Exploring macrophage cell therapy on Diabetic Kidney Disease

Abstract Alternatively activated macrophages (M2) have regenerative properties and shown promise as cell therapy in chronic kidney disease. However, M2 plasticity is one of the major hurdles to overcome. Our previous studies showed that genetically modified macrophages stabilized by neutrophil gelatinase‐associated lipocalin (NGAL) were able to preserve their M2 phenotype. Nowadays, little is known about M2 macrophage effects in diabetic kidney disease (DKD). The aim of the study was to investigate the therapeutic effect of both bone marrow‐derived M2 (BM‐фM2) and ф‐NGAL macrophages in the db/db mice. Seventeen‐week‐old mice with established DKD were divided into five treatment groups with their controls: D+BM‐фM2; D+ф‐BM; D+ф‐NGAL; D+ф‐RAW; D+SHAM and non‐diabetic (ND) (db/‐ and C57bl/6J) animals. We infused 1 × 106 macrophages twice, at baseline and 2 weeks thereafter. BM‐фM2 did not show any therapeutic effect whereas ф‐NGAL significantly reduced albuminuria and renal fibrosis. The ф‐NGAL therapy increased the anti‐inflammatory IL‐10 and reduced some pro‐inflammatory cytokines, reduced the proportion of M1 glomerular macrophages and podocyte loss and was associated with a significant decrease of renal TGF‐β1. Overall, our study provides evidence that ф‐NGAL macrophage cell therapy has a therapeutic effect on DKD probably by modulation of the renal inflammatory response caused by the diabetic milieu.

Some drug therapies targeting inflammation on DKD have been explored with variable success. 7 On the other hand, the potential role of endogenous bone marrow-derived and stem cells in the repair of kidney damage is still under exhaustive investigation. 8 Therefore, interventions such as cell-based therapy are being broadly developed aimed to prevent chronic disease progression 9 although very few may induce renal repair. 10 Among potential therapeutic strategies, macrophage cell-based therapy provides discordant results. 11,12 Macrophages are associated with tissue damage although they have also a critical role in host defense and tissue repair. 13,14 Alternatively Activated Macrophages (AAM) also called as M2, have anti-inflammatory functions and express high levels of mannose receptor (CD206), arginase and IL-10. 15,16 Lipocalin-2 (Lcn-2), also called Neutrophil Gelatinase-Associated lipocalin (NGAL), is able to promote M2 polarization. 17,18 M2 may help to overcome inflammation and through high endocytic clearance capacities mediate wound healing, tissue remodelling and repair. 14 Nevertheless, M2 macrophages are also associated with fibrosis as they can secrete components of the extracellular matrix and produce growth factors that activate epithelial cells and fibroblasts, including TGF-β. 19 M2 macrophages play a controversial role in DKD but studies are scarce. 20 Hence, modulating macrophages to a protective phenotype to reduce kidney injury in chronic kidney disease still needs further investigation. Nevertheless, the mechanisms regulating the distinct functions of macrophages during tissue repair remain largely unclear. 21 We have previously demonstrated that anti-inflammatory effects induced by hepatocyte growth factor (HGF) gene therapy enhanced the presence of bone marrow-derived M2 (BM-ΦM2) macrophages in the glomeruli of diabetic mice and halted DKD progression. Therefore, in this study we sought to investigate whether infusion of modulated ex-vivo BM-ΦM2 macrophages provide a therapeutic effect on DKD in obese db/db mice. Our previous results in UUO mice demonstrated that those BM-ΦM2 macrophages secrete TGF-β, did not diminish inflammation and displayed a phenotypic M1 switch when they faced a pro-inflammatory and pro-fibrotic milieu. 22 To overcome this limitation, we also investigate an alternative macrophage therapy transduced with Neutrophil gelatinase-associated lipocalin (NGAL), which is a M2 phenotypically stabilized cell line overexpressing the anti-inflammatory cytokine IL-10 and showing low TGF-β secretion. 23 2 | MATERIAL AND METHODS

| Animals
The experiments complied with the current legislation on animal experiments in the European Union, and the principles of laboratory animal care were approved by our institution's Ethics Committee for Animal Research. Ten-week-old diabetic female C57BL/Ks/J mice (db/db) and non-diabetic (db/-and C57bl/6J) female mice were purchased from Janvier (Laval, France). Animals were given free access to water and a standard laboratory chow diet.

| Φ-RAW and Φ-NGAL macrophages
Mice RAW 264.7 macrophages (Φ-RAW) were cultured in DMEM/ F12 + GlutaMAX medium supplemented with 10% FBS and 1% P/S. For Φ-NGAL group, RAW 264.7 macrophages were cultured and then transduced ex vivo with the adenoviral vector NGAL as previously described by the group. 24 The adenoviral vector carrying cDNA encoding recombinant NGAL was elaborated, amplified and purified by ViraQuest, Inc (North Liberty, IA, USA). We previously reported the control vector Φ-βGAL as a control vector for the Φ-NGAL transduction. 22 We analysed Φ-βGAL macrophages before infusion and showed no differences compared to Φ-RAW macrophages ( Figure S1), but obviously show a difference regarding their average life expectancy. Tacking this feature into account, we use NGAL modified cell line for chronic models, such as DKD or Unilateral Ureteral Obstruction (UUO).

| Experimental design and study groups
Mice were divided into five treatment groups with their respective controls and followed for four weeks: (a) D+BM-ΦM2 (n = 9), dia-

| Monitoring
Animals were followed from 10 to 21 weeks of age. During this period glucose levels and body weight were measured weekly. Glucose was measured using the ACCU-CHEK Performa blood glucose meter (Roche, Basel, Switzerland). Blood was obtained from the tail vein and samples were collected in order to analyse the serum creatinine at three time points: before cell therapies (15 weeks), when cell therapies were performed (20 weeks) and before mice were killed (21 weeks). Mice were placed in metabolic cages in order to collect 24 hours urine specimens before cell therapies (15 weeks), during cell therapies (18 weeks, 19 weeks and 20 weeks, respectively) and after macrophage cell therapies (21 weeks). Urine samples were collected in order to analyse the urinary albumin concentration. Urine creatinine and serum creatinine were determined following Jaffe's and GLDH reactions (Olympus Autoanalyzer AU400, Hamburg, Germany) in the Veterinary Clinical Biochemistry Laboratory of Universitat Autònoma de Barcelona. Urine albumin excretion was determined using a specific commercially available ELISA kit (Albumine Blue, la Hulpe, Belgium).  Slides were stained with F4/80 for macrophages and CD86 for the M1 subpopulation. Macrophages and M1 subpopulation were evaluated in ten glomeruli per slide and per animal in a blinded manner (n = 9 each group). The percentage of M1+ with respect to F4/80+ macrophages was calculated. Alexa Fluor 488 (green) and Alexa Fluor555 (red) were used as secondary antibodies (Invitrogen). For podocyte assessment, nuclear marker WT-1 (1:30 Santa Cruz Biotechnology, Heidelberg, Germany) and Alexa Fluor 488 as a secondary antibody (green) (1:1000, Invitrogen) were used. In a blinded manner, podocytes were quantified in ten glomeruli per sample and calculated as podocyte/total nuclei ratio. Nuclei were stained blue with DRAQ5.

| Quantitative real-time PCR
RNA was extracted from kidney as previously described. 25

| Cytokine analysis
Serum cytokines were quantitatively measured by FACS Canto with the mouse inflammation kit cytometric bead array (CBA) from BD Bioscience (San Jose, CA, USA). Data were acquired and analysed using BD CBA software.

| Statistical analysis
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 Krustal-Wallis test for non-parametric values. P ≤ 0.05 was considered to be statistically significant. All statistical analyses were carried out using StatView software.

| RESULTS
3.1 | BM-ΦM2 cell therapy does not have therapeutic effect on DKD All experimental diabetic groups (D) showed significantly higher body weight, higher albuminuria and diuresis than the non-diabetic group (ND) (Table 1, Figure 1). There were no significant differences between D+BM-ΦM2 and D+SHAM mice regarding albuminuria (Table 1 and GUITERAS ET AL.  Figure 3A). Of note, no major differences were observed between the D+BM-ΦM2 group and its control, the D+Φ-BM group. We also evaluated the kidney tissue TNF-α protein levels by ELISA ( Figure 3B). Only the D+Φ-NGAL group showed a reduction in TNF-α kidney protein. Its control group, D+Φ-RAW, showed similar TNF-α than D+SHAM controls.
Although serum TNF-α was similar in the D+BM-ΦM2 and D+SHAM groups, kidney TNF-α protein expression, assessed by ELISA, was significantly higher in the D+BM-ΦM2 than in the D+SHAM and showed statistical significance compared to its control group D+Φ-BM ( Figure 3B).
The kidney mRNA CD40 gene expression was up-regulated in diabetic animals and it was marginally reduced after D+Φ-NGAL cell therapy ( Figure 3C). Interestingly, the anti-inflammatory IL-10 kidney gene expression was significantly increased in the D+Φ-NGAL group compared with both diabetic control (D+SHAM) and non-diabetic groups (ND) ( Figure 3D).

| Φ-NGAL macrophage cell therapy reduces the glomerular M1 macrophage ratio
We evaluated the expression of M1 phenotype in glomerular macrophages by confocal microscopy. Macrophages were mainly localized around the glomeruli ( Figure 4A). As shown in Figure 4B Week 16 (pre-treatment)

| DISCUSSION
In this study, we demonstrated firstly that BM-ΦM2 macrophage cell therapy did not exert any therapeutic effect on DKD and secondly that Φ-NGAL macrophage cell therapy was associated with a therapeutic effect on DKD. The lack of therapeutic effect of the BM-ΦM2 macrophage cell therapy is consistent with our previous findings in the CKD UUO model that demonstrates M2 switching to M1 in a pro-inflammatory and pro-fibrotic environment. 22 Hyperglycaemia and the dysregulated metabolic milieu play a key role in DKD. 26 However, in TD2M tight glucose control was not associated with improved renal and non-renal outcome. [27][28][29] Thus, although hyperglycaemia plays a major role in DKD initiation additional effector mechanisms should be involved in the progression of DKD. Among such mechanisms, inflammation is considered as a relevant pathogenic factor in the progression of DKD. 30 Therefore, the reduction in inflammation has been associated with a significant reduction in the extracellular matrix accumulation in DKD. 31 NGAL has been described as a marker and potential positive modulator of inflammation. 32 NGAL display anti-inflammatory properties in different diseases providing protective effects. One of the main acting pathways is by interfering the transcription of NF-kB dependent genes such as TNF-alpha and IL-6. 33 Recently, it has been described NGAL interaction in inflammation by controlling | 847 autophagy 34 and inflammasome. 35 Moreover, in the kidney, NGAL can indirectly reduce inflammation by blocking necrosis and by modulating apoptosis and more specifically by activating DAMP HMGB-1, as demonstrate in the nephrotoxic serum nephritis. 36 Hence, NGAL may also be one of the key potential modulators of macrophage phenotype stabilization and inflammation 37,38 ; thus, promoting renal epithelial cell integrity and renal recovery after injury. 24 Our group has recently described 39 that, in renal epithelial cells, Lcn-2 acts via binding to its specific receptor and triggers downstream pathways of proliferation (activation of the PI3K/Akt-pathway) and inhibition of PPARγ. Previous data from our group suggested that the infusion of Φ-NGAL macrophages, as a result of their IL-10 overexpression, was effective during the inflammatory phase of renal ischaemia and reperfusion injury. 40 We have also recently reported that Φ-NGAL macrophages have reduced cell plasticity and therefore preserve their anti-inflammatory and anti-fibrotic phenotype even when placed in a pro-inflammatory and pro-fibrotic environment. 22 Taking into account this rationale we investigated whether Φ- to the progression of DKD. It has been described that podocyte number correlates with proteinuria and it is one of the best predictors of DKD. 45 In addition to other pathogenic mechanisms as hyperglycaemia-induced ROS generation, 26 inflammation 46 can contribute to podocyte loss by causing podocyte apoptosis and detachment. As podocytes are terminally differentiated cells that cannot be replaced, the remaining podocytes enlarge by cell hypertrophy and finally dedifferentiate to cell type that contribute to extracellular matrix accumulation. In our study, we found that only infusion of Φ-NGAL macrophages exerted a protective effect on podocyte loss and was associated with a significant reduction in albuminuria. As this cell therapy was also the only one that could modulate the inflammatory response in DKD, it has been suggested that Φ-NGAL macrophage cell therapy preserve podocyte number by means of reducing inflammation. Therefore, our results are in-line with other studies suggesting that the reduction in podocyte damage may, in fact, reduce albuminuria and diabetes progression. 47 Albuminuria is a biomarker of DKD and its severity is considered one of the most important predictors of DKD progression. 48 As previously pointed out, the reduction in albuminuria in the Φ-NGAL group may merely by a reflection of podocyte preservation. Nevertheless, albuminuria in pathological states such as DKD, can be also because attenuated tubular reabsorption of albumin. 49 Megalin is a large glycoprotein that is highly expressed in proximal tubular epithelial cells. 50 As we found that Φ-NGAL macrophage cell therapy was associated with preserved megalin gene expression and consequently with tubular epithelial integrity, it is feasible that albumin tubular reabsorption could also play a role to reduce albuminuria in diabetic animals treated with Φ-NGAL.
The fibrogenic cytokine transforming growth factor β (TGF-β) is an important transcriptional regulator of cell transdifferentition promoting epithelial to mesenchymal transition and extracellular matrix accumulation. There is consistent data showing that renal TGF-β expression increases in DKD 6,51 as it is induced by renin angiotensin system activation, metabolic dysregulation and hyperglycaemia.
TGF-β is considered as one of the most relevant pro-fibrotic factors accounting for glomerulosclerosis and interstitial fibrosis in DKD as well as other chronic nephropathies. 52,53 The administration of anti-TGF-β antibodies, antisense TGF-β1 oligodeoxynucleotides or knocking off the downstream Smad3 gene prevent and/or reverse the hypertrophic and pro-fibrotic effects of hyperglycaemia in DKD. 54 Considering the potential benefit of M2 macrophage cell therapy on DKD it is important to note that M2 macrophages have anti-inflammatory properties but are also associated with fibrosis as they can secrete TGF-β. 19,20 Therefore, in our study we found that in mice treated either with BM-ΦM2, Φ-BM and Φ-RAW infusion, the renal TGFβ-1 as well as renal fibrosis were similar to that observed in diabetic non-treated animals. In order to F I G U R E 6 Macrophage gene expression profile and phenotype assessment before infusion. A, Macrophages were cultured and mRNA expression of IL-10, TNF-α, TGFβ-1, NGAL and CD206 were measured by qPCR in fresh cultured BM-фM2, ф-β-GAL and ф-NGAL previously to infusion. B, Phenotype analysis was likewise analysed by FACS CANTO cytometry and BM-фM2 and ф-NGAL showed high expression of CD206 compared to ф-β-GAL. n = 3 in each group. a P ≤ 0.05 BM-фM2 vs ф-NGAL; b P ≤ 0.05 ф-NGAL vs ф-β-GAL switch significantly their phenotype to the inflammatory one M1. 22 Also, Zheng et al 55 in diabetic animals proved that macrophages from the cell therapy were located in spleen immediately after infusion, and then accumulated progressively both in kidney and pancreas.
The immunogenicity of those Φ-RAW macrophages could be interpreted as a potential limitation in our study. Φ-RAW macrophages derive from monocytic leukaemic cell of BALB/c that carries H-2d, and C57Bl/6j carries H-2b. Therefore, adoptive transfer of cells to mice with different H-2 antigen is possibly immunogenic. However, previous studies from our group and others, 22,23,56,57 suggested that infusing these cells in C57Bl/6j was not immunogenic. Consistent with these results, we did not detect any effect associated with the potential activation of an effector alloimmune response against the infused cells in our diabetic mice model.
In summary, our study demonstrates that Φ-NGAL macrophage cell therapy has a therapeutic effect on DKD in diabetic mice probably by modulation of the renal inflammatory response caused by the diabetic milieu. Our study provides evidence about the limitations of macrophage cell therapy in kidney diseases and shows a potential strategy to overcome the inherent M2 plasticity and TGF-β1 overexpression by Φ-NGAL transduction. Thus, our study paves the way to new opportunities in DKD treatment.

CONF LICT OF I NTEREST
All the authors declared no competing interests.