A Novel Two-Hit Murine Model of Viral Primed Sepsis-Associated Acute Kidney Injury

Abstract

Viral priming refers to the host's recognition of viral components, triggering an antiviral response and upregulating pathogen recognition receptors. When followed closely by bacterial infection, this immune activation can provoke a hyperinflammatory response, increasing the risk for secondary hemophagocytic lymphohistiocytosis (sHLH), multiorgan failure, and death. We developed a novel murine two-hit model of viral-primed sepsis-associated acute kidney injury (SA-AKI) using polyinosinic-polycytidylic acid (poly(I:C)) for viral mimicry and lipopolysaccharide (LPS) for bacterial stimulation. Male 8-week-old C57BL/6J mice were primed with poly(I:C) (2.5 mg/kg, intraperitoneal (IP) injection) at 24 hours prior to low-dose LPS (0.5 mg/kg, IP) (time = 0 hours) and compared to vehicle-treated controls for up to 48 hours. Poly(I:C)+LPS mice exhibited a significant reduction in glomerular filtration rate at 4 hours post-LPS, and elevated serum creatinine and urinary kidney injury molecule-1 (KIM-1) at 48 hours, indicating sustained kidney injury. These mice also showed marked increases in plasma ferritin, interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α) at 8 hours post-LPS versus non-primed LPS-treated mice, consistent with a hyperinflammatory state. Poly(I:C) alone induced a rapid type I interferon response, with elevated plasma IFNα, IFNβ, and renal Ifnb1 expression by 4 hours post-injection. This model effectively replicates viral-bacterial co-infection and provides a valuable platform to dissect the mechanisms linking viral priming to dysregulated immune responses and acute tubular injury in sepsis.

New & Noteworthy:

This two-hit murine model of viral primed SA-AKI offers a novel and clinically relevant platform to dissect the immunopathogenic mechanisms linking viral-bacterial co-infections to severe kidney injury. By mirroring features of hyperinflammation and immune dysregulation seen in critically ill patients with SARS-CoV-2, influenza, and other viral infections, this model enables investigation of distinct viral-mediated inflammatory endotypes of SA-AKI, which may guide future efforts to develop targeted, endotype-specific therapies.

Graphical Abstract

INTRODUCTION:

Sepsis and sepsis-associated acute kidney injury (SA-AKI) are highly heterogeneous syndromes triggered by a wide range of infectious insults, leading to diverse and often profound patterns of immune dysregulation that remain incompletely understood(1,2). Clinically, this heterogeneity manifests along a broad spectrum, from mild systemic symptoms with full recovery to multi-organ failure with high mortality. Despite the substantial global burden of sepsis, this biological variability has been a major barrier to translating promising preclinical findings into effective, mechanism-based therapies with consistent success in clinical trials(3,4). In response, contemporary research has shifted focus to establishing subphenotypes and endotypes of sepsis and SA-AKI(5), aiming to overcome the challenge of heterogeneity so that patients can receive a therapy that targets specific biological pathways unique to their underlying pathophysiology(6,7).

Among patients with sepsis, those who develop SA-AKI have an elevated risk of mortality as well as long-term impairments in functional recovery among survivors(8,9). To date, no specific therapies exist to directly abrogate the underlying pathophysiology of SA-AKI, and preclinical SA-AKI models continue to struggle with animal-to-human translation(10). However, a reverse translational approach may provide a novel opportunity to investigate mechanisms of SA-AKI based on an established connection to one of the most severe phenotypes of sepsis characterized by hyperferritinemia, multiple organ failure, and death(11). Recent work by Wang et al. characterized a novel, translationally relevant murine model that recapitulates profound systemic inflammation and hemophagocytosis consistent with secondary hemophagocytic lymphohistocytosis (sHLH) triggered by infection(12). This model leverages sequential administration of polyinosinic-polycytidylic acid (poly(I:C)), a synthetic analog to viral double-stranded RNA, followed by lipopolysaccharide (LPS), a well described glycolipid contained in the outer membrane of gram-negative bacteria and agonist of toll-like receptor (TLR)-4(13,14). Poly(I:C) is known to preferentially bind to TLR3, but also has reported affinity for the intracellular receptors melanoma differentiation-associated protein-5 (MDA-5) and retinoic acid-inducible gene 1 (RIG-1)(15). While this new preclinical sepsis model offers compelling evidence that systemic exposure to poly(I:C) followed by TLR4 stimulation with LPS, but not in reverse order, leads to biological perturbations closely mirroring sHLH, the identification and interrogation of SA-AKI development in this model have not been investigated.

This study aims to refine the murine poly(I:C)-LPS model to determine whether it is a feasible tool to investigate mechanisms of SA-AKI that may be unique to patients presenting with viral-bacterial co-infection. We hypothesize that poly(I:C) followed by low-dose LPS will result in hyperferritinemia, profound systemic inflammation, and evidence of kidney tubular injury and functional declines in renal clearance at early sepsis timepoints.

METHODS:

Experimental Model

Wildtype 8-week-old C57Bl/6J male mice (Jackson Labs, Bar Harbor, ME) were used for the primary data in this study. Mice were viral primed with 2.5 mg/kg poly(I:C) (Invivogen, tlrl-pic) or 0.9% saline vehicle via intraperitoneal (IP) injection at timepoint −24 hours (Fig. 1A). At 0 hours, mice were challenged either with low dose LPS (Cat. No. L2630, Sigma-Aldrich) at 0.5 mg/kg or sterile 1X phosphate buffered saline (PBS) vehicle via IP injection. A small cohort of wildtype 8-week-old C57Bl/6J female mice also underwent sequential administration of poly(I:C) followed by LPS to offer insights into the characterization of this model in both sexes. Mice were housed at the University of Alabama at Birmingham (UAB) and cared for by the UAB Animal Resources Program, an Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC) accredited facility. All mice were exposed to standard 12h light/dark cycles and maintained free access to standard food and water throughout the experiment. At the time of euthanasia, all mice were anesthetized with terminal doses of ketamine (80 mg/kg) + xylazine (10 mg/kg) via IP injection followed by cardiac puncture for blood sample collection. Following cardiac puncture, the abdominal inferior vena cava was transected and 10 mL of phosphate-buffered saline (PBS) was injected via left ventricle in an antegrade fashion to promote exsanguination and tissue clearance prior to collection of tissues. All procedures involving mice were performed in accordance with the National Institutes of Health guidelines regarding the care and use of live animals and were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Alabama at Birmingham (IACUC-22689)(16).

Figure 1. Characterization of Viral Primed SA-AKI Model.

(A) Schematic of two-hit viral primed model. (B) GFR as measured by FITC-labeled sinistrin at 4h and 48h post-LPS. (C) Plasma creatinine 48h post-LPS. (D) Urine KIM-1 48h post-LPS. (E) Kidney expression of Havrc1 48h post-LPS. (F) Western blot of KIM-1 expression in bulk kidney lysate 48h post-LPS. (G) KIM-1 expression by immunofluorescence in the kidney cortex 48h post-LPS (scale bar represents 6 μm). (H) No significant apoptosis was observed at 48h post-LPS per TUNEL staining. Arrows denote sparse TUNEL+ cells. Statistical comparisons were performed using a 1-way ANOVA followed by Dunnett’s multiple comparisons test. **p<0.01, ***p<0.001, ****p<0.0001. SA-AKI, sepsis-associated acute kidney injury; poly(I:C), polyinosinic-polycytidylic acid; GFR, glomerular filtration rate; FITC, fluorescein-isothiocyanate; LPS, lipopolysaccharide; KIM-1, kidney injury molecule-1

Transcutaneous Measurement of Glomerular Filtration Rate (GFR):

Under the anesthesia (Isoflurane, 1.5 – 2%), mice were shaved on the back. A transdermal GFR monitor (Medi Beacon, Germany) was placed on the shaved skin using a double-sided adhesive patch and secured by medical tape. Fluorescein-isothiocyanate (FITC) labeled sinistrin (20mg/mL) was prepared by dissolving in sterile normal saline and administered to mice via tail vein injection. FITC labeled sinistrin clearance was measured for 2h, and monitors were subsequently removed. Data was downloaded from monitors and GFR was determined by calculating kinetics of FITC-sinistrin clearance (Supplemental Fig. 1).

Complete Blood Count (CBC) Profiling:

Whole blood collected by cardiac puncture using a heparinized syringe and analyzed with a Veterinary Hematology Analyzer (Heska, Element HT5) at room temperature. The cellular composition of the blood was measured in absolute counts of white blood cells (WBC), neutrophils, lymphocytes, monocytes, and platelets.

Plasma Analysis:

To isolate plasma, whole blood was centrifuged at 5,000 x g for 10 minutes at 4°C. Ferritin (KT-396, ab157713), IFN-a (MNFAS0, R&D Systems), and IFNβ (MIFNB0, R&D Systems) were then measured by enzyme-linked immunosorbent assay (ELISA) according to manufacturer’s instructions. Creatinine was measured via liquid chromatography-tandem mass spectrometry at the O’Brien Center Core for Acute Kidney Injury at the University of Alabama at Birmingham. Plasma cytokines were measured using a mouse V-PLEX Pro-inflammatory Cytokine Panel I Kit (Meso Scale Discovery) and analyzed with a MESO Sector S600 plate reader (Meso Scale Discovery) per manufacturer’s instructions.

Western Blot:

Kidneys were harvested and lysed in RIPA buffer (Invitrogen/Fisher) with protease and phosphatase inhibitors (Sigma-Aldrich). Lysates were centrifuged at 12,000 × g for 10 minutes at 4°C. The supernatant was collected and assayed for protein quantity using a bicinchoninic acid protein assay (Thermo Fisher Scientific, 23227) following manufacturer’s instructions. Seventy-five micrograms of protein were resolved on a pre-cast 4–12% Bis-Tris polyacrylamide gel (Invitrogen NuPAGE, NP0329BOX) according to manufacturer’s instructions and transferred to a polyvinylidene fluoride membrane. After blocking with 5% w/v nonfat dry milk in Tris-buffered saline with 0.1% v/v Tween-20 (TBST) or 5% bovine serum albumin (BSA) (Tocris, A8531–1VL), the membrane was probed for KIM-1 (Bio-Techne, AF1817, Goat, 1:500), TLR3 (Abcam, Ab13915, Mouse, 1:1,000), MDA-5 (Abcam, Ab315242, Rabbit, 1:1,000), RIG-1 (Abcam, Ab302778, Rabbit, 1:1,000), interferon regulatory factor-3 (IRF3) (Invitrogen, MA532348, Rabbit, 1:1,000), ferritin light chain (FtL) (Santa Cruz Biotechnology, 390558, Mouse 1:1,000), ferritin heavy chain (FtH) (Abcam, Ab183781, Rabbit, 1:500), as well as glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Sigma-Aldrich, MAB347, Mouse, 1:10,000) as a loading control. Primary antibodies were incubated with the membrane overnight at 4°C before washing with TBST and incubating with the appropriate secondary (R1005, R1006, R1007, R1008) for 1h at room temperature. Membranes were then washed with TBST and horseradish peroxidase activity measuring via chemiluminescence KwikQuant detection system (R1002). Images were analyzed using KwikQuant software, and densitometry was measured using ImageJ (v1.54f).

Realtime quantitative PCR:

Template RNA was extracted from murine kidney tissues using an RNeasy Mini Kit (Qiagen, Cat# 74104) according to manufacturer’s instructions. cDNA was generated from template RNA using QuantiTech Reverse Transcription Kit (205311) according to manufacturer’s instructions. RT-qPCR was executed using TaqMan Gene Expression Assays 20x including Havrc1, Il6, Ifnb, Ifnar2, Tlr3, Ifih1, Ddx58, and Gapdh as a housekeeping gene, and corresponding TaqMan Fast Advanced Master Mix (4444963). Samples were assayed using an Applied Biosystems 7500 Real-Time PCR System using standard 96-well plate settings. Results were quantified using the 2^-ΔΔCT method with Gapdh serving as the housekeeping gene.

Urine Analysis:

Urine was collected from mice at time of sacrifice via bladder aspiration using a 31-gauge insulin syringe and assayed for KIM-1 (MKM100) via ELISA according to manufacturers’ instructions.

Immunofluorescence:

Transverse kidney sections (group sample sizes n = 4 for all) were fixed in 10% neutral buffered formalin overnight and transferred to 70% ethanol before embedding in paraffin and cut into 4 μm sections before mounting and heat fixing on slides. The samples were then deparaffinized with xylene and decreasing concentrations of ethanol (100%, 95%) and rehydrated with deionized water. Antigen retrieval was conducted by steaming slides in citrate buffer (pH 6.0) for 35 minutes. After cooling, slides were then blocked using normal horse serum (S-2012–50) or normal goat serum (S-1012–50) for 20 minutes at room temperature in a humidified chamber. Primary antibody against KIM-1 (AF1817, Goat, 1:100) or TLR3 (NBP2–24875, Mouse, 1:100) was then applied respectively and incubated overnight at 4°C. The following day sections were washed with PBS before incubating in species matched secondary for 1h at room temperature in the dark. Slides were then washed in 1X PBS and incubated in FITC-tagged Lotus Lectin for 1h at room temperature in the dark. Endogenous peroxidases were then blocked using an autofluorescence quenching kit (SP-8400) for 5 minutes at room temperature in the dark. Slides were washed in 1X PBS and then incubated in 1:40 Hoescht (33342) in 1X PBS for five minutes before washing once more with 1X PBS and mounting using Prolong Diamond Antifade with DAPI (P36962). The samples were then dried overnight and imaged using X Microscope.

Immunohistochemistry:

Transverse kidney sections (group sample sizes n = 4 for all) were fixed in 10% neutral buffered formalin overnight and transferred to 70% ethanol before embedding in paraffin and cut into 4 um sections before mounting and heat fixing on slides. The samples were then deparaffinized with xylenes and decreasing concentrations of ethanol (100%, 95%) and rehydrated with deionized water. Antigen retrieval was conducted by steaming slides in citrate buffer (pH 6.0) for 35 minutes. After cooling, slides were then blocked using normal horse serum (Vector Laboratories, S-2012–50) for 20 minutes at room temperature in a humidified chamber. Primary antibody against MDA-5 (Abcam, Ab315242, Rabbit, 1:1,000), RIG-1 (Abcam, Ab302778, Rabbit, 1:1,000) Neutrophils (Novus Biologicals, RM0028–3G23, Rat, 1:400) were then applied respectively and incubated overnight at 4C. Primary antibody against F4/80 (Bio-Rad, MCA497, Rat, 1:200) was applied and incubated at room temperature for 1h. Samples were then washed with 1X PBS before labeling with host-matched secondary antibodies (Vector Laboratories) and incubating for 30 minutes at room temperature. The samples were washed with 1X PBS and then developed with 3,3’-diaminobenzidine substrate kit (Vector Laboratories, SK-4100) for 5 minutes. Sections were then dehydrated ethanol (70%, 95%, 100%) and xylene before mounting, drying overnight, and imaging using a Keyence BZ-X800 microscope.

Quantification of Immunohistochemistry:

Neutrophil quantification: For each slide, 6 fields at 40 high-power field (HPF) were analyzed semi-quantitatively. DAB-positive cells were quantified in each field. Neutrophil counts were averaged for each slide.

Macrophage (F4/80+) quantification: For each slide, 6 fields at 40 high-power field (HPF) were analyzed. Using FIJI ImageJ, a uniform threshold was applied, and positive cells counts were then recorded via particle analysis. F4/80 counts were averaged across fields per slide.

TUNEL Staining:

Transverse kidney sections (group sample sizes n = 5 for all) were fixed in 10% neutral buffered formalin overnight and transferred to 70% ethanol before embedding in paraffin and cut into 4 μm sections before mounting and heat fixing on slides. Slides were stained following manufacturer’s instructions (C10618) and imaged using a Keyence BZ-X800 microscope. For each slide, 4 fields at 10 high-power field (HPF) were analyzed. Using FIJI ImageJ, fluorescent images were split by color channel, and Texas red-positive cell counts were recorded and averaged for each slide.

Statistical Approach:

Data are means ± standard error. For comparisons between two independent groups, an unpaired, two-tailed Student’s t-test was performed, assuming equal variances. For comparisons between more than two groups, a 1-way or 2-way analysis of variance (ANOVA) was performed followed by Dunnett’s multiple comparisons test. P < 0.05 was considered significant. All statistics were performed using GraphPad Prism (10.5.0).

RESULTS:

Characterization of Kidney Injury in Viral Primed SA-AKI Model

Sequential administration of poly(I:C)+LPS resulted in significant reductions in GFR by 4h post-LPS (21.37 ± 3.6 μL/min) compared to mice receiving vehicle (221.70 ± 18.04 μL/min, p < 0.001), poly(I:C) alone (169.90 ± 9.56 μL/min, p < 0.001), and low-dose LPS alone (111.60 ± 5.00 μL/min, p < 0.001) (Fig. 1B). This reduction in GFR persisted at 48h post-LPS. Plasma creatinine was statistically elevated in mice receiving poly(I:C)+LPS compared to mice receiving poly(I:C) alone at 48h (Fig. 1C). As a marker of tubular injury, urinary KIM-1 was elevated in mice receiving poly(I:C)+LPS at 48h compared to all other groups (Fig. 1D). Importantly, no kidney injury was observed when the reverse sequential administration of LPS followed by poly(I:C) at 24h was performed (Supplemental Fig. 2). A small cohort of age-matched wildtype female mice demonstrated similar findings after sequential administration of poly(I:C) followed by LPS at 24h (Supplemental Fig. 3).

Expression of KIM-1 (Havrc1) was elevated in bulk kidney tissue (Fig. 1E) and kidney protein lysate (Fig. 1F) at 48h in mice receiving poly(I:C)+LPS. KIM-1 localized via immunofluorescence to proximal tubular epithelial cells in mice receiving poly(I:C)+LPS but not in mice receiving poly(I:C) or LPS alone (Fig. 1G). No differences were seen in apoptosis in any group at 48h per TUNEL staining (Fig. 1H, Supplemental Fig. 4).

Viral Primed SA-AKI is Associated with Hyperferritinemia and Systemic Inflammation

As a marker of hyperferritinemic sepsis and sHLH, plasma ferritin was elevated at 8h post-LPS in mice receiving poly(I:C)+LPS (776.8±117.8 ng/mL) compared to mice receiving LPS (237.8±37.6 ng/mL, p < 0.001) or poly(I:C) (331.9±57.5 ng/mL, p = 0.002) alone (Fig. 2A). Systemic plasma cytokines, including IL-6, TNF-α, and KC/CXCL1, were also elevated at 8h post-LPS in mice receiving poly(I:C)+LPS compared to mice receiving LPS or poly(I:C) (Fig. 2B). With regards to circulating cell counts, poly(I:C)+LPS was associated with a significant reduction in WBC count, lymphocyte count, and platelet count at 8h post-LPS compared to vehicle. There were no differences in circulating monocyte levels among any of the groups (Fig. 2C). Cytokine levels and blood cell counts at 48h post-LPS are included in Supplementary Table 1.

Figure 2. Circulating Inflammatory Cytokines and Immune Cells in Viral Primed SA-AKI Model.

(A) Plasma ferritin levels at 8h post-LPS; (B) Plasma IL-6, TNF-α, and KC/CXCL1 at 8h post-LPS; (C) Circulating cell counts at 8h post-LPS. Statistical comparisons were performed using a 1-way ANOVA followed by Dunnett’s multiple comparisons test. **p<0.01, ***p<0.001, ****p<0.0001. IL-6, interleukin-6; TNF-α, tumor necrosis factor-α; KC/CXCL1, keratinocyte-derived chemokine/CXC-ligand 1 (KC/CXCL1); WBC, white blood cell count; LYM, lymphocyte count; MONO, monocyte count; PLT, platelet count

Poly(I:C) Induces Intrarenal Type 1 IFN Signaling Prior to LPS challenge

To test whether viral primed SA-AKI pathogenesis is dependent on a direct renal response to viral stimuli rather than a mechanism driven by infiltrating immune cells, we aimed to identify a kidney-specific receptor for poly(I:C) in our murine model and characterize the temporal expression of known poly(I:C) receptors following exposure (Fig 3A). Expression levels of Tlr3, Ifih1 (MDA-5), and Ddx58 (RIG-1) were all increased at 4h and 12h post-poly(I:C) in mice receiving poly(I:C) compared to vehicle (Fig. 3B). At the protein level, we could not identify TLR3 in kidney lysate at 4h or 12h post-poly(I:C) (Fig. 3C). However, we were able to detect both MDA-5 and RIG-1 at 4h and 12h post-poly(I:C) compared to mice receiving vehicle (Fig. 3D). Additionally, we confirmed increased expression of TLR4, the cognate receptor for LPS, at 4h following exposure to poly(I:C) (Fig. 3D). By immunohistochemistry, MDA-5 localized to the tubular epithelial cells whereas RIG-1 localized to the glomeruli and peri-tubular endothelial cells (Fig. 3E).

Figure 3. Intrarenal Poly(I:C) Recognition by Various Receptors.

(A) Schematic depicting three candidate receptors for poly(I:C) in the kidney. (B) RT-qPCR of RNA from whole kidney lysate demonstrating expression of Tlr3, Ifih1, and Ddx58 at 4h and 12h post-poly(I:C). (C) Western blot demonstrating absence of TLR3 in bulk kidney lysate. (D) Western blot demonstrating expression of MDA-5, RIG-1, and TLR4 along with quantification by densitometry, normalized to GADPH. (E) Immunohistochemistry demonstrating the renal cortical expression of MDA-5 and (F) RIG-1 at 4h and 12h post-poly(I:C). Statistical comparisons were performed using either a 1-way or 2-way ANOVA followed by Dunnett’s multiple comparisons test as appropriate. **p<0.01, ***p<0.001, ****p<0.0001. N= 4–5. TLR3, toll-like receptor-3; TLR4, toll-like receptor 4; poly(I:C), polyinosinic-polycytidylic acid; MDA-5, melanoma differentiation-associated protein-5; RIG-1, retinoic acid-inducible gene 1; IFN, interferon.

Given the exaggerated severity of kidney injury mediators in mice receiving poly(I:C) prior to low-dose LPS compared to mice receiving low-dose LPS alone, we probed the type 1 interferon pathway to determine the temporal kinetics of interferon signaling prior to LPS administration (Fig. 4A) (17). Both plasma interferon-α (IFNα) and interferon-β (IFNβ) were significantly increased by 4h post-poly(I:C) compared to vehicle (Fig. 4B). Plasma IFNα remained elevated at 12h whereas IFNβ levels returned to baseline by 12h; both type 1 IFNs were back to baseline by 8h post-LPS (32h post-poly(I:C). No significant changes in IFNα or IFNβ were observed in mice following LPS administration (Fig. 4B).

Figure 4. Poly(I:C) Induces the Type 1 Interferon Signaling Pathway in the Kidney.

(A) Timeline depicting post-poly(I:C) investigations and their temporal relationship to a future LPS exposure. (B) Plasma levels of IFNα and IFNβ across experimental timepoints. (C) RT-qPCR of bulk kidney demonstrating expression of Ifnb1 and (D) Irf3 at 4h and 12h post-poly(I:C). (E) Western blot of IRF3 expression, quantified by densitometry with normalization to GAPDH. (F) RT-qPCR from whole kidney lysate demonstrating expression of Ifnar2. Statistical comparisons were performed using either a 1-way or 2-way ANOVA followed by Dunnett’s multiple comparisons test as appropriate. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. N= 4–5. IFN, interferon; poly(I:C), polyinosinic-polycytidylic acid; LPS, lipopolysaccharide; IRF3, interferon regulatory factor-3.

In the kidney, Ifnb1 expression was increased at 4h post-poly(I:C) (Fig. 4C). At 12h post-poly(I:C), Ifnb1 expression returned to normal. As one of the key transcription factors regulating type 1 IFN expression, mRNA levels of interferon regulatory factor-3 (IRF3, Irf3) were no different at 4h post-poly(I:C), but mice receiving poly(I:C) had increased Irf3 at 12h post-poly(I:C) compared to mice receiving vehicle (Fig. 4D). IRF3 was quantified in bulk kidney and was found to be elevated at both 4h and 12h post-poly(I:C) in mice receiving poly(I:C) compared to vehicle (Fig. 4E). As a subunit for the cognate receptor for IFNα and IFNβ, expression of interferon α/β receptor (IFNAR, Ifnar2) was increased at both 4h and 12h following poly(I:C) administration (Fig. 4F).

Intrarenal Inflammatory Signaling at 48h post-LPS in the Viral Primed SA-AKI Model

We investigated the intrarenal inflammatory profile and expression patterns of poly(I:C) receptors at 48h post-LPS given the persistent reduction in GFR and elevations in creatinine and urine KIM-1 among mice receiving poly(I:C)+LPS at this timepoint. Despite the well-established positive feedback in viral pattern recognition receptor expression following poly(I:C) exposure, TLR3 kidney expression could not be localized to the kidney at 48h by immunofluorescence or western blot (Fig. 5A–B) (18,19). mRNA expression of Ddx58 (RIG-1) and Ifnar2 demonstrated had increased kidney expression in mice receiving poly(I:C)+LPS compared to vehicle, whereas Tlr3, Ifih1, Ifnb1 and Il6 demonstrated no differences by 48h post-LPS between poly(I:C)+LPS and vehicle (Fig. 5C). No difference in MDA-5 nor RIG-1 protein abundance was observed in the kidney at 48h (Fig. 5D,F). IRF3 abundance in kidney lysate at 48h was increased in mice receiving poly(I:C)+LPS relative to LPS or vehicle alone, but no difference was observed in IRF3 abundance between mice receiving poly(I:C)+LPS and poly(I:C) alone (Fig. 5E–F).

Figure 5. Intrarenal Inflammatory Signaling post-LPS in Viral Primed SA-AKI Model.

(A) Immunofluorescence and (B) Western Blot continue to show absence of kidney TLR3 at 48h post-LPS. (C) RT-qPCR of bulk kidney demonstrating expression of Tlr3, Ifih1, Ddx58, Ifnar2, Ifnb1, and Il6 at 48h post-LPS. (D) Western blot of MDA-5 and RIG-1 from bulk kidney lysate at 48h post-LPS, along with (E) western blot of IRF3 from bulk kidney lysate at 48h post-LPS. (F) Western blots quantified by densitometry and normalized to GAPDH. Statistical comparisons were performed using a 1-way ANOVA followed by Dunnett’s multiple comparisons test. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. TLR3, toll-like receptor-3; LPS, lipopolysaccharide; poly(I:C), polyinosinic-polycytidylic acid; MDA-5, melanoma differentiation-associated protein-5; RIG-1, retinoic acid-inducible gene 1; IRF3, interferon regulatory factor-3

Immune Cell Recruitment to the Kidney in Viral Primed SA-AKI Model

We sought to characterize changes in neutrophils and macrophages in the kidney during our viral primed SA-AKI model given the early sepsis timepoints under investigation in this study. Poly(I:C) administration was associated with increased kidney neutrophil abundance at 4h and 12h post-poly(I:C) compared to vehicle, whereas no changes were observed in macrophage numbers (Fig. 6A–D). At 48h post-LPS, neutrophil abundance was increased in mice receiving LPS and poly(I:C)+LPS, but not in mice who received poly(I:C) alone (Fig. 6E–F). At 48h post-LPS, macrophage abundance was elevated in the kidneys of mice receiving poly(I:C)+LPS, but not in any of the other conditions (Fig. 6G–H).

Figure 6. Intrarenal Immune Cell Infiltration in Viral Primed SA-AKI.

(A) Presence of kidney neutrophils (anti-neutrophil) at 4h and 12h post-poly(I:C) by immunohistochemistry, (B) quantified by cells per high powered field (HPF). (C) Presence of kidney macrophages (anti-F4/80) at 4h and 12h post-poly(I:C), (D) quantified by cells per HPF. (E) Presence of kidney neutrophils 48h post-LPS, (F) quantified by cells per HPF. (G) Presence of kidney macrophages 48h post-LPS, (H) quantified by cells per HPF. Statistical comparisons were performed using a student’s t test or 1-way ANOVA followed by Dunnett’s multiple comparisons test. *p<0.05, **p<0.01, ***p<0.001. N= 5–8. Poly(I:C), polyinosinic-polycytidylic acid; LPS, lipopolysaccharide; Neu, neutrophil

No Differences are Observed in Intrarenal Iron Handling in Viral Primed SA-AKI

Given the translational association with viral primed SA-AKI and hyperferritinemic sepsis, we attempted to quantify intrarenal iron and ferritin to determine whether this model of viral primed SA-AKI impacts intrarenal iron handling. Upon probing kidney tissue for both ferritin light (FtL) and heavy chains (FtH) via both Western blot (Fig. 7A) and immunohistochemistry (Fig. 7B–C), no differences were noted at 48h post-LPS in mice receiving poly(I:C)+LPS compared to the other groups. Prussian Blue staining confirmed no detectable iron deposition in any experimental group at this timepoint (Fig. 7D).

Figure 7. Assessment of Kidney Iron Storage in Viral Primed SA-AKI.

(A) Western blot for FtL and FtH in bulk kidney lysate at 48h post-LPS, quantified by densitometry and normalized to GAPDH. (B) Immunohistochemistry for FtL and (C) FtH in the kidney 48h post-LPS. (D) Prussian blue staining does not reveal any significant iron staining in the viral primed SA-AKI model 48h post-LPS. Statistical comparisons were performed using a 1-way ANOVA followed by Dunnett’s multiple comparisons test. N= 4 per group.

DISCUSSION:

We developed a novel two-hit murine model of viral primed SA-AKI characterized by hyperferritinemia, early and marked reductions in renal function, and amplification of type 1 interferons with associated increases in intrarenal pattern recognition receptor (PRR) expression. This model was conceptually adapted from a murine model of secondary hemophagocytic lymphohistiocytosis (sHLH), which has demonstrated strong translational validity due to similar features in humans, including hemophagocytosis and high ferritin levels(12). Compared to the sHLH model, our viral-primed SA-AKI model utilizes lower doses of poly(I:C) (2.5 mg/kg) and LPS (0.5 mg/kg), yet still induces robust kidney injury as evidenced by increased serum creatinine, elevated urinary KIM-1 excretion, and significant reductions in GFR, quantified via transcutaneous FITC-sinistrin clearance. Importantly, viral priming appears to induce worsening SA-AKI severity regardless of biological sex. This model provides a promising platform for investigating mechanisms underlying SA-AKI with enhanced relevance to viral-bacterial co-infection in human sepsis.

Poly(I:C), a synthetic analog of viral double-stranded RNA (dsRNA), acts through multiple PRRs, with TLR3 being its most well-established target(20). Poly(I:C) binding to TLR3 initiates downstream antiviral responses via type 1 IFN signaling(21). TLR3 exists in two forms: a membrane-bound version with an extracellular domain and an intracellular form localized to endosomal membranes.(17). TLR3 expression has been documented in various cell types, including myeloid dendritic cells, NK cells, endothelial cells, hepatocytes, lung fibroblasts, neurons, astrocytes, and microglia(21). In the kidney, TLR3 expression has been found in glomerular mesangial cells, proximal and distal tubular epithelial cells, and vascular smooth muscle cells(22–25). Most kidney-related TLR3 studies have focused on lupus or hepatitis C associated glomerulonephritis(25,26). In the context of AKI, one prior study using a murine model of bilateral ischemia-reperfusion injury (BIRI) showed that global TLR3 knockout mice had lower serum creatinine levels and reduced tubular necrosis and apoptosis compared to wildtype controls(24). Although this model did not involve viral stimulation, these findings support a role for TLR3 in AKI pathogenesis. In our viral primed SA-AKI model, we confirmed an increase in Tlr3 mRNA expression in whole kidney tissue by 4h post-poly(I:C); however, we were unable to detect or spatially localize TLR3 protein. This discrepancy may reflect post-transcriptional regulation, targeted TLR3 proteasomal degradation, or TLR3 protein levels below the detection threshold in kidney lysates. However, other studies have also reported increase Tlr3 mRNA in the kidneys without clear demonstration of intrarenal TLR3 protein localization(24).

In addition to TLR3, viral dsRNA analogs are also known to illicit type 1 IFN responses via MDA-5 and RIG-1, two intracellular receptors in the RIG-1-like receptor (RLR) family that are activated by both viral and host RNA and DNA(27). MDA-5 is ubiquitously expressed and generates type 1 IFNs after binding to dsRNA from positive-sense RNA viruses such as picornavirus(28). RIG-1 is expressed in the cytoplasm of most cell types, but it has also been localized to the nucleus(29). Prior studies have demonstrated that MDA-5 preferentially responds to long dsRNA, such as poly(I:C), and RIG-1 detects short dsRNA and single-stranded RNA (ssRNA) with a 5’-triosphate end(15,30). Our data demonstrates that MDA-5 and RIG-1 are both expressed in the kidney, and that administration of poly(I:C) leads to upregulation of both RLR proteins by 4h post exposure. MDA-5 expression localized to tubular epithelial cells in poly(I:C)+LPS mice whereas RIG-1 expression was observed in the glomeruli and basolateral border of tubular epithelial cells, potentially corresponding to renal endothelial cells. Increased mRNA and protein abundance of these RLRs in response to viral priming with poly(I:C) is consistent with prior studies that indicate poly(I:C) signaling leads to increased expression of a variety of viral-related PRRs(18,19).

Following recognition of poly(I:C) by TLR3, MDA-5, and/or RIG-1, there is downstream pathway convergence leading to activation and phosphorylation of IRF3 with subsequent transcription of type 1 IFNs – IFNα and IFNβ (31,32). We demonstrate that plasma levels of IFNα and IFNβ increase by 4h following poly(I:C) administration but return to baseline prior to administration of low-dose LPS at 24 hours. Likewise, kidney expression of Ifnb1 (ligand) and Ifnar2 (receptor) are increased at 4 hours following poly(I:C) administration indicating the rapid promotion of intrarenal antiviral pathways in response to poly(I:C). These data suggest that the type 1 IFN signaling pathway is activated by poly(I:C) with initiation of antiviral responses prior to administration of low-dose LPS at 24h. Other models of AKI have linked type 1 IFN signaling to increased immune cell recruitment, initiation of failed repair and apoptosis, and fibrotic remodeling(33,34). Notably, IFNAR-deficient mice subjected to BIRI are protected from kidney injury, suggesting a pathogenic role for type I IFNs in AKI(35). Our findings extend this concept by identifying type I IFNs as potential mediators of a priming effect that exacerbates kidney injury in the context of viral-bacterial co-infection. Importantly, the production of type 1 IFNs in response to poly(I:C) did not impact macrophage recruitment to the kidney prior to LPS administration, and by the 48h time point there was no difference in neutrophil or macrophage localization to the kidney among mice receiving poly(I:C) versus vehicle. This suggests that the exaggerated severity of kidney injury in response to viral priming is not solely dependent on proinflammatory infiltrating immune cells. One candidate mechanism for the increased severity of SA-AKI identified by our study relates to the upregulation of TLR4 observed within the kidneys by 4h post-poly(I:C). Perhaps, in addition to upregulation of viral PRRs in response to poly(I:C), other anti-microbial related PRRs are also increased within the kidney and produce an amplified response to a secondary bacterial challenge. This mechanism warrants future investigation.

The viral primed model of SA-AKI outlined here has significant translational relevance. Viral co-infection in bacterial sepsis is common but frequently underrecognized. In one study using multiplex PCR on nasopharyngeal swabs from septic adults, 70% of patients with viral positivity would not have been identified based on clinical suspicion alone(36). Among pediatric patients with influenza during the 2009–2010 influenza A pandemic, 33% developed clinical evidence of a bacterial co-infection within 72 hours of admission to a pediatric intensive care unit(37). H1N1 influenza infection was associated with elevations in serum creatinine in 20% of children and the need for dialysis in 5.3% of children, conferring a relative risk for death of 4.5(37). Additionally, viral DNAemia has been associated with hyperferritinemic sepsis in humans, and hyperferritinemia itself is a known risk factor for AKI in sepsis(11,38). Thus, our viral-primed SA-AKI model recapitulates clinically relevant features and provides a platform to investigate this important subphenotype.

This study has limitations. First, the characterization of our viral primed SA-AKI murine model is limited to pathways stimulated by analogs to infectious triggers, not by true infection. This precludes a rigorous investigation into the mechanisms by which individual viruses evade the immune system, develop tissue tropism, and interact with the host immune system. However, the conceptual framework for viral priming followed by bacterial infection has been reported in both humans and in mice(11,12). Future studies incorporating live viral infection followed by bacterial challenge will help validate and refine this model. Second, the precise window by which viral priming exerts a detrimental impact on SA-AKI severity remains unclear. We selected a 24h interval between sequential administration of poly(I:C) and LPS to align with prior literature of increased murine mortality if poly(I:C) is administered between 24–36 hours prior to LPS; this synergistic lethality is significantly blunted if the priming window is extended to 48h(12). Third, although septic shock is known to impair renal perfusion, we did not directly assess hemodynamic parameters. Further studies are needed to determine whether our model induces hemodynamic instability and whether interventions aimed at optimizing renal perfusion can attenuate tubular injury(39–41).

CONCLUSIONS:

We established a novel murine model of viral-primed SA-AKI through sequential administration of poly(I:C) and low-dose LPS. This model induces early and significant markers of kidney injury and is associated with robust activation of intrarenal type I IFN signaling prior to the bacterial stimulus. By recapitulating features of hyperferritinemic sepsis and viral-bacterial co-infection, this model offers a translationally relevant framework to explore the mechanisms driving this severe SA-AKI subphenotype.

Supplementary Material

Supplemental Files: https://doi.org/10.17632/227nxwx653.3

DATA AVAILABILITY:

The data supporting the findings from this study are available within the text and figures of this manuscript. The authors will take all reasonable efforts to make any remaining raw data available by request.