Total RNA was extracted from matched serum-free (SF) control and 1.0 mM hydrogen peroxide (H₂O₂)-treated (18–20 h) cell lines from three separate experiments for cDNA expression profiling. RNA extraction was carried out using the RNeasy Mini Kit (Qiagen) with combined proteinase K digestion as described by the manufacturer. Total RNA (20 μg) from SF RNA (labelled with Cy5) or 1.0 mM H₂O₂ RNA (labelled with Cy3) (reverse labelling reaction was also performed) was mixed with 40U RNasin (Promega), 4 μg oligo d(T15) and 6 μg random hexamer primers (Invitrogen, Carlsbad, CA) and labelled using the amino-allyl (indirect) method according to the manufacturer’s protocol. Hybridisation was carried out in 20 μg of Cot-1 DNA (Invitrogen), 20 μg of poly dA (Sigma) and 80 μL of DIG Easy Hybridisation solution (Roche Diagnostics, Castle Hill, NSW, Australia) for 14–16 h at 37°C in humidified hybridisation chambers (TeleChem International Inc, Sunnyvale, CA). Microarrays were washed twice in a pre-heated (37°C) solution of 1× saline–sodium citrate (SSC) and 0.1% sodium dodecyl sulphate (SDS) for 5 min, then in 1× SSC for 3 min, and finally in 0.1× SSC for 1 min before drying by centrifugation at 100 g for 5 min. Microarray slides were immediately scanned in a GMS-418 Confocal Scanner (Genetic MicroSystems/Affymetrix Inc, Santa Clara, CA) and images imported into ImaGene 5.0 (BioDiscovery Inc, Marina Del Rey, CA) for data extraction. Mean signal pixel intensities and mean background pixel intensities for Cy3 and Cy5 channels were imported into GeneSpring 6 (Silicon Genetics, Redwood City, CA) and normalised using the Lowess algorithm to correct for intensity-dependent bias and data filtering to remove experimental noise and poor data before further analysis. The microarray chips used in this study were from the Clive and Vera Ramaciotti Centre for Gene Function Analysis (University of New South Wales, Sydney, Australia) containing 5535 oligonucleotides spotted onto 1 microarray slide. Gene expression was calculated as the ratio to the SF RNA. Genes with RNA differentially regulated were further segregated into lists of those that were either upregulated or downregulated by 1.0 mM H₂O₂ treatment. Genes were selected on the basis of their known involvement as active constituents, regulators or associated components of characterised signal transduction pathways, in addition to their possible involvement in renal fibrogenesis. Of the 5535 genes spotted onto the Ramaciotti MWG oligo sets, image analysis involving background subtraction and data normalisation indicated that 76 and 81 genes were differentially regulated by ≥ a twofold change (P < 0.05) between SF control and 1.0 mM H₂O₂-treated samples in renal epithelial (NRK52E) and fibroblast (NRK49F) cell lines, respectively. The expression profile of JAK3 indicated that it was downregulated by 1.0 mM H₂O₂ treatment in both NRK52E and NRK49F cells by fold differences of 4.117 and 6.417, respectively (Table 1). JAK1 and JAK2 had negligible alteration. Thus, JAK3 was selected for the following investigation. RT-PCR was used to confirm that JAK3 cDNA levels were reduced in both cell lines; however, no changes were observed with JAK1 and JAK2 (data not demonstrated).

Renal fibroblast (NRK49F) and epithelial (NRK52E) cell lines were grown to 90–100% confluence in 25 or 75 cm² flasks in Dulbecco’s Modified Eagle Medium (DMEM) containing 10% bovine calf serum and 1000 U/mL penicillin and 1000 μg/mL streptomycin. Cells were treated with either DMEM with (control) and without serum (SF), or SF containing 1.0 mM H₂O₂ for 18–24 h. This treatment concentration was selected because it produces a moderate amount of apoptosis and decreased mitosis in vitro, two characteristics of the intrinsic renal cells in diseases in vivo involving oxidative stress (12). This model has been used in other studies to define expression of many pro-fibrotic molecules (11). Following treatment, cells were analysed for apoptosis and proliferation or trypsinised, collected and pellets stored at –80°C prior to extraction of RNA or protein. For RNA analysis, the process was as described previously. For Western blots, established methods were used (14). In brief, cells were lysed directly on the culture plate in cell lysis buffer (0.15 M NaCl, 0.025 M NaF, 0.5 M EDTA, 0.1% SDS, 1.0% Igepal CA-630 in 50 mM Tris-Cl pH7.5) containing phosphatase and protease inhibitors (10 μg/mL leupeptin, 10 μg/mL aprotinin, 100 μg/mL phenylmethanesulphonyl fluoride, 1 mM sodium vanadate). Cell debris was collected by centrifugation at 16,000 g for 15 min at 4°C and protein concentration in the supernatant determined using a Bradford protein assay (Bio-Rad). Excess protein was aliquoted and stored at −80°C.

All experiments were performed with ethics approval from the University of Queensland Animal Experimentation Ethics Committee (Number Path/RBH/732/03/RD). Male Sprague-Dawley rats (N = 6 per group) weighing 180–220 g were used, as previously described, for a period ranging from 0 to 7 days, and results on increasing development of fibrosis over this time have been previously described by us (12). Briefly for UUO, the left kidney was exposed with a mid-abdominal incision and a double ligature placed on the left ureter approximately 1 cm from the renal hilum, before closing the abdominal incision with 4–0 silk and Michel clips. At each time point, each kidney was quickly sliced transversely (2–3 mm slices) to the length of the kidney. The slices closest to the equatorial plane were either fixed in 10% buffered formalin or 4% buffered paraformaldehyde for histology and immunohistochemistry (IHC), or snap frozen in liquid nitrogen for protein or RNA extraction and analysis.

Apoptosis and mitosis were assessed using morphology (15) following hematoxylin and eosin (H&E) staining and microscopy. Data were obtained by counting 10 frames of cells for controls and treatment and calculating the percentage of apoptotic or mitotic cells. The morphological characteristics for apoptosis were as follows: (i) shrunken eosinophilic cells with condensed, marginated nuclear chromatin and intact cell membrane; (ii) discrete apoptotic bodies comprising large, dense, pyknotic nuclear fragments surrounded by a narrow eosinophilic cytoplasm; and (iii) clusters of small apoptotic bodies (assessed as a single apoptotic occurrence). The morphological characteristics used to distinguish mitosis were as follows: (i) formation of mitotic spindles occurring during metaphase and remaining visible in anaphase or (ii) cells in the later stages of mitosis, telophase or undergoing cytokinesis. Morphological assessment has been confirmed as a reliable measure of apoptosis and mitosis, in a similar experimental protocol using molecular biomarkers for apoptosis (ApopTag) and mitosis (PCNA) in parallel with morphology (16).

Disruption of tissue or cells was carried out in ice-cold cell lysis buffer (50 mM Tris-Cl at pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 25 mM NaF, 0.5 M EDTA) containing protease and phosphatase inhibitors (100 μg/mL PMSF, 20 μg/mL leupeptin, 20 μg/mL aprotinin, 1 mM sodium orthovanadate, all Sigma-Aldrich products) using a tissue homogenizer (14). Cell debris was removed by centrifugation at 16,000 g for 15 min at 4°C. Protein concentration was determined in each tissue extract by a Bradford protein assay (Bio-Rad) and spectrophotometry at 595 nm; 40 μg of total protein were electrophoresed on a 10% SDS polyacrylamide gel using a Bio-Rad mini-protean unit, transferred to a polyvinylidene difluoride (PVDF) membrane and blotted routinely with JAK3 (Neomarkers/Lab Vision, Fremont, CA) (1:100), ERK (1:1000) or pERK (1:1000) (both from Cell Signaling Technology, Beverly, MA) primary antibodies and then appropriate horse-radish peroxidase conjugated secondary antibodies diluted (1:2000–1:5000) in 5% powdered milk in Tris-buffered saline. Phosphorylated tyrosine (pTyr) and phosphorylated STAT3 (Invitrogen; 1:400) were analysed for JAK activation. Protein bands were visualised using enhanced chemiluminescence. X-ray film was scanned using a Hewlett Packard ScanJet 3200 C at 300 dpi, and Scion Image (β4.0.2) software was used to quantify the density of the bands in arbitrary densitometry units. Membranes were routinely stained with Coomassie Blue (Sigma-Aldrich), or actin immunoblots were used, to verify equal protein loading of lanes.

Sub-serial sections of paraffin-embedded kidney slices were cut, numbered, deparaffinised and rehydrated, before staining with the peroxidase–anti-peroxidase IHC method published by us previously (17). Non-specific binding of peroxidase or antibodies was blocked with 0.3% H₂O₂ in 4% skim milk powder followed by incubation in diluted normal rabbit or goat serum. Primary antibodies were JAK3 (mouse monoclonal, Neomarkers/Lab Vision, Fremont, CA; 1:100) or phosphorylated ERK (pERK; rabbit polyclonal, New England Biolabs, Beverly, MA, USA; 1:400 dilution) and PCNA (mouse monoclonal, DAKO, Glostrup, Denmark; 1:50 dilution). The appropriate secondary antibody was applied at a dilution of 1:200 to 1:400 in Tris buffer. Sections were developed using the peroxidase–anti-peroxidase reaction plus the chromogen diaminobenzidine tetrahydrochloride (DAB) for IHC. Negative controls were prepared without primary antibody or with non-specific serum. Positive tissue control sections from unrelated studies were included. Sections were lightly counterstained with haematoxylin and then dehydrated in ethanols, cleared in xylene and mounted in Depex.

Using paraffin-embedded sections, double antibody IHC was performed with pERK (1:400) and JAK3 (1:100). Staining procedures were identical to previous IHC methods with the exception of the chromogen. For pERK IHC, sections were developed with the NovaRED™ substrate kit for peroxidase detection (red label) (Vector Laboratories; catalogue number SK-4800), and JAK3 was developed with DAB.

To assess any temporal association between altered JAK3 mRNA (Table 1) and apoptosis or mitosis, percentage change in these parameters was investigated in control, SF and 1.0 mM H₂O₂-treated cultures of NRK49F fibroblast and NRF52E tubular epithelial cells at 24 h of treatment. Figure 1 demonstrates percentage change in apoptosis (A) and mitosis (B). In Figure 1A, the low levels of apoptosis in both cell lines for control or SF medium were significantly increased compared with oxidative stress (1.0 mM H₂O₂). The increase was 10-fold greater in fibroblasts and fivefold greater in epithelial cells compared with controls. In Figure 1B, levels of mitosis seen in control and SF cultures were significantly decreased (P < 0.05) compared with 1.0 mM H₂O₂. Mitosis was almost negated in NRK52E cells with oxidative stress. Temporally, these changes in apoptosis and mitosis with oxidative stress were associated with the decrease in JAK3 mRNA in both cell lines.

To determine if the changes in mRNA in JAK3 were translated to the protein, extracts from control, SF and 1.0 mM H₂O₂-treated cultures were subjected to Western blotting. Figure 2A demonstrates JAK3 protein expression in cell culture experiments, for NRK49F and NRK52E cells for control, SF medium and medium containing 1.0 mM H₂O₂ for 18–20 h. These were the same experiments used for the mRNA analyses (Table 1) and for histological analysis of apoptosis and mitosis (Figure 1). JAK3 protein expression (125 kDa approximately) was not significantly altered from controls after H₂O₂-treatment in either cell line, although JAK3 expression for NRK52E in the SF treatment tended to be lower than the normal medium controls, and the H₂O₂ treatment (Figure 2A). To determine if JAK3 protein expression was altered in vivo, Western immunoblotting was performed on protein extracts from whole kidney tissue taken from UUO rats at 0 (control), 6 h, 1, 2, 4 and 7 days. Significant levels of oxidative stress and tubulointerstitial fibrosis had been reported by us previously at 4 days of obstruction (12). Figure 2B demonstrates that there was no alteration in whole kidney JAK3 protein from controls in UUO kidneys from animals with increasing duration of UUO (6 h, 24 h, 2, 4 and 7 days).

The differences between mRNA and protein from in vitro Western blot results of JAK3 suggested post-transcriptional control on JAK3. To investigate if JAK3 activation (phosphorylation) was altered, pTyr of JAK3 and phosphorylated STAT3 (pSTAT3) were analysed (Figure 3). In Figure 3A, similar to JAK3 protein expression (Figure 2A), pTyr of JAK3 in tubular epithelial NRK52E cells was significantly increased after oxidative stress. In NRK49F cells, although JAK protein expression was stable (Figure 2A), JAK3 activation (p-Tyr) decreased with oxidative stress (Figure 3B), thereby correlating with the microarray results of decreased JAK3 mRNA for these cells. In Figure 3C, pSTAT3 expression was investigated in an attempt to determine its post-transcriptional association with JAK3. STAT3 phosphorylation was negligible in both H₂O₂-treated cell lines. The SF controls for the NRK49F cells also had negligible pSTAT3. The pSTAT3 results were counterintuitive (pSTAT3 often increased in fibrosis) but may be the result of loss of pSTAT3-stimulatory cytokines after SF and oxidative stress treatments. This requires further study.

To determine whether localisation of JAK3 altered during the development of kidney fibrosis during UUO (proven association with oxidative stress) from that seen in control kidneys, IHC analysis was conducted on paraffin-embedded sections using the same antibody to JAK3 as was used for Western immunoblots. Activation of JAK3 was not investigated. Figure 4A shows negligible JAK3 in the cortex of control kidneys. Figure 4B demonstrates localisation of JAK3 to vascular endothelium (red blood cells visible in the vascular lumens), but little expression otherwise, in the outer stripe of the medulla of controls. With increasing duration of UUO, there was positive JAK3 staining of tubular epithelial cells in the cortical regions (4C, 7 days UUO demonstrated). In all cases, subcellular localisation of JAK3 was cytoplasmic. Tubular epithelium with apoptotic cells had little to no JAK3 expression. There were very few interstitial cells (fibroblasts) that stained positive for JAK3.

Double IHC staining confirmed that JAK3 was mainly localised to tubular cells at the later stages of UUO (4–7 days UUO), but not associated with apoptotic epithelium in UUO. It is possible that JAK3 expression was in apoptotic cells of the epithelium and that the shrinkage of apoptotic cells during the death process meant the JAK3 positivity was not detected. This would influence the lack of co-localisation of JAK3 and pERK (Figure 5), with one interpretation being that JAK3 did not localise to apoptotic cells that occur in this structure (12) and the other interpretation being that the apoptotic cell structure precluded detection of JAK3 in the cytoplasm.