Introduction

Diabetes and its associated complications, including retinopathy, neuropathy and nephropathy, are an increasing worldwide burden on healthcare systems. Although diabetic nephropathy is the leading cause of end-stage renal disease in developed nations as well as a major risk factor for cardiovascular disease, the pathological mechanisms leading to its development and progression remain to be fully elucidated. It has been well established, however, that both metabolic and haemodynamic factors are involved in the development of diabetic nephropathy in the context of genetic susceptibility [1].

Infiltrating macrophages are a major feature of the kidney inflammatory response in patients with diabetic nephropathy [27], and are thought to become activated through specific receptors. Studies in experimental animal models of diabetes have identified a functional role for both macrophages and the receptor for AGEs (RAGE) in the pathogenesis of diabetic nephropathy [811]. RAGE is expressed on macrophages and other circulating and resident tissue cells, and has a physiological role in antigen presentation, dendritic cell homing to lymph nodes and T cell priming [12]. RAGE signal transduction via interaction with ligands such as S100 calgranulin (S100A)8/9, high mobility box protein B-1 (HMGB-1) and AGEs, is thought to induce a proinflammatory response, resulting in the production of chemoattractant molecules, such as CC chemokine ligand 2 (CCL2), and adhesion molecules [13]. More recently, the interaction of RAGE with ligands S100A8/9 has been demonstrated as critical for the proliferation of bone marrow myeloid progenitor cells [14]. The deletion of RAGE from bone marrow cells through chimerism also protects against the development of atherosclerosis in both the presence [14] and the absence [15] of experimental diabetes.

Hence, in this present study, we used bone marrow transplantation from either wild-type (WT) control or RG mice to generate chimeras. We then induced diabetes in these chimeric mice to study the contribution of RAGE in bone-marrow-derived cells to the development of experimental diabetic nephropathy.

Methods

Mouse lines

Male mice deficient in Ager (encoding the protein RAGE; RG on C57BL/6J background) and their WT C57BL/6J controls were bred and maintained at the Alfred Medical Research and Education Precinct (AMREP) Animal Services, Melbourne, VIC, Australia. RG mice had been previously backcrossed to C57BL/6J for nine generations after generated by insertion of enhanced green fluorescent protein (eGFP) into intron 3 of Ager [16]. The protocols followed for animal handling and experimentation were in accordance with ethical guidelines of the AMREP Animal Ethics Committee, Monash University Animal Ethics Committee and the National Health and Medical Research Council of Australia.

Generation of chimeric mice

To generate chimeric mice, groups of recipient male WT (CD45.2+) mice (6 weeks of age; n = 8/group) were lethally irradiated with two total body exposures of 5.5 Gy, 3 h apart. Recipient mice were then intravenously injected with 5 × 106 cells from 6-week-old male donor mice within 24 h of irradiation. WT (CD45.2+) recipient mice were transplanted with bone marrow from either RG donors expressing GFP (RG → WT; expressing eGFP) or WT CD45.1 congenic donors (WT → WT; expressing CD45.1). Bone marrow reconstitution occurred over 8 weeks and successful chimerism (94 ± 1%) was established by collecting tail vein blood and performing flow cytometry analysis for CD45.1 and eGFP.

Experimental diabetes induction

Chimeric mice were rendered diabetic 8 weeks after bone marrow reconstitution via intraperitoneal injection of low-dose streptozotocin (55 mg kg−1 day−1; MP Biomedicals, Eschwege, Germany) or vehicle (sodium citrate buffer, pH 4.5) for 5 consecutive days as previously described [9, 10]. Throughout the duration of the study, mice were given standard mouse chow and water ad libitum while kept on a 12 h day/night cycle. After 24 weeks, mice were anaesthetised and exsanguinated by cardiac puncture. Kidneys were removed and were either snap frozen for further analysis or processed for histological analyses.

Measurement of physiological and biochemical variables

During week 23 of the study, mice were individually placed into metabolism cages (Tecniplast, Buguggiate, VA, Italy), for a period of 24 h. A blood sample was taken, body weight, food and water intake monitored and urine collected. Glycated haemoglobin (GHb) was measured in whole blood by HPLC (CLC330 GHb Analyser; Primus, Kansas City, MO, USA) as previously described [9, 10].

Measurement of renal function

Urinary AER (UAER) was determined from 24 h urine collections by ELISA (Bethyl Laboratories, Montgomery, TX, USA) as per the manufacturer’s instructions. Creatinine clearance, corrected for body surface area, was calculated in timed urine and plasma samples following processing with acetyl nitrile and centrifugation at 4°C. The supernatant fractions were then dried in a Speedivac, resuspended in 10 mmol/l ammonium acetate, pH 3.2 and injected into a C18 column (Waters Division, Millipore, Marlborough, MA, USA). HPLC was used for detection of creatinine at 235 nm using a Hewlett Packard photodiode array (PDA) detector [10].

Histological assessment of kidney injury and leucocyte infiltration

Formalin-fixed kidney sections (2 μm) were stained with periodic acid–Schiff’s reagent and haematoxylin for the assessment of glomerulosclerosis by semi-quantitative scoring of 20 glomerular cross sections, and tubulo-interstitial fibrosis quantified by point counting using a 100 point grid [17]. Immunohistochemistry for Wilm’s tumour antigen (WT-1; 1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA) was performed on 4 μm methyl Carnoy’s fixed sections. Frozen sections (5 μm) from kidneys fixed in periodate lysine paraformaldehyde (PLP) were stained for infiltrating CD4+, CD8+, CD19+, CD11c+ or CD68+ (BD Biosciences, Franklin Lakes, NJ, USA) leucocytes as well as CD169 for activated monocytes/macrophages (1:100; Serotec, Oxford, UK). Cells staining for CD45.1 and CD45.2 were also visualised in frozen sections (1:100, Abcam, Cambridge, UK, or eBioscience, San Diego, CA, USA). Collagen IV (1:20; Millipore, CA, USA) and α smooth muscle actin (α-SMA), for myofibroblast accumulation (1:1,000; Millipore), were localised using immunofluorescence, visualised by confocal microscopy and quantified using Image J software (Rockville, MD, USA).

Renal fractionation

In total, 50 mg renal cortex was homogenised (Polytron PT-MR2100, Kinematica, Lucerne, Switzerland) in extraction buffer (20 mmol/l HEPES buffer, pH 7.2, 1 mmol/l EGTA, 210 mmol/l mannitol, 70 mmol/l sucrose) to isolate cytosol and membranous fractions as previously described [18, 19]. Total protein was determined by the bicinchoninic acid method (Pierce, Rockford, IL, USA), according to the manufacturer’s protocol.

Real-time RT-PCR

Total RNA was extracted from kidney cortex using Trizol (Invitrogen) and reverse transcribed with random primers using the Superscript First-Strand Synthesis kit (Invitrogen). Real-time PCR was performed on a StepOne Real-Time PCR System (Applied Biosystems, Scoresby, VIC, Australia) with thermal cycling conditions of 37°C for 10 min, 95°C for 5 min, followed by 50 cycles of 95°C for 15 s, 60°C for 20 s and 68°C for 20 s. The primer pairs and carboxyfluorescein-labelled minor-groove-binding probe used are described in electronic supplementary material (ESM) Table 1. The relative amount of mRNA was calculated using the 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). Each of the primer/probe sets were pre-tested and determined to have equivalent PCR amplification efficiencies and are expressed compared with 18S.

Chemotactic and inflammatory markers

Renal cortical IL-6 (R&D Systems, Minneapolis, MN, USA), CCL2 (RayBiotech, Norcross, GA, USA), macrophage inhibitory factor (MIF; USCN Life Sciences, Hubei, China) and S100A8/9 (Immundiagnostik, Bensheim, Germany) were measured according to the manufacturer’s instructions. Serum S100A8/9 (Immundiagnostik) as well as plasma mouse macrophage-colony stimulating factor (M-CSF; R&D Systems), were assayed by ELISA according to the manufacturer’s instructions. Urinary and plasma AGE-modified (carboxymethyllysine [CML]) albumin were assayed via ELISA as previously described [9].

Harvesting of bone-marrow-derived macrophages

Bone marrow was isolated aseptically from the femurs of 8-week-old male donor WT (C57BL/6J) and RG mice. Isolated bone marrow was washed, counted and stained for the expression of CD3 (T cells), CD11b (macrophages/monocytes), CD19 (B cells) and CD11c (dendritic cells) with pre-conjugated antibodies for flow cytometry (BD Biosciences). A minimum of 10,000 events were acquired on the FACScanto and were analysed by Flowlogic (Inivai, Mentone, VIC, Australia).

For differentiation studies, bone marrow cells were harvested, washed, counted and then resuspended in RPMI containing M-CSF (106 U/ml) and cultured for 5 days. The medium was then collected from control WT- and RG-derived bone-marrow macrophages and expression of F4/80 and CD169 on cells was determined using flow cytometry. Bone-marrow-derived macrophages were then incubated for 10 h in the presence of AGE-BSA (100 μg/ml) or BSA. At the end of 10 h, the conditioned medium from bone-marrow-derived macrophages was placed onto cultured proximal tubules cells (PTCs; HK-2 cells) for 6 h and collagen IV expressed by PTCs was examined by confocal microscopy. Cell culture medium from macrophages exposed to AGE-BSA was also incubated ex vivo with the AGE inhibitor alagebrium chloride for 2 h pre-exposure to PTCs (40 μmol/l) [17].

Statistical analysis

Data are expressed as means ± SD, unless otherwise stated. Analyses of data were performed using Student’s t tests. Data for albuminuria were not normally distributed and were therefore analysed after logarithmic transformation. A p value of less than 0.05 was considered statistically significant.

Results

Physiological and metabolic variables

GHb and fasting blood glucose concentrations in all chimeric mice were within diabetic ranges following injection of streptozocin (Table 1). GHb was further increased by week 24 of diabetes (Table 1) compared with week 12 (Table 1), which was consistent in all chimeric groups. Body weights and kidney weight:body weight ratios were not different between diabetic mouse groups (Table 1), but were elevated by diabetes.

Table 1 Physiological and biochemical variables in non-diabetic control and mouse chimeras at week 3, 12 and 24 of diabetes

Renal function and structure

Polyuria was evident by week 24 of diabetes in WT → WT mice compared with RG → WT and RG and WT control mice (Table 1). Consistent with this, diabetic WT → WT mice had increased glomerular filtration rates (creatinine clearance [CrCl]; Fig. 1a) and increased urinary albumin excretion over 24 h at week 12 (Table 1) and week 24 compared with control mice (Fig. 1b); the increases were attenuated in diabetic RG → WT mice (Table 1; Fig. 1a, b). At week 24, CrCl was decreased in RG compared with WT mice (Fig. 1a). There was also increased expansion of the tubulo-interstitium seen with diabetes in WT → WT mice, which was significantly less in diabetic RG → WT mice (Fig. 1c, g, h). Glomerulosclerosis was not different between diabetic mouse groups but was increased by diabetes (Fig. 1d). The renal cortical expression of mouse Tim-1 gene (also known as Havcr1) encoding kidney injury molecule (KIM-1) was elevated by diabetes in WT → WT mice but to a lesser degree in diabetic RG → WT mice (Fig. 1e). WT-1+ cells within a glomerular cross section were also decreased with diabetes in WT → WT mice, and this was attenuated in diabetic RG → WT mice (Fig. 1f). There was no change in renal cortical activation of the profibrotic cytokine TGF-β in diabetic RG → WT when compared with WT → WT mice (WT → WT, 43.9 ± 27.7 pg/ml; RG → WT, 85.4 ± 83.3 pg/ml vs WT, 30.9 ± 5.7 pg/ml; RG, 26.6 ± 10.1 pg/ml).

Fig. 1
figure 1

Renal functional and structural variables at week 24 of diabetes: (a) CrCl by HPLC, corrected for body surface area; (b) 24 h UAER; (c) tubulo-interstitial area by morphometric quantification as percentage of total cortical area; (d) glomerulosclerosis; (e) renal cortical gene expression of Tim-1 (encoding kidney injury molecule); and (f) WT-1 in podocytes. (g, h) Photomicrographs of periodic acid–Schiff’s reagent staining of renal cortex (×400): (g) diabetic WT → WT; and (h) diabetic RG → WT. Arrows highlight areas of tubulo-interstitial accumulation. White bars, non-diabetic control; grey bars, diabetic WT → WT; black bars, diabetic RG → WT. *p < 0.05 vs WT; p < 0.05 vs WT → WT. GCS, glomerular cross section; GSI, glomerulosclerosis

Collagen IV was increased in the tubulo-interstitium of diabetic WT → WT mice compared with WT and RG mice, but this increase was attenuated in kidney cortices from RG → WT chimeras (Fig. 2a–c). However, there was no significant decrease in glomerular collagen IV in diabetic RG → WT mice (WT → WT, 10.4 ± 2.5%; RG → WT, 8.9 ± 5.7% glomerular tuft area).

Fig. 2
figure 2

Morphometric analyses of confocal images (×630) of immunofluorescence for (ac) tubulo-interstitial collagen IV: (a) quantification as percentage of total cortical area; (b) diabetic WT → WT; (c) diabetic RG → WT. (df) Similar analyses for α-SMA: (d) quantification; (e) diabetic WT → WT; (f) diabetic RG → WT. White bars, non-diabetic control mice; grey bars, diabetic WT recipients reconstituted with WT bone marrow; black bars, diabetic WT mice reconstituted with RG bone marrow. In (a) *p < 0.05 vs WT; p < 0.05 vs WT → WT; in (b), p = 0.065 for WT vs WT → WT, and p = 0.075 for WT → WT vs RG → WT

Cortical α-SMA was not increased with diabetes (p = 0.065, Fig. 2d–f) nor was there a significant difference seen between diabetic groups (p = 0.075).

Circulating and renal markers of inflammation

Renal cortices from diabetic WT → WT mice had a greater number of interstitial CD68+ macrophages compared with non-diabetic and diabetic RG → WT mice (Fig. 3a, h, i). There were also more CD169+ CD68+ macrophages in the tubulo-interstitium with diabetes in WT → WT mice, with lower levels in non-diabetic mice as well as diabetic RG → WT mice (Fig. 3b–d). However, there was a modest increase in glomerular macrophages in mice that received RG bone marrow compared with bone marrow from WT donors (Fig. 3e). There were no significant differences among groups for other immune cell types as assessed by immunohistochemistry (Table 2). Renal cortical gene expression of Cd68 relative to 18S (Fig. 3f) and the M2 marker Fizz-1 (also known as Retnlb) to Cd68 expression (Fig. 3g) were significantly elevated by diabetes in WT → WT but lower in diabetic RG → WT mice. There were no changes seen in the gene expression of other M2 markers (arginase, iNos [also known as Nos2]) between diabetic mice (data not shown).

Fig. 3
figure 3

Macrophage accumulation within renal cortices. Control and diabetic chimeric mice were analysed at week 24. (a) Tubulo-interstitial accumulation of CD68+ cells. (bd) Immunofluorescence for CD169: (b) quantification of tubulo-interstitial CD169+ cells; (c) CD169+ cells in WT → WT; and (d) CD169+ in RG → WT. (e) Glomerular CD68+ cells per glomerular cross section. (f) Renal cortical Cd68 gene expression relative to 18S expression. (g) Renal cortical Fizz-1/Cd68 expression. (hm) Photomicrographs of CD68 immunohistochemistry as well as CD-45.1 and eGFP immunofluorescence in renal cortical sections: (h) CD68 in WT → WT (×200); (i) CD68 in RG → WT (×200). Photomicrographs of infiltrating bone-marrow-derived cells labelled with CD45.1 (red, WT; magenta, CD45.2) at: (j) × 630; (k) × 200; and eGFP fluorescence at (l) × 630; and (m) × 200 (RG, green) within renal cortices at week 24 of diabetes. Arrows indicate bone-marrow-derived cells. White bars, non-diabetic control; grey bars, diabetic WT → WT; black bars, diabetic RG → WT. *p < 0.05 vs WT; p < 0.05 vs WT → WT. GCS, glomerular cross section

Table 2 Morphometric quantification of renal cortical lymphocytes in chimeric mice at week 24 of diabetes

Infiltration of bone-marrow-derived cells into the diabetic kidney of WT → WT mice was tracked using immunofluorescence for CD45.1, which showed significant infiltration of cells into both glomeruli and the tubular compartments (Fig. 3j, k). Given that nuclei from Ager-deficient mice contained an eGFP insert, RG bone-marrow-derived cells were also tracked into the kidneys during diabetes. Within diabetic RG → WT (Fig. 3l, m) eGFP+ cells were also present both within the glomeruli and tubulo-interstitium.

Diabetic WT → WT had higher levels of several chemokines and cytokines within renal cortices compared with control mice. These included CCL2 (Fig. 4a), MIF (Fig. 4b) and IL-6 protein (Fig. 4c), levels of which were lower in RG → WT diabetic mice. Renal cortical concentrations of the RAGE ligands S100A8/9 (Fig. 4d) and AGEs (Fig. 4e) were increased in diabetic WT → WT mice, but attenuated in the diabetic RG → WT group. By contrast, renal cortical concentrations of the RAGE ligand HMGB-1 (Fig. 4f) were lower in all diabetic mice compared with controls. Plasma M-CSF concentrations were similar in both diabetic mouse groups (WT → WT, 614 ± 51 pg/ml; RG → WT, 649 ± 86 pg/ml). Circulating concentrations of HMGB-1 and AGEs were decreased by diabetes and there was no difference in circulating S100 A8/9 between WT → WT and RG → WT mice (data not shown).

Fig. 4
figure 4

Cytokines, chemokines and RAGE ligands. Renal cortices taken from diabetic mice were analysed by ELISA at week 24 for protein expression of: (a) CCL2; (b) MIF; (c) IL-6; (d) S100A8/9; (e) AGEs (CML-albumin); and (f) HMGB-1. Black bars, diabetic WT recipients; grey bars, diabetic RG recipients. *p < 0.05 vs WT; p < 0.05 vs WT → WT. ND, not detected

In bone marrow from WT mice, there was a greater number of F4/80+ macrophages differentiated in the presence of M-CSF (Fig. 5a; 65.5%) when compared with Ager −/− mice (Fig. 5b; 47.2%). Bone marrow derived from Ager −/− mice had fewer CD11b+ cells (Fig. 5c), and increased numbers of CD11c+ cells, but there was no difference between the total cell numbers harvested from the bone marrow (WT 2.34 × 107 cells vs RG 2.28 × 107). A greater number of WT bone-marrow-differentiated macrophages were also F4/80+CD169+ following incubation with AGE-BSA (25.6%) compared with BSA (Fig. 5d, 19.3%). PTCs incubated with conditioned media collected from WT bone-marrow-derived macrophages exposed to AGE-albumin showed an increased expression of collagen IV (Fig. 5e) compared with conditioned media from macrophages exposed to unmodified albumin (Fig. 5f) or to AGE-albumin pre-incubated with alagebrium chloride (Fig. 5g). PTCs incubated with conditioned media collected from Ager −/− bone-marrow-derived macrophages exposed to AGE-albumin (Fig. 5h) also had a reduced expression of collagen IV when compared with conditioned media obtained from AGE-albumin-exposed WT bone-marrow-derived macrophages.

Fig. 5
figure 5

(a, b) F4/80+ bone-marrow-derived macrophages after differentiation for 5 days with M-CSF. (a) WT mice, 65.5% F4/80+; (b) Ager −/−, 47.2% F4/80+. (c) Flow cytometry characterisation of WT and Ager −/− bone marrow at transplantation. Data are expressed as fold change above WT (C57BL/6 J). (d) CD169+F4/80+ bone-marrow-derived macrophages following incubation with AGE-albumin (AGE) or unmodified albumin (BSA; 100 μg/ml) for 10 h. (e–h) Confocal images (×630) of collagen IV immunofluorescence of HK-2 PTCs after 6 h exposure to conditioned media harvested from bone-marrow-derived macrophages from: (e) WT incubated with AGE-modified albumin (100 μg/ml); (f) WT mice incubated with unmodified albumin (100 μg/ml); (g) WT incubated with AGE-modified albumin and AGE-lowering pharmacotherapy alagebrium chloride (40 μmol/l) pre–exposure; (h) Ager −/− mice incubated with AGE-modified albumin. B220, B cells; CD3, T cells; CD11b, monocytes/macrophages; CD11c, dendritic cells. White bars, WT; black bars, RAGE-deficient bone marrow cells. *p < 0.05 vs WT by Student’s t test. Cont, control; max, maximum

Discussion

Within the present study, reconstitution using RG bone marrow (RG → WT) attenuated the development and progression of kidney disease when mice were rendered diabetic. Indeed, decreases in albuminuria were even evident at a relatively early time point (after 12 weeks of diabetes), as well as by the study endpoint at week 24, with concomitant retention of podocytes as well as less tubulo-interstitial injury and macrophage infiltration and activation. This protection was associated with better renal function in RG → WT diabetic mice. We also demonstrated that bone-marrow-derived macrophages from WT mice exposed to AGEs can induce greater collagen IV expression in PTCs in vitro, which was not seen in macrophages from Ager-knockout mice. Taken together, these findings suggest that one of the major defects seen as a result of increases in RAGE expression on infiltrating immune cells into the diabetic kidney is tubulo-interstitial fibrosis, coupled to changes in renal function.

Given that the transplanted bone marrow from RG mice expressed GFP and bone marrow from WT mice expressed CD45.1, we followed reconstitution and renal infiltration during diabetes. The images suggested that bone-marrow-derived cells derived from both donor types had migrated to recipient kidneys. There was no evidence that bone-marrow-derived cells with RAGE deficiency do not undergo monocytosis/myelopoiesis and migrate into the kidney tubulo-interstitium as suggested previously [4, 5, 2022]. There were, however, lower concentrations of the chemotactic molecule CCL2 in kidneys from these mice. Further, a number of infiltrating cells in diabetic mouse renal cortices from mice reconstituted with WT bone marrow were CD68+ monocytes/macrophages or cells that had migrated to glomeruli, where there was no difference in macrophage number between diabetic groups. Indeed, there were no significant changes in other infiltrating immune cell types identified between diabetic mouse groups in the present study. This is consistent with a previous report from our group [23] suggesting that macrophages are the major immune cells contributing to renal functional changes and tubulo-interstitial fibrosis in experimental diabetes. There was evidence of reconstitution of glomerular cells by bone marrow from both donor groups as has been previously shown in diabetes [24], but the relative contribution of these remains to be delineated in future studies.

Global RAGE deficiency has been shown to provide protection against renal injury in experimental models of diabetes [9, 10]. However, these studies were not able to determine if the deletion of RAGE on infiltrating cells per se was responsible for the improved renal function and structure as identified in the present study. In another study, bone-marrow-derived macrophages taken from RG mice produced lower levels of cytokines such as IL-6 [25] on stimulation, which implies that RAGE may be required for macrophage activation and cytokine production. Further, the renal parenchyma from diabetic mice reconstituted with RG bone marrow contained fewer activated macrophages and lower concentrations of various cytokines. This is further supported by another experimental murine study, where deletion of RAGE from bone marrow mediated immune cell infiltration into atherosclerotic lesions in a non-diabetic context [15]. In addition, we have shown that F4/80+ cells derived from RG bone marrow stimulated less collagen IV production from human PTCs following stimulation with the ligand AGE-BSA, which links the improvements seen in the tubulo-interstitium to the presence of RAGE on the interstitial infiltrate. As tubulo-interstitial injury is the most accurate determinant of progressive loss of renal function in chronic kidney disease, including diabetic nephropathy in humans, it is likely that the deletion of RAGE from infiltrating cells contributed directly to the changes in renal function that were observed. It is also plausible that the reduction in RAGE ligand formation that occurred as a result of less renal parenchymal damage in RG → WT diabetic mice also prevented damage to PTCs via RAGE expressed on intrinsic renal cells. Not surprisingly, a number of RAGE ligands are known pathological mediators of proximal tubular damage in diabetic nephropathy [1] and blockade of ligands such as AGEs is traditionally associated with structural and functional improvements in both experimental [17] and human diabetic nephropathy [26].

Overall, this group of studies suggests that changes in RAGE in interstitial infiltrates may be an important modulator of tubulo-interstitial disease and likely podocyte injury in the diabetic kidney. This concept is also supported by a previous study showing that functional blockade of macrophage infiltration in established diabetic nephropathy results in improved renal function, with reductions in interstitial inflammation and fibrosis [27]. Furthermore, deletion of RAGE in bone marrow had significant beneficial effects on the levels of other RAGE ligands, which are pathologically increased in the diabetic kidney, alleviating damage by preventing ligation to RAGE on intrinsic renal cells such as podocytes and PTCs. Hence, targeted dampening of the expression of RAGE on bone-marrow-derived cells appears renoprotective and therefore manipulation of this pathway warrants further investigation in diabetic nephropathy and likely other complications where interstitial infiltration by monocytes and macrophages is a hallmark.