A Mouse Model for “Definitive” Polyomavirus Nephropathy with End-Organ Injury

Abstract

Mouse models for “definitive” polyomavirus nephropathy with lytic viral replication and end-organ disease (PyVN) do not exist. We aimed at defining a PyVN model in Black-Swiss mice (n = 126) by injecting newborn animals with murine polyomavirus (strain A2; 1 × 10⁵ plaque forming units) that led in all mice to a productive intrarenal infection with genetically stable, episomal MuPyV lacking latency. Animals were monitored and morphologic, immunohistochemical, molecular, genetic, and immunological analyses were conducted. Results: Within 3–6 weeks peak PyVN developed characterized by acute tubular injury, lytic replication in up to 14% of tubules (mainly collecting ducts/distal nephrons), high viral gene coverage (up to 3589 viral DNA reads/cell equivalent) and RNA expression (up to 9317 VP1-RNA reads/10⁷ RNA reads), inflammation, DNAemia, and-uria. MuPyV doubling times were high early post-infection (urine: 0.17 day–1.61 day) followed by steady slow viral clearance post day 28 (urine, half-life: 9.90 days). By 54 weeks PyVN had morphologically cleared (no chronic tissue injury) with only qPCR evidence of residual MuPyV in 17% of kidneys/mice. Infection induced an IgM/IgG response (peak plasma IgG titer at 7–30 weeks 1:20,480; low IgG titers in 92% of mice at end of follow-up after one year). During infection, episomal MuPyV remained genetically stable, without significant alterations that could have modified the course of PyVN. The mouse model resembles “definitive” PyVN in humans. It is suited for research on the pathogenesis of PyVN including virally induced tubular injury and immunologic interactions. It facilitates in vivo studies of therapeutic interventions aimed at blocking lytic intrarenalPyV replication.

1. Introduction

Since the mid-1990s, when tacrolimus and mycophenolate mofetil were introduced as immunosuppressive drugs, polyomavirus nephropathy has emerged as a serious infectious complication in kidney transplantation [1,2,3,4,5,6]. The “definitive” form of polyomavirus nephropathy, defined by morphologic evidence of lytic viral replication, tubular damage and associated kidney injury (PyVN), is primarily caused by BK polyomavirus [4,7]. It affects 3–5% of kidney transplant recipients. In the absence of specific antiviral therapy, allograft fibrosis and dysfunction remain common with two-year graft loss rates of 8–30% [6,8,9,10,11,12].

Most current knowledge of PyVN comes from human transplant studies, which established disease classifications and management strategies [5]. However, interpretation is complicated by confounding factors, such as immunosuppression, rejection, and concurrent kidney disease, such as arterionephrosclerosis obscuring the natural history and pathology of PyVN.

In animals, lytic PyV replication with end-organ injury has been reported only sporadically in non-human primates and horses [13,14]. Few systematic efforts have modeled PyVN with end-organ injury (reviewed in [15]). For example, viral replication in squirrel monkeys occurred predominantly in endothelial cells, contrasting with the tubular tropism characteristic of human disease [16]. Atencio and colleagues (1990s) described a two-hit rodent model combining latent PyV infections with acute kidney injury that resulted in lytic viral activation and tubular damage. This approach, however, was never further developed into a “disease model” [17,18]. More recently, mouse kidney transplant models introduced by Emory University showed only limited resemblance to human PyVN, as they lacked hallmark lytic PyV replication and were confounded by severe organ rejection [19,20,21,22]. Thus, no well-defined animal model has yet captured florid lytic PyV replication with kidney injury akin to human PyVN with end-organ disease. This lack constitutes an unsolved obstacle in PyVN research.

Here, we present a novel rodent model of PyVN in native kidneys featuring extensive lytic viral replication in tubular epithelial cells, inflammation, and renal injury. Long-term follow-up (up to 12 months) allowed detailed morphologic, molecular, and immunologic characterization. This model closely parallels human PyVN with end-organ disease and provides a valuable platform for studying disease mechanisms and evaluating potential therapies.

2. Materials and Methods

2.1. Animals and Infection with Murine Polyomavirus

Black Swiss Mice (NIH Genetic Resource: N:NIH Swiss outbred mice/albino crossed to C57BL/6N, black, non-agouti mice and backcrossed to N:NIH-S) free of infection were purchased from Charles River Laboratories, Wilmington, MA, USA. Mice for outbreeding were housed (1–2 adult animals per cage) under infection-free conditions with ad libitum access to food/water. For model purposes outbreeding was chosen to (i) limit tumorigenesis after PyV infection and (ii) increase genetic diversity, thereby reducing genetically influenced bias and enhancing broader biological relevance and external validity.

A hallmark of murine polyomavirus (MuPyV) infection in mice is its age-dependent outcome: neonates are highly susceptible to productive infection, whereas adults typically control the virus and remain disease-free [23,24].

On the day of birth (D0), study cohort mice (n = 126) were infected intraperitoneally with 50 µL of MuPyV strain A2, containing 1–2 × 10⁸ viral genome equivalents (qPCR) or 1 × 10⁵ PFU (kindly provided by Prof. M. Fluck, Michigan State University). Control neonates (n = 20) received vehicle alone. The overall protocol for inducing PyVN with end-organ lytic replication was adapted from Atencio et al. [17,25] Litters were weaned at approximately day 21 (D21) and monitored for up to 54 weeks (Wk) post-infection. At defined time points, animals were euthanized individually (CO₂ chamber; fill rate 30–40% per minute), and urine, blood/plasma, and tissues (particularly kidneys) were collected for molecular, genetic, immunologic, functional, and morphologic analyses. Renal function was evaluated using serum creatinine and blood urea nitrogen (BUN) levels and by urine dipstick testing for leukocytes, protein, and blood (Vet-10, Jorgensen Laboratories, Loveland, CO, USA, cat# J0630X). Blood was obtained via cardiac puncture/exsanguination, and plasma stored at −20 °C. Urine was fixed 1:1 in 4% paraformaldehyde and stored at 4 °C for long term storage and preservation of the viral capsid architecture as previously described [26,27]. After exsanguination, kidneys were excised, blotted dry, and bisected: one half was fixed in 10% neutral buffered formalin, and the other snap-frozen in liquid nitrogen and stored at −80 °C.

2.2. Sample Analysis

Tissue sections (formalin fixed and paraffin embedded; FFPE) were stained with H&E, Periodic-acid Schiff (PAS) and trichrome for light microscopy (LM). Transmission electron microscopy (EM) was conducted according to standard protocols.

Immunohistochemistry: Fixed tissue was incubated in AR-10 antigen retrieval buffer for 30 min (microwaving) to expose epitopes followed by incubation with primary antibodies (Ab) directed against PyV large-T (LT) and VP-1 (VP) antigens. LT: rat monoclonal anti-LT Ab (Abcam, Waltham, MA, USA,# ab18187; dilution 1/200) followed by a preabsorbed polyclonal secondary goat-anti-rat HRP Ab (Abcam#7097), visualization with 3,3′-diaminobenzidine (DAB) chromogen (Abcam, cat# 64238), and DAB enhancer (Abcam; cat# 675). VP: rabbit polyclonal anti SV40PyV-VP1 Ab (Abcam, cat# 53977) diluted 1:2000 followed by secondary anti-rabbit-HRP envision plus detection system (Dako, Santa Clara, CA, USA,; cat#-K4003), visualized with DAB plus substrate chromogen (Dako; cat#-K3468) and DAB enhancer (Dako; cat#-S1961). Scoring of MuPyV signals in renal compartments: IHC signals for VP were used to evaluate either the presence (renal pelvis) or extent (various renal compartments) of MuPyV replication.

Mapping MuPyV replication onto specific nephron segments (triple IHC staining procedure): Abs directed against Calbindin (1st staining step; distal tubular marker, brown chromogen), PyV large-T antigen/LT (2nd staining step; MuPyV replication, blue chromogen), and Aquaporin (3rd staining step; proximal tubular marker, red chromogen) were used in IHC triple staining protocols. Antigen retrieval: AR-Citra for 30 min in a steamer (Biogenex, Fremont, CA, USA; cat# HK087-20K). Calbindin 1st step: primary Ab rabbit recombinant monoclonal anti calbindin (Abcam cat# ab229915; dilution 1:16.000) followed by anti-rabbit envision detection system HRP (Dako; cat# K4003) and chromogen/DAB enhancer (Dako; cat# S196131). Signal: brown. LT 2nd step: primary Ab rat monoclonal anti-LT Ab (Abcam# ab18187; dilution 1:100) followed by biotinylated mouse-absorbed goat anti-rat IgG (Vector, Newark, CA, USA; cat# BA-9401; dilution 1:250), chromogen ABC-alkaline phosphatase kit (Vector/Vectastain; cat# AK5000), and BCIP/NBT substrate kit (Vector; cat#SK-5400). Signal: blue. Aquaporin 3rd step: primary Ab rabbit polyclonal anti-aquaporin-1 (Thermo Fisher Scientific/Invitrogen, Waltham, MA, USA; cat# PA5-78805: dilution 1:500) followed by biotinylated solid phase absorbed goat anti rabbit IgG (Vector; cat# BA-1000; dilution 1:500), chromogen ABC-alkaline phosphatase kit (Vector/Vectastain; cat# AK5000), and Vector Red substrate kit (Vector; cat# SK-5100). Signal: red. The slides were digitally scanned at 20× magnification using Aperio AT2 (Aperio Technologies, Vista, CA, USA) and uploaded to the Aperio eSlideManager database (Leica Biosystems Inc., Deer Park, IL, USA) at the Pathology Services Core at UNC; Images-cope (64_v12.4.6.7001, Leica Biosystems Inc.) was used for further analysis and quantitative assessment.

Scoring of mapping results: In each kidney specific cortical nephron segments (proximal—red; distal—brown; collecting duct—no staining) and corresponding LT expression were manually counted in 10 random 20 × 20 μm fields lacking glomeruli, dense inflammation, and arteries.

Typing of interstitial cell elements/IHC staining protocols: IHC was carried out in a Leica Bond III auto stainer. Primary Abs were directed against CD3 (dilution 1:1000), CD8 (dilution 1:200), F4/80 (a murine macrophage marker; dilution 1:2000), and alpha smooth muscle actin (SMA, a myofibroblast marker; dilution 1:250—all Abs from Cell Signaling, Leiden, The Netherlands; cat# 78588S, 98941, 70076, 19245) and directed against CD4 (dilution 1:2000; Abcam; cat# 183685). Heat-induced antigen retrieval was performed for 20 min at 100 °C in Bond-epitope retrieval solution 2, pH 9.0 (Leica; cat# AR9640) for the detection of CD4 antigen or in Bond-epitope retrieval solution 1 pH 6.0 (Leica; cat# AR9961) for the detection of the remaining cell-specific antigens. Following incubation with the primary Abs, the samples were subsequently incubated with the Novocastra post primary (Leica; cat# RE7260-CE) and the Novolink (Leica) polymer secondary Abs followed by detection with DAB using the Bond Intense R detection system (Leica; cat# DS9263). Results were reported as semi-quantitative observational data.

qPCR: Tissue (2–25 mg wet weight) was lysed in proteinase K enzyme solution; subsequently RNase-A (Qiagen, Hilden, Germany; cat# 19101) was added. DNA was extracted from 200 microliter tissue lysate or plasma, utilizing Qiagen DNeasy Blood and Tissue Kits (cat #: 69504 and 69506) and QiAamp MiniElute Virus spin Kits (Qiagen; Cat# 57704). Fixed urine samples were diluted in deionized water (1:50–1:1000 depending on the expected MuPyV load) for PCR analysis. Probes and primers for qPCR were designed to target sequences in the large-T region of MuPyV (forward primer: 5′AGG TGG AAG CCA TGC CTT AA 3′; reverse primer: 5′GGA AGC CGG TTC CTC CTA GA 3′; Taqman probe: 5′CAG GAA TTG AAC AGT CTC 3′). DNA concentrations and purity of each 1–2 microliter tissue/plasma DNA sample were determined by the ratio of absorbances at 260 nm and 280 nm in a Nanodrop spectrophotometer. Normalized DNA samples were then loaded in qPCR wells in triplicate alongside MuPyV standards to achieve quantitative linearity of the MuPyV assay from 10 to 10⁹ MuPyV gene copies. Test results were reported as MuPyV-gene equivalents/copy numbers per mL (blood/urine) or mg wet tissue weight (tissue/kidney).

DNA/RNA sequencing:

(a)

Library preparation and sequencing: Library preparation was carried out post extraction from FFPE tissue using the Watchmaker Genomics DNA Kit for whole genome sequencing (WGS) and Watchmaker Genomics RNA Kit for RNAseq (Watchmaker Genomics Inc., Boulder, CO, USA). The kits utilize a ligation-based approach to construct DNA libraries suitable for downstream sequencing applications. The following steps were performed for library preparation: isolated genomic DNA was sheared using the Covaris L system to obtain DNA fragments within the desired size range of 200–400 base pairs. Fragmented DNA was treated with the NEBNext FFPE DNA Repair v2 kit (New England Biolabs, Ipswich, MA, USA), which is an optimized cocktail of enzymes designed to repair DNA retrieved from FFPE tissue. RNAseq depletion of ribosomal gene expression was performed using the Polaris Depletion kit (Watchmaker Genomics) according to manufacturer recommendations, followed by cDNA generation. The DNA/cDNA fragments were subjected to end repair and A-tailing using the Watchmaker Genomics DNA or RNA Kit components. This process involved enzymatic treatment to ensure blunt-ended fragments and the addition of adenine nucleotides to the 3′ends. Illumina-compatible “Stuby” adapters (ITD) were ligated to the A-tailed DNA fragments using T4 DNA ligase followed by purification and removal of unligated adapters. PCR amplification was performed to enrich the DNA fragments containing the ligated adapters selectively. The number of PCR cycles was optimized to minimize over-amplification risk and ensure library complexity. Primers used for PCR reaction had barcoding and platform-specific sequences required for cluster generation and sequencing. The quality and quantity of the constructed libraries were assessed using Qubit (ThermoFisher, Waltham, MA, USA) and TapeStation (Agilent, Santa Clara, CA, USA). Libraries were analyzed to confirm the appropriate size distribution and absence of contamination or artifacts. The individual libraries were normalized to equimolar concentrations and pooled to create a sequencing library pool. The pooled library was then denatured and diluted according to the manufacturer’s guidelines to achieve the desired final loading concentration for sequencing. The prepared library pool was sequenced using NextSeq6000 S4 (Dispendix, Ashland, MA, USA) flowcell at the University of North Carolina High Throughput Sequencing Facility. Paired-end Sequencing was performed with read lengths of 150 to ensure comprehensive coverage of the genomic regions of interest.

(b)

Bioinformatics: The raw DNA sequencing data and RNAseq were processed using tools from CLC Genomic Workbench to perform quality filtering, read mapping to a reference genome, variant calling, and downstream analysis. Raw sequencing (SRA) data were recorded under bioproject accessioning number: PRJNA1393053.

(c)

Assessment of MuPyV location—episomal versus integration: The physical state ofMuPyV in PyVN was assessed using complementary evidence from WGS and RNAseq. In the WGS data set, reads were aligned to the mouse reference genome (MuPyV reference A2 strain PLY2CG, Gene Bank: J02288 [28]) and the MuPyV reference sequence. Focus was placed on evidence of canonical signatures of viral integration, including discordant paired-end reads with one mate mapping to MuPyV and the other to the mouse genome, as well as split (chimeric) reads spanning virus–host junctions.

2.3. Antibody Response to MuPyV Infection in Plasma

Murine polyomavirus multiplex fluorescent immunoassays (MFI) to analyze anti-MuPyV IgM and IgG responses were conducted by Idexx Laboratories (Greensboro, NC, USA). A MFI was created by covalently coupling purified mouse polyomavirus antigen to 1 × 10⁶ carboxylated polystyrene microspheres (Luminex Corp., Austin, TX, USA) according to manufacturer’s recommended protocols. Microspheres were stored at 4 °C in the dark until use. Testing of plasma samples for anti-MuPyV Ab was performed using the Luminex LX200 (Luminex Corp.). Microspheres coupled to MuPyV antigen were suspended by vortexing and sonication. Approximately 1500 microspheres in 100 μL PBS-BSA (PBS, 1% BSA, 0.05% sodium azide; Sigma-Aldrich, St. Louis, MO, USA) along with the test serum at a final dilution of 1:200 were added to each well of an AcroPrep 96-well filter-bottom plate (Pall Corporation, Port Washington, NY, USA). Plates were covered and incubated for 60 min on an orbital shaker at 400 rpm at room temperature in the dark. Each well was washed four times by adding 100 μL of PBS-BSA, shaking at 900 rpm for 1 min, and then removing the fluid with a vacuum manifold. The microspheres were suspended in 100 μL of PBS-BSA containing 1:500 diluted biotinylated F(ab′)₂ fragment goat anti-mouse immunoglobulin G (IgG) (heavy plus light chains [H + L]) secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) or 1:200 diluted biotinylated F(ab′)₂ fragment goat anti-mouse immunoglobulin M (IgM) (µ chain specific) secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA). The plate was covered and incubated for 60 min on an orbital shaker at 400 rpm at room temperature in the dark. Plates were washed four times as described above, and microspheres were suspended in 100 μL of fresh PBS-BSA solution containing 1:750 diluted streptavidin RPE (MOSS, Inc., Pasadena, MD, USA). The plate was covered and incubated for 60 min on an orbital shaker at 400 rpm at room temperature in the dark. Following incubation, plates were washed four times as described previously, resuspended in 100 μL of fresh PBS-BSA solution, and analyzed on the Luminex workstation. For each serum sample, the median fluorescent intensity of 100 MuPyV coated microspheres was recorded. Samples with IgG and IgM MFI values < 2500 were deemed negative (using a previously established MFI cutoff value), and no further testing occurred. These samples are listed in the study as an endpoint titer value of 0. Samples yielding MFI > 2500 but <16,000 were further serially diluted 1:2 and tested for antibody reactivity until the MFI signal dropped below the negative threshold. Samples with antibody reactivity of >16,000 were diluted at an initial dilution of 1:640 and then serially diluted 1:2 until the MFI signal dropped below the negative threshold. All positive samples are reported with an endpoint titer corresponding to the last dilution at which the samples remained above the negative cutoff value.

2.4. MuPyV Replication and Clearance Rates

Mathematical models and “rule of thumb” estimates on PyV turnover rates were used as proposed by Funk and colleagues [29,30,31]. MuPyV growth, i.e., the exponential rates of MuPyV increase, and clearance, i.e., the exponential rate of MuPyV decrease, in urine, tissue/kidney and plasma were studied.

• Viral growth/clearance rates: a straight line/regression line was fitted to log-transformed PCR viral load levels to obtain the slopes
• r = exponential rate/slope of viral increase/growth
• c = exponential rate/slope of viral decrease/clearance

  $$\mathrm{S}{\mathrm{E}}_{{\mathrm{t}}_2}={\displaystyle \frac{\mathrm{l}\mathrm{n}\left(2\right)}{r^2}}\times \mathrm{S}{\mathrm{E}}_r$$

  $$\mathrm{S}{\mathrm{E}}_{{\mathrm{t}}_{1/2}}={\displaystyle \frac{\mathrm{l}\mathrm{n}\left(2\right)}{c^2}}\times \mathrm{S}{\mathrm{E}}_c$$

3. Results

In all study animals (n = 126), the injection of MuPyV reliably induced a systemic infection with renal injury and time-dependent characteristic intratubular lytic viral replication. Infected animals exhibited growth retardation (Figure 1) but showed no behavioral abnormalities, changes in food or water intake, or premature death. Fewer than 10% of mice developed post-infection malignancies; these animals were excluded from the study cohort.

Urinalysis (n = 42 animals) performed between weeks 3–8 post-infection—when peak intrarenal lytic viral replication occurred—did not reveal diagnostic hematuria, proteinuria, or leukocyturia (dipstick: 0–trace). Blood chemistry (n = 34 animals) conducted between D4 and Wk10 showed normal serum creatinine levels (0.1–0.2 mg/dL) in all cases. BUN values were within the normal range (<29 mg/dL) in 27/34 animals, while 7/34 animals exhibited marginal elevations (30–50 mg/dL) [32]. At euthanasia, all kidneys appeared grossly unremarkable.

All control animals (n = 20), injected with vehicle only and sacrificed at 8, 15, and 21 days—time points at which all infected study animals showed marked morphologic and IHC evidence of MuPyV replication—were unremarkable by gross inspection, histology, and IHC evaluation.

3.1. Monitoring Disease Progression

Post infection, animals were monitored using morphologic, immunologic, molecular, and genetic approaches. Histologic assessment of viral replication in kidneys included IHC staining for VP and LT, both established diagnostic markers of “definitive” PyVN with end-organ disease (

Table 1. MUPyV load levels and plasma antibody responses.

	Urine PCR (Copies/mL)	Plasma PCR (Copies/mL)	Kidney PCR (Copies/mg of Tissue)	IHC MuPyV-VP Positive Tubules	Anti-MuPyV Ab Plasma IgG Titers	Anti-MuPyV Ab Plasma IgM Titers
Observational Time Bracket One (D1–D3—initial infection)
Median	0	3.14 × 107	2.48 × 107	0	0	0
Minimum	0	1.90 × 105	2.94 × 106	0	0	0
Maximum	1.67 × 106	3.31 × 1010	2.86 × 1010	0	0	0
Observational Time Bracket Two (D4–13—evolving-PyVN)
Median	7.80 × 105	2.50 × 1010	3.46 × 1010	0	0	0
Minimum	0	3.10 × 108	1.17 × 109	0	0	0
Maximum	1.28 × 1010	3.73 × 1011	1.65 × 1013	45	10	20
Observational Time Bracket Three (Wk 3–6—peak PyVN)
Median	3.12 × 1011	3.71 × 109	2.56 × 1010	100	240	320
Minimum	7.98 × 107	7.33 × 106	5.58 × 108	71	20	40
Maximum	2.17 × 1012	5.67 × 1010	1.22 × 1012	100	20,480	5120
Observational Time Bracket Four (Wk 7–30—diminishing PyVN)
Median	1.14 × 1010	6.03 × 107	9.92 × 108	26	20,480	0
Minimum	3.23 × 107	5.31 × 105	5.36 × 107	5	5120	0
Maximum	3.94 × 1012	2.94 × 109	4.09 × 1010	81	20,480	5120
Observational Time Bracket Five (Wk 47–54—post PyVN)
Median	0.00 × 100	0.00 × 100	0.00 × 100	0	40	0
Minimum	0.00 × 100	0.00 × 100	0.00 × 100	0	0	0
Maximum	9.03 × 106	1.36 × 105	8.26 × 107	0	10,240	0

Legend Table 1: See Supplementary Tables S1 and S2 (Supplementary pages 12–18) for detailed PCR, IHC, and plasma Ab data of individual study animals.

, Tables S1 and S2 and Figure S1) [7,8,9,33]. In mice, depending on the time post-infection, various cell types presented with granular nuclear and/or cytoplasmic staining for VP capsid protein. Occasional extracellular granular VP signal was also observed, particularly during observation bracket two (see below). LT expression, predominantly nuclear with rare cytoplasmic staining, was detected in fewer cells than VP (Figure 2 and Figure S2).

To capture the dynamics of intrarenal MuPyV infection relevant for the PyVN model, the observation period (D0–Wk54) was divided into five brackets. These were defined by the occurrence and extent of virus-induced kidney injury, i.e., acute tubular injury (ATI) and IHC detection of VP in tubular cells, that is by hallmarks of “definitive” PyVN.

• Bracket 1 (D1–D3): initial infection—no tubular IHC-VP signal
• Bracket 2 (D4–13): evolving PyVN—minimal tubular IHC-VP signal
• Bracket 3 (Wk3–6): peak PyVN—marked tubular IHC-VP signals
• Bracket 4 (Wk7–30): diminishing PyVN—decreasing tubular IHC-VP signals
• Bracket 5 (Wk47–54): post-PyVN—no tubular IHC-VP signal

The following sections describe model-relevant findings and PyVN-associated morphologic, molecular, and immunologic changes in brackets 1–5.

Bracket one (D1–D3—initial infection; LM/IHC n = 6 animals; qPCR/anti-MuPyV Ab response in plasma n = 9 animals, Table 1; additional data in Supplementary Information, Figure S1 and Tables S1 and S2): LM: Kidneys were unremarkable, showing partially immature parenchyma. IHC-VP: MuPyV replication/VP-staining was detected in very few, scattered interstitial cell elements and very rare glomerular tufts (2/6 mice, 33%); no signal was noted in the renal pelvis (sampled in 5 animals). qPCR: Viral gene copies were detected in all tested mice in the kidney (median: 2.48 × 10⁷/mg tissue) and plasma (median: 3.14 × 10⁷/mL). In the urine qPCR signals were detected in 2/9 (22%) mice (median in 2 animals: 1.67 × 10⁷/mL); no MuPyV gene sequences were found in urine samples of the remaining 7/9 (78%) animals. Plasma anti MuPyV IgM and IgG responses: not detected.

Bracket two (D4–13—evolving-PyVN; LM/IHC n = 10 animals; qPCR/anti-MuPyV Ab response in plasma n = 12 animals, Table 1; additional data in Supplementary Information, Figures S1 and S2 and Tables S1 and S2): LM: Transient interstitial changes (edema, spindle and inflammatory cells, scattered neutrophils, and some apoptotic debris) evolved at the cortico-medullary junction with finger-like extensions reaching deep into the cortex. In addition, the interstitium showed scattered cell elements with enlarged, sometimes vesicular nuclei often located adjacent to distal nephron segments or adjacent to medullary rays. Some distal nephrons appeared immature (Figure S2). IHC-VP: There was nuclear and/or cytoplasmic expression of VP noted in different cell types as well as focally in the extracellular space. Note: MuPyV located in the cytoplasmic compartment (compare to EM) showed a dotted/granular brown staining pattern by IHC and a dotted/granular blue staining pattern in corresponding H&E stained sections (Figure 2). Cellular distribution of VP-signals: spindle cell elements at the cortico-medullary junction (10/10 mice; nuclear signals in >50% of spindle cells in 6/10 mice) and other scattered interstitial cells. All animals (10/10) showed rare glomeruli (<10% glomeruli/animal) with scattered VP-expressing cells. Few animals (4/10, 40%) presented with distinct nuclear IHC-VP signals in cortical tubules (positivity in 3/4/5/45 cortical tubules/animal, respectively). Few animals (3/9, 33% of those with evaluable medulla) presented with nuclear epithelial cell staining in medullary ducts (positivity in <5 medullary ducts/animal). IHC expressions in immature distal nephron segments were difficult to distinguish from interstitial IHC signals; any equivocal IHC expression was not used for tubular counts. In transitional cells lining the renal pelvis (sampled in six mice) MuPyV replication was noted in 1/6 (17%) of mice. IHC-LT expression was less abundant (nearly exclusively limited to nuclei with only sporadic cytoplasmic signals) (Figure 2 and Figure S2). qPCR: Viral gene copies were detected in all tested animals (12/12, 100%) in the kidney parenchyma (median: 3.46 × 10¹⁰/mg tissue), plasma (median: 2.50 × 10¹⁰/mL), and urine (median: 7.8 × 10⁵/mL). Plasma anti-MuPyV IgM and IgG responses: very low IgG and IgM titers (1:10 and 1:20, respectively) were detected in 1/12 (8%) animals only.

Bracket three (Wk 3–6—peak PyVN; LM/IHC = 14 animals; qPCR/anti-MuPyV Ab response in plasma n = 14 animals; Table 1; additional data in Supplementary Information, Figure S1 and Tables S1 and S2):

LM: Kidneys showed patchy severe cortical ATI with epithelial cell necrosis/apoptosis, segmental denudation of tubular basement membranes, and accumulation of intra tubular debris (Figure 3). Injured tubular cell nuclei were enlarged with a “vesicular appearance”, expressed VP, and LT by IHC and resembled PyV-induced type 4 nuclear changes described in human PyVN.

Figure 3. Marked lytic intrarenal viral replication (D22, observational time bracket 3—peak PyVN). Productive intrarenalMuPyV infections cause multifocal acute cortical tubular injury (circles in (A), injured tubules labeled “T” in (B,D)). Severe virally induced host cell lysis and tubular basement membrane denudation (arrow heads) are illustrated in (C) (the circled area in (C) highlights focal tubulitis). In mouse PyVN the formation of relatively well-formed nodular mononuclear cell aggregates encircling arterioles (labeled “A” in (A,B)) and/or injured tubules (small circle in (A), labeled “T” in (B)) is a typical phenomenon. H&E stain (A,B), PAS stain (C), and trichrome stain (D), 10× original magnification (A), 20× original magnification (B,D), 40× original magnification (C) Ground glass type 1 viral inclusion bodies were scarce (Figure 4) [8,34]. Most injured cortical tubules were surrounded by edema and a mixed mononuclear and plasma cell infiltrate focally causing tubulitis. In addition, in 14/14 (100%) of the animals, scattered, somewhat nodular mononuclear/plasma cell aggregates had formed in the cortex constituting a typical presentation pattern of PyVN in the mouse. These so-called “tertiary lymphoid organs” [35] were occasionally associated with virally injured tubules, or they encircled arterioles (A,B).

Figure 3. Marked lytic intrarenal viral replication (D22, observational time bracket 3—peak PyVN). Productive intrarenalMuPyV infections cause multifocal acute cortical tubular injury (circles in (A), injured tubules labeled “T” in (B,D)). Severe virally induced host cell lysis and tubular basement membrane denudation (arrow heads) are illustrated in (C) (the circled area in (C) highlights focal tubulitis). In mouse PyVN the formation of relatively well-formed nodular mononuclear cell aggregates encircling arterioles (labeled “A” in (A,B)) and/or injured tubules (small circle in (A), labeled “T” in (B)) is a typical phenomenon. H&E stain (A,B), PAS stain (C), and trichrome stain (D), 10× original magnification (A), 20× original magnification (B,D), 40× original magnification (C) Ground glass type 1 viral inclusion bodies were scarce (Figure 4) [8,34]. Most injured cortical tubules were surrounded by edema and a mixed mononuclear and plasma cell infiltrate focally causing tubulitis. In addition, in 14/14 (100%) of the animals, scattered, somewhat nodular mononuclear/plasma cell aggregates had formed in the cortex constituting a typical presentation pattern of PyVN in the mouse. These so-called “tertiary lymphoid organs” [35] were occasionally associated with virally injured tubules, or they encircled arterioles (A,B).

Figure 4. MuPyV-induced acute tubular injury and nuclear inclusions/changes (D22, observational time bracket 3—peak PyVN). Productive MuPyV infections in renal tubules (labeled “T” in (A–C)) result in severe epithelial cell injury, host cell lysis, and denudation of tubular basement membranes (arrow heads in (A)). Tubular epithelial cell nuclei are enlarged with mostly vesicular/spider-web-like changes (termed type-4 PyV-induced nuclear changes in human PyVN [8,34]; broad arrows in (A–C)). Smudgy ground glass-type intra-nuclear viral inclusion bodies (termed type-1 PyV-induced nuclear changes in human PyVN [8,34]; open arrows in (A–C)) are uncommon in the mouse. In panel (C) the infected tubule/corresponding tubular basement membrane is highlighted by arrow heads. The long arrow in (C) points at an injured tubular cross section containing virions and debris. H&E stain (A) and trichrome stain (B,C), 40× original magnification.

Figure 4. MuPyV-induced acute tubular injury and nuclear inclusions/changes (D22, observational time bracket 3—peak PyVN). Productive MuPyV infections in renal tubules (labeled “T” in (A–C)) result in severe epithelial cell injury, host cell lysis, and denudation of tubular basement membranes (arrow heads in (A)). Tubular epithelial cell nuclei are enlarged with mostly vesicular/spider-web-like changes (termed type-4 PyV-induced nuclear changes in human PyVN [8,34]; broad arrows in (A–C)). Smudgy ground glass-type intra-nuclear viral inclusion bodies (termed type-1 PyV-induced nuclear changes in human PyVN [8,34]; open arrows in (A–C)) are uncommon in the mouse. In panel (C) the infected tubule/corresponding tubular basement membrane is highlighted by arrow heads. The long arrow in (C) points at an injured tubular cross section containing virions and debris. H&E stain (A) and trichrome stain (B,C), 40× original magnification.

At the cortico-medullary junction, edema and spindle cell elements that were prominent in time bracket 2 had largely vanished. Glomeruli, blood vessels, the renal medulla, and the pelvis were without significant changes. Fibrosis, tubular atrophy, and so-called chronic tissue injury were absent. IHC-VP: VP capsid protein expression was detected in 14/14 (100%) mice in many injured cortical tubular cells (positivity in >100 tubules/kidney in 13/14 (93%) animals). Occasionally, also intra luminal tubular VP staining was detected (Figure 5). VP capsid protein was only minimally expressed in medullary ducts (5/10 mice, 50% of those with evaluable medulla; positivity in <5 medullary ducts/kidney).

In glomeruli only scattered VP-expressing cells within capillary tufts were found in 10/14 (71%) animals; signals in parietal epithelial cells lining Bowman’s Capsule were exceptionally rare. No IHC staining was seen in arterioles, arteries, or veins. In transitional cells lining the renal pelvis (sampled in 11/14 mice), MuPyV was noted in 2/11 (18%) animals. IHC LT: In comparison to VP expression, less IHC-LT staining was found. EM: Virions of approximately 40–50 nanometers were primarily detected in tubular epithelial cells, occasionally in the interstitial compartment/extra cellular space, and in few interstitial cell elements. Virions, some aggregated, were also noted in the lumen of severely injured tubules (“PyV-Haufen”, Figure 6).

qPCR: Viral gene copies were detected in all animals (14/14, 100%) in the kidney (median: 2.56 × 10¹⁰/mg tissue), plasma (median: 3.71 × 10⁹/mL) and urine (median: 3.12 × 10¹¹/mL). Plasma anti-MuPyV IgM and IgG responses: IgG and IgM were detected in 14/14 (100%) animals; IgM: median titer 1:320, 5/14 (36%) animals with IgM titers > 1:1000; IgG: median titer: 1:240; 1/14 (7%) animals with an IgG titer > 1:1000.

Bracket four (Wk 7–30—diminishing PyVN; LM/IHC = 15 animals; qPCR/anti-MuPyV Ab response in plasma = 20 animals; Table 1; additional data in Supplementary Information, Figure S1 and Tables S1 and S2).

During follow-up the overall LM and IHC pattern of PyVN remained unchanged; however, the degree of MuPyV replication with ATI and interstitial inflammation decreased. Compared to time bracket three, VP capsid protein expression by IHC was detected in fewer cortical tubules (5–81 tubules/animal; 15/15 (100%) of mice), in rare glomerular tufts (2/15 (13%) mice), and only scattered interstitial cell elements. In the medulla rare tubules expressed VP (10/15 (67%) mice; positivity in <5 medullary ducts/animal). No animal presented with IHC-VP signals in the renal pelvis (14 mice evaluable). In all animals LM demonstrated well-formed nodular mononuclear cell aggregates (tertiary lymphoid organs) in cortical regions. Fibrosis, tubular atrophy, and so-called chronic tissue injury were absent. qPCR: Viral gene copies were detected in all animals (20/20, 100%) in the kidney (median: 9.92 × 10⁸/mg tissue), plasma (median: 6.03 × 10⁷/mL), and urine (median: 1.14 × 10¹⁰/mL). Compared to time bracket three, values had decreased. Plasma anti-MuPyV IgM and IgG responses: IgG was detected in 20/20 (100%) animals with a median titer much higher than in time bracket three (1:20,480). IgM was not detected in 15/20 (75%) mice; the remaining 5/20 animals had a median IgM titer of 1:2560.

Bracket five (Wk 47–54: post PyVN; LM/IHC = 14 animals; qPCR/anti-MuPyV Ab response in plasma = 12 animals; Table 1; additional data in Supplementary Information, Figure S1 and Tables S1 and S2):

One year post infection there was no LM or IHC evidence of MuPyV replication in the kidneys. The cortical parenchyma/medulla/renal pelvis looked largely unremarkable without any signs of fibrosis, tubular atrophy, or so-called chronic injury. Only few animals (2/14, 14%; animals#66 and #67 in Table S1) presented with scattered cortical nodular inflammatory cell aggregates, focal minimal interstitial inflammation including rare plasma cells, and mild ATI (Figure 7).

qPCR: Animals#66 and #67 (2/12, 17%) demonstrated MuPyV gene sequences in the kidney (median in 2 mice: 4.13 × 10⁷/mg tissue), urine (median: 4.32 × 10⁷/mL), and plasma (median: 1 × 10⁵/mL; Table S1). In 10/12 mice (83%) no qPCR signals were detected. Plasma anti MuPyV IgM and IgG responses: IgG was found in 11/12 (92%) animals in relatively low concentrations; median titer: 1:40 (IgG titer in animal#66:1:10,240; in animal#67: 1:20). No animal (0/12; 0%) presented with an IgM response.

3.2. Mapping MuPyV Replication onto Specific Nephron Segments (Figure 8)

In 13 animals (various stages of PyVN between Wk3–Wk22 post-infection) a total of 3556 tubules were counted in 130 cortical squares (10 squares/animal; 400 μm²/square). In 6.0% of all tubules (201/3556), IHC-LT expression was noted (range per animal: 13.6% of all tubules during peak PyVN—0.4% during dissolving PyVN). Evidence of MuPyV replication was mainly seen in cortical collecting ducts (54% of IHC-LT signals, 108/201), less in distal tubules (35%, 71/201), and least in proximal tubules (11%, 22/201).

Figure 8. Mapping MuPyV replication onto specific nephron segments. MuPyV replication in specific cortical nephron segments was assessed using triple IHC staining (Ab against Aquaporin, proximal nephron—red; Ab against Calbindin, distal nephron—brown; cortical collecting duct—no staining; MuPyV-LT expression (nuclear expression)—blue). In randomly selected cortical fields (each 20 × 20 μm; boxed area), specific nephron segments and LT expression were counted (the arrow indicates LT-expression in a distal nephron segment). Tubules outside the boxed were not considered in this analysis. Triple immunohistochemistry, 20× original magnification.

Figure 8. Mapping MuPyV replication onto specific nephron segments. MuPyV replication in specific cortical nephron segments was assessed using triple IHC staining (Ab against Aquaporin, proximal nephron—red; Ab against Calbindin, distal nephron—brown; cortical collecting duct—no staining; MuPyV-LT expression (nuclear expression)—blue). In randomly selected cortical fields (each 20 × 20 μm; boxed area), specific nephron segments and LT expression were counted (the arrow indicates LT-expression in a distal nephron segment). Tubules outside the boxed were not considered in this analysis. Triple immunohistochemistry, 20× original magnification.

3.3. Typing of Interstitial Cell Elements

Cell infiltrates in the cortex were analyzed in 12 infected animals (D8-Wk54) and 4 controls.

D8 (Figure S3): In edematous regions (at the cortico-medullary junction plus extensions into subcapsular zones; see above and Figure S2) approximately 80% of cells expressed SMA (mainly spindle cell elements), 25–40% CD3-CD8, and 10–20% CD4, respectively. Expression of the macrophage marker F4/80 was largely limited to hot spots of inflammation and apoptosis. In the superficial cortex, and especially in subcapsular zones, marker expression (mainly CD3–CD8 less CD4 and F4/80) was seen in scattered individual cells and small cell clusters. In the superficial cortex myofibroblasts were prominent, focally surrounding tubules in a chicken-wire fashion.

D15: With evolving prominence of lytic tubular viral replication and the gradual disappearance of edematous interstitial changes at the cortico-medullary junction, inflammatory cells started to cluster in the cortex. They were mainly but not exclusively located around injured tubules and in subcapsular zones (CD3–CD4 > CD8-F4/80). Myofibroblasts were focally detected in a chicken-wire fashion surrounding tubules.

D22 (Figure S4): During florid PyVN (observational time bracket three) mononuclear nodular cell aggregates formed in the cortex. These nodules were composed of approximately 60–70% CD3/CD4 and 25–40% CD8 positive cells, whereas macrophages (F4/80) and myofibroblasts were uncommon cell elements (<10% and <5%, respectively). Along cortical tubules and in subcapsular zones, all markers were focally expressed in the interstitium either in single cells or frequently in small cell clusters (“chicken-wire pattern”; SMA and F4/80 both > CD3/CD4 > CD8).

Wk16: With decreasing overall signs of PyVN, including diminished evidence of acute tubular injury in observational time bracket four, the extent of marker expression also decreased, while the overall distribution pattern (in nodules, in the interstitium, and in subcapsular regions) remained largely unchanged.

Wk50—Wk54 (Figure S5): At the tail end of PyVN in observational time bracket five, only few animals (14%, see description above) presented with scattered cortical nodular cell aggregates (marker profile: 30–40% CD3/CD4, 10% CD8, <5% F4/80; no SMA). In those animals presenting with nodular cell aggregates, focally accentuated minor inflammation was also noted in the cortex (with expression of all markers). In those animals (lacking cortical nodular cell aggregates (86%), marker expression was rare (mainly limited to scattered single cell elements) and similar to patterns seen in controls.

Controls: Control kidneys only showed very rare, scattered marker-expressing interstitial cells.

3.4. DNA/RNA Sequencing of MuPyV

WGS was conducted on eight mice euthanized between D6 and Wk30 post infection. Across all samples in the WGS data set, virus–host discordant pairs, split reads, or regional accumulation of breakpoint-supporting reads on the MuPyV genome were not detected that would indicate a recurrent integration junction. RNAseq demonstrated robust viral transcription, including expression of the VP1 capsid gene, which is consistent with transcription from an episomal viral template typical of productive MuPyV activity rather than expression driven by a host–genome integration event. Together, the absence of virus–host junction evidence in WGS and the viral expression profile in RNAseq supported an episomal state and ruled out detectable viral integration during PyVN.

Gene sequencing data were compared to the MuPyV reference A2 strain PLY2CG (Gene Bank: J02288) [28]. In 8/8 animals 31 single-/multiple-nucleotide variants were detected in the viral genome (Table S3). One variant at position 5251 bp in the replication origin beta showed a frequency of 20% and 37% on D6 and D7 and less than 20% thereafter. All other variants (30/31) were detected with relative consistent frequencies (mostly between 87% and 100%) over time. In addition, an 18 bp duplication was observed in the NCCR region of the replication origin. This duplication, ranging from position 5118 bp to 5135 bp (replication origin alpha element), was consistent across all time points. De novo alterations in the MuPyV genome were not observed. All detected genetic variants lacked distinct predicted functional consequences.

MuPyV DNA: MuPyV exhibited the highest coverage with 3589 viral gene copies/DNA reads per kidney cell equivalent 4 weeks post infection, when histologic evidence of PyVN peaked. The coverage gradually decreased in parallel to decreased morphologic/IHC evidence of viral replication and reached 55 and 20 MuPyV copies/DNA reads per kidney cell equivalent by Wk22 and Wk30, respectively (

Table 2. MUPyV-DNA and RNA Expression Levels.

	Viral DNA Reads/Copy Equivalents * per Mouse Cell Equivalent
	D17 **	Wk4 **	Wk14 **	Wk15 **	Wk22 **	Wk30 **
	731	3589	345	275	55	20
MUPyV Genes	RNA Reads aligned to MUPyV Genes in Kidneys (normalized to a total of 10 million RNA reads)
	D17 **	Wk4 **	Wk14 **	Wk15 **	Wk22 **	Wk30 **
small, middle, large T	185	894	11	59	1	2
VP1	4069	9317	355	888	4	25
VP2, VP3	520	1543	59	181	0	4
total RNA reads ***	13.972.524	10.936.602	13.327.792	6.248.636	9.691.776	14.655.336

Legend Table 2: * MUPyV strain PLY2CG is used as analytical reference strain, ** D: days post infection; Wk: weeks post infection, *** total RNA reads unaligned to mouse genome.

). The overall mouse reference genome’s coverage ranged between 10× and 20×.

MuPyV RNA: In the kidney RNA sequencing revealed highest early and late viral gene expression levels at Wk4 post-infection: ‘small/middle/large-T’ RNA reads 894; ‘VP1’ RNA reads 9317 (levels calculated per total of 10⁷ RNA reads). Subsequently, gene expression gradually decreased, with only minimal reads at the end of the observational time span at Wk30: ‘small/middle/large-T’ RNA reads 2; ‘VP1’ RNA reads 25 (Table 2).

3.5. MuPyV Replication and Clearance Rates

MuPyV replication and clearance rates in urine, tissue/kidney, and plasma were calculated in animals euthanized between D1–D357 post-infection (as proposed by Funk et al. [29,30,31],

Table 3. Calculated MUPyV growth and clearance rates.

	Time Period	Units	Value	SE
Urine			(mean)
viral growth rate—r	Day 4–Day 8	1 copy/day	3.97	0.419
viral doubling time—(t2)	Day 4–Day 8	days	0.17	0.018
viral growth rate—r	Day 15–Day 20	1 copy/day	0.43	0.173
viral doubling time—(t2)	Day 15–Day 20	days	1.61	0.649
viral clearance rate—c	Day 28–Day 357	1 copy/day	0.07	0.003
viral half-life—(t1/2)	Day 28–Day 357	days	9.90	0.424
Plasma
viral growth rate—r	Day 4–Day 8	1 copy/day	0.59	0.256
viral doubling time—(t2)	Day 4–Day 8	days	1.17	0.510
viral growth rate—r	Day 15–Day 20	1 copy/day	0.32	0.159
viral doubling time—(t2)	Day 15–Day 20	days	2.17	1.076
viral clearance rate—c	Day 28–Day 357	1 copy/day	0.05	0.003
viral half-life—(t1/2)	Day 28–Day 357	days	13.86	0.832
Kidney
viral growth rate—r	Day 4–Day 6	1 copy/day	3.17	0.521
viral doubling time—(t2)	Day 4–Day 6	days	0.22	0.036
viral growth rate—r	Day 15–Day 20	1 copy/day	0.02	0.145
viral doubling time—(t2)	Day 15–Day 20	days	34.66	251.266
viral clearance rate—c	Day 28–Day 357	1 copy/day	0.06	0.004
viral half-life—(t1/2)	Day 28–Day 357	days	11.55	0.770

Legend Table 3: Calculations based on cross-sectional data collection.

and Figure S6). Viral growth rates, i.e., the exponential rates of MuPyV increase, and doubling times were calculated from PCR data during (a) initial rapid (urine/plasma: D4–D8; tissue: D4–D6) and (b) subsequent slow (urine/plasma/tissue: D15–D20) increases in viral load levels. MuPyV doubling times t₂ in days (mean ± SE) in (a) initial window rapid growth: urine 0.17 (±0.018); plasma 1.17 (±0.510); tissue/kidney 0.22 (±0.036); and (b) subsequent window slow growth: urine 1.61 (±0.649); plasma 2.17 (±1.076); tissue/kidney 34.66 (±251.266). In the kidney no significant increase in exponential growth was noted post D15. Viral clearance rates, i.e., the exponential rate of MuPyV decrease, and half-lives were calculated from data collected between D28 and D357 during a time of continuous viral clearance (up to undetectable levels). MuPyV half-life t_1/2 in days (mean ± SE): urine 9.90 (±0.424); plasma 13.86 (±0.832); tissue/kidney 11.55 (±0.770). Since longitudinal data collection from individual animals could not be conducted, current calculations based on cross-sectional analyses carry the risk of greater variations.

4. Discussion

Current knowledge of PyVN primarily stems from immunosuppressed kidney transplant recipients. However, studying viral nephropathy in this context is challenging due to confounding factors such as immunosuppression and alloimmune injury/rejection—all of which alter PyVN disease presentation, morphology, and progression. To date, mouse models reproducing characteristic lytic intrarenal PyV replication and definitive end-organ disease have not been established, representing a major obstacle to understanding PyV-induced kidney injury and developing targeted therapies [20,22].

Our goal was to establish a robust mouse model of “definitive” PyVN with intrarenal tissue injury, suitable for studies on pathogenesis, immune responses, and therapeutic interventions. Building on previous reports of MuPyV kidney infections, we defined optimal strategies to induce end-organ injury and characterized the natural course of viral nephropathy in the native mouse kidney.

All Black Swiss mice infected with MuPyV strain A2 developed kidney injury/”definitive” PyVN within three weeks. Florid PyVN showed multifocal marked lytic viral replication in up to 14% of cortical tubules, mainly in collecting ducts and distal nephrons, with nuclear inclusions, severe tubular damage, T cell–rich inflammation, viral DNAemia/-viruria, and a humoral antibody response closely resembling human Banff class 2 PyVN (

Table 4. Mouse and human PyVN: similarities and differences.

	Polyomavirus Nephropathy
	Human	Mouse Model
Episomal PyV location	+	+
Latent PyV	+	−
Activation of latent/dormant intrarenal PyV	+	−
Lytic PyV replication	+	+
Prominent replication in cortical collecting ducts/distal nephrons	+	+
Replication medulla/renal pelvis	+	(+)
Replication in non-epithelial cells	−	+
Viral inclusions/nuclear changes	+	+
Viral protein expression (IHC)	+	+
IHC Large-T nuclei	+	+
IHC Large-T cytoplasm	−	(+)
IHC VP nuclei	+	+
IHC VP cytoplasm	(+)	+
Severe tubular injury/TBM denudation	+	+
Interstitial inflammation	+	+
Chronic tissue injury (fibrosis/atrophy)	+	−
Urinary PyV-Haufen as biomarkers for disease *	+	+
Criteria of Banff PyVN disease class 2 **	+	+
Renal dysfunction	+	−
Coinciding allograft injury (rejection/pre-existing donor disease/hypertension with arterionephrosclerosis etc.)	+	NA
Associated systemic infection	−	+
Humoral IgM/IgG Immune Response in Plasma	+	+

Legend Table 4: * ref. [33], ** ref. [9,11,37], (+): minimal/minor occurrence, NA: not applicable; PyVN in native kidneys lacking comorbidities; + present; − absent

) [9,11,36,37]. PyVN peaked between weeks 3 and 7, followed by slow viral clearance over five months. At one year, 86% of mice fully recovered, while 14% retained minimal inflammation and low-level molecular viral persistence, mimicking late-stage human disease [37]. Throughout infection episomal MuPyV remained genetically stable.

Although during PyVN SMA-expressing myofibroblasts were recruited to sites of virally induced acute kidney injury, chronic changes such as fibrosis and tubular atrophy did not develop in the mouse model. In contrast, human PyVN often progresses to parenchymal scarring and Banff class 3 disease [9,11,38]. Whether this progression in human transplants reflects alloimmune injury or smoldering rejection—absent in the native mouse model—remains to be determined [20].

In both mice and humans, antibody responses during active disease fail to rapidly clear lytic PyV replication or resolve tubular injury. However, in mice, pre-existing neutralizing antibodies can block viral replication and prevent PyVN (manuscript in preparation) [39]. These findings, supported by early human data, suggest that preventive approaches such as vaccination could protect at-risk patients from viral nephropathy [40,41].

This novel mouse model uniquely enables in vivo studies of end-organ disease caused by lytic PyV replication and tissue injury. To date, no comparable rodent model has been available. The florid stage of PyVN—marked by peak viral replication and maximal tissue damage over a four-week interval in observational time bracket three—provides a window to examine mechanisms of virally induced tubular injury, viral replication, inflammation, and immune responses. It also serves as a platform to test antiviral therapies, with efficacy assessed against the natural course of morphologic recovery spanning several months (time bracket four). Time bracket four, diminishing PyVN, is particularly well suited for studies on mechanisms of viral clearance, modulation of the viral life cycle, and immunologic dynamics during late-stage PyVN. Because PyVN develops in native kidneys without confounding factors such as alloimmune injury, immunosuppression, or other comorbidities, the model of ‘uncomplicated’ PyVN’ allows focused investigations of pure viral pathogenesis and host defense. In this context targeted immunosuppression, such as the administration of drugs affecting B- or T-cell function, can offer unique opportunities to worsen the course of viral nephropathy under controlled experimental conditions [42,43] and to study alterations in anti-viral immune responses, viral clearance, and chronic tissue injury as seen in human transplant recipients (studies under way). Thus, the mouse model of PyVN offers unique insights into the intrinsic mechanisms of viral pathogenesis and the corresponding host responses.

Several model-specific features must be considered when using the PyVN mouse model. First, unlike human PyVN, which results from reactivation of latent infections, disease in the mouse kidney follows a primary infection—though this difference is functionally and for model purposes irrelevant once lytic replication drives tissue injury. Second, disease severity is strain dependent, robust in Black Swiss and mild in 129/sv (personal observation). In general, the mouse strain is a well-established determinant of disease severity, kinetics, and pathology in virtually all experimental disease models including oncogenesis and infections—the strain is a biological variable that must be carefully considered when designing studies on PyVN. Third, MuPyV shows some intra cytoplasmic replication, suggesting species-specific viral dynamics. Fourth, disease susceptibility is age dependent—neonates develop injury, whereas adults mount protective immune responses [23,24,44] Fifth, because infection clears without prominent latency, the model is unsuitable for studying latent PyV infection.

Here we describe a mouse model of “definitive” PyVN in native kidneys, featuring genetically stable episomal MUPyV infections leading to lytic viral replication with extensive tubular epithelial lysis, inflammation, and urinary PyV-Haufen shedding (Table 4 and Figure 9) [11,27,33,45]. Mice were followed for up to 12 months, enabling comprehensive morphologic, molecular, genetic, and immunologic analyses. This model closely parallels florid human PyVN and provides a platform for studying PyV-induced end-organ disease and pathogenesis and testing antiviral therapies.