Introduction
Chronic kidney disease (CKD) is a major health problem worldwide [
1]. CKD is a risk factor for cardiovascular diseases and mortality. Diabetic nephropathy is the leading cause of CKD in developed countries [
2]. Diabetic nephropathy is characterised by renal hypertrophy, glomerulosclerosis, arteriolar hyalinosis and tubulo-interstitial fibrosis [
3]. As in other diseases evolving to renal scarring, growth factors such as TGF-β1 and connective tissue growth factor (CTGF) are involved in the progression of renal damage [
4]. There are very few therapies that induce renal repair in chronic nephropathies [
5]. Hepatocyte growth factor (HGF) is a mesenchyme-derived cytokine with anti-fibrotic and regenerative properties in some experimental models of chronic renal damage [
6,
7]. HGF inhibition exacerbates renal fibrosis, whereas HGF supplementation reverses progression [
8‐
11]. We have previously studied the effect of
HGF gene therapy on diabetic kidney disease in rats and have demonstrated HGF-supplementation-induced regression of glomerular sclerosis [
12]. The reparative role of HGF has been robustly recognised in many studies [
11]. The majority of HGF-established reparative mechanisms are directly related to a local effect of TGF-β1 inhibition [
13], by increasing extracellular matrix degradation [
14] or by inhibiting epithelial to mesenchymal transition [
15]. However, there is a lack of information about the potential role of bone-marrow-derived cells on the regression of CKD induced by HGF. Kollet et al [
16] demonstrated that, after liver injury, HGF facilitated bone marrow haematopoietic stem cell (HSC) motility,
CXCR4 expression and stromal-cell-derived factor-1 (SDF-1)-mediated directional migration of these cells to the damaged liver. More recently, Higashiyama et al [
17] studied spontaneous liver regeneration after cirrhosis in wild-type and transgenic animals overexpressing
Hgf. Animals overexpressing
Hgf healed more rapidly because resident as well as bone-marrow-derived cells infiltrating the liver were producing some metalloproteases that mainly contributed to regression of liver fibrosis.
In order to study whether bone-marrow-derived cells are involved in the reparative mechanisms of HGF on diabetic kidney disease, we have created chimeric diabetic db/db mice. We performed bone marrow transplantation (BMT) using transgenic C57BL/6J mice producing the enhanced green fluorescent protein (EGFP) as donors to db/db mouse recipients.
Methods
Animals
The experiments complied with current legislation on animal experiments in the European Union, and the principles of laboratory animal care and were approved by our institution’s Ethics Committee for Animal Research. Female C57BLKS mice (db/db), 8 weeks old, were purchased from Janvier (Laval, France). Transgenic C57BL/6J mice producing EGFP were kindly provided by J. Barquinero (Unitat de Diagnòstic i Teràpia Molecular, Centre de Transfusió i Banc de Teixits, Barcelona, Spain).
Bone marrow transplantation
Study groups
Chimeric animals were divided into four treatment groups and followed for 4 weeks: (1) db/db–BMT (n = 11), diabetic animals with BMT; (2) db/db+HGF (n = 12), diabetic animals with BMT and treated by HGF; (3) db/db+granulocyte-colony stimulating factor (G-CSF) (n = 10), diabetic animals with BMT and treated by G-CSF and (4) db/db+HGF+G-CSF (n = 11), diabetic animals with BMT and treated by HGF and G-CSF. We used db
/− (n = 10), non-diabetic animals, and db/db (n = 10), diabetic animals, as age-matched control groups.
Therapeutic interventions
Monitoring
Animals were followed from 8 to 33 weeks of age. During this period glucose levels and body weight were measured weekly (glucose was measured using the Glucocard G+meter GT-1820 [Menarini, Barcelona, Spain]). BMT was performed when animals were 24 weeks old. Chimerism was analysed 5 weeks after BMT, and the animals were killed in week 33. Urine and blood samples were collected in order to analyse urinary albumin and serum creatinine concentrations at three time points: before BMT (24 weeks of age), when chimerism was analysed (29 weeks of age) and before the mice were killed (33 weeks of age).
Mice were placed in metabolism cages in order to collect 24 h urine specimens before BMT (week 24), and thereafter before and after therapeutic interventions (at weeks 29 and 33, respectively). Blood was obtained from the tail vein. Serum creatinine and urine creatinine were determined by an autoanalyser (Olympus AU400; Olympus, Hamburg, Germany). Urine albumin excretion was determined using a specific commercially available ELISA kit (Albumin Blue Fluorescent Assay Kit; Active Motif, La Hulpe, Brussels, Belgium).
Circulating haematopoietic stem cells
Analysis of HSCs was performed using a BD FACSCanto II (BD Biosciences, San Jose, CA, USA). Peripheral blood was collected at baseline and then 2 and 5 days after each
HGF administration. Whole blood was incubated with allophycocyanin-conjugated anti-lineage (Lin) cocktail, phycoerythrin (PE)-Cy7-conjugated anti-cKit (also known as CD117) and PE-conjugated anti-stem cell antigen (Sca)-1 antibodies (BD Biosciences, Aalst, Belgium). Blood was then incubated with lysis buffer in order to discard erythrocytes. PBMCs were finally incubated with 7-amino-actinomycin D for dead cell staining. PBMCs were considered HSCs when they were Lin
− cKit
+ Sca-1
+ as previously described [
20].
Optical microscopy, immunohistochemical, immunofluorescence and confocal studies
Renal slices were fixed in 10% (vol./vol.) formalin and embedded in paraffin. Histological cross sections of 3 μm thickness were stained with haematoxylin and eosin, periodic acid–Schiff and Masson’s trichrome for optical microscopy assessment. Fibronectin (1:500) and SDF-1 (1:25) (Abcam, Madrid, Spain) were stained using the immunohistochemical technique described previously [
21]. Anti-collagen IV (1:100; Millipore, Livingston, UK), anti-green fluorescent protein (GFP; 1:500, Abcam), anti-α-smooth muscle actin (SMA) (1:100, Sigma-Aldrich, Madrid, Spain), anti-claudin-I (1:19, Invitrogen, Paisley, UK), anti-WT-1 (1:30, Santa Cruz Biotechnology, Heidelberg, Germany), anti-F4/80 (1:50, LabClinics, Barcelona, Spain), anti-galectin-3 (1:100, Abcam) and anti-mannose-receptor (1:50, Abcam) antibodies were analysed using immunofluorescence. Slides were stained with: collagen IV and fibronectin to assess glomerulosclerosis; α-SMA to stain mesangial cells; claudin-I for epithelial cells from the Bowman’s capsule; WT-1 for podocytes; F4/80 for macrophages; CD90, CD73, CD105 for mesenchymal cells; CD34 for endothelial cells; and galectin-3 and mannose receptor for the M2 macrophage subpopulation.
Double immunolabelling was performed using anti-GFP with anti-α-SMA, anti-claudin-I, anti-WT-1 or anti-F4/80 antibodies; and anti-F4/80 with anti-galectin-3 and anti-mannose receptor. Optical microscopy, fibronectin, SDF-1 and collagen IV were evaluated (from 0 to +3) by a pathologist (A. Vidal) who was blinded to group assignment. Macrophages and the M2 subpopulation were evaluated in ten glomeruli for every sample in a blinded manner. Macrophage counts were expressed as macrophage/glomeruli. The percentage of M2+ with respect to F4/80+ macrophages was calculated (n = 4 per group).
Additional renal slices were embedded and frozen in optimal cutting temperature compound. Subsequently, cross sections of 5 μm thickness were fixed with 4% (wt/vol.) paraformaldehyde at 4°C for 20 min and incubated with the following primary antibodies: anti-CD90; anti-CD73; anti-CD105; and anti-CD34 (1:50, BD Pharmingen, Aalst, Belgium). Also, the amount of EGFP in renal frozen tissue was directly viewed with microscopy and automatically quantified using Leica Confocal Software. Finally, confocal studies (using a Leica TCS-SL confocal espectral microscope [Mannheim, Germany]) were performed in a blinded manner on paraffin-embedded and frozen renal tissues. Alexa Fluor647 (far red), Alexa Fluor488 (green), Alexa Fluor546 (red) and Alexa Fluor555 (red) were used as secondary antibodies (Molecular Probes, Invitrogen, Madrid, Spain). In a blinded manner, podocytes per glomerular tuft were quantified in ten glomeruli per sample and calculated as podocyte/total nuclei ratio. Nuclei were stained blue with DRAQ5 and counted using Fiji Is Just ImageJ software (
http://fiji.sc).
Gene expression analysis
RNA was extracted from kidney with PureLink RNA Mini Kit (Invitrogen, Madrid, Spain), using a Trizol reagent (Invitrogen, Madrid, Spain) to lyse the tissues, chloroform to separate the aqueous phase and organic phase (where the RNA remains), and ethanol to purify the RNA. RNA purity was analysed on a NanoDrop (NanoDrop ND-1000V3.3, Wilmington, DE, USA) and was considered pure when the absorbance ratio 260/280 nm was lower than 1.75. A total amount of 1,000 ng RNA was used to perform the reverse transcription using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Warrington, UK). Thermal cycling conditions were 10 min at 25°C, 120 min at 37°C, 5 min at 85°C and finally held at 4°C. Tissue expression levels of HGF, TGFβ (also known as TGFB1) and CTGF were quantified by TaqMan real-time PCR (ABI Prism 7700, Applied Biosystems) using the comparative Ct method (Applied Biosystems).
Cytokine analysis
Serum cytokines were quantitatively measured by FACSCanto with two different cytometric bead array (CBA) kits: the CBA mouse inflammation kit (IL-6, IL-10, MCP-1, IFN-γ, TNF and IL-12p70) and the CBA mouse Th1/Th2 cytokine kit (IL-4, IL-5, IFN-γ and TNF), both from BD Biosciences (San Jose, CA, USA). Data were acquired and analysed using BD CBA software (San Jose, CA, USA). If a sample had a cytokine concentration below the detection limit of the assay, a value of 0 was assigned for that particular cytokine concentration.
Statistical analyses
All data are presented as mean±SE. Group means were compared with either the Student’s t test or ANOVA for parametric values, or the Mann–Whitney U test or Kruskal–Wallis test for non-parametric values. All p values were two-tailed, and a p value of less than 0.05 was considered statistically significant. All statistical analyses were analysed using StatView software.
Discussion
In the present study we show that HGF gene therapy enhances renal expression of SDF-1 and is also associated with an increasing number of bone-marrow-derived cells getting into the injured diabetic kidney. These cells are mainly macrophages, which may fuse with resident epithelial cells from the Bowman’s capsule and participate in renal repair and regeneration.
We have created chimeric diabetic db/db mice by BMT with bone marrow from donor transgenic EGFP+ mice. Thus, we obtained a diabetic mouse model with easily identifiable bone-marrow-derived cells in blood and peripheral tissues. The wild-type C57BL/6J mice express leptin receptor and may have distinct properties from their db/db counterparts. Despite bone marrow EGFP+ cells came from the wild type, we proved BMT was not associated with any significant modification of diabetic nephropathy, as there were no differences between db/db and db/db–EGFP+ chimeric mice regarding hyperglycaemia, body weight or albuminuria. We were also able to reproduce previous findings from our group concerning the functional and histological benefits of HGF gene therapy on diabetic nephropathy in this model. Thereafter, in this work we have investigated whether some bone-marrow-derived cells are in some way involved in HGF-mediated renal repair.
We found that both HGF and G-CSF increased the presence of bone-marrow-derived cells in the kidney, especially surrounding the glomeruli, by different and unconnected mechanisms. We observed that G-CSF consistently induced HSC mobilisation in peripheral blood, whereas
HGF had no effect on HSC mobilisation. Nevertheless,
HGF gene therapy enhanced SDF-1 production in renal tissue, which is a major ligand for the C-X-C chemokine receptor 4 (CXCR4) molecule expressed on HSCs [
23]. Enhanced SDF-1 in renal tissue has been reported to induce homing of CXCR4
+ cells to the kidney after ischaemic injury [
24]. Accordingly,
HGF enhanced renal SDF-1 and led to an increased number of EGFP
+ cells getting into the kidney, despite the lack of induced HSC mobilisation. The combination of
HGF gene therapy with HSC mobilisation by G-CSF increased the amount of bone-marrow-derived EGFP
+ cells in the diabetic kidney since G-CSF induced their mobilisation and HGF induced their recruitment. The cells located around the glomeruli were mainly macrophages. In kidney disease, it has been suggested that monocyte-derived macrophages infiltrating the kidney cause injury and fibrosis in renal tissue [
25], but can also facilitate kidney repair and regeneration, depending on the renal milieu [
26]. Despite finding similar amounts of macrophages in diabetic kidneys from mice treated with
HGF or G-GSF, we observed consistent histological benefit only in those receiving
HGF gene therapy. The concomitant administration of G-CSF in
HGF-treated animals did not provide any additional histological improvement. We have previously described how HGF inhibits pro-inflammatory cytokine expression via inhibition of nuclear factor κB signalling [
26]. In agreement, we found
HGF gene therapy reduced pro-inflammatory cytokines and was associated with a relatively high proportion of reparative M2 macrophages in glomeruli. Therefore, our findings suggest
HGF gene therapy may induce a renal microenvironment that might promote macrophage-mediated renal tissue repair and regeneration.
We tested whether there were any differentiated glomerular cells co-expressing EGFP. There was no co-expression regarding mesangial, endothelial and podocyte markers. Nevertheless, we found a small number of Bowman’s capsule PECs that were also EGFP
+. Interestingly, this only happened in mice that received
HGF gene therapy. These highly specialised cells, which cover the glomerular capillary tuft, are crucial for podocyte repair after injury [
27]. Accordingly, we found the percentage of podocytes was well preserved in
HGF-treated compared with non-treated diabetic animals. It has been suggested that some of these PECs are actually renal stem cells that could migrate along the Bowman’s capsule and transition to the tuft to areas of injury and differentiate into podocytes [
28]. Sagrinati et al [
28] demonstrated that PECs derive from a resident renal population during kidney development rather than from the bone marrow. Our data suggest these PECs expressing EGFP derive from cell fusion. First, we found extensive macrophage infiltrates around the capsule. Second, macrophages have the ability to undergo cell–cell fusion with themselves or other cell types as renal resident cells, particularly in response to inflammatory stimuli [
29]. Therefore, our data suggest HGF attracts bone-marrow-derived cells around the renal capsule, and these can fuse with PECs and probably help repair. Altogether, our findings add new knowledge and open new opportunities for inducing renal regeneration in diabetic nephropathy.
Acknowledgments
We thank B. Torrejón (Scientific and Technical Services, UB, Barcelona, Spain) for help with confocal microscopy; M.L. Angelotti (Laboratorio interdipartimentale di Nefrologia Cellulare e Molecolare, Università degli studi di Firenze, Florence, Italy) and C. Varela and M. C. Díaz (Nephrology Laboratory, Departament de Ciències Clíniques, IDIBELL, Barcelona, Spain) for technical help; J. Barquinero (Unitat de Diagnòstic i Teràpia Molecular, Centre de Transfusió i Banc de Teixits, Barcelona, Spain) for kindly providing the EGFP transgenic mice; and C. Gutierrez (Oncology and Radiotherapy, ICO, Barcelona, Spain) for helping us to implement the irradiation procedure.