Macrophage Overexpressing NGAL Ameliorated Kidney Fibrosis in the UUO Mice Model

Progressive tubulointerstitial fibrosis is the final common pathway for all kidney diseases leading to chronic renal failure [1]. Kidney fibrosis results by ongoing tissue injury and inflammation or if kidney repair is insufficient or suppressed [2]. A common feature in many chronic kidney diseases (CKD) is the renal interstitial inflammatory cell infiltration at early stages [3, 4]. While inflammation as a whole contributes significantly to the engagement of fibrogenesis, evidence in recent years also highlights the antifibrotic effects of various lymphocyte and macrophage subsets, providing novel potential avenues for antifibrotic therapies [1]. Macrophages are highly heterogeneous cells that are divided into subpopulations based on their distinct functions. Two well defined phenotypes are commonly referred to as classically activated macrophages (M1), produced by exposure to lipopolysaccharide (LPS) or interferon gamma (IFN-γ) which are pro-inflammatory and alternatively activated macrophages also called as M2, produced by Th2 cytokines such as IL-4 and IL-10 [2, 5]. M2 macrophages can be further subcategorized into at least three subgroups: M2a induced by IL-4 and/or IL-13 (wound healing macrophages), M2b induced by immune complexes and regulatory macrophages known as M2c (anti-inflammatory macrophages) [6-8]. However, in vitro classification of macrophages does not necessarily reflect their true phenotypes in vivo; hence, a newly suggested classification closely reflects phenotypes of in vivo macrophages, and has been defined as pro-inflammatory anti-inflammatory fibrotic and fibrolytic [2]. Nevertheless, these proposed macrophage phenotypes deserve to be well defined.

A pathogenic role of macrophages in kidney disease has been demonstrated by depletion of kidney macrophages by liposomal clodronate (LC) in ischemia reperfusion injury (IRI) and unilateral ureteral obstruction (UUO) models [8, 9]. Nevertheless, there is increasing evidence showing that macrophages also play a key healing role during the reparative phase after kidney injury [10, 11]. Some studies revealed that depletion of macrophages at the time of injury phase (Ml predominantly) and during repair phase (when M2 are predominant) attenuated kidney injury and delayed kidney repair [2]. Further studies suggested that macrophages change their phenotype in response to signals from the local kidney milieu although the mechanisms by which they are polarized are not well understood [12, 13]. Ex vivo programmed macrophages displaying anti-inflammatory or reparative phenotype have been used as a macrophage cell therapy in distinct injured kidney mice models including IRI, lupus nephritis and Adriamycin nephropathy [12, 14, 15] with discordant results, suggesting thatphenotypic stability of transferred cells in vivo is critical to achieve therapeutic success.

In a murine model of diabetic kidney disease, we have previously demonstrated that hepatocyte growth factor (HGF) gene therapy increases the presence of bone marrow-derived M2 macrophages, which seem to contribute to the renal tissue repair and regeneration [16]. However, it is unclear what the role of "M2" macrophages are in UUO renal fibrosis, though some descriptive studies would suggest that "M2" macrophages are associated with fibrosis in the UUO context [9]. Although most studies showed association of M2 infiltration with renal fibrosis, the information regarding adoptive transfer of M2 in vivo is scarce. Therefore, in the present study we sought to examine the therapeutic effects of alternatively activated bone marrow-derived macrophages (φ-IL-4/IL-13) cell therapy in the UUO mice model, a well-established model of renal fibrosis. As we did not observe any therapeutic effect, we further characterized the phenotype and the functional plasticity of infused φ-IL-4/IL-13 macrophages isolated from the obstructed kidney, showing that the lack of effect was associated with transition to Ml-like phenotype. Finally, we provided evidence that this limitation could be overcome by the infusion of genetically modified and stable φ-NGAL macrophages.

Φ-IL-4/IL-13 isolation and culture. Primary cultures of murine macrophages were obtained from bone marrow of C57M/6J mice purified by CDllb+ negative selection (EasySep, Stemcell technologies, Grenoble, France). Bone marrow-derived monocytes were matured in DMEM/F12 + GlutaMAX (Gibco, Waltham, USA) medium supplemented with 10% fetal bovine serum, 1% penicillin and 20 ng/ml macrophage colony-stimulating factor (M-CSF) (ProSpec, East Brunswick, NJ, USA) and cultured the last 3 days with 10 ng/ml of IL-4 and IL-13 (Invitrogen, CA, USA) to become M2 macrophages. All cells were maintained at 37°C in a 5% CO₂ incubator. Cells were harvested and the macrophage phenotype and purity were assessed by FACS analysis.

Mice RAW 264.7 macrophages were transduced with either adenoviral vector NGAL (φ-NGAL) or the control vector β-galactosidase (φ-βGAL) as previously described by the group [17]. Adenoviral vectors were applied in order to transduce macrophages ex vivo. The adenoviral vector carrying cDNA encoding recombinant NGAL and β-GAL were elaborated, amplified, and purified by ViraQuest, Inc (North Liberty, IA, USA). Transfected cells were incubated for 24 hours at 37°C in a CO₂ incubator.

To analyze the purity of cultured macrophages for the cell therapy, cells were stained with PE-conjugated anti-mouse F4/80 (eBioscience, San Diego, CA, USA), APC-conjugated anti-mouse CDllb (BD Bioscience, San Jose, CA, USA), FITC-conjugated anti-mouse mannose receptor (CD206) and CD86 both from Biolegend (San Diego, CA, USA). Flow cytometry was performed using a FACS CANTO II cytometer (BD Bioscience, San Jose, CA, USA) and data was acquired and analyzed using FACS DIVA software.

Eight-weeks-old C57M/6J male mice were purchased from Janvier (Laval, France). All procedures followed the Guidelines of the European Community Committee on Care and Use of Laboratory Animals and Good Laboratory Practice. Mice were monitored daily for body weight and were housed at constant temperature with a 12 h-light/12 h-dark cycle. They were given water and standard chow ad libitum.

UUO surgery was performed under continuous inhaled isofluorane anesthesia (2%). Mice were placed on a thermostatically controlled operating table. The left flank was opened through a small incision and whereby the bowel gently displaced to one side and covered with sterile saline-soaked gauze; the left ureter was exposed and ligated with 6-0 silk suture (Ethicon, Edinburgh, UK). The bowel was then laid back in place and muscle and skin were then closed with sterile surgical of 4-0 nylon (B. Braun, Melsungen, Germany). For the UUO time-course study tissue were harvested on day 1, 3, 5, 7, 9,11,13, and 15 (n = 3).

UUO mice were divided into the following groups according to the killed day (n = 7 each group): UUO day 9, UUO+Φ-IL-4/IL-13 day 9, UUO+φ-NGAL day 9, UUO+φ-βGAL day 9, UUO day 15, UUO+Φ-IL-4/IL-13 day 15 and UUO+φ-NGAL day 15, UUO+φ-βGAL day 15. Wild type mice and contralateral kidneys were used as controls. For the cell therapy, one million (1 x 10⁶) of Φ-IL-4/IL-13, φ-NGAL or φ-βGAL macrophages were infused into UUO mice via the tail vein at day 7 after the UUO surgery.

In order to assess whether adaptive transfer of exogenous macrophages reached obstructed kidneys, a macrophage tracking in vivo study was performed. Macrophages were cultured and then stained with a red fluorescent membrane label Vivotrack680 (Vertex, Barcelona, Spain) according to manufacturer's instruction. One million (1 x 10⁶) labeled macrophages were injected into mice by a single tail-vein injection at day 7 after UUO. Mice were killed 48h after macrophage injection (day 9 after UUO surgery) and were assayed and quantified (n = 3). Living Image® Software was used to analyze the images.

To assess the macrophage renal cell sorting and the further phenotypic characterization, labeled macrophages were injected at day 7 after UUO surgery and mice were killed 48h after macrophage cell therapy. Kidneys were carefully removed and cut into 1-2 mm³ pieces, excluding any adjoining renal fat. Slices were incubated in 1 ml of 2 mg/ml Collagenase I (Sigma, Steinheim, Germany) for 30 min at 37°C. Gently cell suspension was filtered in BD cell strainer (70 µm) and re-suspended in 2ml of DMEM medium with 20% of FBS. Suspension was centrifuged at 1500 rpm for 5 min. Pellet was re-suspended with 1 ml of Lysis Buffer and suspension was allowed to settle briefly (5 min). 1 ml of DMEM medium with 20% of FBS was added to stop the lysis buffer reaction and then centrifuged again at 1500 rpm for 5 min. Supernatant was decanted and 2 ml of PBS was added for counting cells to estimate the number of total cells in the suspension.

To analyze both exogenous and endogenous macrophages phenotype in the UUO treated mice, a cell sorting experiment with a MoFlo Astrios Cell Sorter (Beckman Coulter, Pasadena, CA, USA) was performed. Firstly, two parameters were used to select macrophages from the entire kidney (both obstructed and contralateral): positive VivoTrack680 fluorescence cells, which are the macrophages from the cell therapy, and the negative Vivotrack680 cells, which are the endogenous cells. Exogenous macrophages (Vivotrack680 positive) were labeled with FITC-conjugated anti-mouse mannose receptor (CD206) (M2 marker) and CD86 (PE) (Ml marker) antibodies to analyze their phenotype, whereas endogenous macrophages were detected with anti-mouse F4/80 PE-Cy7 (Affymetrix eBioscience, CA, USA) and anti-mouse CD lib APC-Cy7 (BD Bioscience, San Jose, CA, USA) antibodies from the entire Vivotrack680 negative population. Detection of Ml macrophages with anti-mouse CD86 PE (BD Bioscience) and M2 macrophages with FITC-conjugated anti-mouse mannose receptor (CD206) (Biolegend San Diego, CA, USA) was evaluated for endogenous macrophages. After labeling, samples were re-suspended in 500 µl of FACS buffer (PBS with 2mM EDTA and 0.5mM BSA) and filtered through a 40 µM micron filter to remove any clamps or aggregates prior to analysis.

After UUO, mouse kidneys were removed under anesthesia and rapidly decapsulated and chopped into 2-3 mm pieces. Renal slices were fixed in 10% (vol./vol.) formalin and embedded in paraffin. All pieces contained both cortex and medulla. Histological cross sections of 3µm thickness were stained with Haematoxylin and eosin (H&E), Sirius Red, Fibronectin and anti-α-smooth muscle actin (α-SMA) for optical microscopy assessment. Fibronectin (1:500, Abeam, Madrid, Spain) and α-SMA (1:100, Thermo Scientific, CA, USA) were stained as described previously [16].

Fibronectin, Sirius Red and α-SMA staining were quantified using Image J software in each non-overlapping cortical field from the cortical region. To evaluate the degree of renal injury, a magnification of x200 was assessed to quantify histological sections. Points within brown and red areas were considered positives. Scores from 5 fields per kidney section were averaged and mean scores were quantified. Values are obtained as relative stained area (%). Pathological evaluation of interstitial fibrosis and tubular atrophy (IFTA) was also assessed following Banff criteria [18] and graded as follows: 0, no damage; 1, Mild interstitial fibrosis and tubular atrophy (<25% of cortical area); 2, Moderate interstitial fibrosis and tubular atrophy (26-50% of cortical area) and 3, Severe interstitial fibrosis and tubular atrophy/loss (>50% of cortical area).

Confocal studies (using a Leica TCS-SL confocal spectral microscope, Mannheim, Germany) were performed in a blinded manner on paraffin-embedded renal tissues. Anti-mouse F4/80 (1:50, LabClinics, Barcelona, Spain) and double immunolabeling of F4/80 and anti-mannose-receptor (CD206) (1:50, Abeam) antibodies were analyzed using immunofluorescence. Slides were stained with: F4/80 to assess macrophage infiltration and mannose receptor for the M2 macrophage subpopulation. Nuclei were stained blue with DRAQ5. F4/80-positive cells located in the interstitial area were analyzed in 20 random fields in each sample and assessed using Image J software. Values are obtained as relative stained area (%). CD206 positive macrophages were numerically quantified of the entire sample. Alexa Fluor488 (green) and Alexa Fluor555 (red) were used as secondary antibodies (Invitrogen, CA, USA).

RNA was extracted from kidney with PureLink RNA Mini Kit (Invitrogen, CA, USA) as previously described [16]. A total amount of 400 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 minat 85°C and finally held at4"C. The mRNA expression levels of TGF-βl, MMP-9, CD40, MCP-1, IL-10, N-GAL, iNOS, E-cadherin, Aquaporin-2, Megalin, Heme Oxygenase-1 (HO-1) and Wnt7b were quantified by TaqMan real-time PCR (ABI Prism 7700, Applied Biosystems, Waltham, USA) using the comparative 2^{–[delta][delta]Ct} method (Applied Biosystems).

Sorted macrophages were collected and RNA was extracted as described previously in the quantitative real-time PCR paragraph. RNA was necessarily pre-amplified before the real-time reverse transcriptase PCR was performed. Renal macrophage expression levels of TGF-βl‚MMP-9, IL-10,Wnt7b, HO-1 and TNF-α were quantified by TaqMan real-time PCR (ABI Prism 7700, Applied Biosystems, Waltham).

All data are reported 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 analyzes were carried out using StatView software.

First of all, we confirmed that renal macrophage infiltration correlates with interstitial fibrosis development in the UUO model. Fibrosis was assessed and scored using Fibronectin and Sirius Red staining (Fig. 1B-E). As observed in Fig. 1, an increasing number of interstitial macrophages infiltrating the obstructed kidney were detected (Fig. 1A). Macrophage infiltration began at day 3 and increased exponentially after day 7. Therefore, we selected day 7 as appropriate timing for macrophage cell therapy administration in order to reduce fibrosis and inflammation. More detailed information regarding the evolution of kidney fibrosis in the UUO model is provided in Fig. 2. As expected, pro-inflammatory and pro-fibrotic cytokines such as TGF-βl, MCP-1, CD40 and MMP-9 were upregulated in the obstructed kidney.

Then, to ascertain whether the infused macrophages reached the obstructed tissue, we performed in vivo cell tracking and specific assessment in several organs using Vivotrack680 NIR fluorescent agent (Fig. 3). The macrophages used for the in vivo cell tracking yielded a >96% of double F4-80 and CD lib-positive. Further characterization confirmed their alternative M2 phenotype, showing almost 100% of CD2 06 positive marker (M2 phenotype) and CD86 negative (Ml marker) (Fig. 4A). Our results indicated that those infused macrophages were higher in the obstructed kidney than in the contralateral non-obstructed kidney (Fig. 3D-E) probably due to the inflammatory recruitment signals present in the damaged tissue. Moreover, our results showed that the liver was the main organ retaining exogenous macrophages (Fig. 3B-C), as previously described by other authors[14, 19, 20]. Once the detection of the infused macrophages in the kidney was confirmed, we next isolated them by cell sorting two days after the infusion to analyze their phenotype and function.

In both therapies, around 2% of infused macrophages were detected in the obstructed kidney and 0.01% in the contralateral kidney. In the case of Φ-IL-4/IL-13 therapy Vivotrack680 positive cells comprised 2.19 ± 0.32 % of the entire viable cells detected in the whole obstructed kidney (Fig. 5A), which displayed 44.69 ± 3.06 % of CD86+ (Ml marker), 2.13 ± 0.31 % of CD206+ (M2 marker) and 53.2 ± 2.71 % of double positive (both Ml and M2 markers). On the other hand, in the case of φ-NGAL therapy, Vivotrack680 positive macrophages were the 1.71 ± 0.22 % (Fig. 5A), which displayed 30.45 ± 4.05 % of CD86+ (Ml marker), 4.91 ± 0.53 % of CD206+ (M2 marker) and 64.64 ± 4.21% of double positive. Therefore, a higher Ml-percentage of exogenous macrophages was observed in Φ-IL-4/IL-13 therapy, whereas the M2 proportion was nearly double in φ-NGAL (Fig. 5B). Regarding endogenous renal macrophages, they represented the 25 ± 0.7 % of all UUO kidney cells in both therapies, with 15.34 ± 0.32 % of Ml, 6.34 ± 1.58 % of M2 and 78.32 ± 1.65 % of double stainingin Φ-IL-4/IL-13 and 26.34 ± 0.23 % of Ml, 6.79 ± 0.24 % of M2 and 66.88 ±0.28 % of double stainingin φ-NGAL (Fig. 5C).

In the contralateral kidney, the amount of exogenous (0.01%) and endogenous (0.63 ± 0.06 % in ΦTL-4/IL-13 and 0.72 ± 0.0 1% in φ-NGAL) macrophages was too low to take their characterization into account (Fig. 5A-D).

We then measured the expression of several pro-reparative and pro-fibrotic genes in the isolated endogenous and exogenous macrophages. Gene expression of pro-fibrotic as well as pro-reparative markers from endogenous and exogenous macrophages is shown in Fig. 6. Exogenous macrophages from Φ-IL-4/IL-13 treated mice expressed high levels of TGF-βl and less MMP-9. By contrast, φ-NGAL showed higher levels of Wnt7b and NGAL, whereas HO-1 expression was not modified.

Endogenous macrophages from obstructed kidneys without cell therapy exhibited high mRNA levels of MMP-9, TGF-βl and TNF-α and reduced levels of NGAL, Wnt7b and H0-1. Cell therapy treatment induced a decrease of MMP-9, TGF-βl and TNF-α, and an increase of Wnt7b and HO-1 gene expression in the endogenous macrophages. Figure 4B shows the gene expression profile of the macrophages prepared for the therapy at baseline, that is, before infusion.

First, we found that Φ-IL-4/IL-13 cell therapy did not improve the severity of IFTA, neither fibrosis, evaluated by Fibronectin and Sirius Red stain, or α-SMA renal expression at day 9. In contrast, φ-NGAL treated mice showed reduced kidney IFTA, reduced fibrosis and decreased α-SMA expression in comparison to the UUO group (Fig. 7), Φ-βGAL treated mice showed no amelioration of IFTA, fibrosis and α-SMA expression compared to UUO non-treated mice both at day 9 and 15 (Fig. 8). Similar results at day 15 for Φ-IL-4/IL-13 and φ-NGAL treated mice are shown in Fig. 9.

Then, as showed in Table 1, we observed that φ-NGAL cell therapy was associated with a significant reduction of some pro-inflammatory and pro-fibrotic renal genes as well as with a higher expression of some genes related to kidney epithelial integrity.

Finally, we examined the amount of M2 macrophages (mannose receptor) in the obstructed kidney (Fig. 10). We found no differences among UUO control mice and Φ-IL-4/IL-13 mice. Interestingly we observed a high number of positive CD206 macrophages in the φ-NGAL treated animals.

Regenerative medicine is in preclinical development in nephrology. Among potential strategies [21-23], macrophage cell-therapy provides promising but discordant results [12, 14, 15, 19, 20, 24]. Macrophage precursors can be isolated from the bone marrow or spleen, alternatively activated in vitro and then infused to test their effect on acute and chronic kidney disease models. In humans, bone marrow and blood cells are commonly used as cellular therapy [12]. Cell therapy using autologous macrophages seems safer than some stem cell based therapies, at least in terms of tumor development risk [25] and immunogenicity [26]. Interestingly, although M2 macrophages are required for the kidney repair after injury, their fate or long-term effect still remains uncertain [27]. Recently, Harris and Wang [2] proposed that in chronic kidney disease Ml macrophages increase in kidney tissue whereas M2 macrophages coexist in small numbers or are absent due to persistent inflammatory kidney milieu. Therefore, there is a rationale to test whether M2-macrophage cell therapy is able to induce regression of the persistent inflammation and fibrosis in chronic kidney disease.

We demonstrate that macrophages from cell therapy although mainly retained in the liver, are capable of reaching the damaged kidney parenchyma. Our findings are in agreement with two studies performed on murine Adriamycin nephropathy [12, 14]. Therefore, infused macrophages could exert their protective effect on the UUO kidney directly, by a paracrine mechanism or both [26].

Nevertheless, we found that, between the two tested therapies, Φ-IL-4/IL-13 macrophage cell-therapy was unable to induce any kidney protective effect. In fact, histological damage and the expression of some pro-inflammatory and pro-fibrotic genes in the obstructed kidney were similar in treated and non-treated groups. Cao et al. [12] reported similar findings by Φ-IL-4/IL-13 macrophage cell therapy on murine Adriamycin nephropathy. Moreover, we found that the proportion of M2 macrophages in the UUO kidney was the same in Φ–IL-4/IL-13 and non-treated control groups. Taking together, these results suggest that Φ-IL-4/IL-13 macrophages could switch their phenotype depending on the kidney environment [2]. On the other hand, we performed a cell therapy group of genetically modified stabilized cell line that produce SIP-dependent neutrophil gelatinase-associated lipocalin (NGAL/Lcn-2) [28], trying to overcome the plasticity of ΦIL-4/IL-13 macrophages. Previously, we have reported that φ-NGAL macrophages, due to their IL-10 over-expression, were effective during the inflammatory phase of renal ischemia and reperfusion injury [28]. With this macrophage therapy we found that kidney pro-inflammatory genes such as MCP-1 and CD40 as well as the kidney fibrosis markers TGFβ-1 and MMP-9, were significantly reduced. Furthermore, after φ-NGAL infusion, the proportion of M2 macrophages in the UUO kidney was higher than after Φ-IL-4/IL-13 therapy. These results suggest that φ-NGAL macrophages could preserve their phenotype when they reach the damaged kidney.

As there is lack of evidence about this macrophage phenotypic switch in vivo [2], we sought then to provide new knowledge by isolating the infused macrophages from the obstructed kidneys to analyze their phenotype and functional activity in comparison with baseline macrophage cell therapy and resident kidney macrophages. Of note, we found that UUO kidney Φ-IL-4/IL-13 macrophages lost their M2 phenotype and showed either the Ml or the Ml and M2 double positive transition phenotype. Thus, we demonstrate that nearly all Φ-IL-4/IL-13 macrophages switch their canonical phenotype when localized into the UUO kidney. In addition to phenotype switching transcriptional analysis confirmed that the infused Φ-IL-4/IL-13 macrophages switched to functional Ml. Although our findings strongly suggest that those changes on Φ-IL-4/IL-13 were induced by the inflammatory milieu present in the UUO kidney, unfortunately we were not able to prove such hypothesis since the number of infused macrophages isolated from the contralateral normal kidney was too low to allow further characterization.

Matrix metalloproteinases (MMPs) were previously thought to be anti-fibrotic because of their ability to induce extracellular matrix remodeling. However, recent studies suggest that MMPs are involved in the initiation and progression of kidney fibrosis in chronic kidney disease [29]. The precise mechanism by which MMP-9 contributes to renal fibrosis still is a matter of debate. Our results agreed with those reported by Tan et al. suggesting that MMP-9 contributes to the pathogenesis of renal fibrosis via macrophage recruitment [30]. Furthermore, Harris and Wang [2] suggested that Ml macrophages induce renal fibrosis by MMP-9 secretion. In summary, the previous study on Adriamycin nephropathy showing functional switching of ΦTL-4/IL-13 macrophages to Ml [12] and our study, extend current knowledge about macrophage plasticity in vivo by showing for a first time both phenotype and functional change of ΦTL-4/IL-13 macrophages isolated from the UUO kidney.

There are several approaches to overcome this significant limitation. One is provided by Cao et al. [12] by using spleen macrophages which show lower proliferative ability. Alternatively we used in vitro stabilized NGAL macrophages. In agreement with the significant reduction of kidney fibrosis after φ-NGAL therapy we found that φ-NGAL macrophages isolated from the UUO kidney displayed predominantly the M1-M2 double positive phenotype with a significantly higher proportion of macrophages with preserved M2 phenotype in comparison to the Φ-IL-4/IL-13 findings. Interestingly the φ-NGAL macrophages showed lower TGF-βl and TNF-α expression compared to the Φ-IL-4/IL-13 ones. Moreover, the φ-NGAL macrophages isolated from the UUO kidneys showed high Wnt7b gene expression, consistent with their favorable phenotype [2]. Lin et al. [31] described that macrophages in response to renal injury produce and release the Wnt ligand Wnt7b that acts on injured tubular epithelial cells (TECS) to promote regeneration of the tubule basement membrane, thus re-establishing renal architecture and reducing renal fibrosis. Φ-IL-4/IL-13 macrophages also express Wnt7b and Lcn2/NGAL genes but at low level, since they just did not get induced as high as the RAW 264.7 cells transduced with Lcn2/NGAL. Altogether, this data reveals that transformed macrophages were able to better preserve their antinflammatory and antifibrotic phenotype when were eventually located in the obstructed kidney.

Additionally, information of endogenous macrophages gene expression, although indirectly, may provide further evidence of the adoptive cell transfer effect. Therefore, φ-NGAL therapy could modify the functionality of the endogenous macrophages in the obstructed kidney, as suggested by the reduction of some pro-inflammatory (TNF-α) and pro-fibrotic (MMP-9 and TGF-βl) genes in comparison to the endogenous macrophages from non-treated UUO mice.

Macrophage plasticity resulted in confusion regarding the identity of individual macrophage populations [32] although it is a critical feature of these cells to regulate the balance of renal injury, inflammation, repair and fibrosis [2, 33]. Moreover, this plasticity is a major limitation to macrophage cell therapy. Infusion of Φ-IL-4/IL-13 macrophages in the UUO model may act as adding fuel to the fire since more cells become Ml and thus amplifying renal inflammation and fibrosis. Understanding macrophage heterogeneity is critical because these cells can be biomarkers and have also the potential to be used as therapy. Therefore, more studies are needed before using this kind of therapy in humans [32].

There is one limitation in our study. RAW264.7 is a cell line derived from monocytic leukemic cell of BALB/c that carries H-2d whereas C57BL/6 carries H-2b. Adoptive transfer of cells to mouse with different H-2 antigen is possibly immunogenic. The mice strain usually used to study chronic kidney damage is C57BL6 based on its susceptibility to renal fibrosis [34]. On the other hand, to achieve a stable φ-NGAL transduction, we used the commercial cell line RAW264.7. Previous studies from our group and others [28, 35, 36], suggested that infusing these cells in the C57BL/6 is and even in other species was not immunogenic, at least in the short-term. In agreement with these experiences, we did not observe any effect related to the potential activation of an effector alloimmune response against the infused cells in our model, although we could not rule out that repeated infusions during a more prolonged follow up could be challenging. In conclusion, our study provides further evidence that macrophage plasticity is a critical limitation for macrophage cell therapy in CKD and shows that it can be partially overcome by inducing ex vivo stable and anti-fibrotic macrophage.

We thank Esther Castaño for her help with sorter analysis, Beatriz Barroso for her help with cytometer and Benjamin Torrejón for his help with confocal microscopy (Scientific and Technical Services, UB, Barcelona, Spain). We also thank to Núria Bolaños and Cristian Varela (IDIBELL, Barcelona, Spain) and Angeles Muñoz (ΠBB-CSIC-IDIBAPS) for their excellent technical support.

This research was supported by Spanish Government Instituto de Salud Carlos III (ISCIII) grant PI12/01427, PI15/00638 and PI12/00720 Confunded by FEDER funds/European Regional Development Fund (ERDF)-a way to Build Europe-//, AMGEN grant under the auspices of The Red de Investigación Renal (European Regional Development Funds ISCIII Red Temática de Investigación Cooperativa en Salud Red de Investigación Renal; RD16/0009/0003) and SENEFRO (awarded to AS). SAF 2015-67770 (awarded to GH). AS is supported by Miguel Servet contracting system (CP08/00138).

All the authors declared no computing interests.

R. Guiteras and A. Sola contributed equally to this work.