly binding to TLR4 promoter. In LPS-stimulated MAVS-KO macrophages, c-jun mRNA level and c-fos protein expression were significantly higher (p < 0.05) (Fig. 5B–G). These findings suggest that AP-1 might be a direct regulator of TLR4, mediated by MAVS activation through the LPS-induced IKK and MAPK pathways.

MAVS-KO BMDMs increase mitochondrial ROS production by LPS stimulation

We then studied ROS production in BMDMs using MitoSOX staining and flow cytometry to investigate the influence of LPS-induced ROS generation on TLR4 gene expression.

We used forward scatter area and side scatter area as gating strategies to identify live cells (Fig. 6A). After gating the positive control group and nonstaining group, the percentage of stained MitoSOX cells in the B6 and MAVS-KO BMDMs after LPS treatment could be evaluated (Fig. 6B). Results showed an increased MitoSOX signal after LPS treatment in macrophages (p < 0.01) (Fig. 6C), with a higher trend of positive percentages in MAVS-KO groups. Our data suggest that MAVS has a role in regulating TLR4 expression by mediating mitochondrial ROS production.

FIGURE 6.

Mitochondrial ROS production in LPS-stimulated BMDMs analyzed by FACS.

BMDMs were treated with/without 100 ng/mL LPS for 24 h and then stained with the mitochondrial probe MitoSOX Red (5 μM, 20 min, 37°C) and analyzed by flow cytometry. (A) Gating strategy to select living macrophages via forward scatter area (FSC-A) and side scatter area (SSC-A), stained MitoSOX, positive control (treated with H₂O₂), and negative control (unstained). (B) Representative FACS image of treated/untreated BMDMs stained with MitoSOX. (C) Percentages of positive cell numbers, as defined by MitoSOX-positive cells (n = 5). Data represent median with 95% CI. *p < 0.05, **p < 0.01.

To disclose the mechanism of how MAVS interferes with LPS-induced ROS regulation, we continued evaluating the ROS-producing upregulation-related factors and antioxidant factors. The data indicated that MAVS might be involved in ROS generation rather than oxidative stress alleviation, as supported by an increase in antioxidant gene expression in MAVS-KO (Supplemental Fig. 3B).

Upregulated ROS production in MAVS-KO BMDMs after LPS treatment was induced by Nox2 and electron transport complexes

Various mechanisms generate cellular ROS, with phagosomal NADPH oxidase and the mitochondrial electron transport chain being the primary sources in macrophages.

Therefore, the transcriptional expression of representative NADPH oxidases was analyzed in LPS-activated macrophages, focusing on Nox1, Nox2, and Nox4. Results demonstrated that Nox2 expression increased in MAVS deficiency BMDMs more than in B6 cells (p < 0.05) (Fig. 7A), whereas Nox1 and Nox4 mRNA were not expressed in both B6 and MAVS BMDMs with a Ct value >30 (data not shown). This finding indicates that Nox2 might stimulate TLR4 signaling through ROS generation, and MAVS could mediate its transcription.

FIGURE 7.

Expression of ROS production regulators in LPS-stimulated B6 and MAVS-KO BMDMs.

BMDMs were treated with/without 100 ng/ml LPS for 24 h, after which mRNA levels were analyzed by qPCR or protein relative intensity was evaluated by Western blot. (A and B) mRNA levels of ROS regulating genes including (A) NADPH oxidase 2 (Nox2), (B) mitochondrial genes of complex I (CI) (mtND1, mtND2, mtND4), complex II (CII) (succinate hydrogenase [Sdha, Sdhb]), complex III (CIII) (cytochrome b [Cytb]), complex IV (CIV) (cytochrome c oxidase subunit I [mtCo1]), and complex V (CV) (ATPase 6 [mtATP6]). β-Actin was used as reference gene. (C) Western blot analysis of SDHB and HIF-1α expression in LPS-treated B6 and MAVS-KO cells. (D) Quantitative densitometry of the Western blot analysis. (E) Effect of atpenin A5 (CII inhibitor) in LPS-activated macrophages on transcriptional expression of hypoxia-inducible factor 1α (HIF-1α), IL-1β, and TLR4. β-Actin is used as reference gene. Data represent median with 95% CI. *p < 0.05, **p < 0.01.

The mitochondrial reverse electron transport (RET), one of the main sources of sepsis-induced ROS production in macrophages, was subsequently evaluated. Gene expression of representative components in mitochondrial respiratory complexes was analyzed, and a decrease in mRNA transcripts was observed after LPS stimulation (p < 0.05) (Fig. 7B). However, the MAVS-KO group still showed a higher trend of mitochondrial gene expression compared with B6 cells. Remarkably, the expression levels of mt-ND4, a vital element of complex I (CI), showed a significant increase even in normal conditions in the absence of MAVS. Additionally, the abundance of mt-ND4 mRNA remained notably higher in comparison with other mitochondrial genes (Fig. 7B). This suggests a strong correlation between MAVS and mt-ND4 transcription regulation, which leads to the distinctive upregulation of mt-ND4 relative to other mitochondrial genes within the same polycistronic mitochondrial DNA (mtDNA).

The data revealed that MAVS-KO BMDMs expressed higher mRNA levels of mt-ND4 (CI), Sdha, and Sdhb (CII) (p < 0.01, p < 0.05, and p = 0.052, respectively), indicating the crucial role of mitochondrial CI and CII in LPS-induced ROS generation (Fig. 7B). MAVS might contribute to mitochondrial gene suppression to prevent ROS overproduction. MAVS deficiency could lead to uncontrolled activation of mitochondrial functional genes and excessive oxidative stress, which was consistent with the higher trend of superoxide detection in LPS-treated MAVS-KO macrophages (Fig. 6).

HIF-1α, a transcription factor and an important product of LPS-induced RET, was reported to mediate the inflammation response to sepsis by stimulating TLR4 transcription and enabling inflammatory cytokine expression, especially IL-1β transcription. Mitochondrial ROS and succinate dehydrogenase (CII) could regulate HIF1A (gene encoding for HIF-1α) expression. Of note, although HIF1A expression without MAVS activity was significantly increased after LPS treatment (p < 0.01), inhibiting CII activity with a specific inhibitor (AA5) remarkably reduced HIF1A expression in MAVS-KO cells (p < 0.05) but not in B6 cells (Fig. 7C).

These data are also consistent with the transcriptional expression of TLR4 and IL-1β, two downstream products of HIF-1α activation. The significant difference in TLR4 mRNA expression between treated MAVS-KO and B6 cells became insignificant, with a trend of reduced transcriptional expression in MAVS-KO cells. The role of ROS/CII-induced HIF-1α in mediating inflammatory response under MAVS activity was also confirmed by IL-1β mRNA expression. It is also noteworthy that the protein analysis conducted through Western blot of SDHB and HIF-1α (Fig. 7C) was consistent with the gene expression data, revealing a stronger protein expression in treated MAVS-KO BMDMs.

In summary, these results propose that there could be an association between MAVS protection role and cellular ROS-related stimulation factors, including Nox2 and mitochondrial respiratory complexes in LPS-activated macrophages (Fig. 8).

FIGURE 8.

Proposed mechanism for protective role of MAVS in correlation with LPS-stimulated TLR4 signaling pathway in BMDMs.

LPS triggers inflammatory responses in macrophages through both TLR4/MyD88-dependent and -independent pathways. It also regulates cellular ROS production by controlling NADPH oxidase-2 transcription and mediating mitochondrial ROS generation. Reverse electron transport is the key process that generates mitochondrial ROS in LPS-stimulated macrophages. CI and CII play indispensable roles in this process, affecting the accumulation of succinate, a substrate of succinate dehydrogenase (CII), which activates HIF-1α and promotes TLR4 mRNA transcription. The resulting robust mitochondrial and cellular ROS activity increases the NF-κB/MAPK signaling pathway and elevates the transcriptional expression of proinflammatory cytokines. However, to prevent excessive inflammatory responses and oxidative stress, the transcriptional expression of important signaling components, including respiratory complexes and TLR4, is downregulated. The signaling adaptor protein MAVS is activated by cellular ROS generation and could be directly activated by LPS. Activated MAVS helps balance the cellular redox state by suppressing the gene expression of Nox2 and respiratory complexes, particularly CI and CII, thereby downregulating ROS production in proinflammatory macrophages.

Discussion

In this study, we report that MAVS signaling was involved in the pathogenesis of the LPS-associated murine AKI model. LPS induced more severe kidney injuries in MAVS-KO mice compared with B6 mice. Inflammatory macrophage/neutrophil accumulation was increased in the kidneys of MAVS-KO mice. When exposed to LPS, MAVS-KO macrophages exhibited heightened gene and protein expression levels of cytokines/chemokines. Downregulation of the TLR4 gene was inhibited in LPS-stimulated MAVS-KO macrophage, and the expression of major components associated with MyD88-dependent and TRIF-dependent pathways were found to be upregulated in these cells. MAVS-KO macrophages had increased mitochondrial ROS and respiratory complex–associated gene expression, key stimulators of TLR4 gene expression. Mitochondrial CII inhibitor effectively decreased TLR4 as well as HIF1A gene expressions in LPS-stimulated MAVS-KO macrophages. Collectively, MAVS deficiency skewed M1 macrophage polarization via TLR4 signaling, leading to severe kidney damage in the LPS-induced AKI model.

This research aligns with previous studies that have found more severe inflammatory responses in MAVS-KO macrophages stimulated by bacterial pathogens (15, 23). However, Sun et al. (24) reported that the innate response of MAVS-deficient cells to bacterial challenges such as LPS or infection with Listeria monocytogenes is not affected, and there was no difference between MAVS-KO and B6 mice in IL-6 and IFN production. One possible explanation for this discrepancy is that Sun et al. used a very high dose of LPS (10 μg/ml, 24h) to stimulate BMDMs, and the specific types of LPS used were not mentioned. Prior studies have confirmed that macrophage immune responses to LPS are both dose- and type-dependent, and other than macrophages, different MAVS-deficient cell types exhibit varying levels of responses to LPS stimulation (25–27). It is noteworthy that Listeria monocytogenes, a Gram-positive pathogen that lacks LPS, does not signal the innate immune system through TLR4 (28). Therefore, the inflammatory responses to this bacteria may differ from those of other Gram-negative pathogens, such as E. coli in this study.

Macrophages play a crucial role in the pathogenesis of AKI. We have reported that sustained proinflammatory (M1) macrophage accumulation mediated defective kidney repair from AKI in lupus-prone mice (20). Conversely, CSF-1 promotes immune regulatory (M2) macrophages in the repair phase of AKI (29). In LPS-induced AKI, M1 macrophages contribute to the pathogenesis of kidney damage, and inhibiting or modulating their activation can improve kidney injury (10, 30). Consistent with these findings, the renal M1 macrophages accumulated after LPS administration, especially in MAVS-KO mice, which had a higher frequency of those expressing high levels of Ly6C (CD45⁺CD11b⁺F4/80^loLy6C^hi). In contrast, the frequency of M2 macrophage (F4/80^hi) was lower in MAVS-KO kidney compared with B6 kidney after LPS administration. In accordance with results in the present study, some previous studies emphasized the essential role of the proinflammatory CD11b⁺Ly6C^hi macrophage population in sepsis-induced kidney injury, which is tightly regulated by TLR4 (31, 32). Our in vitro assays also showed that LPS stimulation induced higher proinflammatory cytokine/chemokine gene expressions in MAVS-KO BMDMs as compared with B6 BMDMs. These imbalances of M1/M2 macrophages might lead to more severe kidney damage in MAVS-KO mice than in B6 mice.

LPS/TLR4 signaling is a strong inducer of M1 macrophages (33). Inhibited downregulation of TLR4 gene expression might be one reason for skewed M1 polarization of MAVS-KO macrophages in our LPS stimulating model. Indeed, the expression of some essential downstream components in TLR4 signaling was more upregulated in MAVS-KO macrophages at both the transcriptional and protein levels, suggesting that the TLR4 signaling pathway was highly activated in MAVS-KO macrophages (Fig. 8). Supporting this notion, studies have shown that downregulation of TLR4 by small interfering RNA inhibited inflammatory cytokine/chemokine production (34), whereas upregulation of TLR4 expression promoted inflammatory responses under LPS stimulation in RAW264.7 cells (35). These findings indicate that the expression levels of TLR4 might regulate M1 polarization of macrophages.

The TLR4 expression level is influenced by the balance of biosynthesis, trafficking, and degradation (36, 37). Our data showed that degradation-related genes of TLR4 were upregulated, implying the increased degradation of TLR4 in MAVS-KO BMDMs. This prompted us to focus on investigating the biosynthesis pathway that regulates TLR4 expression. While TLR4 signaling is known to increase ROS production, elevated intracellular levels of ROS/oxidative stress can also promote TLR4 expression (38–41). Interestingly, in our study, MAVS-KO BMDMs demonstrated a higher expression of LPS-induced mitochondrial ROS than did B6 BMDMs. Accordingly, to determine how MAVS contributes to the regulation of LPS/TLR4-induced inflammation, we further examined the expression of ROS regulation factors. ROS in LPS-stimulated macrophages originate from various sources, but the two main ones are the cytosolic source (NADPH oxidase complexes) and the mitochondrial source (respiratory complexes) (42).

NADPH oxidase is a major oxidative stress mediator of macrophages (42). Although the Nox transmembrane enzyme family consists of various isoforms, only Nox1, Nox2, and Nox4 have been identified in phagocytes (43). Our study found upregulated expression of the Nox2 gene with LPS stimulation, consistent with other studies (44, 45). However, we observed very low and unstable transcriptional expression of Nox1 and Nox4, which is different from other studies (46, 47), and could be explained due to differences in model and cell types used in the study. Indeed, Nox2 is the most well-characterized NADPH oxidase in phagocytes, and its ROS-generating activity after LPS stimulation was described to be independent of the activity of mitochondrial respiratory complexes (44). Thus, our findings suggested that LPS could induce cytosolic ROS through Nox2 regulation in BMDMs, and Nox2 might be a target gene regulated by the LPS-activated MAVS signaling pathway to enhance TLR4 mRNA expression (Fig. 8).

Another potential source of ROS in endotoxin-activated macrophages is reverse electron transport, with the critical role of CI and CII of the canonical electron transport chain in producing ROS (48, 49). In this model, LPS suppressed the oxidative phosphorylation in M1 macrophage, leading to a reprogramming of the mitochondrial respiratory system. As a result, cells switch to glycolysis and produce ATP, leading to elevated mitochondrial membrane potential and the succinate oxidation-dependent highly reduced CoQ pool. This, in turn, drives electrons in reverse to CI to produce ROS (49, 50). The results of this study, especially the dramatically downregulated expression of the mitochondrial complex gene, supported the RET model as the main ROS source in LPS-activated BMDMs, with a significant contribution of CI and CII. Surprisingly, in MAVS-deficient cells, mRNA levels of most mitochondrial respiratory functional enzymes could remain remarkably higher than in B6 cells, suggesting that MAVS could maintain mitochondrial homeostasis under LPS suppression (Fig. 8). It is noteworthy that all mitochondrial genes are located on the same polycistronic mtDNA (51). mt-ND4 is indispensable in mitochondrial CI structure, which enables ATP synthesis of mitochondria (52). Therefore, the mt-ND4 gene showed a significantly distinct increase in expression, especially when compared with mt-ND1 and mt-ND2 (Fig. 5B), suggesting its crucial role in the model and a unique mechanism related to MAVS activity. Although the correlation between mt-ND4 and MAVS remains ambiguous in the literature, several possible explanations exist based on the reported mechanisms of mitochondrial gene regulation (53–55). First, MAVS might directly bind to mtDNA or collaborate with another regulatory molecule to inhibit the mRNA synthesis of downstream genes at the position after mt-ND2 location on mtDNA, which generated short transcripts. This postulated mechanism is based on the location order of mitochondrial genes on mtDNA (51) and the observation that only downstream genes starting from mt-Co1 showed greater upregulation in LPS-treated MAVS-KO BMDMs (Fig. 7B). Other proposed mechanisms could be the existence of a noncanonical promoter that transcribes mt-ND4 exclusively and the potential role of microRNAs (miRNAs) in posttranscriptional regulation of mitochondrial gene expression in correlation with MAVS. Some studies have revealed that specific miRNAs can target particular mitochondrial genes, including direct regulation of mt-ND4 (56, 57). It seems possible that LPS-stimulated MAVS, alone or in combination with other regulatory molecules as miRNAs, might increase the suppression of mitochondrial target gene mRNA levels, particularly mt-ND4. However, further and separate studies are required to fully understand the significance of mt-ND4 and to confirm our hypothesis about the complex relationship between MAVS and mitochondrial gene expressions in correlation with LPS.

Previous studies have shown that MAVS depletion results in mitochondrial dysfunction by inducing oxidative stress and impaired mitophagy even in the absence of stress (14, 58). However, this study highlights a distinct role of MAVS in protecting cells from overaccumulation of LPS-enhanced ROS by inhibiting the transcriptional expression of respiratory complexes. It provides considerable insight into the various functions of MAVS depending on diverse pathological and physiological conditions.

HIF-1α is a master transcriptional regulator of the adaptive response to hypoxia (59, 60). Interestingly, our study confirmed that even under normal pO2 conditions, LPS could still stimulate the HIF1A gene, similar to the “pseudohypoxia” state reported by other authors (61). Additionally, in the absence of LPS-activated MAVS, HIF-1α expression was significantly increased compared with treated wild-type cells. HIF-1α was previously described as a connecting point between the mitochondrial RET-ROS/SDH activity and important inflammatory components such as TLR4 and proinflammatory cytokines, especially IL-1β (62, 63). Thus, it is likely that such connections also exist in our study to elucidate how MAVS regulates the TLR4 and inflammatory cytokine mRNA expression after LPS treatment in macrophages (Fig. 8). To validate this correlation, we employed AA5, a ubiquinone-binding site mitochondrial CII inhibitor, to block CII function. As expected, this suppression downregulated gene expression of inflammatory-related molecules, including HIF1A, TLR4, and IL-1β. Furthermore, this study found that inhibiting SDH alone did not completely mitigate the inflammatory responses caused by sepsis-induced oxidative stress. This suggests that beyond CII, there might be other chief targets of MAVS and that the interaction between these proteins protects host cells from sepsis-induced oxidative stress.

In addition, cytosolic ROS and mitochondrial ROS were previously reported to enhance LPS/TLR4 downstream inflammatory signaling axes, including NF-κB and MAPK signaling (64–66). AP-1, a downstream transcriptional factor of these pathways composed of members of the Fos and Jun families, could directly interact with the TLR4 promoter to enhance TLR4 transcription (67). Therefore, AP-1 could also be a potential candidate to link the MAVS-dependent ROS generation and LPS-activated TLR4 mRNA regulation (Fig. 8).

In the traditional approach, MAVS was only emphasized for its role in the antiviral inflammatory response and could induce kidney damage in some disease models due to its ROS upregulation (68). Conversely, ROS could induce MAVS activation independently of viral infection stimulation (13). Aside from its harmful role, MAVS was reported to have a renoprotective role in diabetic kidney disease (15), and the current study is, to our knowledge, the first to report the renoprotective role of MAVS in the murine sepsis/AKI model by regulating ROS in kidney-recruited inflammatory macrophages. Our study also uncovers a (to our knowledge) novel mechanism by which MAVS could contribute to anti-inflammatory defense in bacterial infection by maintaining redox homeostasis balance via regulating related gene expression in correlation with LPS/TLR4 signaling.

It is also noteworthy that some previous studies focused on the responses of renal cells in the sepsis-associated AKI model (15). However, our study took a different approach by examining the macrophage contribution, as we found more similarities in LPS-induced gene and protein expression between BMDMs and in vivo analysis compared with primary tubular epithelial cells (Figs. 1, 2, 4, 7C, Supplemental Fig. 1). Although our study sheds light on the role of macrophages in sepsis-associated AKI, there is still much to learn about the roles of other components, such as kidney cells and other infiltrated immune cells, to fully reveal the mechanism of sepsis-associated AKI, which is a limitation of this study. Another limitation is that our study has yet to reveal a complete picture of the interaction between MAVS and other principal redox-related molecules or the insightful mechanism of how LPS co-operates with MAVS in mediating mitochondrial activity. For this reason, our future work will concentrate on addressing these limitations.

In conclusion, this study extends our knowledge of the multitasking role of MAVS and the complex relationship between sepsis, MAVS deficiency, ROS production, and AKI. Therefore, controlling redox homeostasis through MAVS signaling could be a potential therapeutic target for sepsis-associated AKI.

Disclosures

The authors have no financial conflicts of interest.