Spleen tyrosine kinase inhibition is an effective treatment for established vasculitis in a pre-clinical model.

The anti-neutrophil cytoplasm antibody (ANCA)-associated vasculitides (AAV) are a group of life-threatening multi-system diseases characterized by necrotising inflammation of small blood vessels and crescentic glomerulonephritis. ANCA are thought to play a direct pathogenic role. Previous studies have shown that spleen tyrosine kinase (SYK) is phosphorylated during ANCA-induced neutrophil activation in vitro. However, the role of SYK in vivo is unknown. Here, we studied its role in the pathogenesis of experimental autoimmune vasculitis, a pre-clinical model of myeloperoxidase-ANCA-induced pauci-immune systemic vasculitis in the Wistar Kyoto rat. Up-regulation of SYK expression in inflamed renal and pulmonary tissue during early autoimmune vasculitis was confirmed by immunohistochemical and transcript analysis. R406, the active metabolite of fostamatinib, a small molecule kinase inhibitor with high selectivity for SYK, inhibited ANCA-induced pro-inflammatory responses in rat leucocytes in vitro. In an in vivo study, treatment with fostamatinib for 14 days after disease onset resulted in rapid resolution of urinary abnormalities, significantly improved renal and pulmonary pathology, and preserved renal function. Short-term exposure to fostamatinib did not significantly affect circulating myeloperoxidase-ANCA levels, suggesting inhibition of ANCA-induced inflammatory mechanisms in vivo. Finally, SYK expression was demonstrated within inflammatory glomerular lesions in ANCA-associated glomerulonephritis in patients, particularly within CD68+ve monocytes/macrophages. Thus, our data indicate that SYK inhibition warrants clinical investigation in the treatment of AAV.

L iv e A n n e x in + A n n e x in + /P I+

SYK inhibitors
R406, the active small molecule inhibitor of SYK, and its oral prodrug, fostamatinib disodium (R788), were provided by Rigel Pharmaceuticals (South San Francisco, California) and AstraZeneca (UK). The details of these molecules have been reported previously 1,2 .
Fostamatinib was provided as the disodium salt. For in vivo experiments, it was reconstituted in vehicle formulation (0.1% carboxymethylcellulose, 0.1% methylparaben sodium, 0.02% propylparaben sodium, in distilled water, pH 6.5) to a concentration of 8mg/ml, sonicated for 15-20 minutes until dissolved, and then pH re-adjusted to 6.5 with 0.1M HCl. This preparation was stored at 4°C for a maximum of seven days prior to use.
Based on a previous dose-ranging study in nephrotoxic nephritis in Wistar Kyoto (WKY) rats 3 , treated animals received 20mg/kg or 30mg/kg body weight, administered by twice daily oral gavage. Control animals received an equivalent volume and schedule of vehicle formulation. Animals were dosed by group (vehicle/20mg/30mg) and in numerical order within group.
For in vitro experiments, R406 was reconstituted in 0.02% dimethylsulphoxide (DMSO; Sigma-Aldrich, Poole UK) in cell culture media as detailed below.

Animal study approval
All animal procedures were licensed by the Home Office Science Unit (personal license 70/23027 and project license 70/7104) and conducted in accordance with the UK Animals (Scientific Procedures Act) 1986.

Animal husbandry
Rats were purchased from Charles River (Margate, UK) and maintained in a pathogen-free animal facility at the Central Biomedical Services Unit, Imperial College London, Hammersmith Hospital campus, in individuallyventilated cages with free access to water and standard laboratory diet. Cage occupancy varied between two to five rats per cage, depending on body weight. Whilst undergoing gavage, animals were provided with wet laboratory diet in case of oesophageal irritation. Animal well-being was assessed daily and monitored weekly by measurements of body weight (Table 1). Immunised animals also received 500ng of pertussis toxin (Invitrogen, Life Technologies) in PBS intraperitoneally on day zero and day two. For control experiments, animals were immunised with CFA, Mycobacterium butyricum, and pertussis toxin using a comparable immunisation schedule.

Urine collection
For collection of urine samples, rats were housed overnight, with free access to water and standard laboratory diet, in individual metabolism cages. At the end of the collection, total urine volume was measured, and aliquots were centrifuged at 1500rpm for five minutes to sediment macroscopic debris, and stored at -80°C until use.
There was a minimum 48 hour interval between urine collections, in accordance with the terms of the project license.

Serum collection
For collection of serum samples at non-terminal time-points, rats were placed in a warming chamber at 30°C to promote vasodilation, then placed under isofluorane anaesthesia on a warming mat. Whole blood (maximum volume 0.5ml) was collected by superficial tail vessel puncture using a 23 gauge needle. Samples were centrifuged at 1500rpm for five minutes, serum removed and stored at -80°C until use.

Terminal processing of blood and tissues
At the end of the study protocol, animals were placed under isofluorane anaesthesia for final collection of blood and tissues. Blood was obtained either by cardiac puncture using a 10ml pre-heparinised syringe and a 23 gauge needle, or by transection of the great vessels and exsanguination into both plain and heparinised collection tubes.
Tissues were dissected out after the animal had been sacrificed by exsanguination. Lung haemorrhage score was documented by inspection of the lung surfaces before dissection of organs. For each animal, a coronal mid-pole section of kidney, and a sample of lung, spleen and liver tissue was placed in a histology cassette and transferred to 10% neutral buffered formalin for fixation. A remaining kidney pole was placed in OCT (optimal cutting temperature compound) embedding matrix (ThermoFisher Scientific, Waltham, Massachusetts) on a cork disc, immersed in isopentane and then snap frozen in liquid nitrogen, and stored at -80°C until use. The remaining kidney material, and a sample of lung tissue and spleen tissue were snap frozen in individual cryovials and stored at -80°C until use for PCR or Western blot.

Biochemical analysis of serum and urine
Creatinine concentrations in serum samples were kindly measured by Dr Olatunji Rowland and Dr John Morris in the Department of Clinical Biochemistry, Hammersmith Hospital UK, using an AU700 analyser (Olympus, Southend, UK).
Proteinuria was quantified using sulphosalicylic acid method. Urine samples, diluted between 1:3 and 1:100 in water, were added to a 96-well microtitre plate in triplicate, to which 10μl of 25% sulphosalicylic acid was added to two of three replicates, and 10μl of water to the third replicate (providing a 'blank' reading for that sample).
Known quantities of bovine serum albumin (BSA; Sigma) in water were used to define a standard curve, and absorbance read at 450nm on microplate reader and dedicated software (Biotek EL800; Gen5 Analysis Software; both Biotek Instruments Ltd, Potton, UK). The protein concentration in each sample was calculated from a regression equation described by the standard curve, and the 24 hour protein excretion rate was calculated by multiplying by the total urine volume.

Haematological analysis
Haemoglobin concentrations and white blood cell and platelet counts were measured in heparinised whole blood samples using an automated analyser (XE-2100, Sysmex, Milton Keynes, UK) in the Department of Clinical Haematology, Hammersmith Hospital UK.

Renal histology
Kidney tissue collected at the time of sacrifice was fixed in 10% neutral buffered formalin overnight, then transferred to 70% ethanol and processed to paraffin blocks. 4μm sections were cut on a rotary microtome and stained with haematoxylin and eosin (H&E) and periodic acid Schiff (PAS) for assessment of renal injury. For quantification of renal injury in EAV, 50 consecutive glomeruli were assessed by light microscopy in a blinded fashion, and graded as normal or abnormal. Results are expressed as the percentage of abnormal glomeruli per animal.

Immunohistochemistry for leucocyte markers
Immunostaining

Macroscopic assessment of lung injury
A semi-quantitative scoring system was used to grade the severity of lung haemorrhage by visual inspection at the time of cull. Lungs were graded as follows: 0 points -normal macroscopic lung appearances; one point if fewer than ten petechiae were visible; two points if ten to 20 petechiae were visible; three points if more than 20 petechiae were visible; four points if large areas of infarction or haemorrhage were evident. Representative lung appearances are shown in Figure 1. Lung tissue was collected for histological analysis and processed as per kidney tissue (section 3.4). 4μm paraffinembedded sections were stained with Perls' Prussian blue without counterstain, to enable identification of haemosiderin-laden cells. These were quantified by a blinded observer using automated image analysis software (ImagePro Plus, Media Cybernetics, Rockville, Maryland) to measure the proportion of Perls' stained cells across five random high-power fields of lung sections from each animal, and expressed as the mean proportion per high power field per animal.

ELISA for anti-MPO antibodies
Anti-MPO antibody levels in rat sera were measured using direct ELISA 4,5 . Wells were coated with purified human MPO (Calbiochem) at a concentration of 1.33μg/ml in carbonate buffer and stored at 4°C overnight. The following day, plates were washed three times in 0.1% PBS/T, and non-specific binding sites blocked by incubating the wells with 1% BSA in 0.1% PBS/T for one hour. After three washes in 0.1% PBS/T, 100μl of serum samples (diluted 1:100 to 1:100000 in PBS/T plus 1% BSA) were applied in duplicate, with a negative control 'blank' that did not contain serum. Plates were incubated at 37°C for one hour and then washed three times in 0.1% PBS/T. 100μl of rabbit anti-rat IgG with ALP conjugate in 0.1% PBS/T plus 1% BSA (dilution 1:1000) was then added for one hour at 37°C. After a wash cycle in 0.1% PBS/T, 100μl of p-nitrophenyl phosphate was applied to each well. The colour change was monitored, and then quantified on a microplate reader at 405nm. A coefficient of variance of 10% between sample replicates was accepted. Historical pooled sera were used to develop a standard curve from which subsequent samples could be assigned an interpolated value in arbitrary units.For quantification of IgG subclasses, a modification of this protocol was used, substituting secondary antibodies specific for rat IgG1, IgG2a, IgG2b, and IgG2c (BioLegend, San Diego CA).

Direct immunofluorescence for deposited antibodies
IgG deposition in glomeruli was assessed by direct immunofluorescence using snap frozen renal tissue obtained at the time of cull. Frozen kidney sections were cut on a cryostat at 5µm thickness and placed on poly-L-Lysine coated slides (Leica Biosystems, Milton Keynes, UK). After fixation in acetone for ten minutes, the slides were air dried. If not used immediately, the slides were stored at -80°C in sealed boxes containing silica crystals. The slides were blocked with 20% normal rabbit serum (Dako Agilent, Santa Clara CA) for 30 minutes at room temperature. After washing in PBS, the slides were incubated with FITC (fluorescein isothiocyanate)-conjugated rabbit anti-rat IgG (Sigma) at 1:100 dilution in PBS for one hour at room temperature in a humidified chamber.
Following two ten minute washes in PBS, the sections were mounted in PBS/Glycerol (Citifluor, London, UK). For quantification, 20 consecutive glomeruli on each section were inspected by fluorescence microscopy in a blinded fashion, and intensity graded 0/1+/2+/3 in arbitrary units, and expressed as the mean intensity per glomerulus for each animal. Positive control tissue taken from animals after induction of nephrotoxic nephritis 6 were used for this analysis.

Derivation of undifferentiated bone marrow cells
After sacrifice as described in section 2.4.3, healthy WKY rat femurs were isolated and transferred to a tissue culture hood in ice-cold sterile HBSS solution. The bones were cleaned of hair and soft tissue, washed twice in 70% ethanol, and rinsed again in HBSS. Both ends of the cleaned bones were then cut, and bone marrow cells flushed out with a 20 gauge needle using 10ml of cold HBSS per bone, and collected into a 50ml centrifuge tube.
Cells were then centrifuged at 1500rpm for five minutes at 4°C, and re-suspended in 10ml HBSS buffer. The 50ml centrifuge tube was then placed in a CO2 incubator in a horizontal position for 10min to hypotonically lyse red blood cells. The cells were then centrifuged at 1500rpm for five minutes at 4°C, the supernatant aspirated, and the cell pellet re-suspended in serum-free cell culture medium (RPMI supplemented with 2% penicillin and streptomycin, and 2mM L-glutamine).

Derivation of MPO-ANCA and control rat IgG
Rat serum collected from historical animals six weeks after induction of EAV was diluted in a 1:1 ratio in binding buffer (20mM sodium phosphate, pH 7.0) and sterile filtered (0.2 nm filter), then passed through a 'HiTrap' protein G sepharose column (GE), and eluted by 0.1M glycine HCl (pH 2.7) into Tris HCl (pH 9). Total IgG concentrations were quantified by spectrophotometry (absorbance 260nm) and aliquots with highest concentration were pooled, dialysed overnight in a 10,000 MW cut-off cartridge against sterile PBS, and total IgG concentrations then re-quantified as above.

Stimulation of bone marrow cells
Undifferentiated bone marrow cells were suspended in serum-free cell culture medium to a final concentration of 20 million cells per ml in 24 well cell culture plate, and primed by 30 minutes exposure to rat TNFα (final concentration 4ng/ml; Sigma), and then stimulated with either MPO-ANCA IgG (100μg/ml) or control IgG (100μg/ml). For SYK inhibitor studies, cells were pre-incubated with vehicle (0.02% DMSO) or 0.2μM, 1μM and 2μM R406 in vehicle for 30 minutes (after TNFα-priming and prior to IgG stimulation). Cell culture supernatants were collected after four hours and stored at -80°C until used. Experiments were conducted in technical (cell culture) triplicate and with four biological replicates.

Annexin-PI assay
Annexin PI (Biolegend) viability assay was carried out on cells following stimulation as detailed. Cells were recovered, washed twice in cold cell staining buffer and resuspended in binding buffer at 1x10 7 cells/ml. 5µl of Annexin V and 10 µl of PI solutions were added to 100 µl of cell suspension and incubated at room temperature for 15 minutes followed by addition of excess binding buffer and analysis by flow cytometry.

Measurement of monocyte chemoattractant protein 1 (MCP-1)
MCP-1 concentrations in cell culture supernatants were measured using a commercially available sandwich Cambridge Bioscience, Cambridge UK) and hydrogen peroxide was prepared, and 100µl added to each well. The plate was incubated in the dark for approximately ten minutes before stopping the reaction by adding 50µl of 2M sulphuric acid. The absorbance was measured at 450nm on a microplate reader. Adjusted absorbance was calculated by subtracting the background absorbance of the 'blank' sample. A best-fit standard curve was generated from the absorbances of the recombinant MCP-1 reference series using four variable non-parametric regression modelling. Concentrations of MCP-1 in samples could then be calculated from the raw blanksubtracted absorbances using the resultant regression equation. A co-efficient of variance of <10% between sample replicates was accepted. As absolute amounts of MCP-1 production were variable between each biological replicate experiment, results were standardised and expressed as a percentage maximum MCP-1 production for a given experiment. Tert-butyl hydroperoxide.

Pre-treatment of paraffin sections
All IHC in this study was performed on formalin-fixed paraffin embedded tissue sections. Paraffin wax embedding and sectioning of formalin-fixed animal tissues was performed by Ms Lorraine Lawrence (Department of Leukocyte Biology, Imperial College London). Sections were cut on a rotary microtome to a thickness of 4µm.
Sections were placed in two sequential baths of xylene, to remove wax, before passage through graded ethanol concentrations and finally water. De-waxed sections underwent heat-induced epitope retrieval in 0.1M sodium citrate buffer, pH 6.0. The sections were then sequentially blocked for endogenous peroxidase activity by submerging in 0.3% hydrogen peroxide from ten minutes, and rinsed in PBS. A 20% serum solution was applied for 30 minutes (See Table 1 for details) to block non-specific binding of subsequent antibodies.

Primary antibodies
The blocking antibody was tapped off, and primary antibodies added at the appropriate dilution in sufficient volume to cover the tissue section, typically 50-200μl. Slides were placed in a covered, humidified staining chamber to minimise evaporation and left to incubate for the appropriate duration (Table 1). For SYK IHC, negative controls comprised omission of primary antibody and, where available, use of primary antibody preincubated with the cognate immunising peptide at a 1:1 ratio. Omission of primary antibody was used for IHC protocols that were already established in our laboratory (e.g. ED-1).

Detection
Three methods were used for detection of primary antibodies in tissue sections: (i) A commercial secondary antibody detection system (EnVision, Dako) was used for the detection of the total and phosphorylated SYK antibodies (Table 1). Primary antibody was tapped off and slides washed two times for five minutes in PBS. Sufficient EnVision reagent was then added to completely cover the tissue section and the slide placed in the humidified incubation chamber for 30 minutes. The reagent was then tapped off and a further two PBS washes performed.
(ii) For rat ED1 staining, excessive non-specific background staining within glomeruli was observed using the EnVision system. In these instances, a biotinylated rabbit anti-mouse secondary antibody (as detailed in Table   1) was incubated with the sections for one hour. The sections were then washed twice in PBS, and incubated with an Extravadin-HRP conjugate (Sigma) at a dilution of 1:100 for 30 minutes, then washed thrice in PBS. A 3,3'-diaminobenzidine (DAB) chromagen solution was prepared by mixing DAB+ chromagen with DAB+ substrate buffer in a ratio of 1:50 (Dako). This was applied to tissue sections for 30 seconds to five minutes depending on the primary antibody. The reaction was terminated by placing slides in PBS. Slides were then rinsed in water and counterstained by immersion for 30 seconds in filtered Harris haematoxylin (CellPath, Powys, UK). After rinsing, excess stain was removed by brief (<1 second) immersion in acid-alcohol solution (1% HCl in 70% ethanol). Sections were then washed in water and dehydrated by sequential passage through ascending alcohol concentrations to xylene. Slides were then mounted with DPX (Distrene, Plasticiser, Xylene; VWR International, Lutterworth, UK) and glass coverslips and allowed to dry before examination.

Double staining
For double staining for SYK and CD68/ED-1, in both rat and human tissue, sections were first stained for T-SYK as described above. Following development in DAB, the sections were sequentially rinsed, re-subjected to heat-

RNAScope™
RNAscope™ in situ hybridization was carried out on FFPE tissue sections according to the manufacturer's instructions (Advanced Cell Diagnostics, Newark CA). Freshly cut sections were baked for 1 hour at 60°C then de-waxed as described for IHC. Antigen retrieval was for 15 minutes in boiling TR reagent. Sections were then incubated with hydrogen peroxide for 10 minutes at room temperature followed by Protease Plus at 40°C in a HybEZ™ oven. Sections were then sequentially incubated with hybridization probed for SYK, PPIB (positive control) or DapB (negative control), followed by six amplification probes; these steps were performed at 40°C in a HybEZ™ oven. Signal was detected using 2.5 HD red detection reagents. Sections were washed in distilled water to stop the reaction and counterstained using Mayer's haematoxylin. Sections were dried for one hour at 60°C, briefly dipped in xylene and mounted using aqueous mounting media.

Polymerase chain reaction
Approximately 50mg of rat renal cortex (from animals with EAV, animals with EAV treated with fostamatinib or vehicle, andcontrol rats immunised with CFA alone (n=4/group minimum)) was homogenised in 1ml of TRIzol   anti-mouse IgG HRP-conjugated secondary antibody (1:5000, BioRad) for one hour followed by further washes and signal detected using ECL detection reagents (Thermofisher) and photographic film (Amersham).
Membranes were then stripped of antibody by incubating with stripping buffer (15% glycine, 1% SDS, 0.1% Tween-20, pH 2.2) and re-probed for T-SYK (Santa-Cruz) and GAPDH (R+D) as above but with 5% marvel milk solution in place of BSA and secondary antibodies of the appropriate species.

Statistics
Graphs were constructed and statistical analysis conducted using Prism 5.0 (GraphPad Software Inc., San Diego, California