Introduction

The accumulation of inflammatory cells in renal biopsies of diabetic patients is associated with tissue damage and a progressive decline in renal function [1]. In animal models, the use of immunosuppressants, neutralising antibodies and genetic deficiencies has shown that reduction of leucocyte accumulation and activation in diabetic kidneys suppresses the development of renal injury [2]. Most of this research has focused on the contribution of macrophages, which are the major infiltrating immune cell in diabetic kidneys. However, many of these previous studies have used anti-inflammatory strategies that also suppress accumulation of lymphocytes in diabetic kidneys, indicating that lymphocytes may also contribute to disease progression.

Lymphocytes may play a role in the development of diabetic nephropathy, despite any clear evidence of antigen involvement. T cells are known to play a significant role in renal injury induced by non-immune insults including ischaemia or toxins (e.g. adriamycin) [3, 4]. Activated T cells can cause injury directly through cytotoxic effects and indirectly by recruiting and activating macrophages. In addition, kidney autoantigens may develop during chronic diabetic renal injury and, if this occurs, B cells could present these antigens to T cells to promote their activation. Furthermore, diabetic patients have increased levels of serum immunoglobulins, which include antibodies against proteins modified by glycoxidation or lipoxidation [5, 6]. These circulating antibodies can form immune complexes [6, 7], which may deposit in glomeruli and promote activation of complement or macrophages via receptor interactions.

The development of diabetic renal injury has been associated with T cell activation in the circulation and T cell accumulation in the kidney. Type 1 diabetic patients with mild proteinuria have higher levels of circulating activated lymphocytes than non-proteinuric patients; these levels decline when patients develop nephrotic-range proteinuria [8]. This suggests that lymphocyte activation may play a role in early diabetic nephropathy. A recent clinical study also established that T cell infiltration of the juxtaglomerular apparatus is common during the early stages of type 1 diabetes and correlates with an increased filtration surface per glomerulus and a lower albumin excretion rate [9]. This suggests that T cells in the juxtaglomerular apparatus could delay the onset of albuminuria in type 1 diabetic patients. In addition, rodent models of type 1 and type 2 diabetic nephropathy have demonstrated that CD4+ and CD8+ T cells accumulate in the glomeruli and interstitium of diabetic kidneys, but on a much smaller scale than macrophages [10–12], suggesting that these cells may interact with macrophages to regulate inflammation and renal injury. In comparison, there is no direct evidence that B cells infiltrate the kidney during diabetes. However, the involvement of B cells is indicated by studies showing glomerular deposition of immunoglobulin in experimental diabetic nephropathy [11, 13–15].

Elements of the diabetic milieu can directly or indirectly activate T cells in diabetic kidneys. CD4+ T cells express the receptor for AGEs and can respond to AGEs by producing IFN-γ, which could exacerbate inflammation in the diabetic kidney [16]. In addition, hyperglycaemia induces macrophage production of IL-12, which can also stimulate CD4 cell production of IFN-γ [17]. In diabetic kidneys of NOD mice, accumulation of CD8+ cells, which also express the receptor for AGE, is associated with increased expression of genes encoding perforin and granzyme B, and with co-localisation in immunostaining for perforin [14]. This suggests that CD8+ cells may exert a cytolytic function in the diabetic kidney. It is also noteworthy that some CD4+ cells in diabetic kidneys could be T-regulatory cells and play a role in suppressing inflammation, although this has yet to be investigated.

It is important to identify the functional role of lymphocytes in diabetic renal injury, as researchers would then be able to determine the relative contribution of innate and adaptive immune components. This would help in the development of anti-inflammatory strategies to treat diabetic nephropathy. Mice deficient in Rag1 (Rag1 −/−) do not produce mature T and B lymphocytes [18] and can be used to monitor the development of kidney disease in the absence of lymphocytes [19]. In the current study, we used streptozotocin to induce equivalent diabetes in wild-type (Rag1 +/+) and Rag1 −/− mice, and examine the impact of lymphocyte deficiency on the development of type 1 diabetic nephropathy.

Methods

Animal model

Rag1 −/− mice with a C57BL/6 background and wild-type controls were obtained from the Animal Resource Centre (Perth, WA, Australia) and were bred under pathogen-free conditions at the Monash Medical Centre Animal Facility (Clayton, VIC, Australia). Groups of 8-week-old male Rag1 −/− and wild-type mice (n = 8–10) were given intraperitoneal injections of streptozotocin (Sigma, St Louis, MO, USA) at day  0 (100 mg/kg) and day  5 (125 mg/kg). Using this protocol, both strains were equally susceptible to the development of streptozotocin-induced diabetes and displayed equivalent hyperglycaemia. Mice with significant diabetes (blood glucose >16 mmol/l at 2 weeks after the first streptozotocin injection) were followed for 20 weeks. Age-matched non-diabetic Rag1 −/− and wild-type mice (n = 7) were used as controls. Mice were housed in pathogen-free micro-isolator cages and maintained on a normal diet. Random morning blood glucose (tail vein sample) and body weight were measured weekly and mice with glucose levels >30 mmol/l were given 0.5 units of protaphane insulin (Novo Nordisk, Sydney, NSW, Australia) subcutaneously three times a week to maintain body weight. Urine was collected at weeks 8, 12, 16 and 20 to assess urine albumin excretion. HbA1c was measured from blood samples at weeks 8 and 20 to confirm that hyperglycaemia remained equivalent in both diabetic groups. Tissues were collected at week 20 and fixed in 4% (vol./vol.) formalin, 2% (wt/vol.) paraformaldehyde–lysine–periodate (PLP) and methyl Carnoy’s fixative (60% methanol, 30% chloroform, 10% glacial acetic acid; vol./vol.), or snap-frozen and stored at −80°C. All animal experiments were approved by the Monash Medical Centre Animal Ethics Committee.

Biochemistry

Urine was collected from mice housed in metabolism cages for 24 h. Whole blood was collected by cardiac puncture in anaesthetised animals and stored as serum or heparinised plasma. ELISA kits were used to determine urine albumin (Bethyl Laboratories, Montgomery, TX, USA), urine monocyte chemoattractant protein 1 (MCP-1), and serum and urine IL-6 (OptEIA; BD Biosciences, San Diego, CA, USA) as per manufacturer’s instructions. HbA1c, and serum and urine creatinine were measured by HPLC. Creatinine clearance was adjusted to body surface area (ml min−1 m−2).

Histopathology

Staining with periodic acid–Schiff’s (PAS) reagent and haematoxylin was performed on 2 µm sections of formalin-fixed kidney for structural assessment and morphometric analysis. Glomerular volume and PAS+mesangial matrix fraction were analysed by computer image analysis (magnification ×400) [20], while total cellularity was determined by counting nuclei in 20 hilar glomerular cross sections (gcs) per animal. Tubular injury was assessed by counting injured tubules (dilated, atrophic, necrotic) as a percentage of total tubules in ten cortical fields (magnification ×250). All scoring was performed on blinded slides.

Antibodies

The primary antibodies used in this study were: rabbit anti-Wilm’s tumour antigen 1 (WT1) (Santa Cruz Biotechnology, Santa Cruz, CA, USA); goat anti-collagen IV (Southern Biotechnology, Birmingham, AL, USA); rabbit anti-cleaved caspase 3 (ASP175), rabbit anti-phospho-p38 mitogen-activated protein kinase (Thr180/Tyr182) and rabbit anti-phospho-c-Jun amino-terminal kinase (Thr183/Thy185; Cell Signaling, Beverly, MA, USA); rat anti-mouse Ki-67 (TEC-3; Dako, Carpinteria, CA, USA); rat anti-CD68 and rat anti-CD169 (Serotec, Oxford, UK); rat anti-mouse forkhead box P3 (FOXP3; eBiosciences, San Diego, CA, USA); mouse anti-α-tubulin, fluorescein conjugated mouse anti-α-smooth muscle actin (SMA) and rabbit anti-podocin (Sigma-Aldrich, St Louis, MO, USA); biotinylated mouse anti-CD22.2 (BD Pharmingen, San Diego, CA, USA); and rat anti-CD4 (GK1.5) and anti-CD8 (YTS169) were produced by cell culture of hybridomas obtained from the American Tissue Culture Collection (Manassas, VA, USA).

Immunohistochemistry

Immunostaining for WT1, α-SMA, cleaved caspase-3 and Ki-67 was performed on 4 µm formalin-fixed paraffin sections. Immunostaining for collagen IV was performed on 4 µm sections fixed in methyl Carnoy’s solution. Immunostaining for CD4, CD8, CD68 and CD169 was performed on 5 µm PLP-fixed cryostat sections, while immunostaining for FOXP3 and CD22 was performed on 6 µm snap-frozen sections. For retrieval of antigens in formalin-fixed sections (except α-SMA), slides were heated in a pressure cooker (high setting 20 min, full pressure 5 min) in a 10 mmol/l sodium citrate buffer (pH 6). Endogenous peroxidise was quenched by immersing slides in 3% (vol./vol.) H2O2/PBS (formalin-fixed sections) or 0.5% (vol./vol.) H2O2/methanol (PLP and snap-frozen sections). Antigens were labelled by overnight incubation with primary antibody, followed by biotinylated secondary antibodies (Vector Laboratories, Burlingame, CA, USA). Detection was done using a standard peroxidase-ABC system (Vector) and development with 3,3-diaminobenzidine (Sigma). For detection of FITC-conjugated anti-α-SMA monoclonal antibody, sections were incubated with peroxidase-conjugated sheep anti-fluorescein IgG (Roche Biochemicals, Mannheim, Germany) prior to development with 3,3-diaminobenzidine. For FOXP3 and CD22 staining, frozen sections were fixed in ice-cold acetone for 10 min and quenched prior to incubation with primary antibody. Normal rabbit and goat serum or isotyped-matched irrelevant IgGs were used as negative controls.

Immunofluorescence for mouse IgG and podocin

For detection of IgG, frozen sections (6 µm) were fixed in ethanol at 4°C for 10 min. After blocking with 10% (vol./vol.) sheep serum, sections were incubated with FITC-conjugated sheep anti-mouse IgG Fab fragments (Silenus, Melbourne, VIC, Australia) at 1:50 dilution for 30 min. For detection of podocin, frozen sections (6 µm) were fixed in acetone on ice for 20 min. Sections were then blocked with 10% (vol./vol.) normal swine serum and incubated overnight with 1:1000 dilution of rabbit anti-podocin (Sigma) in 1% (wt/vol.) BSA at 4°C. The next day sections were incubated with biotinylated swine anti-rabbit IgG (1:500; Dako, Glostrup, Denmark) for 1 h and then with streptavidin-FITC (1:100; Zymed, San Francisco, CA, USA) for 30 min. After immunofluorescence labelling, sections were mounted in anti-fade media containing 5% (wt/vol.) 1,4-diazabicyclo [2,2,2,] octane (Sigma) and examined with a fluorescence microscope (Leica, Wetzlar, Germany). The glomerular area expressing podocin was assessed in 20 glomeruli/section using image analysis.

Quantification of immunohistochemistry

Immunostained cells were counted as cells/gcs or cells/mm2 in the interstitium of the renal cortex in each animal. WT1+ podocytes were counted in 25 hilar gcs. CD4+, CD8+, CD68+, CD169+ and Ki-67+ cells were counted in 25 hilar gcs and 50 cortical fields (magnification ×400). Cleaved caspase-3+ apoptotic cells, FOXP3+ regulatory T cells and CD22+ B cells were counted in 50 gcs and the entire kidney cortex. α-SMA+myofibroblast accumulation was assessed by computer image analysis of the percentage area stained in ten cortical fields (magnification ×250). Glomerular collagen IV accumulation was assessed as the percentage of area stained in 25 hilar gcs, while tubulointerstitial collagen IV was assessed by measuring the area stained in ten cortical fields (magnification ×250), excluding glomeruli and blood vessels.

Real-time PCR

Total RNA was extracted from whole-kidney samples using a reagent (RiboPure; Ambion, Austin, TX, USA) and reverse-transcribed using a kit (Superscript First-Strand Synthesis; Invitrogen, Carlsbad, CA, USA) with random primers. Real-time PCR was performed (Rotor-Gene 3000 system; Corbett Research, Sydney, NSW, Australia) with thermal cycling conditions of 37°C for 10 min and 95°C for 5 min, followed by 50 cycles at 95°C for 15 s, 60°C for 20 s and 68°C for 20 s. The primer pairs and FAM-labelled MGB probes used are shown in the Electronic supplementary material (ESM) Table 1. The relative amount of mRNA was calculated using comparative Ct (ΔΔCt) method. All specific amplicons were normalised against 18S rRNA, which was amplified in the same reaction as an internal control using commercial assay reagents (Applied Biosystems, Scoresby, VIC, Australia). Each of the primer/probe sets was pre-tested and determined to have equivalent PCR amplification efficiencies.

Statistical analysis

Statistical differences between two groups were analysed by the unpaired Student’s t test. Those between multiple groups were analysed by one-way ANOVA with Tukey’s multiple comparison post-test. Correlations were performed using Pearson’s correlation coefficient. Data were recorded as mean ± SD, with p < 0.05 considered to denote significance. All analyses were performed using GraphPad Prism 5.0 (GraphPad, San Diego, CA, USA).

Results

Development of diabetes in Rag1−/− mice

Preliminary studies (data not shown) led to our development of a streptozotocin-model which induced equivalent diabetes in wild-type and Rag1 −/− mice. Using this protocol, we found that both groups of diabetic mice showed a similar profile in body weight and progression of hyperglycaemia over the 20 week examination period (Fig. 1a, b). Both diabetic groups also demonstrated an equivalent increase in HbA1c levels at weeks 8 and 20 (Fig. 1c).

Fig. 1
figure 1

Lymphocytes did not affect the development of streptozotocin-induced diabetes in the experimental mice. Following treatment with streptozotocin, Rag1 −/− and wild-type C57BL/6 mice became diabetic and demonstrated similar profiles of (a) body weight and (b) blood glucose over a 20-week period (black circles, non-diabetic wild-type mice; white circles, non-diabetic Rag1 −/− mice; black squares, diabetic wild-type mice; white squares, diabetic Rag1 −/− mice). c Assessment of HbA1c confirmed that diabetic Rag1 −/− and wild-type mice had equivalent hyperglycaemia at weeks 8 and 20 (white bars, non-diabetic; black bars, diabetic). Data are means ± SD, n = 7–10; ***p < 0.001 vs non-diabetic control

Full size image

Lymphocyte accumulation in diabetic kidneys

Immunostaining identified a two- to threefold increase in glomerular and interstitial CD4+ and CD8+ T cells in diabetic wild-type compared with non-diabetic wild-type kidneys (Fig. 2a, b, d, e). There was also a fivefold increase in interstitial FOXP3+ regulatory T cells in diabetic compared with non-diabetic wild-type mice (Fig. 2f), which accounted for approximately 15% of the CD4+ infiltrate; however, no FOXP3+ cells were detected in wild-type glomeruli. In comparison, CD22+ B cells were rarely detected by immunostaining, only accounting for about 1% of total lymphocytes within the wild-type kidney. A small but significant increase in glomerular B cells was observed in diabetic compared with non-diabetic wild-type mice (6.1 ± 2.5 vs 2.8 ± 2.1 cells/30 gcs, p = 0.03). However, interstitial B cells were not significantly increased in diabetic wild-type kidneys (0.46 ± 0.26 vs 0.26 ± 0.12 cells/mm2, p = 0.21). As expected, no T- or B-lymphocytes were identified in the kidneys of Rag1 −/− mice (Fig. 2c, B cell data not shown).

Fig. 2
figure 2

T cell accumulation in the kidney of diabetic mice. Representative immunostaining of lymphocytes demonstrated a few CD8+ T cells in the cortex of (a) a normal mouse kidney, but these numbers were increased approximately threefold in (b) the kidney cortex of a wild-type diabetic mouse. As expected, no CD8+ lymphocytes were detected in (c) the diabetic kidney of a Rag1 −/− mouse. Magnification (a–c) ×100. The total numbers of CD4+ and CD8+ in (d) glomeruli and (e) the interstitium of the kidney cortex were assessed at week 20 in diabetic (black bars) wild-type mice and their non-diabetic (white bars) controls. Immunostaining also revealed a marked increase in (f) interstitial FOXP3+ T cells in the kidney cortex of diabetic compared with non-diabetic wild-type mice. CD4+ and FOXP3+ T cells were absent in Rag1 −/− mice (not shown). Data (d–f) are means ± SD, n = 7–10; *p < 0.05, **p < 0.01, ***p < 0.001 vs non-diabetic control

Full size image

Lymphocyte deficiency reduces albuminuria and prevents some glomerular changes in diabetic kidneys Rag1 −/− mice had a higher urine AER than non-diabetic wild-type controls, although this just failed to reach statistical significance (52.9 ± 7.0 vs 30.7 ± 5.7 µg/24 h, p = 0.05). As we found no differences in any of the glomerular and tubular variables examined in the non-diabetic strains, the cause of this divergence is unclear. To compensate for this baseline difference, we compared urine AER in diabetic mice in terms of the fold-increase in each strain relative to their non-diabetic levels (i.e. baseline = 1). Accordingly, diabetic wild-type mice showed a progressive rise in urine AER after week 8, which was increased more than sixfold at week 20 compared with non-diabetic wild-type mice (Fig. 3). In contrast, urine AER in RAG1 −/− diabetic mice was increased twofold at 8 weeks, but did not progress between weeks 8 and 20 (Fig. 3). The protection against increasing albuminuria seen in Rag1 −/− mice was also validated by measuring the urine albumin/creatinine ratio at week 20 (Table 1). Examination of WT1+ immunostaining demonstrated that podocyte numbers and glomerular podocin levels were reduced in diabetic wild-type kidneys at week 20; however, diabetic Rag1 −/− mice were partially protected from loss of podocytes and fully protected from loss of podocin production (Table 1). In this model, podocyte numbers showed a strong correlation with urine AER (r = −0.72, p < 0.001).

Fig. 3
figure 3

Lymphocyte deficiency suppresses albuminuria in diabetic mice. Diabetic mice showed an early twofold increase in urine AER at 8 weeks compared with control mice. Despite a progressive increase in urine AER in diabetic wild-type mice from week 8 to week 20, urine AER in diabetic Rag1 −/− mice remained static, indicating that lymphocyte deficiency abrogated progressive albuminuria in diabetic nephropathy. Data are means ± SD, n = 7–10; **p < 0.01 and ***p < 0.001 vs non-diabetic control; ††† p < 0.001 vs diabetic wild-type. Black circles, non-diabetic wild-type mice; white circles, non-diabetic Rag1 −/− mice; black squares, diabetic wild-type mice; white squares, diabetic Rag1 −/− mice

Full size image
Table 1 Assessment of glomerular injury
Full size table

Immunofluorescence analysis identified moderate IgG deposition in the glomeruli of diabetic wild-type mice, which was completely absent in Rag1 −/− mice (Fig. 4). The numbers of total glomerular CD68+ macrophages were increased two- to threefold in diabetic compared with non-diabetic wild-type mice, a development not significantly altered by Rag1 knockout (Table 1). The number of CD169+-activated macrophages was increased three- to fourfold in the glomeruli of diabetic compared with non-diabetic wild-type mice; however, these numbers were markedly reduced in Rag1 −/− mice (Table 1). The degree of urine AER in diabetic mice also correlated with glomerular macrophage activation (r = 0.50, p = 0.04), but there were insufficient data pairs to determine correlations with lymphocyte counts. Despite these positive findings, lymphocyte deficiency had no effect on several other histological changes seen in diabetic glomeruli, including increased glomerular cellularity, hypertrophy, proliferation, apoptosis and matrix deposition (Table 1).

Fig. 4
figure 4

Lymphocyte deficiency prevents glomerular IgG deposition in diabetic mice. a A faint fluorescence signal for IgG was detected in the glomeruli of non-diabetic control mice. b The glomeruli of diabetic wild-type mice had moderate fluorescence signal intensity, indicating increased deposition of IgG. c A complete absence of fluorescence signal was noted in Rag1 −/− kidneys, consistent with absence of IgG within glomeruli. Magnification ×100

Full size image

Lymphocyte deficiency does not protect diabetic kidneys from hypertrophy, tubulointerstitial injury or renal dysfunction

Diabetic wild-type and Rag1 −/− mice both exhibited a 40% increase in the kidney/body weight ratio compared with non-diabetic mice, indicating similar renal hypertrophy (Table 2). Diabetic wild-type mice also demonstrated increased tubular atrophy, apoptosis and proliferation, which were unaffected by Rag1 knockout (Fig. 5a–c, Table 2). Equal levels of interstitial fibrosis were observed in diabetic wild-type and Rag1 −/− mice, including increased myofibroblast accumulation (α-SMA+cells) and tubulointerstitial deposition of collagen IV (Fig. 5d–f, Table 2). In addition, significant tubulointerstitial damage was indicated by increased gene expression of Kim-1(also known as Havcr1), Fn1 (also known fibronectin1) and Pai-1 (also known as Serpine1) in diabetic compared with non-diabetic wild-type mice, a development unchanged by Rag1 knockout (Fig. 5g–i). The development of histological damage in diabetic wild-type kidneys at week 20 was associated with a 35% decline in creatinine clearance. A similar value was observed in diabetic Rag1 −/− mice, demonstrating equivalent renal dysfunction (Fig. 6).

Table 2 Assessment of tubulointerstitial injury
Full size table
Fig. 5
figure 5

Lymphocytes do not promote the progression of tubulointerstitial injury in diabetic mice. Compared with control mice (a, d), the kidneys of diabetic wild-type (b, e) and Rag1 −/− (c, f) mice demonstrated significant tubular atrophy in PAS staining (a–c) and interstitial fibrosis (collagen IV) (d–f), which were unaffected by lymphocyte deficiency; magnification ×250. When compared with non-diabetic mice (white bars), diabetic mice (black bars) showed a (g) 15-fold increase in Kim-1 mRNA, (h) threefold increase in Fn1 mRNA and (i) fivefold increase in Pai-1 mRNA. The gene transcript levels of these markers of tubulointerstitial injury were not reduced in lymphocyte-deficient Rag1 −/− mice. Data are means ± SD, n = 6–10; ***p < 0.001

Full size image
Fig. 6
figure 6

Lymphocyte deficiency does not protect diabetic mice from renal dysfunction. At 20 weeks, both wild-type and Rag1 −/− diabetic mice (black bars) had suffered a 35% decline in creatinine clearance (CrCl) compared with non-diabetic controls (white bars). Data are means ± SD, n = 7–10; ***p < 0.001

Full size image

Lymphocyte deficiency does not reduce interstitial inflammation in diabetic kidneys

After 20 weeks of diabetes, the total number of interstitial macrophages (CD68+ cells) was increased three- to fourfold and interstitial accumulation of the CD169+-activated macrophage subset was increased 17-fold in diabetic compared with non-diabetic wild-type kidneys. Equivalent increases were also found in diabetic Rag1 −/− mice (Fig. 7a–c, Table 2). Urine analysis revealed that excretion of IL-6 and MCP-1 were increased about threefold in diabetic compared with non-diabetic wild-type mice and were unchanged by Rag1 knockout (Fig. 7d, e). Similarly, expression of Ccl2 was increased tenfold in diabetic compared with non-diabetic wild-type kidneys and was unaffected by the absence of lymphocytes (Fig. 7f). In contrast, kidney gene expression of the T cell-related cytokines Il12 (also known as Il12b), Ifn-γ (also known as Ifng) and Il10 was not elevated in diabetic compared with non-diabetic wild-type mice and did not change with Rag1 knockout (data not shown). In addition, serum levels of IL-6 were also unchanged in diabetic and non-diabetic mouse strains (data not shown).

Fig. 7
figure 7

Lymphocyte deficiency does not reduce tubulointerstitial inflammation in diabetic mice. Macrophages (CD68+ cells, brown) were rarely observed in the cortex of (a) a non-diabetic wild-type mouse kidney, but were increased to a similar degree in (b) a diabetic wild-type mouse kidney and (c) a diabetic Rag1 −/− mouse kidney. Sections were counterstained with haematoxylin, magnification ×400. d Urinary excretion of the inflammatory cytokine, IL-6 was increased threefold in diabetic mice (black bars) compared with non-diabetic (white bars) mice. e Urinary excretion of MCP-1 was increased in diabetic mice, corresponding to (f) upregulated transcription of Ccl2 in the kidney. The increased urinary excretion of IL-6 and MCP-1, and increased kidney expression of Ccl2 were not dependent on lymphocytes. Data (d–f) are means ± SD, n = 7–10; *p < 0.05 and ***p < 0.001 vs non-diabetic control

Full size image

Discussion

Our investigation showed that the development of early diabetic renal injury is associated with significant lymphocyte infiltration. We identified an increase in glomerular and interstitial CD4+ and CD8+ T cells in our mouse model of type 1 diabetes, which is consistent with other rodent models of diabetes [10–12]. We also found that FOXP3+ T-regulatory cells accumulate in the interstitium of diabetic kidneys. In comparison, B cells were rarely detected in any part of the diabetic kidney, but IgG was deposited in the glomeruli of diabetic mice. A genetic deficiency (knockout) of Rag1 resulted in mice (and kidneys) devoid of these lymphocytes, but did not alter the development of diabetes in our model, which enabled us to determine the impact of the total lymphocyte infiltrate on diabetic renal injury. However, this study did not examine whether particular subsets of lymphocytes, whose activity may be regulated by interactions with other types of lymphocytes, played discrete roles. In addition, we should also point out that this study did not use the common multiple low-dose model of streptozocin induction of diabetes [21], because Rag1 −/− mice were resistant to the pancreatic injury induced by this protocol. Instead, we used a previously described two-dose model of streptozotocin-induced diabetes, which does not cause acute kidney damage [22].

Lymphocyte deficiency had a selective effect on glomerular inflammation. Kidney macrophage numbers, which were approximately tenfold greater than lymphocytes in diabetic wild-type mice, were unchanged in Rag1 −/− diabetic mice. In addition, expression of Ccl2 and urine excretion of MCP-1 were equally increased in wild-type and Rag1 −/− diabetic kidneys, indicating that the major mechanism for kidney macrophage recruitment had not been affected by the absence of lymphocytes. Urine levels of IL-6 were also increased by similar amounts in both diabetic strains, suggesting that the kidney inflammatory response was comparable. Surprisingly, Rag1 knockout reduced the number of CD169+-activated macrophages to near normal levels in diabetic glomeruli, but not in the interstitium. This suggests that lymphocytes help promote glomerular macrophage activation during diabetes. However, our examination of wild-type and Rag1 −/− diabetic kidneys found no changes in expression of Ifn-γ, which is known to promote macrophage activity, suggesting that T cells may have little involvement in regulating macrophage activity in diabetic kidneys.

Analysis of diabetic Rag1 −/− mice showed that lymphocytes promoted the progression of albuminuria between weeks 8 and 20 of diabetes. The loss of podocytes and glomerular podocin production at week 20 correlated with the development of albuminuria and was significantly inhibited in the absence of lymphocytes, suggesting that lymphocytes play a role in podocyte injury that leads to greater urinary albumin excretion during diabetes. The mechanisms by which lymphocytes induce podocyte injury during diabetes are unclear, but may involve activation of glomerular macrophages, which was markedly reduced in diabetic Rag1 −/− mice. T cells found in glomeruli (CD4+, CD8+) may stimulate neighbouring macrophages to become activated by increasing the local levels of activating cytokines. T cell secretion of proinflammatory cytokines can activate glomerular macrophages directly and also indirectly by stimulating mesangial cell production of colony stimulating factor-1 and MCP-1 [23]. Another possibility is that glomerular macrophages are stimulated by the glomerular deposition of IgG, which is increased during diabetes and was absent in Rag1 −/− mice. Macrophages possess receptors for the Fc component of IgG and for C3, which can both activate macrophages [24, 25]. Increased levels of circulating antibodies to proteins modified by diabetes may form immune complexes, which deposit in glomeruli and potentially activate complement. Circulating immune complex levels have been associated with the development of albuminuria and proliferative retinopathy in diabetic patients suggesting that they could promote tissue injury [6, 7, 26]. However, the presence of glomerular Ig and C3 in patient biopsies cannot be clearly attributed to diabetes and is more commonly considered to be a form of immune complex disease co-existing with diabetic nephropathy [27, 28]. More convincing evidence is seen in experimental animal models, where glomerular deposition of Ig and C3 is due to diabetes alone and correlates with renal injury [11, 13–15]. Our findings in Rag1 −/− mice support this concept by demonstrating that glomerular Ig deposition is a complication of diabetes that may play a role in the evolution of diabetic albuminuria.

Another consideration is that lymphocytes may also promote diabetic albuminuria by enhancing tubular hypertrophy and injury. Tubular hypertrophy is known to stimulate glomerular blood flow through a tubular glomerular feedback mechanism [29]. Proximal tubules can also endocytose and degrade some of the albumin filtered by glomeruli, a process possibly reduced by tubular injury [30]. However, in the current study, we found that diabetic Rag1 −/− were not protected from renal hypertrophy, tubular damage (atrophy, apoptosis, proliferation) or expression of Kim-1, a sensitive marker of tubular injury. These findings suggest that lymphocytes are not essential for inducing the tubular changes seen in early diabetic renal injury, thus arguing against lymphocyte interactions with tubules as a mechanism that promotes albuminuria in diabetes.

Rag1 knockout also had no impact on the development of glomerular and interstitial matrix accumulation in diabetic mice. This observation, in conjunction with a lack of lymphocyte effect on tubular injury, suggests that lymphocytes have no significant role in the progression of tubulointerstitial damage in diabetic kidneys. The development of tubulointerstitial fibrosis is considered to be a strong predictor of the decline in renal function in diabetic nephropathy [31]. Therefore, it was not surprising that both wild-type and Rag1 −/− diabetic mice showed similar increases in interstitial fibrotic responses (myofibroblast accumulation, collagen IV deposition and gene expression of Fn1 and Pai-1) and a similar decline in creatinine clearance.

In conclusion, our study has shown that lymphocytes play a role in the progression of albuminuria, but do not promote renal hypertrophy, increased matrix deposition, tubulointerstitial damage or the decline in renal function in a mouse model of early type 1 diabetic renal injury. Previous evidence has demonstrated that inflammation is critical for the development of diabetic nephropathy. However, our current study suggests that this inflammatory process is primarily due to an innate immune response, which is induced by diabetes and dominated by macrophage-mediated effects. In comparison, adaptive immunity appears to play a more limited role, which includes promotion of glomerular macrophage activation, podocyte injury and albuminuria, in the early development of diabetic renal injury. These results do not preclude a possible role for lymphocytes in the later development of diabetic nephropathy. We expect these findings to be helpful in the development of novel therapies aimed at specifically targeting diabetic renal inflammation.