A rat model of Immunologic and hypertensive kidney injury

Abstract

Non-surgical rodent chronic kidney disease (CKD) models for both glomerular and tubular injuries are currently limited. The current study aimed to develop a rat model of CKD by combining anti-Fx1A with N(ω)-Nitro-L-Arginine Methyl Ester (L-NAME) administrations. Rats were assigned to groups receiving L-NAME, anti-Fx1A, anti-Fx1A + L-NAME, or vehicle. Renal function, stiffness, renal injury biomarkers, histopathology and renal genome-wide transcriptomic changes were evaluated. Protein and renal injury biomarker levels in urine were elevated in the anti-Fx1A alone and combination group. Shear wave elastography revealed increased stiffness of the kidneys in all treatment groups. Histopathological evaluation revealed glomerular injury, characterized by enlarged glomeruli with increased hyaline materials in both anti-Fx1A groups and tubular degeneration/regeneration in the renal cortex of all treated groups with the highest incidence and severity in the combination group. These tubular changes were sometimes accompanied by interstitial mononuclear cell infiltrates and interstitial fibrosis. Proteinuria and mild changes in blood, urine renal injury biomarkers and imaging endpoints were noted in association with these histopathologic changes. The concurrence and higher incidence and/or severity of glomerular and tubular injuries in the combination group indicates that this would be a useful and relevant CKD model suitable for mechanistic, pharmacologic and toxicologic investigations.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-23316-0.

Introduction

Chronic kidney disease (CKD) is a major and growing challenge for healthcare systems worldwide. The prevalence of CKD appears to be increasing globally^1,2 and is likely to increase further as a consequence of ageing populations and increased prevalence of type II diabetes mellitus and hypertension. Patients with hypertension are at higher risk for the development of CKD³ and end-stage renal disease (ESRD)⁴. Hypertensive nephropathy is a common complication of hypertension and the second leading secondary cause of CKD and ESRD globally^5,6. Pre-clinically, animal models of CKD are crucial for investigating new mechanistic pathways and validating potential therapeutic interventions prior to clinical trials. Although mice have provided valuable model systems for nephrology research, rats remain the preferred preclinical model due to their physiology, anatomy and larger blood and tissue volume. However, non-surgical CKD models for rats are limited. Commonly used experimental renal models for basic research and drug development include the unilateral ureteral obstruction-induced fibrosis mouse model, the 5/6 nephrectomy rat model, the folic acid-induced renal tubular injury model, and the adriamycin-induced mouse model⁷. While these models still have a place in CKD research, they do have some disadvantages in that they either need invasive surgical procedures, exhibit strain-specific sensitivity, or induce undesirable systemic toxicity, or even mortality. Although non-surgery hypertension-induced CKD models are available (e.g. DOCA/NaCl rat model⁸, they might not fully capture the complexity of human hypertension due its neurogenic nature⁹. Strain-related variability in susceptibility is another significant limitation to many of the existing mouse CKD models particularly in renal fibrosis induction. The anti-Fx1A induced Passive Heymann Nephritis (PHN) model¹⁰ has been widely studied and utilized in various rat strains for membranous glomerulonephritis and podocyte injuries¹¹. L-N^G-Nitro arginine methyl ester (L-NAME), an antagonist of nitric oxide synthase (NOS), has been used to induce hypertension in rodents¹². It also induces progressive CKD with systemic hypertension^13,14. These animals exhibit massive macrophage infiltration in the renal cortex¹⁴, which has been suggested to lead to interstitial fibrosis¹⁵.

Furthermore, the 3Rs principle (replacement, reduction, and refinement) is crucial in animal-based research. One reduction strategy is to combine assays when suitable, which can greatly decrease the number of animals used in experiments. We hypothesized that merging the PHN model with the L-NAME model can lead to a substantial reduction in the use of animals. This approach can also lower costs and shorten timelines in the pre-clinical phase of drug development. Nonetheless, it’s essential to acknowledge that combining different models may lead to interactions due to overlapping biological mechanisms. For example, local renal vascular dysfunction is one of the first pathological events that occurs in diabetic kidney disease¹⁶ due to afferent arteriolar vasodilation and efferent vasoconstriction¹⁷, resulting in higher intraglomerular pressure, and thus glomerular hyperfiltration. This effect is thought to be a result of endothelial dysfunction due to nitric oxide (NO) deficiency¹⁸, which is a major player in regulating blood pressure. Indeed, in rats, long-term inhibition of NO synthase is associated with a mild degree of renal failure, as evidenced by a decrease in glomerular filtration rate (GFR), proteinuria, and glomerular sclerotic injury, and an aggravated course of nephritis through nonimmunological mechanisms¹⁹. In humans, endothelial dysfunction and hypertension are strongly linked with diabetic nephropathy²⁰, and it is known that genetic knockout of endothelial NOS in mice makes them prone to developing diabetic nephropathy²¹.

The purpose of this study was to investigate whether combining anti-Fx1A and L-NAME could accelerate the manifestation of both glomerular and tubular injuries, along with fibrosis. For the assessment of fibrosis, the hallmark of CKD, semiquantitative histologic assessment of biopsy or animal tissues is currently the most common method for fibrosis quantification in pre-clinical animal studies. While liquid renal fibrosis biomarkers, such as degraded collagen fragments are being tested^22,23, non-invasive imaging of extracellular matrix (ECM) has only rarely been used for kidney fibrosis monitoring^24–27, although it has shown promise in the liver²⁸. Therefore, the other aim of this study was to develop a translatable non-invasive imaging biomarker of renal fibrosis by using shear wave elastography (SWE)²⁹.

Materials and methods

Animals, reagents, and materials

Male Sprague Dawley rats (Crl: CD[SD]^®, Charles River Laboratories, Raleigh, NC, USA), 7–8 weeks old, weighing 225–350 g, were used in the study. The animals were group-housed in Tecniplast cages in a Pfizer facility where temperature and humidity were maintained between 20–26° C, and 30– 70%, respectively, with a 12-hour light/12-hour dark cycle. Rats were acclimatized for 3 days prior to onset of study activities. Rats were supplied ad libitum with municipal drinking water and a certified rat diet (#2916 C, Envigo Teklad Global Diet). The study protocol was reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at Pfizer. Rats were randomly divided into the following four groups: vehicle control (Group 1, n = 8): received water, anti-Fx1A (Group 2, n = 10): immunized (once at 2.5 ml/kg, iv), L-NAME (Group 3, n = 10): received L-NAME in drinking water at 20 mg/kg body weight/24 h, and the combination group (Group 4): received L-NAME at 20 mg/kg body weight/24 h and were immunized with anti-Fx1A (once at 2.0 ml/kg). The treatment duration for L-NAME was 7 weeks and in the combination group L-NAME treatment was started once the animal were immunized. Glomerulonephritis in Group 2 and 4 animals was induced following the PHN model³⁰ using anti-Fx1A (commercial sheep anti-rat Fx1A serum; PTX-002 S, Probetex, Inc, San Antonio, TX, USA). L-NAME was obtained from Sigma-Aldrich. Nine tenth percent (0.9%) sodium chloride was used as a vehicle for injections.

Shear wave elastography imaging

An ultrasound system (Aixplorer, SuperSonic Imagine, Aix-en-Provence, France) was used with an 8-MHz probe under the principles of supersonic shear imaging as described elsewhere^31,32. Briefly, a vibration force was generated by four successive focusing ultrasound beams at different depths with 5-mm spaces. Each focused beam consisted of a 150 µs burst at 8 MHz. Propagating shear waves were imaged at a frame rate up to 20,000 frames/s and raw radiofrequency data were recorded. Using a speckle tracking correlation technique, movies of displacements induced in tissues by the shear wave were calculated. Then the shear wave speed (c) was locally deduced using a time-of-flight algorithm. From the shear-wave speed in locally homogeneous soft tissues, the elastic modulus, the so-called Young’s modulus (E), is deduced from the following equation:

where µ is the shear modulus, ρ is the density of the area, c is the velocity of the shear wave.

Spatial resolution used was 0.3 mm² for the ultrasound image (B-mode) and 1.2 mm² for the elastic image. Three measurements were made in the left kidney for each animal at each session.

Enzyme-linked immunosorbent assay (ELISA)

Blood samples were collected in serum separator tubes, centrifuged at 10,000 × g for 10 min at 4 °C, and were stored at -20 °C. Rat serum Collagen Type 1 Alpha 1 (cat no. NEP2-75840 Novus Biologicals, CO), Actin Alpha 2, Smooth Muscle (cat no. NEP2-66430, Novus Biologicals), and Fibronectin (cat no. NBP2-67958, Novus Biologicals) were determined using commercially available kits, per the manufacturer’s instructions.

Western blotting

For protein extraction, frozen kidneys (30–50 mg) were homogenized in 1000 µL RIPA lysis with Pierce protease and phosphatase Inhibitor Cocktails (Thermo Fisher Scientific) on TissueLyser II and were shaken for 2 min. The tubes were removed and placed on ice for 10 min. Following centrifugation at 12,000 × g for 30 min at 4 °C, the supernatants were collected and used for protein quantification using Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific). The following antibodies were used: Cell Signaling (Cell Signaling Technology): α-smooth muscle actin (No. 19245, 4 µg/µL, 1:10 dilution), Vinculin (No. 13901, 1 µg/µL, 1:100–200 dilution); Abcam (Abcam): Collagen type 1 (No. ab34710, 4 µg/µL, 1:100 dilution) and Fibronectin (No. ab268020, 1 µg/µL, 1:200 dilution). To quantify target proteins, an automated Western blot was performed on a Simple Western™ Automated Western Blot System (Bio-Techne, Minneapolis, MN, USA) according to the manufacturer’s instructions.

Blood and urine biomarkers

Serum urea nitrogen (BUN) and creatinine (CREA) were measured using an ADVIA 1800 chemistry analyzer (Siemens Healthcare Diagnostics, Tarrytown, NY, Version # 2.03e00.03). Urine samples were collected at ~ 6 AM on each scheduled collection days (Days − 2, 14, 28 and 49) in the metabolic cages for up to 6 h. These urine samples were then centrifuged (1500 rpm for 5 min) and immediately stored at -80℃. Measured urinary biomarkers included CREA, microalbumin (MALB), N-Acetyl-β-D-Glucosaminidase (NAG), and total protein (PRO) as well as kidney injury molecule 1 (KIM-1), neutrophil gelatinase-associated lipocalin (NGAL), and osteopontin (OPN), which were analyzed using the rat kidney injury panel-1 purchased from Mesoscale Diagnostics (cat no. K15162C-1, Meso Scale Discovery, Gaithersburg, MD, USA) per manufacturer’s protocol. All the renal biomarker values were normalized to urine creatinine.

Renal histopathology and image analysis

All rats were euthanized by CO₂ inhalation followed by exsanguination on day 49 (except for 1 animal in Group 4 and 2 animals in Group 3) after vehicle, anti Fx1A and/or L-NAME administration. Sections of kidneys were collected, weighed, and fixed in 10% neutral buffered formalin (NBF). Five-micron formalin-fixed-paraffin-embedded sections were prepared for staining with hematoxylin and eosin (H&E), and seven-micron sections for Picrosirius Red (PSR) staining. Briefly, slides were deparaffinized, hydrated, and then mordanted in Bouin’s Fluid (American MasterTech, cat no. FXBOULT, Lodi, CA). Slides were then rinsed and stained on a Tissue-Tek Prisma^® Plus (Sakura Finetek USA Inc, Torrance, CA) with the following steps: 1% Phosphomolybdic Acid for 5 min (Electron Microscopy Sciences, cat no. 26693-08, Hatfield, PA); 0.1% Sirius Red in saturated Picric Acid for 90 min (Rowley Biochemical Inc., cat no. F-357-2, Danvers, MA); 30 s wash in 0.5% Acetic Acid twice. Slides were then automatically dehydrated and mounted with a permanent mounting medium. H&E and PSR slides were scanned on a Leica/Aperio AT2 whole slide digital scanner (Leica Biosystems, Buffalo Grove, IL), using the 20× magnification setting. The H&E slides were evaluated and reviewed by board-certified veterinary pathologists. Individual microscopic findings for each rat were graded using the following 0–5 grading scheme: 0 = within normal limits, 1 = minimal, 2 = mild, 3 = moderate, 4 = marked, and 5 = severe. PSR images were analyzed using custom artificial intelligence (AI)-enabled algorithms developed in Visiopharm software (V2022.09). The regions of interest (ROI) were detected using a custom Bayesian tissue detection App that was applied to each stained section to detect and segment the following features: cortex, whitespace, large vessels, and papilla. The cortex segment was used as the ROI for analysis. ROIs were manually QCed and any incorrectly classified areas were corrected. Custom classification algorithms were used to detect the following endpoints:

In additional to a conventional veterinary pathologist assessment, H&E images of kidney sections were analyzed using a custom AI App designed to detect 13 different types of abnormal microscopic features in the kidney. A composite pathology score was also calculated. The App outputted the total area of each abnormal feature in the image along with a heatmap. For each feature, we calculated the abnormal area% which is given by:

RNAseq profiling

RNA was extracted using the Qiagen RNeasy Lipid Tissue (Qiagen, #74804) protocol on a QIAcube according to the manufacturer’s instructions. At the end of extraction, RNA was eluted in 50 µL of RNase free water. Nucleic acid concentrations were quantified using the Qubit RNA BR Assay Kit (ThermoFisher #Q32854) using the Qubit 4 Fluorometer. RNA quality was determined by the RNA Integrity Number (RIN), measured on the Tape Station 4200 using RNA Screen Tape reagents (Agilent #5067–5576), following the manufacturer’s protocol. Library preparation was performed with the Quantseq 3’ mRNA-Seq Library Preparation Kit FWD for Illumina (Lexogen #015.96) from 500 ng of total RNA input according to the manufacturer’s recommendations. Briefly, first strand synthesis was performed by reverse transcription with oligo-dT priming followed by RNA removal. During second strand synthesis the library was converted to dsDNA and purified. The double-stranded cDNA library fragments were then amplified for 14 PCR cycles and labeled with unique dual indices before a second purification. Libraries were eluted in 17 µL of Elution Buffer. Final library concentrations were quantified using the Qubit dsDNA HS Assay Kit (ThermoFisher #Q32854) and library size was determined on the Agilent 4200 Tape Station instrument using the High Sensitivity D1000 ScreenTape (Agilent #5067–5584). Equimolar pools of the prepared libraries were sequenced on an Illumina Nextseq 500 Sequencing System (Illumina, San Diego, CA) using a 75-cycle single-end read flow cell at an average depth of 9.3 million reads per library.

Statistical analyses

Data are expressed as mean ± the standard error of the mean (SEM) or standard deviation (SD). The significance of the differences between mean values was evaluated by one-way or two-way ANOVA in GraphPad Prism software (version 9.0). A parametric or non-parametric test (Kruskal Wallis with Dunn’s multiple comparison for histologic image analysis) was performed to test for significant difference between the different groups at a 95% level of confidence. A p-value < 0.05 was considered significant. For the image analysis, .tsv files were exported from Visiopharm then formatted in Microsoft Excel for subsequent analyses. For RNAseq data analysis, reads were aligned to the Rattus norvegicus reference genome version 6.0 using Salmon version 1.4.0³³. For expression data normalization and processing, normalization factors were determined using the edgeR package (version 3.40.2)^34,35. For differential expression analysis, pairwise comparisons were run, defined by treatment. Benjamini-Hochberg adjusted p-value of 0.05 was used as the threshold for significance. PCA plots and heatmaps were generated using an internal application “MegaOmics”. Box and whisker plots were generated using ggplot2. Pathway analysis was performed using Ingenuity Pathway Analysis (Ingenuity Systems; Mountain View, CA, USA, v2021). Evaluation of overlap between the current RNAseq data and published transcriptomic datasets was performed using a Fisher’s exact test in GraphPad Prism software (version 10.4) following download of the specified published datasets from Gene Expression Omnibus (GEO) and pathway analysis in Ingenuity alongside the rat data.

Ethics

All procedures and animal experiments were performed in full compliance with the ARRIVE guidelines. Ethical permit was approved by the Institutional Animal Care and Use Committee (IACUC) at Pfizer in accordance with the Guide for the Care and Use of Laboratory Animals³⁶.

Results

Animal tolerance

All animals in the vehicle and anti-Fx1A, as well as 8 out of 10 animals in the L-NAME alone groups and 9 out of 10 animals in the combination group survived until the scheduled necropsy on day 49 without significant body weight loss (Fig. S1A) or obvious clinical signs. However, two rats (#23 and #25) in the L-NAME group exhibited approximately 10% weight loss on Days 45 and 43, respectively. To prevent potential overnight mortality and preserve tissue integrity for pathology, we proactively euthanized the animals and collect tissues on the same days. One rat (#33) in the combination group was reported to appear uncomfortable and had stopped drinking L-NAME water on Day 26 during routine health checks. The following day, its body weight had decreased by 13% compared to three days earlier. On Day 30, the animal was found to be quiet, alert, and responsive, with delayed skin turgor and a roughened haircoat. Its response to handling was abnormal. A proprioceptive deficit in both forelimbs and a head tilt were observed, while hind limb placement remained within normal limits. Porphyrin staining was noted at the nostrils and around the eyes. The animal had ceased drinking L-NAME water. Consequently, the rat was euthanized on Day 30, and tissues were collected. The daily average drinking water volume in all 4 groups were generally comparable, but slightly decreased over time (Fig. S1B), similar to that reported in the literature for Wistar (female) rats¹⁹. At the end of the experiment, anti-Fx1A at 2.5 ml/kg alone or at 2.0 ml/kg combined with L-NAME resulted in increases in the kidney weight/body weight ratio (KW/BW) compared with control rats, but this increase did not reach statistical significance (p = 0.09, Fig. S1C).

Glomerular injuries

After 7 weeks of treatment, minimal to mild glomerular changes were observed microscopically in both anti-Fx1A treated groups with a comparable incidence and severity (Table 1). Typically, these changes exhibited a diffuse distribution although in some animals they were multifocal. These glomerular alternations were characterized by enlarged glomeruli with glomerular tufts thickened by increased amounts of eosinophilic matrix material usually exhibiting diffuse involvement of the glomerular tuft but occasionally it was segmental (Fig. 1A & B). A correlating increase in urinary protein was observed 14 days following anti-Fx1A administration. Urinary protein levels peaked on day 28 and declined afterwards. On day 49 the protein remained elevated and statistical significantly (p < 0.05, Fig. 1C) compared with controls. Elevated protein creatinine ratio (PCR) and albumin creatine ratio (ACR) also peaked on day 14 and 28, respectively (Fig. 1D & E). L-NAME co-administration (2.0 ml/kg anti-Fx1A + 20 mg/kg L-NAME) enhanced the proteinuria, but the increase did not reach statistical significance. These results suggest L-NAME did not hinder, and may have even enhanced, the glomerular injuries caused by anti-Fx1A.

Table 1.

Incidence and severity of pathological findings.

Group/Findings	Vehicle	Anti-Fx1A	L-NAME	L-NAME + Anti-Fx1A
Group size	8	10	10	10
Glomerular injury	0	8	1	10
Grade 1	-	6	1*	9
Grade 2	-	2	-	1
Tubular degeneration/regeneration	0	5	5	9
Grade 1	-	4	3	3
Grade 2	-	-	2	4
Grade 3	-	1	-	2
Tubular hyaline casts	0	2	4	7
Grade 1	-	2	4	6
Grade 2	-	-	-	1
Tubular dilatation	0	1	1	5
Grade 1	-	-	1	5
Grade 2	-	1	-	-
Interstitial fibrosis (H&E)	0	1	2	2
Grade 1	-	-	2	2
Grade 2	-	1	-	-
Interstitial fibrosis (PSR)	0	1	3	7
Grade 1	-	-	3	6
Grade 2	-	-	-	1
Grade 3	-	1	-	-
Vascular Degeneration	0	0	3	3
Grade 1	-	-	3	3

*: only a few glomeruli were affected.

Fig. 1.

Morphological and functional changes in glomeruli after administration of anti-Fx1A, L-NAME or their combination. Representative image of normal (A) and an enlarged glomerulus with thickened glomerular tuft (B, anti-Fx1A administration at 2.0 mL/kg). Anti-Fx1A administration at 2.0 and 2.5 mL/kg caused increased total urinary protein (C), protein creatinine ratio (PCR, D) and albumin creatine ratio (ACR, E) starting from Day 14 (p < 0.05 on day 14, 28 and 49, two-way ANOVA). Anti-Fx1A administration at 2.0 mL/kg combined with L-NAME appeared to cause greater proteinuria than Anti-Fx1A alone, but the increase did not reach a statistically significant level. Values of Y-axis in C, D and E are means ± SEM. Legends are appliable for C, D and E.

Tubular injuries

Seven weeks after a single dose of anti-Fx1A and continuous dosing of L-NAME, the predominant histopathologic change was multifocal tubular degeneration/regeneration that was mostly minimal to mild in severity and localized to the cortex although occasional animals had moderate severity tubular degeneration/regeneration (Fig. 2A). These tubular changes were often accompanied by hyaline tubular casts (Fig. 2B) and tubular dilatation (Fig. 2C; Table 1). Affected tubular epithelium had basophilic cytoplasm, enlarged nuclei and occasional mitotic figures indicative of a regenerative response as well as modest numbers of degenerate epithelial cells. Basement membranes around affected tubules were sometimes thickened and accompanied by minimal mononuclear cell infiltrates around the periphery of the tubule (Fig. 2A - C). The animals in the combination group had larger areas of tubule abnormalities (trend) and more hyaline casts (p < 0.05) than that seen in other groups (Fig. 2D & E). Although the general renal functional biomarkers BUN and CREA were not altered during the dosing phase with or without the combination of compounds (data not shown), the specific urinary tubular injury biomarkers KIM-1, NGAL, and OPN elevated significantly in L-NAME and anti-Fx1A dosed animals by Day 49 compared to vehicle controls (Fig. 2F and Fig. S2). The elevation of renal biomarkers was observed at all timepoints measured in the experiment (Fig. 2F and S2). On day 49, proximal tubule specific biomarker KIM-1, proximal and distal tubule markers NGAL and more universal OPN were elevated in the combination L-NAME and anti-Fx1A group with statistical significance (p < 0.05) relative to the control group (Fig. 2F and S2). Individual treatment of each compound had only a minimal to mild impact on the expression of these renal markers. These findings highlight that a combination treatment may increase the severity of tubular injuries.

Fig. 2.

Representative images of and renal injury biomarker changes in tubules after administration of anti-Fx1A, L-NAME or their combination. Compared with vehicle-treated animals, L-NAME administration with or without anti-Fx1A for 7 weeks caused tubule degeneration and regeneration (A), increased casts (B) and tubular dilatation (C). AI-enabled image analysis revealed trend of increase in the total abnormal tubular areas (D) and small increased areas with hyaline casts (E, p < 0.05) in the combination group animals compared with vehicle control group. Likewise, the OPN biomarker significantly (p < 0.05) increased on day 49 in the combination group animals compared with the vehicle group (F).

Renal fibrosis

Microscopic examination revealed that affected tubules sometimes had associated fibrosis in the adjacent interstitium. This was predominantly minimal to mild in severity and, with the exception of one animal, this effect was limited to the L-NAME alone and combination groups (Table 1; Fig. 3A). Further analysis with AI revealed that only the combination group had significantly larger PSR-stained area compared with the vehicle control group (p < 0.01, Fig. 3B & C). Interestingly, animals in all treatment groups showed increased in vivo renal stiffness compared with the vehicle control animals on day 44 but not on day 23 (all p < 0.0001, Fig. 3D). Animals treated with both anti-Fx-1 A (at 2 ml/kg) and L-NAME had greater renal stiffness compared with those treated with anti-Fx1A alone (at 2.5 ml/kg, p = 0.052, Fig. 3D). The increase in the renal stiffness was also confirmed by ex vivo scan immediately after the euthanasia on day 49 (Fig. S3). In addition, the protein expression of alpha smooth muscle actin (αSMA), the hallmark of interstitial myofibroblasts, was examined using Western blot analyses on kidney tissue extracts. This revealed increased expression of αSMA in the kidney of animals in all the treatment groups, but only the mean increase in the combination group reached statistical significance compared with the vehicle control (p < 0.05, Fig. 3E). Furthermore, we determined the amount of collagen content in circulating blood by measuring serum Col1A1 concentration and this was also increased in the combination group animals compared with other treatment groups, although without statistical significance. These results suggest that the combination treatment may have a more significant impact on the development of renal fibrosis.

Fig. 3.

Representative morphological, SWE imaging and molecular changes in tubules after administration of anti-Fx1A, L-NAME or their combination. Compared with vehicle treated animals, L-NAME administration alone or with anti-Fx1A for 7 weeks caused fibrosis in the interstitium as revealed in H&E (A) and PSR (B) stained renal sections. (C) AI-enabled image analysis revealed increased total PSR-stained areas (p < 0.05) in the combination group compared with vehicle control group. In vivo renal stiffness was significantly (p < 0.05) higher in all three treatment groups on day 44 but not on day 23 (D). ELISA revolve that fibroblastic marker αSMA increased significantly (p < 0.05) on day 49 in the combination group animals compared with the vehicle group (E).

Vascular injuries

All groups administered L-NAME exhibited minimal multifocal degeneration of small to medium sized arteries within the renal parenchyma.

Transcriptional profiling

Kidney tissue was profiled for whole transcriptome analysis using RNAseq. A principal component analysis (PCA) showed a clear separation of the treatment groups from the vehicle group, but abundant overlap between anti-Fx1A and L-NAME, and between L-NAME and anti-Fx1A + L-NAME (Fig. S4). Compared with vehicle treatment, anti-Fx1A treatment resulted in 90 differentially expressed genes (DEGs), L-NAME resulted in 1275 DEGs, and anti-Fx1A + L-NAME resulted in 4007 DEGs (adjusted p < 0.05). Comparing the DEGs between treatment groups revealed substantial overlap, with 61 DEGs in common between all three treatment groups (Fig. 4). The top 9 DEGs are plotted in Fig. 5. Col6a3, Dock2, Fus, Hoxc11, Mta2 and Pcdhga5 were increased with treatment, while Eif4g2, Hspd1 and Tyro3 were decreased with treatment. In all cases, anti-Fx1A + L-NAME showed the most extreme difference, followed by L-NAME, consistent with the overall pattern of gene expression suggested by the PCA plot.

Fig. 4.

RNAseq profiling and pathway analysis. (A) Venn diagram showing the differentially expressed genes in common between the 3 treatment groups. (B) Heatmap of genes overlapping between all 3 treatment groups.

Fig. 5.

Expression plots for the top differentially expressed genes in common amongst all three treatment groups. The center line represents the median. The boxes represent the first to third quartiles, and whiskers extend to 1.5 times the interquartile range. Asterisks denote significant differential expression versus vehicle (padj < *0.05, **0.01, ***0.001).

To better understand the pattern of DEGs in common genes, the list of genes differentially expressed in two or more groups was analyzed using Ingenuity Pathway Analysis. Many of the top canonical pathways suggested possible oxidative stress (e.g. mitochondrial dysfunction, and detoxification of oxygen species, Table 2). Mitochondrial dysfunction was identified from Ingenuity Toxicity lists as well, which included multiple gene lists previously associated with renal damage. The top 6 genes associated with renal damage were significantly altered by L-NAME and anti-Fx1A + L-NAME: Bgn, Igfbp3 and Emp2 were increased, and Cycs, E2f6 and Man2a1 were decreased (Fig. S5). Ingenuity also suggested potential upstream regulators based on the DEGs, including predicted activation of up-regulation of P53. estrogen receptors, Il-4 and Myc (Table 2).

Table 2.

Top pathways enriched by differentially expressed genes in common between all 3 treatment groups.

	-log(p-value)	Ratio	z-score	Upstream Regulator	z-score
Ingenuity Canonical Pathways
Mitochondrial dysfunction	6.14	0.10	-2.48	TP53	2.09
Oxidative phosphorylation	4.76	0.13	3.87	KRAS	0.08
Nitric oxide signaling in the cardiovascular system	4.40	0.13	2.31	beta-estradiol	3.84
Detoxification of reactive oxygen species	4.28	0.22	2.83	ESRRA	5.10
Apelin muscle signaling	4.23	0.19	2.00	STAT6	0.91
Relaxin signaling	4.17	0.11	3.00	IL4	4.81
Estrogen receptor signaling	4.17	0.08	4.15	HNF4A	1.71
Apelin endothelial signaling	4.14	0.11	3.00	MYC	4.90
Cardiac β-adrenergic signaling	3.86	0.10	0.38	TGFB1	2.97
Huntington’s disease signaling	3.77	0.08	1.51	TEAD1	4.34
Oxytocin in spinal neurons signaling	3.59	0.20	2.65	APP	1.68
Ephrin-B signaling	3.50	0.14	2.12	HTT	0.86
Signaling by Rho family GTPases	3.36	0.08	3.87	SPP1	4.14
Thrombin signaling	3.11	0.08	3.16	methylprednisolone	1.75
Mitochondrial translation	3.06	0.12	3.32	CPT1B	-1.51
Thrombin signalling through proteinase activated receptors (PARs)	2.99	0.19	2.45	lipopolysaccharide	3.21
mTOR Signaling	2.95	0.08	1.27	EGF	3.94
Platelet homeostasis	2.88	0.12	3.16	pirinixic acid	3.42
Nonsense-mediated decay (NMD)	2.86	0.10	3.46
Protein ubiquitination	2.85	0.08	4.58
Ingenuity Toxicity Lists
Mitochondrial dysfunction	7.24	0.10
Renal necrosis/cell death	5.65	0.07
Genes downregulated in response to chronic renal failure	3.48	0.40
RAR activation	2.41	0.06
Cell cycle: G1/S checkpoint regulation	1.85	0.10
Renal Proximal Tubule Toxicity Biomarker Panel	1.77	0.15
Oxidative stress	1.73	0.11
Fatty acid metabolism	1.52	0.08
NRF2-mediated oxidative stress response	1.47	0.06
Increases renal damage	1.44	0.07
TR/RXR activation	1.31	0.07
Increases renal proliferation	1.30	0.06

-log(p-value) of enrichment represents strength of association, ratio = number of genes affected in the dataset versus total number of genes in the pathway, z-score ascribes directionality (positive = predicted activation, negative = predicted inhibition).

To evaluate these RNAseq data in the context of human CKD, the following publicly available kidney biopsy transcriptomic datasets were downloaded for comparison: RNAseq data from patients with glomerulonephritis³⁷, microarray data from patients with histopathologically-confirmed CKD³⁸, microarray data from patients with autosomal dominant polycystic kidney disease³⁹, and single-cell RNAseq data from patients with acute or chronic kidney disease⁴⁰. DEGs from rats with L-NAME + anti-Fx1A were compared with DEGs in patients with CKD relative to healthy control tissue. The overlap between rat and human DEGs was significant (Fisher’s exact p < 0.0001) for each dataset, with the rat model DEGs overlapping with 24% of the Park et al. 2022 dataset, 19% of the Nakagawa et al. 2015 dataset, and 22% of the Song et al. 2009 dataset. In the Lake et al. single cell RNAseq data, 91% of a gene list characterizing CKD was represented in the rat DEGs, and 15% of a gene list characterizing acute kidney disease showed overlap. At the pathway level, the rat dataset showed significant overlap with these datasets as well, with 65–70% overlap between the rat and human pathways in Ingenuity Pathway Analysis. The top 10 canonical pathways and upstream regulators in common between the L-NAME + anti-Fx1A rats and the 3 bulk transcriptomic human datasets is shown in Table 3. This list is dominated by inflammatory processes, but also includes fibrosis and extracellular matrix/cell surface interactions and RHO GTPase cycle.

Table 3.

Top pathways in common between rat combo treatment and published human transcriptomic profiles.

	L-NAME + Anti-Fx1A		Park 2022		Nakagawa 2015		Song 2009
	-log(p-value)	z-score	-log(p-value)	z-score	-log(p-value)	z-score	-log(p-value)	z-score
Canonical Pathways
Pathogen Induced Cytokine Storm Signaling Pathway	6.10	6.06	5.06	5.73	4.90	6.94	2.59	5.19
Phagosome Formation	4.83	4.31	4.00	5.58	7.38	8.32	5.02	3.15
Molecular Mechanisms of Cancer	8.48	3.95	3.81	4.20	12.70	7.47	7.61	5.08
Multiple Sclerosis Signaling Pathway	4.58	3.92	5.11	6.62	5.85	6.41	1.16	3.45
Extracellular Matrix Organization	2.92	4.60	7.10	5.90	4.01	3.32	17.76	5.88
Pulmonary Fibrosis Idiopathic Signaling Pathway	6.18	2.77	3.88	4.51	7.89	4.83	17.11	6.55
S100 Family Signaling Pathway	3.07	3.10	1.35	3.64	9.07	8.69	3.10	3.15
RHO GTPase Cycle	9.44	1.98	11.61	5.74	7.98	5.16	12.44	5.68
Integrin Cell Surface Interactions	3.27	4.69	6.51	4.87	2.66	3.66	13.12	4.73
Neuroinflammation Signaling Pathway	2.04	3.89	5.04	4.84	6.16	5.47	4.21	3.41
Upstream Regulators
Lipopolysaccharide	49.33	8.50	52.86	6.98	24.85	9.90	67.49	10.81
Interferon alpha	10.66	8.07	30.01	8.00	4.91	6.68	18.55	7.51
TNF	35.26	7.45	32.60	4.88	27.92	9.06	72.23	8.16
Poly rI: rC-RNA	17.81	7.06	38.09	5.53	4.79	8.11	20.65	7.27
IFNG	36.42	8.53	41.79	5.98	16.03	5.66	42.21	7.33
AGT	22.62	5.74	22.16	5.15	15.67	6.06	57.82	8.56
CSF2	11.17	5.82	23.00	4.61	14.69	8.54	11.82	3.85
Tetradecanoylphorbol Acetate	12.11	5.19	12.41	1.95	23.97	6.49	22.17	8.06
OSM	5.05	4.75	12.17	4.63	18.81	6.26	40.52	5.82
Alpha Catenin	10.01	-5.93	9.98	-4.67	2.53	-3.75	21.01	-7.00

-log(p-value) of enrichment represents strength of association, z-score ascribes directionality (positive = predicted activation, negative = predicted inhibition).

Discussion

This study demonstrates enhanced glomerular and tubule injuries as well as renal fibrosis following co-administration of L-NAME on PHN model induced noninvasively by anti-FX1A in adult rats. This model combines hypertension related renal injury with glomerulonephritis, two key elements that can lead to the development of chronic kidney injury in humans. In drug discovery and development, this model could be used to evaluate drug efficacy in a CKD model that has both glomerular injuries.

The rat anti-Fx1A-induced PHN model is a classical animal model of human idiopathic membranous glomerulonephritis. The pathogenesis of the PHN model is divided into a heterologous and autologous phases⁴¹. During the heterologous phase, the injected heteroantibodies recognize and bind to the antigens in glomerulus causing transient proteinuria. In the following days, the recipient rats develop an immune response to the heterologous antibodies, which incites a new phase of nephritis due to the autologous antibodies binding to heterologous antibodies in the glomerulus. Our first concern was whether vascular changes in the glomerular vasculature caused by L-NAME administration could impact distribution of anti-Fx1A in this structure, especially in the heterologous phase, since it was reported that acute administration of L-NAME reduced proteinuria and inhibited mesangiolysis to a large extent in Thy-1 model of glomerulonephritis⁴². In our experiment, L-NAME administration covered both phases given that L-NAME was administered in drinking water at the time of anti-Fx1A administration with continued dosing for 7 weeks. As expected, we observed significant proteinuria starting from day 14 and this lasted to day 49 in both the anti-Fx1A alone and anti-Fx1A and L-NAME combination groups, with the latter appearing to have more severe proteinuria than the former but no statistical significance, possibly due to the anti-Fx1A dose difference (2.0 mL/kg in the combination group vs. 2.5 mL/kg in the anti-Fx1A alone group). This observation provides evidence that L-NAME induced glomerular hemodynamic changes do not interfere with glomerulonephritis induction and maintenance. L-NAME alone did not cause proteinuria as has been reported¹⁹. In the L-NAME alone group only one animal had minimal glomerular pathology, but given the singular occurrence and minimal severity this may have just been a spontaneous finding. Alternatively, this may have been suggestive of an emerging effect on day 49, since others have reported focal glomerular sclerosis after two months of moderate hypertension following L-NAME administration in rats⁴³.

Direct blood pressure measurement was not performed in the current study due to previous research that has already reported significant increases in mean arterial pressure and decreased renal blood flow in rats and mice with chronic or acute administration of L-NAME^44,45 resulting in significantly increased mean arterial pressure and a 13– 30% decrease in renal blood flow. L-NAME is known to induce systemic hypertension in association with progressive renal tissue injury^14,43. These animals exhibit massive infiltration of macrophages in the interstitium of the renal cortex, which is thought to leads to interstitial fibrosis⁴⁶. While the current study observed minimal mononuclear cell infiltrates around the degenerative tubules, there was a significant increase in urinary OPN, a chemotactic and adhesive factor for macrophages, in the combination group animals. Upregulation of OPN expression in renal tubules has been reported to strongly associated with macrophage infiltration subsequent to tubulointerstitial injury both in experimental models⁴⁷ and in patients with kidney disease⁴⁸. In normal rat kidneys, OPN is primarily present in the glomeruli and descending thin limbs of the loop of Henle in the outer medulla⁴⁹. As such, the increase of OPN in the combination dosing group of the current study may be indicative of multiple sources or mechanisms of glomerular and tubule-interstitial injury in this model. The proximal and distal tubule biomarkers were altered significantly at the end of dosing period, indicating enhanced injury in L-NAME and anti-Fx-1 A combination dosing group animals. Again, our original concern or hypothesis on the potential interference of PNH on L-NAME effect is not supported by our results. This is aligned with an earlier report that hypotensive action is preserved in PHN rats⁵⁰.

Assessment of renal fibrosis would be of utmost importance not only for therapeutic and prognostic decisions in clinical practice but also for advancing preclinical translational research. Interstitial ECM expansion is the well-documented histological hallmark of chronic kidney injury and is usually the best structural correlate of the degree of renal functional loss and a strong predictor of progression risk⁵¹. At present, the clinical gold standard to assess renal fibrosis is kidney biopsy. However, there is a clear need for non-invasive biomarkers to monitor for the degree of renal fibrosis. In the current study, we applied the SWE imaging technique in vivo and ex vivo to measure renal stiffness. We also semi-quantitively analyzed the degree of fibrosis in kidney sections using the PSR stain technique and these findings were consistent with the increased levels of collagen type 1 chain (COL1A1) and fibronectin protein in the serum. AI-enabled image analysis of the PSR-stained kidney sections demonstrated a statistically significant increase in fibrosis in the combination group, but not in other groups, suggesting that combination treatment enhanced interstitial fibrosis in this cohort of animals, although the changes are small in magnitude. Although upregulation of fibronectin⁵², collagen I and III⁵³ in kidney tissue is considered an early event in renal fibrosis, we did not find significant increase in these proteins in the animals in any treatment group. Despite the absence of strong ECM accumulation in renal tissue, we did find significantly increased expression of the myofibroblast marker αSMA in the combination group animals. Altogether, our results demonstrate that fibrosis developed in these animals, especially in those treated with both anti-Fx1A and L-NAME. Similar to Derieppe and co-workers’ measurements⁵⁴, our SWE imaging revealed increased stiffness in the kidneys of animals treated with anti-Fx1A, L-NAME or both on day 44, but not on day 23. It is well accepted that during the renal fibrosis process proximal tubular epithelial cells⁵⁵ and glomerular podocytes⁵⁶ can undergo mesenchymal-to-epithelial transition, resulting in matrix-producing fibroblasts that contribute to the development of renal fibrosis. ECM is not only changed in a quantitative but also qualitative manner. For example, tissue-transglutaminase 2 can cross-link various proteins, including collagens and fibronectin⁵⁷. This crosslinking of ECM-proteins leads to increased stability of ECM and resistance to proteolytic degradation. In diabetes, advanced glycation end-products were shown to increase stiffness of collagens by crosslinking⁵⁸. The kidney has a complex structure, with three structural distinct compartments: glomeruli, tubulo-interstitium, and vasculature, each possessing different types of basal-membrane and non-basal membrane (interstitial) ECMs, all of which could be affected by fibrotic process as we observed here⁵⁹. Our results demonstrate that SWE imaging might have more sensitivity than conventional biomarkers or histopathology in detecting each of these processes although further investigation of the spatial characterization of SWE would be needed to confirm this.

Whole transcriptome profiling revealed significant overlap in the gene expression changes produced by anti Fx-1 A and L-NAME, indicative of invoking common kidney damage mechanisms in these models. Mitochondrial dysfunction and oxidative stress, which occur in many degenerative disorders and injury states, including CKD^60,61, rose to the top of a pathway analysis of common genes amongst the treatments. Several of the top up-regulated genes altered by all three treatments have been previously described in kidney disease. Col6a3 was upregulated in multiple patient cohorts with diabetic kidney disease⁶². Dock2 has been identified as a biomarker of renal fibrosis in two different patient cohorts⁶³. Hspd1 was downregulated by all three treatments in the present study. Inhibition of this gene in kidney epithelial cells has been shown to promote oxidative stress, and down-regulation has been observed in subjects with IgA nephropathy and renal fibrosis⁶⁴. Estrogen receptor signaling was identified as both a canonical pathway and potential upstream regulator of the DEGs for overlapping treatments in the present study, and this pathway has been implicated in a variety of kidney diseases and kidney repair⁶⁵. Furthermore, direct comparison of the rat CKD model data to published human CKD transcriptomic datasets show significant overlap, particularly in regard to inflammation. Comparison to a single cell RNAseq dataset³⁷ showed higher concordance to CKD than to acute kidney injury. Thus, these data provide further evidence of shared molecular underpinnings of kidney injuries and fibrosis produced by anti-Fx1A and L-NAME to processed observed in human CKD.

Our model boasts several strengths. First, it allows animals to develop renal injuries and interstitial fibrosis within 7 weeks without significant body weight loss or clinical signs. Second, it is useful for assessing the mechanisms of both glomerular and tubule injuries, particularly the hypertension-related CKD, which is more commonly studied using more complicated models such as Cyclosporine A or deoxycorticosterone acetate (DOCA)-salt nephropathy model⁶⁶ which require the use of special diet. In the neurogenic DOCA model, the administration of DOCA has to occur in combination with a high-salt diet and unilateral nephrectomy in order to induce moderate to severe hypertension, low renin levels, and renal injury⁶⁷. The DOCA-salt and L-NAME rat models are both used to study CKD and hypertension, but they induce these conditions through different mechanisms and exhibit distinct characteristics. DOCA-salt hypertension is characterized by increased blood pressure, proteinuria, and renal injury, with a significant role for mineralocorticoid receptors and the renin-angiotensin system. L-NAME-induced hypertension, on the other hand, is primarily mediated by nitric oxide synthase inhibition, leading to increased oxidative stress and different patterns of renal damage and cardiovascular remodeling. In our current model, the addition of administration of anti-Fx-1 A to L-NAME adds or enhances immunoglomerular injuries, characterized by subepithelial immune complex deposits, thickening of the glomerular basement membrane, effacement of podocyte foot processes⁴¹. ZSF1 is a spontaneous model of diabetic nephropathy with metabolic and hemodynamic factors driving renal damages⁶⁸. While this model replicates the slow and progressive nature of human type 2 diabetes, the breeding and maintenance are expensive. Transgene renin rats carrying the mouse Ren-2 renin gene (TGR(mRen2)27) are also a well-established model for studying hypertension and its associated pathologies, including kidney injuries, including arteriolar and arterial wall thickening, glomerular lesions, and sclerosis⁶⁹. This genetic rat kidney injury model shares similar limitation with ZSF1 rats, e.g. genetic manipulation-induced compensatory changes in other genes or pathways, resource-intensive and time-consuming, and interspecies difference in clinical translation. Third, our model can generate mild to moderate fibrosis suitable for imaging and pathologic evaluations in therapeutic and toxicity studies. In spite of these, we observed body weight decrease and subsequent clinical signs in 2 animals in the L-NAME group and 1 animal in the combination group on Day 43–45 and day 26, respectively. It appears that the animal welfare issue might be related to this particular animal strain, since L-NAME was dosed in drinking water at 50 mg/kg/day to 10-week-old male Wistar rats for 12 weeks without reported clinical manifestations⁷⁰., Additionally, the combination model warrants further testing in female animals and for reversibility or recovery of CKD.

In conclusion, given the concurrence and more severe incidence and/or severity of glomerular and tubular injury in the anti-Fx1A/L-NAME combination group, the current study presents a more valuable and relevant CKD model for mechanistic, pharmacologic and toxicologic research. This model may be considered as a useful rat CKD model for glomerular diseases, hypertensive nephropathy, and interstitial fibrosis⁹ in pharmacology studies to test prophylactic efficacy or early treatment invention of drug candidates. For mechanistic research, this model might be useful for the following mechanisms: Diffuse, focal or crescentic glomerulonephritis; focal and segmental glomerulosclerosis; membranous nephropathy, urinary-tract infections, stones, obstruction, renal dysplasia and drug-induced renal injuries.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: Chang-Ning Liu and Seo-Kyoung Hwang Data curation: Seo-Kyoung Hwang, Christopher Houle, Kodihalli C. Ravindra, Leah Newman, Jessie Qian, Sarah Vargas, Thomas A. Lanz, Timothy Coskran, Chang-Ning LiuFormal analysis: Seo-Kyoung Hwang, Christopher Houle, Kodihalli C. Ravindra, Leah Newman, Jessie Qian, Sarah Vargas, Thomas A. Lanz, Timothy Coskran, Chang-Ning LiuInvstigation: Seo-Kyoung Hwang, Jeffrey Morin, Kodihalli C. Ravindra, Leah Newman, Jessie Qian, Sarah Vargas, Jennifer Olson, Thomas A. Lanz, Timothy Coskran, Chang-Ning LiuMethodolgy: Seo-Kyoung Hwang, Jeffrey Morin, Kodihalli C. Ravindra, Leah Newman, Sarah Vargas, Jennifer Olson, Chang-Ning LiuProject administration: Seo-Kyoung Hwang, Jessie Qian, Thomas A. Lanz, Timothy Coskran, Chang-Ning LiuResources: Christopher Houle, Kodihalli C. Ravindra, Leah Newman, Jessie Qian, Thomas A. Lanz, Timothy Coskran, Chang-Ning LiuSupervision: Christopher Houle, Jessie Qian, Thomas A. Lanz, Timothy Coskran, Chang-Ning LiuValidation: Seo-Kyoung Hwang, Christopher Houle, Kodihalli C. Ravindra, Leah Newman, Jessie Qian, Sarah Vargas, Jennifer Olson, Thomas A. Lanz, Timothy Coskran, Chang-Ning LiuVisulization: Seo-Kyoung Hwang, Christopher Houle, Jessie Qian, Sarah Vargas, Jennifer Olson, Thomas A. Lanz, Timothy Coskran, Chang-Ning LiuWriting – Original draft Seo-Kyoung Hwang, Christopher Houle, Kodihalli C. Ravindra, Jessie Qian, Sarah Vargas, Jennifer Olson, Thomas A. Lanz, Timothy Coskran, Chang-Ning LiuWriting Review & editing Seo-Kyoung Hwang, Jeffrey Morin, Christopher Houle, Kodihalli C. Ravindra, Leah Newman, Jessie Qian, Sarah Vargas, Jennifer Olson, Thomas A. Lanz, Timothy Coskran, Chang-Ning Liu.

Data availability

The RNAseq datasets generated and analysed during the current study are available in the NCBI Gene Expression Omnibus repository, and accession number is GSE283627.The datasets used and/or analyzed during this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.