Characterisation of an enhanced preclinical model of experimental MPO-ANCA autoimmune vasculitis

Experimental autoimmune vasculitis (EAV) is a model of antineutrophil cytoplasm antibody (ANCA)-associated vasculitis (AAV) induced by immunisation of susceptible rat strains with myeloperoxidase (MPO). Animals develop circulating MPO-ANCA, pulmonary haemorrhage and glomerulonephritis, although renal injury is mild and recovers spontaneously without treatment. In this study we aimed to augment the severity of glomerulonephritis. Following induction of EAV on day 0, a sub-nephritogenic dose of nephrotoxic serum (NTS) containing heterologous antibodies to glomerular basement membrane was administered on day 14. This resulted in a significant increase in disease severity at day 28 compared to MPO immunisation alone – with more urinary abnormalities, infiltrating glomerular leucocytes, and crescent formation that progressed to glomerular and tubulointerstitial scarring by day 56, recapitulating important features of human disease. Importantly, the glomerulonephritis remained pauci- immune, and was strictly dependent on the presence of autoimmunity to MPO, as there was no evidence of renal disease following administration of sub-nephritogenic NTS alone or after immunisation with a control protein in place of MPO. Detailed phenotyping of glomerular leucocytes identified an early infiltrate of non-classical monocytes following NTS administration that, in the presence of autoimmunity to MPO, may initiate the subsequent influx of classical monocytes which augment glomerular injury. We also showed that this model can be used to test novel therapeutics by using a small molecule kinase inhibitor (fostamatinib) that rapidly attenuated both glomerular and pulmonary injury over a four-day treatment period. We believe that this enhanced model of MPO-AAV will prove useful for the study of glomerular leucocyte behaviour and novel therapeutics in AAV in the future.

Keywords: MPO, ANCA, vasculitis, monocytes, glomerulonephritis, experimental vasculitis

Introduction

Anti-neutrophil cytoplasm antibody (ANCA) associated vasculitis (AAV) is a rare systemic autoimmune disease, which can cause life-threatening lung haemorrhage and end stage kidney disease (ESKD) [1]. The typical renal lesion in AAV is pauci-immune crescentic glomerulonephritis. Circulating ANCA are directed to myeloperoxidase (MPO) or proteinase- 3 (PR3), which are present in the granules of neutrophils and lysosomes of monocytes [2,3], and a number of experimental and clinical observations indicate they have a directly pathogenic role in disease pathogenesis [4,5].

Studies in animal models have been critical for understanding disease mechanisms, and several rodent models of anti-MPO vasculitis have been developed [6]. These include passive transfer of anti-MPO antibodies, raised in MPO-deficient mice, to naïve wild-type mice, causing glomerulonephritis (GN) and pulmonary capillaritis [5–7]. Models of active autoimmunity in mice have also been developed; mice immunised with mouse MPO develop anti-MPO antibodies at low titre, but these are not sufficient to cause GN, and a ‘second-hit’ is required to induce disease. For example, a subsequent injection of heterologous anti-mouse glomerular basement membrane (GBM) globulin results in transient neutrophil recruitment to the glomerulus. This ‘planting’ of the MPO autoantigen (derived from retained neutrophils) results in recruitment of MPO-specific CD4+ T cells, neutrophils and macrophages to the glomerulus, and the development of crescentic GN [8,9]. The disease triggered by anti-GBM globulin is dependent on MPO; disease does not occur in MPO-deficient mice. A limitation of this model is that the response to anti-GBM globulin is itself nephritogenic; by 4 days mice develop an autologous immune response to deposited anti-GBM antibody, leading to severe GN, even in the absence of pre-immunisation with MPO. As such, study of anti-MPO mediated disease is limited to early time points, meaning that therapeutics can only be tested in preventative studies, not after the development of vasculitis [9,10].

A model of experimental autoimmune vasculitis (EAV) in the susceptible Wistar–Kyoto (WKY) rat strain was previously described by Little et al in our laboratory. Rats immunised with human MPO develop polyclonal MPO-ANCA cross-reactive to rat MPO expressed in neutrophils and monocytes, and subsequently small vessel vasculitis, pauci-immune GN and haemorrhagic pulmonary capillaritis [11–14]. A limitation of the model is that renal disease is relatively mild; glomerular lesions are mainly proliferative, crescent formation is rare, and disease spontaneously resolves from six weeks post-induction. It has been shown in other studies in rats that addition of a low-dose of NTS, containing heterologous anti-rat GBM antibodies, after immunisation with MPO, results in increased disease severity [15,16], similar to the approach described in mice. However, this approach is not well characterised in rats, and in these reports the administration of NTS alone, in the absence of autoimmunity to MPO, induced GN, and immunoglobulin deposits could be detected in glomeruli.

In this study, we aim to augment disease severity of EAV by the addition of a truly sub- nephritogenic dose of NTS that is insufficient to induce disease in the absence of autoimmunity to MPO. To investigate possible mechanisms by which sub-nephritogenic NTS increases disease severity, we isolate and phenotype glomerular leucocytes using flow cytometry. Finally, we show that an enhanced model can be used to test therapeutic approaches in a preclinical study.

Materials and methods

Animal husbandry

WKY and Lewis rats were purchased from Charles River (Saffron Walden, UK) and maintained in a pathogen-free animal facility at the Central Biomedical Services unit, Hammersmith Hospital Campus, Imperial College London. All procedures were carried out in accordance with the regulations of the UK Animals (Scientific Procedures) Act (1986) and ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines.

Experimental autoimmune vasculitis/NTS administration

NTS was prepared in rabbits as previously described [17]. For the NTS titration experiment, male WKY rats aged 8–9 weeks (n=4/group) were immunised intravenously (IV) with 100 μl of NTS (or serial dilutions in sterile PBS) and maintained until 10 days post-immunisation.

EAV was induced by immunising 7–8 week old male WKY or Lewis rats (n=4–10/group) intramuscularly with 800 µg/kg human MPO (Calbiochem, Merck Millipore, Darmstadt, Germany), or human serum albumin (HSA; Sigma, Poole, UK) as a control human protein, emulsified in complete Freund’s adjuvant supplemented with Mycobacterium butyricum (Sigma). Intraperitoneal pertussis toxin (Sigma) was administered on days 0 and 2 [13].

For ‘double’ immunisation studies, on day 14 after initial immunisation with MPO, 100 μl of 1:100 dilution of NTS or normal rabbit serum (NRS) was administered IV. Urine was collected weekly by housing animals in individual metabolism cages overnight, and blood collected every 14 days by lateral tail vein bleed. Animals were sacrificed after 28, 42 or 56 days.

For additional experiments to assess glomerular cell infiltrate at early time points, neat or 1:100 NTS was given IV to 8–9 week old male WKY rats (n=4/group) and animals sacrificed after 3 h, 24 h or 7 days.

Assessment of renal disease

Proteinuria was measured by the sulphosalicylic acid method and haematuria by urine dipstick (Multistix 8; Siemens, Munich, Germany) [18]. At the end of each experiment animals were exsanguinated under terminal anaesthesia. Renal tissue was fixed in 10% neutral buffered formalin, transferred to 70% ethanol, and processed to paraffin blocks. Sections were stained with periodic acid Schiff (PAS), haematoxylin and eosin, and Jones’ silver stain. Fifty consecutive glomeruli were assessed for crescents ± necrosis, segmental necrosis, or minor changes such as segmental endocapillary proliferation, by a blinded observer, and results expressed as mean proportion of glomeruli for each animal. Immunoperoxidase staining was carried out using CD68 (ED-1, Bio-Rad, Watford, UK; dilution 1:500), CD8 (OX-8, Bio-Rad; dilution 1:50), CD3 (1F4, Bio-Rad; dilution 1:400), and for smooth muscle actin (M0851, Dako, Ely, UK; dilution 1:100). Number of positive cells was quantified using ImageProPlus software to measure percentage area staining in 20 consecutive glomeruli, or 20 consecutive high-powered field (HPF) for interstitial staining, and results expressed as mean percentage for each animal. Smooth muscle actin immunostaining was used to quantify the area proportion of fibrous/fibrocellular crescents using assessment of 50 consecutive glomeruli.

Glomeruli were isolated by serial sieving of whole kidney tissue as described previously [19], then digested with 1 mg/ml Type IV collagenase (Sigma), 0.5 mg/ml trypsin (Sigma) and 0.1 mg/ml Type I DNase (Roche, Welwyn, UK) for 20 min at 37 °C with gentle agitation. Cells were washed and used for cell surface staining with antibodies against CD172a (OX-41, FITC, Biorad, 1:5), CD45 (OX-1, V450, BD Biosciences, Oxford, UK, 1:40), CD3 (eBioG4.18, PE, eBiosciences, Hatfield, UK, 1:40), B220 (HIS24, PE, eBiosciences, 1:40), CD161a (3.2.3, PE, BioLegend, San Diego, USA, 1:40), granulocyte marker antibody (HIS48, biotin, eBiosciences, 1:40) and CD43 (W3/13, AlexaFluor647, BioLegend, 1:40) followed by streptavidin-PECy7 (BioLegend, 1:300) secondary. Cells were analysed on a BD LSRFortessa flow cytometer with standard lasers, and gating strategy and analysis performed using FlowJo v10 software [20]. For quantification of cell numbers, precision count beads (BioLegend) were used.

Assessment of lung injury

Severity of lung injury was graded by visual inspection of the lungs using a semi-quantitative scoring system which graded lungs as: 0- normal; 1- less than 10 petechiae; 2- 10–20 petechiae; and 3 if > 20 petechiae were seen. Lung tissue was also collected and processed as for kidney tissue. Perls’ Prussian blue staining was used to quantify haemosiderin-laden macrophages using ImageProPlus software by measuring proportion of stained cells in 5 HPF.

Assessment of autoantibody response

Anti-MPO antibodies were assayed in serum using a direct ELISA. Plates were coated with 1.33μ g/ml of hMPO overnight, blocked with 3% BSA followed by rat serum, and standards diluted in PBS. A goat anti-rat IgG-ALP conjugate (1:1000, Sigma) was used as a secondary antibody and plates were developed with p-nitrophenyl-phosphate solution (Sigma).

IgG binding to rat leucocytes was assayed using flow cytometry and indirect immunofluorescence (IIF). For flow cytometry, whole blood was collected via cardiac puncture, red cells were lysed (1X RBC Lysis Buffer, eBioscience) and cells were used for cell surface staining with antibodies against CD3 (eBioG4.18, PE), B220 (HIS24, PE), and CD161a (3.2.3, PE), all 1:40. Cells were fixed and permeabilised then incubated with 1:1000 dilution of rat serum followed by an Alexa647 conjugated anti-rat IgG secondary (1:1000, Biolegend). Cells were analysed on a BD LSRFortessa flow cytometer with standard lasers, and gating strategy and analysis performed using FlowJo v10 software. The mean fluorescence intensity (MFI) of PE-negative cells was used to quantify IgG binding. For IIF rat bone marrow was applied to microscope slides using a cytospin at 300 rpm for 3 min. Cells were fixed in 95% ice cold ethanol for 10 min, blocked for 20 min in 20% goat serum and incubated with rat serum diluted 1:10 in PBS. Bound IgG was detected using FITC-conjugated goat anti-rat IgG (1:1000, Sigma).

Deposited rat and rabbit IgG was assessed using direct immunofluorescence on frozen kidney sections using FITC-labelled antibodies and quantified by examining 20 glomeruli and scoring each as 0 to 3+, with results expressed as mean per animal. Indirect immunofluorescence for C3 was carried out using anti-C3 (1:200, 12E2, Abcam) with anti-mouse FITC secondary (1:200, Vector Laboratories, Peterborough, UK). Direct immunofluorescence for MPO was carried out using anti-MPO FITC (Abcam, 2D4) and to aid visualisation of glomeruli, tomato- lectin-DyLight 594 conjugate was added (Vector). For electron microscopy, renal tissue was collected in 2.5% glutaraldehyde. Processing of tissue sections and imaging was carried out by North West London Pathology.

SYK inhibitors

Fostamatinib disodium (R788) was a gift from Rigel Pharmaceuticals (South San Francisco, California). It was reconstituted in vehicle formulation (0.1% carboxymethylcellulose, 0.1% methylparaben sodium, 0.02% propylparaben sodium, in distilled water, pH 6.5). Based on a previous dose-ranging study in nephrotoxic nephritis (NTN) in WKY rats, animals received 30 mg/kg, administered by twice daily oral gavage [21]. Control animals received an equivalent volume and schedule of vehicle formulation.

Statistical Analysis

Statistical analysis was conducted using Prism 8.0 (GraphPad Software Inc., San Diego, CA, USA). Unless otherwise stated, all data are reported as median with interquartile range. Where appropriate, Mann–Whitney U and Kruskal–Wallis tests were used to assess the difference
between 2 or >2 groups, with Dunn’s post hoc test to compare individual groups.

Results

Administration of a sub-nephritogenic dose of NTS in EAV increases renal injury in the presence of autoimmunity to myeloperoxidase, and disease remains pauci-immune
In the conventional NTN model, WKY rats are immunised with 100 µl neat (undiluted) rabbit anti-rat NTS by intravenous (IV) injection, resulting in rapid glomerular deposition of rabbit IgG, urinary abnormalities by day 4, deposition of autologous rat IgG by day 6, and severe crescentic GN by day 10 [17].

To identify a sub-nephritogenic dose of NTS, rats were immunised with serial dilutions of NTS from neat to 1:100 (n=4 per group) and assessed at day 10. After immunisation with 1:50 or lower dilutions of NTS, there was no detectable glomerular injury, and there was no evidence of either deposited rabbit or autologous rat IgG within the kidney by direct immunofluorescence (supplementary material, Figure S1A–J). Thus, a 1:100 dilution of NTS was selected as a sub-nephritogenic dose for all subsequent experiments.

To assess whether the addition of this sub-nephritogenic dose of NTS could augment disease severity in EAV, WKY rats were immunised with human MPO (or human serum albumin (HSA) as control) on day 0, followed by 1:100 NTS (or 1:100 normal rabbit serum (NRS) as control) on day 14 (n=4–6/group). This time point (day 14) after MPO immunisation was selected because rats have developed circulating MPO-ANCA but no urinary abnormalities or glomerular injury. By 7 days after administration of NTS (21 days after MPO immunisation) animals began to develop urinary abnormalities. At day 28 after initial immunisation with MPO, the addition of 1:100 NTS caused significant increases in haematuria (Figure 1A; median dipstick haematuria 3,0,0,0 for MPO/NTS, MPO/NRS, HSA/NTS, HSA/NRS respectively, p<0.0001) and proteinuria (Figure 1B; median proteinuria 137.0, 5.0, 6.0, 3.4 mg/day for MPO/NTS, MPO/NRS, HSA/NTS, HSA/NRS respectively, p<0.0001).

Histological assessment at day 28 showed glomerular necrosis and crescents in all animals immunised with MPO/NTS, with around 60% abnormal glomeruli, including 30% with crescents. Disease was similar to that seen in AAV, with focal disease, crescents, and segmental necrosis the main features (Figure 1C,I). In animals immunised with MPO/NRS there were mild proliferative changes and occasional crescent formation, in keeping with early conventional EAV (supplementary material, Figure S2). Animals immunised with HSA and NTS/NRS had near normal glomerular histology (supplementary material, Figure S2). There was a significant increase in cell infiltration into glomeruli at 28 days in MPO/NTS animals when assessed by immunostaining and this was predominantly CD68/ED-1+ monocyte/macrophages (Figure 1D,I; median % staining/glomerular cross section (GCS) 5.65, 0.04, 0.12, 0.01 for MPO/NTS, MPO/NRS, HSA/NTS, HSA/NRS respectively, p<0.0001). A smaller infiltrate of CD8+ (Figure 1E,I; median % staining/glomerular cross section (GCS) 0.84, 0.01, 0.02, 0 for MPO/NTS, MPO/NRS, HSA/NTS, HSA/NRS respectively, p=0.005) and CD3+ (Figure 1F,I; median % staining/glomerular cross section (GCS) 0.92, 0.04, 0.08, 0.03 for MPO/NTS, MPO/NRS, HSA/NTS, HSA/NRS respectively, p=0.01) cells were also present. The glomerular cell infiltrate was further phenotyped using flow cytometry; this showed an increase in both non-classical and classical monocytes in animals immunised with MPO/NTS, and a small increase in non-classical (NC) monocytes in animals immunised with HSA/NTS (Figure 1G,H; median 17.4, 7.3 and 8.2 NC monocytes/glomerulus for MPO/NTS, MPO/NRS, HSA/NTS respectively, p=0.002).

Importantly, in animals immunised with sub-nephritogenic NTS, disease remained pauci- immune with no detection of deposited autologous rat IgG or C3 at day 28 in any group by indirect immunofluorescence (Figure 2A–F). There were no deposited immune complexes seen using electron microscopy in any group (Figure 2G,H). Cellular crescents and cells interacting with the GBM were seen by electron microscopy in animals immunised with MPO/NTS, but not in other groups (Figure 2G,H and supplementary material, Figure S3A– D).

Susceptibility to GN in this model was limited to the WKY rat strain. Despite developing robust auto-immunity to MPO, Lewis rats did not develop urinary or glomerular abnormalities following immunisation with MPO and low-dose NTS (supplementary material, Figure S4A– F). Immunostaining for monocytes/macrophages using CD68/ED-1 identified a small cellular infiltrate (supplementary material, Figure S4G,H). This was also seen on flow cytometry phenotyping of glomerular cellular infiltrate which identified a small infiltrate of non-classical monocytes (supplementary material).
Addition of a sub-nephritogenic dose of NTS has no effect on lung injury or circulating autoantibodies. There was no difference in lung haemorrhage severity in animals given 1:100 NTS in addition to MPO; both by visual inspection (Figure 3A; median lung haemorrhage score 1,1,0,0 for MPO/NTS, MPO/NRS, HSA/NTS, HSA/NRS respectively) and by Perls’ staining for haemosiderin-laden macrophages (Figure 3B,C; median Perls’ stain 0.13, 0.11, 0, 0 au for MPO/NTS, MPO/NRS, HSA/NTS, HSA/NRS respectively). The degree of lung injury was in keeping with that seen in our previous studies of EAV without additional NTS/NRS [14]. There was no difference in circulating MPO-ANCA levels between the groups of rats immunised with MPO either by ELISA using human MPO, or by flow cytometry to assess IgG binding to rat leucocytes (Figure 3D,E). IIF using normal rat bone marrow cells confirmed that sera from rat immunised with hMPO (+/-NTS) resulted in perinuclear staining in cells with neutrophil nuclear morphology.

Rats immunised with MPO followed by a sub-nephritogenic dose of NTS develop glomerular and tubulointerstitial scarring at 6 and 8 weeks

In the conventional EAV model, disease spontaneously resolves beyond six weeks. To assess disease phenotype at later time points in this enhanced model, WKY rats (n=5 or 6/group) were immunised with the protocol described above for examination at days 42 and 56. We did not examine control animals (MPO/NRS, HSA/NTS, HSA/NRS) at these time points as no significant disease was present in these groups at 28 days. Proteinuria decreased steadily after day 28 in 10/11 rats. One animal developed persistent high levels of proteinuria until day 56 (Figure 4A; median proteinuria 69.9, 54.5, 55.8, 37.9 and 40.1 mg/day at day 28, 35, 42, 49 and 56, respectively). All rats continued to have 3+ haematuria until time of sacrifice (Figure 4B). Cellular crescents at day 28 progressed to a mixture of cellular and fibrocellular crescents at day 42, with further progression at day 56 (Figure 4C,H). Fostamatinib Glomerular cell infiltrate decreased sequentially at days 42 and 56 (Figure 4D,E). CD68/ED-1+ cell infiltrate was predominately tubulointerstitial and peri-glomerular at both time points (Figure 4E,G; median glomerular CD68/ED-1 staining 1.0 and 0.6 %/GCS at week 6 and week 8, respectively). Smooth muscle actin (SMA) staining was used to identify myofibroblasts and the development of fibrocellular crescents, with predominantly interstitial staining seen at day 42, and glomerular and interstitial staining by day 56 (Figure 4F,G,J; median tubulointerstitial staining 0.23, 2.7 and 6.1%/HPF at day 28, 42 and 56 respectively, p=0.0004; median glomerular staining 0.03, 1.2 and 10.1%/GCS at day 28, 42 and 56 respectively, p<0.0001). There was continuing evidence of lung haemorrhage in most animals, more severe at day 42 than day 56, in keeping with the natural history of conventional EAV (supplementary material, Figure S5A–C; median Perls’ stain 1.4 and 0.4 at day 42 and 56 respectively). Circulating MPO-ANCA levels peaked at day 42 after immunisation (supplementary material, Figure S5D). There was no evidence of deposited glomerular rat or rabbit IgG at days 42 or 56.