Ferritin heavy chain floxed mice (FtH^fl/fl) and myeloid-specific ferritin heavy chain knockout mice (FtH^LysM–/–) were used for this study and have been previously described (35). Briefly, myeloid-specific FtH knockout mice, on a C57BL/6 background were generated by breeding FtH^fl/fl and LysM Cre mice. All mice were humanely treated, and all methods were approved by the Institution Animal Care and Use Committee of the University of Alabama at Birmingham (Approval – 21255). Studies were executed in accordance with all National Institutes of Health guidelines.

Two separate protocols were used to induce rhabdomyolysis: (1) single dose of glycerol administration, (2) unilateral nephrectomy followed by glycerol administration. Glycerol mediated rhabdomyolysis was induced in age matched FtH^fl/fl and FtH^LysM–/– mice as previously described (13). Briefly, mice were deprived of water for 16 h prior to glycerol administration. Mice were anesthetized with isoflurane and injected with 50% glycerol in water, 7.5 ml/kg body weight, with half the volume delivered into each anterior thigh muscle. A cohort of mice underwent unilateral nephrectomy and 2 weeks later, the mice were deprived of water and administered glycerol as described above. Mice were sacrificed at various timepoints following injection and kidneys were analyzed for gene and protein expression. In an additional cohort of mice, blood was collected via facial vein from the same animals at multiple time-points. Blood was collected via cardiac puncture or facial vein bleed, and serum was isolated for measurement of creatinine and other injury markers. Urine was collected at the time of harvest via bladder puncture and analyzed for markers of kidney injury.

ELISA analyses for KIM-1 (Cat no. MKM100, R&D Systems), NGAL (Cat no. MLCN20, R&D Systems), Cystatin C (Cat no. MSCTC0, R&D Systems), and Albumin (Cat no. E99-134, Bethyl Laboratories) were performed on serum or urine following the manufacturer’s instructions. Urine and serum creatinine were measured by LC-MS/MS. Data were expressed as nanograms per milliliter or were normalized to urine creatinine and expressed as nanograms per milligram creatinine.

Total RNA was isolated from tissues using TRIzol reagent (Invitrogen), cDNA was generated from the total RNA using the Maxima H Minus First Strand cDNA Synthesis Kit (Cat no. K1651, Thermo Scientific). SYBR Green-based real-time PCR was performed on the cDNA using PowerUp SYBR Green Master Mix (Cat no. A25742, Applied Biosystems). Relative mRNA expression was quantified using the ΔΔCt method. All values were normalized to GAPDH as an internal control. All reactions were performed in triplicate and melting curve analysis was used to confirm specificity. All real-time primers used are listed in Table 1.

Harvested tissues were homogenized and protein was isolated as previously described (35). Briefly, tissues were lysed in protein lysis buffer (10 mM/1 Tris–HCl, 5 mM/1 EDTA, 150 mM/1 NaCl, 10% Non-idet P-40, and 10% Triton-X) with protease (Sigma-Aldrich) and phosphatase (Biotool) inhibitors. Protein was isolated and quantified with the bicinchoninic acid protein assay (Cat no. 23227, ThermoFisher). Protein (50 μg) was resolved on 12% Tris-glycine sodium dodecyl sulfate polyacrylamide gel electrophoresis or 4–12%, BIS-TRIS polyacrylamide gel (ThermoFisher) and transferred to a PVDF membrane (EMB Millipore). Membranes were blocked with 5% non-fat dry milk in TBST for 1 h and then incubated with the following antibodies: a mouse anti-FtH (Cat no. sc517348, Santa Cruz, 1:1,000), a rabbit anti-phospho-JNK (Cat no. 4668, Cell Signaling Technology, 1:1,000), a rabbit anti-JNK (Cat no. 9252, Cell Signaling Technology, 1:1,000), a rabbit anti-phospho-p44/42 MAPK (Cat no. 9101, Cell Signaling Technology, 1:1,000), a rabbit anti-p44/42 MAPK (Cat no. 4695, Cell Signaling Technology, 1:1,000), a rabbit anti-fibronectin (Cat no. F3648, Sigma-Aldrich, 1:10,000), a mouse anti-smooth muscle actin (Cat no. A6228, Sigma-Aldrich, 1:5,000), a mouse anti-4-HNE (Cat no. MA33249, R&D Systems, 1:1,000), a goat-anti NGAL (Cat no. AF1857, R&D Systems, 1:2,000), a goat-anti KIM-1 (Cat no. AF1817, R&D Systems; 1:1,000), a mouse anti-GAPDH (Cat no. MAB314, Millipore, 1:10,000) followed by an incubation with peroxidase-conjugated anti-mouse (Kindle BioSciences, Greenwich, CT, United States), or anti-rabbit (Kindle BioSciences), or donkey anti-goat (Cat no. R1002, Kindle BioSciences). Horseradish peroxidase activity was detected using the enhanced chemiluminescence KwikQuant detection system (Kindle BioSciences). Membranes were stripped and reprobed with a mouse anti-GAPDH (Cat no. MAB314, Millipore, 1:10,000) to confirm loading and transfer. Densitometry analysis was performed using ImageStudioLite.

Transverse kidney sections were fixed in 10% neutral buffered formalin for 24 h then embedded in paraffin. Tissues sections were deparaffinized with xylene and rehydrated before antigen retrieval using citrate buffer. Tissues were blocked in normal horse serum, followed by incubation with Ki67 (Cat no. 16667, ABCAM) or 4-HNE (Cat no. MA33249, R&D Systems) or NGAL (Cat no. AF1857, R&D Systems). Sections were washed and then incubated with anti-rabbit (or anti-goat or anti-mouse) secondary antibody (Vector Laboratories) following the manufacturer’s instructions. Sections were washed again and exposed to peroxidase substrate, 3,3-diaminobenzidine (DAB) (Vector Labs), washed in distilled water, dehydrated, mounted, and then visualized by light microscopy. Images were analyzed using ImageJ software (National Institutes of Health). The ApopTag Plus Peroxidase In Situ Apoptosis Kit (Cat no. S7101, Sigma-Aldrich) was also used to perform a terminal deoxynucleotidyl dUTP nick end labeling (TUNEL) stain on kidney sections following the manufacturer’s protocol.

Collagen deposition was measured using picrosirius red staining as previously described (36). Briefly, tissues were fixed in 10% neutral buffered formalin for 24 h then embedded in paraffin. Tissues were sectioned and deparaffinized with xylene and ethanol washes. The sections were then incubated in picrosirius red for 1 h, washed with acidified water, dehydrated, mounted, and then visualized by light microscopy. Collagen deposition was imaged under polarized light on a Leica DMi8 microscope, and percent area was quantified using Leica LASX software.

The capsules were removed from the harvested kidneys before weighing each kidney. Kidneys were then minced and enzymatically digested using 2 ml 1.67 Wunsch Units per ml Liberase DL (Roche) in DMEM medium (Gibco) for 30 min at 37°C. Cold PEB (PBS, 2 mM EDTA, 1% BSA) was added, and the suspensions were passed through an 18 g needle three times before being filtered through a 40 μm filter (Fisher Scientific). Red blood cells were lysed using the ACK lysis buffer. Samples were incubated with Live/Dead Fixable Violet Dead Cell Stain Kit (Invitrogen) according to the manufacturer’s protocol. After washing with staining buffer (PBS, 0.5% BSA, 0.01% Sodium Azide), samples were incubated in anti-mouse CD16/32 for 10 min prior to incubation with specific primary antibodies. Cells were stained with the following panel for 30 min in the dark at 4°C: Brilliant Violet 650 anti-CD45.2 (BioLegend, clone 104), Super Bright 600 anti-CD11b (eBioscience, clone M1/70), PerCp anti-CD8a (BD Biosciences, clone 53-6.7), FITC anti-CD4 (eBioscience, clone GK1.5), Super Bright 436 anti-Ly6C (eBioscience, clone HK1.4), Alexa Fluor 700 anti-Ly6G (BioLegend, clone 1A8), APC-eFluor 780 anti-F4/80 (eBioscience, clone BM8). Samples were washed in staining buffer, centrifuged at 400 g for 5 min, and the supernatants were aspirated. Samples were then fixed and permeabilized with Foxp3 Transcription Factor Staining Buffer (eBioscience, clone FJK-16s) following the manufacturer’s protocol and stained with APC anti-FoxP3, washed and resuspended in 500 μl of staining buffer for flow cytometry. Data were acquired using a Cytek Northern Lights spectral cytometer with associated SpectroFlo software.

Data are presented as mean ± SEM. For comparisons involving three or more groups, the ANOVA test was used. For repeated measures on the same animal at multiple timepoints, ANOVA test (post-test Dunnett’s, Tukey or Šídák multiple comparison test) was used. Unpaired two-tailed Student’s t-test was used for comparison between two groups. A p-value of less than 0.05 was considered significant.

Given the significance of FtH in the safe sequestration of the potentially detrimental iron, we hypothesized that heme-mediated AKI would be exacerbated in the absence of myeloid FtH. To this end, we administered glycerol to wild-type (FtH^fl/fl) and myeloid FtH deficient mice (FtH^LysM–/–) and found similar kidney injury in each genotype during the acute phase (1- and 3-days post-glycerol injection), as evident by a significant rise in serum creatinine (Figure 1A). Similar results were observed in kidney injury markers measured in the serum (NGAL, KIM-1) and urine (NGAL, KIM-1, albumin) at various timepoints following glycerol injection (Figures 1B–H). Of note, repeated blood collections from an additional cohort of mice at various time-points following injury showed an increase in BUN, cystatin C, and creatinine during the acute phase. However, there were no significant differences between the genotypes (Figures 1G,H and Supplementary Figure 1). Renal histology further corroborated these findings with comparable structural damage between the genotypes following injury (Figure 2). An increase in necrotic tubules and tubular casts were identified in the kidney cortex and medulla starting at 1-day post-glycerol injection and persisted until day 7. Interestingly, although fewer in number, casts were present even 28 days post-glycerol administration in both groups.

Intra-renal KIM-1 gene expression was rapidly induced in both genotypes and remained persistently elevated until day 7 following glycerol injection (Figure 3A). While NGAL gene expression was markedly induced in FtH^fl/fl kidneys after 1 day of glycerol administration, this induction was lower in myeloid FtH deficient kidneys, an effect that persisted until day 7 (Figure 3B). Intrarenal FtH mRNA transcripts were reduced early after injury in both genotypes but were restored by day 28 in both groups (Figure 3C). Concomitantly, we observed an increase in FtH protein expression at the earlier timepoints, suggesting that the decline in mRNA may correspond with enhanced translation of FtH. Corroborating our gene expression data, rhabdomyolysis was associated with an increase in intrarenal NGAL and KIM-1 protein expression (Figures 3D,E). NGAL was predominantly expressed in the tubular epithelium in a punctate pattern in the cytoplasm (Supplementary Figure 2). Induction of KIM-1 was higher in mice that were deficient for myeloid FtH at day 1 (Figure 3D). We further determined that activation of MAPK pathways, JNK and ERK, were not differentially regulated in the absence of FtH during injury (Figures 3D,E). As rhabdomyolysis is associated with excess heme burden and oxidative stress, we examined kidneys for lipid peroxidation. As shown in Figure 3D, glycerol administration is associated with an increase in lipid peroxidation in both genotypes, as evident by an increase in expression of 4-hydroxynonenal. Immunohistochemical staining of 4-HNE revealed an intense tubular expression during the acute phase of rhabdomyolysis which reduced by day 7 (Supplementary Figure 3).

Cell death, as determined by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay was increased at day 1 in both genotypes compared to later timepoints and was not different between the genotypes (Figure 4A). Cell proliferation, as evaluated by Ki67 immunostaining, was only found to be significantly elevated at day 3 and no significant difference between the genotypes was observed (Figure 4B). These data collectively indicated that loss of myeloid FtH did not impact early response to rhabdomyolysis-induced cell death and cell proliferation.

While the initial damage associated with heme burden was not differentially affected by the absence of myeloid FtH, we identified that interstitial fibrosis after 4 weeks of injury was markedly lower in mice deficient for FtH. As shown in Figures 5A–D, picrosirius red staining was increased at day 7 following injury and continued to progress in FtH^fl/fl mice but this progression was not observed in the myeloid FtH deficient mice. Quantification of percent PSR-positive area under polarized light further confirmed these findings (Figures 5B,C). At the transcriptional level, we determined that fibrotic remodeling began as early as 3 days post-injury and peaked at 7 days, as demonstrated by increased gene expression of collagen and α-smooth muscle actin, with no significant changes in fibronectin expression (Figure 6A). At the protein level, we demonstrate notable induction of fibrotic markers at day 7 and 28 after injury (Figure 6B). Corroborating the histological data, myeloid FtH deficiency was associated with lesser induction of TGF-β at 7 days and consequently, lower expression of α-SMA at 28 days after injury (Figures 6B,C). In addition to fibrotic markers, FtH knockout mice also expressed significantly lower levels of S100A8, a potent immunomodulatory gene as well as TNF-α, a prominent pro-inflammatory cytokine during the peak of fibrogenic signaling.

To determine whether loss of kidney mass by unilateral nephrectomy would impact injury and fibrosis outcomes in the absence of myeloid FtH, we subjected control and myeloid FtH deficient mice to unilateral nephrectomy followed by glycerol administration. We found that glycerol induced kidney injury was similar between the genotypes, as evident by a temporal increase in levels of BUN, cystatin C, and NGAL during the early timepoints (Figures 7A–C). At 28 days following glycerol administration, expression of the fibrotic marker, α-SMA was significantly higher only in the wild-type mice (Figures 7D–F). Additionally, picrosirius red staining for fibrosis further confirmed the increase in fibrosis in the wild-type mice, an effect that was dampened in the absence of myeloid FtH (Figures 7G–I).

To determine whether the difference observed in fibrosis could be attributed to immune cell populations, we next assessed intrarenal lymphoid and myeloid cell populations after glycerol administration (Figure 8). Relative to quiescent controls, both genotypes experienced a similar rise in percent CD45+ cells that remained elevated until day 28. We also observed a temporal response in the lymphoid and myeloid populations after injury. Specifically, myeloid cells dominated the leukocyte population at early timepoints whereas, lymphoid cells were the dominant leukocyte population at day 28. We further analyzed the myeloid cell populations and determined that loss of myeloid FtH did not impair trafficking of neutrophils and infiltrating macrophages into the kidneys during injury and also did not influence accumulation of resident macrophages. Upon further characterization of the lymphoid populations, we determined that rhabdomyolysis was associated with a temporal increase in CD4+, CD8+, and regulatory T cells, with maximal abundance identified on the day 28 after injury. Overall, difference in fibrotic remodeling in the FtH deficient mice could not be attributed to alterations in absolute numbers of intra-renal immune cell populations during rhabdomyolysis.