YB‐1 orchestrates onset and resolution of renal inflammation via IL10 gene regulation

Abstract The Y‐box‐binding protein (YB)‐1 plays a non‐redundant role in both systemic and local inflammatory response. We analysed YB‐1‐mediated expression of the immune regulatory cytokine IL‐10 in both LPS and sterile inflammation induced by unilateral renal ischaemia–reperfusion (I/R) and found an important role of YB‐1 not only in the onset but also in the resolution of inflammation in kidneys. Within a decisive cis‐regulatory region of the IL10 gene locus, the fourth intron, we identified and characterized an operative YB‐1 binding site via gel shift experiments and reporter assays in immune and different renal cells. In vivo, YB‐1 phosphorylated at serine 102 localized to the fourth intron, which was paralleled by enhanced IL‐10 mRNA expression in mice following LPS challenge and in I/R. Mice with half‐maximal expression of YB‐1 (Yb1 +/−) had diminished IL‐10 expression upon LPS challenge. In I/R, Yb1 +/− mice exhibited ameliorated kidney injury/inflammation in the early‐phase (days 1 and 5), however showed aggravated long‐term damage (day 21) with increased expression of IL‐10 and other known mediators of renal injury and inflammation. In conclusion, these data support the notion that there are context‐specific decisions concerning YB‐1 function and that a fine‐tuning of YB‐1, for example, via a post‐translational modification regulates its activity and/or localization that is crucial for systemic processes such as inflammation.


Introduction
The YB-1 is a transcriptional and translational factor, which regulates many cellular processes such as cell proliferation, DNA repair, cellular stress response, cell differentiation, embryonic development and inflammation [1][2][3]. In both, systemic and localized inflammatory responses to infectious stimuli [4] as well as during 'sterile' inflammation [3,4], YB-1 regulates expression of pro-inflammatory mediators. In the onset of inflammation, YB-1 up-regulates the expression of pro-inflammatory factors such as interleukin (IL)-6 [5] and CCL5 [4,[6][7][8]. CCL5 is an important inducer of immune cell infiltration, which is transcriptionally and translationally [4] regulated by YB-1 in monocytes/macrophages [8] as well as in human arterial smooth muscle cells [7]. However, anti-inflammatory properties of YB-1 via a trans-repressive capacity on the Ccl5 promoter upon macrophage differentiation are also described [8]. Thus, YB-1 influences the earlyphase of inflammation and also seems to contribute to its termination in the later phase. However, so far, it is unknown whether this dual role of YB-1 also encompasses the regulation of proteins with antiinflammatory properties such as IL-10.
Experimental renal ischaemia-reperfusion (I/R) allows to analyse the onset of acute kidney injury (AKI) and late sequelae including renal regenerative processes [9,10]. Following lack of blood supply with hypoxia (ischaemia) and restoration of the circulation (reperfusion), reversible renal inflammation occurs. Experimental I/R models one of the leading cause of AKI in both native and transplant kidneys [11,12], which is still associated with high mortality and morbidity [13,14]. In I/R-induced AKI, an immune response, is initiated following the ischaemic insult via the extravasation of immune cells through the disrupted endothelium. Early inflammatory events occur in the microvascular and tubulointerstitial compartment of the medulla with leukocytes and platelets blocking renal microcirculation. Subsequently, this strong inflammatory response further spreads to the renal cortex and chemokine expression is still high during the repair phase [10].
Both, the early-and the late-phase of kidney I/R are characterized by infiltration of immune cells including different subtypes of T lymphocytes [15], macrophages and dendritic cells [16] that can facilitate injury but also promote repair.
Immune regulatory mediators, such as IL-10, have been described as renoprotective, as they alleviate kidney damage in nephritis [17,18]. IL-10 suppresses activation and accumulation of innate immune cells. Recently, regulatory T cell (Treg)-derived IL-10 has been described to exert an important protective role in murine models of I/R [17] and crescentic glomerulonephritis [18]. Beyond immune cells, resident renal cells, that is mesangial [19] and tubular [20] cells, are also able to produce IL- 10. In this study, we unravel molecular mechanisms underlying YB-1induced IL-10 production during immune responses to microbial PAMPs as well as during 'sterile' inflammation and clarify the general capacity of YB-1 to mediate pro-and anti-inflammatory processes in these models in wild-type (WT) and Yb1 +/À mice. As it controls IL-10 expression, we demonstrate that YB-1 orchestrates both onset and resolution of the inflammatory responses.

Cell culture
Rat mesangial cells (rMCs) and human T lymphocytes (HUT78T) were cultured in RPMI 1640 medium, and human tubular epithelial cells (HK-2) were incubated in low-glucose DMEM medium with 1% MEM-NEAA and 10 lg/ml insulin. All media were supplemented with 10% FCS, 100 units/ml penicillin and 100 units/ml streptomycin. All cell lines were maintained at 37°C in humidified air with 5% CO 2 . Some cells were incubated with ochratoxin A (OTA, Santa Cruz Biotechnology, Heidelberg, Germany, 20 lM) for the indicated times or with 10 ng/ml LPS (Sigma-Aldrich, Steinheim, Germany) or with PBS alone for 6 hrs. To prevent phosphorylation of YB-1, cells were pre-incubated with 10 lM Ly294002 (Calbiochem, Darmstadt, Germany) for 2 hrs. Results were confirmed in at least three independent experiments.

Plasmids
Plasmids encoding the luciferase gene solely or the luciferase gene fused with the IL10 fourth intron (630 bp) were described before [21]. A fulllength YB-1 expression plasmid (YB-1-pSG5) was kindly donated by J.Ting (Lineberger Comprehensive Center, University of North Carolina).

Transient transfection
Transient transfections in HK-2 cells were performed using calcium phosphate precipitates as described previously [22]. rMCs (2 9 10 5 cells/well) were transiently transfected with the lipid-based transfection reagent FuGENE 6 HD (Promega, Madison, WI, USA) according to the manufacturer's instruction. Human HUT78T cells were transfected by means of electroporation; 2.5 9 10 6 cells resuspended in 1 ml of RPMI 1640 medium supplemented with 20% FCS were added to electroporation cuvettes (0.4 cm gap, Bio-Rad, Hercules, CA, USA) together with a total amount of 20 lg of plasmid DNA, respectively. The mixture was incubated for 5 min. on ice, and cells were electroporated at 250 V/1100 lF in a Gene Pulser II electroporation system (Bio-Rad). Cells were incubated another 5 min. on ice prior to resuspension in 2 ml of RPMI 1640 medium supplemented with 20% FCS, transferred to six-well tissue culture plates and incubated at 37°C and 5% CO 2 .

Luciferase reporter assay
Luciferase reporter constructs were introduced into HUT78T, HK-2 cells and rMCs by the indicated transfection protocols mentioned above. Luciferase assay using the Dual-Luciferase â Assay System (Promega) was performed according to the manufacturer's instruction and as described before [6]. Results were confirmed in at least three independent experiments.

Animal experiments
The local review board approved all animal studies according to prevailing guidelines for scientific animal experimentation. Animals were held in cages with constant temperature and humidity with drinking water and food ad libitum. Mice heterozygously targeted for a disruption of the Yb1 locus (Yb1 +/À , C57BL/6 background) were kindly donated by Timothy J. Ley (Section of Stem Cell Biology, Division of Oncology, Washington University Medical School, St. Louis, MO). Homozygous YB-1-deficient mice are not viable and die before birth due to neuronal disorders, which precludes their use [23].

Ischaemia/Reperfusion (I/R)-induced inflammation model
For the unilateral renal I/R-induced inflammation model, 13-to 16-week-old male Yb1 +/À mice and their age-matched WT littermates (Yb1 +/+ , WT) were anaesthetized with i.p. injection of 0.01% xylazine hydrochloride and 1% ketamine hydrochloride mixture diluted in 0.9% NaCl acclimatized on a warmed electrical plate for 15 min. (constant plate temperature of 37°C). The abdominal cavity was exposed through left lateral incision, and the left renal pedicle was carefully isolated. The pedicle occlusion was performed using non-traumatic vascular clamps for 30 min. or 25 min., and the efficacy of occlusion was confirmed by colour changing in the entire kidney. Sham-operated mice underwent the same procedure without clamping the vessels. During the whole procedure, mice were maintained at 37°C using a warmed electrical plate. Mice were killed by cervical dislocation on days 1, 5 and 21 after renal reperfusion. The left kidney tissues were taken for immunohistochemical analyses, protein and mRNA extraction. 3495 LPS (1.5 mg/kg body weight, 6 and 12 hrs) dissolved in PBS or PBS alone as control by i.p. injection as described before [4].
For ChIP analyses, real-time qPCR was performed using a qPCR Core Kit for SYBR Green I (Eurogentec). The conditions for all Taqman PCRs were 50°C for 2 min., followed by 40 cycles of 95°C for 15 sec. and 60°C for 1 min.

Chromatin immunoprecipitations (ChIPs)
ChIP assay was performed using SimpleChip Plus Enzymatic Chromatin IP Kit (Cell Signaling, Danvers, MA, USA) according to the manufacturer's protocol and as described before [24]. Antibodies specific to histone H3 (Cell Signaling) as a positive control, YB-1 CSD (Eurogentec) and p-YB-1 S102 (Cell Signaling), were added for immunoprecipitation and compared to control non-specific IgG (Cell Signaling). A proportion (20%) of the samples was kept as 'Input' to represent the PCR amplification of the total sample. The real-time qPCR primer sequences used to amplify the regions within the murine Il10 fourth intron were as follows: 5'-TGTGGGAACC-CAGCAAATG-3'; 5'-TGTAACTGAGGTGGTGGCTTTCTA-3'. The amount of immunoprecipitated DNA was subtracted by the amplified DNA that was bound by the non-specific IgG antibody and calculated relative to the respective input DNA. Further, real-time qPCR products were separated on a 3% agarose gel containing 0.01% GelRed (Biotium, Hayward, CA).

Electrophoretic mobility shift analyses
The biotinylation of synthetic DNA probes corresponding to the antisense strands of the human IL10 fourth intron sequences was performed with the Biotin 3' End DNA Labeling Kit (Thermo Fisher Scientific, Rockford, Illinois, USA) according to the manufacturer's instructions. Nuclear cell extract preparation and EMSA with Light-Shift chemiluminescent EMSA kit (Thermo Fisher Scientific, Rockford, Illinois, USA) were performed as described before [6].
For supershift analyses, the following antibodies were incubated with nuclear proteins 20 min. prior to addition of probes: IgG as a negative control (Millipore, Temecula, CA, USA), C-terminal YB-1 (Sigma-Aldrich)-and YB-1 CSD (Eurogentec)-specific antibodies. Supershifted complexes were indirectly seen from decreased band intensities on the blot, whereas bands should not change when control non-specific IgG was added to the binding reactions.

Cortical kidney lysates and Western blot analysis
Cortical kidney lysates were prepared as described before [4]. Protein concentrations were determined using BC Assay Protein Quantitation Kit (Uptima Interchim, Montluc ßon, France) according to the manufacturer's protocol; 10 to 30 lg protein of kidney extracts was subjected to SDS-PAGE. Proteins were subsequently transferred to nitrocellulose membranes and visualized by Lumi-Light and Lumi-Light PLUS Western blotting substrate (Roche Diagnostics GmbH, Mannheim, Germany), respectively. The primary antibodies anti-YB-1 against the protein C-terminus (Sigma-Aldrich) and anti-p-YB-1 against phosphorylated Ser-102 (Cell Signaling) and anti-mouse NGAL (R&D systems, USA) were used. To ensure equal protein loading, the blots were additionally incubated with a monoclonal GAPDH-specific antibody (Novus Biologicals, Littleton, CO).

Immunohistochemistry/immunofluorescence
Immunohistochemistry was performed in 1-lm-thick paraffin sections of methyl Carnoy's-fixed specimens using the Vectastain Avidin/Biotin System (Vector Laboratories, Burlingame, CA), as described before [25]. Renal tissues were stained using the following primary antibodies: antihuman Col1A1 (Southern Biotech, Birmingham, AL), anti-Kim1 (R&D systems, USA), anti-mouse F4/80 (Serotec, D€ usseldorf, Germany), antiproliferating cell nuclear antigen (PCNA) (Leinco Technologies, St. Louis, MO) and anti-Ly6G (BD Biosciences, San Jose, CA). Negative controls for the immunohistochemical procedures consisted of substitution of the primary antibody with non-immune IgG. Quantification of Ly6G was described before [4]. Total number of PCNA-stained nuclei in renal tubuli was counted in 25 randomly selected fields at 9200 magnification. For staining quantification of collagen and F4/80, 20 consecutive images of renal cortex were taken per section. The percentage of the positively stained area was extracted for intensity using ImageJ software (Wayne Rasband, NIH) and a mean area was calculated. Periodic acid-Schiff (PAS) staining was performed as described before [25], and the tubular injury was scored on a scale of 0-4: 0 = none; 1 = 0-25%; 2 = 25-50%; 3 = 50-75%; 4 = more than 75%. The total score is the calculated average of all tubular scores.

Statistical analysis
Values are expressed as means AE S.D. Statistical significance was evaluated using the Student's t-test or ANOVA and Bonferroni post hoc test, whenever more than two groups were compared, with significance accepted when P value was less than 0.05. All in vitro experiments were performed at least in triplicate.

YB-1 regulates IL10 expression via a regulatory element within the IL10 gene locus
To examine the impact of YB-1 on IL10 expression, we inspected the fourth intron (4 th intron), which constitutes an important intronic enhancer element within the human IL10 gene locus [21], for potential YB-1 binding sites. We detected a putative Y-box with an adjacent inverted repeat motif (indicated by asterisks in Fig. 1A) starting at 4346 bp within the IL10 gene. This element displayed strong homologies with known YB-1-binding motifs within the matrix metalloprotei-nase2 (Mmp2) [26] and DNA polymerase-a (DPA) [27] gene loci (Fig. 1B).
Electrophoretic mobility supershift assays (EMSA) were performed with oligonucleotides that encompass the Y-box within the fourth intron of the human IL10 gene. Indeed, we observed protein-DNA complexes (indicated by arrowheads in Fig. 1C) when nuclear extracts of renal cells, namely rat mesangial cells (rMCs, left) and human tubular cells (HK-2, right) that are known to produce IL-10 [19,20], were included. Specifically, addition of an antibody against the C-terminus of YB-1 protein confirmed its binding to this site. Consequently, YB-1 overexpression enhanced the IL10 fourth intron activity in rMCs (Fig. 1D) and HK-2 cells (Fig. 1E).
For some genes, Akt-mediated YB-1 phosphorylation at serine residue 102 (p-YB-1 S102 ) [28] is a prerequisite for its trans-activating capacities [6,29] and previous studies have revealed YB-1 phosphorylation in LPS-challenged primary MCs and in a model of acute LPS-triggered nephritis in vivo [4]. Thus, we next asked whether targeting Akt/PKB kinase activity also affects IL10 expression upon exposition to microbial PAMPs. Pre-incubation with kinase inhibitor Ly294002 resulted in a complete blockade of LPS-triggered elevated expression of IL-10 ( Fig. 2A) and activation of the fourth intron measured by a reporter assay in rMCs (Fig. 2B). Comparable results were obtained in T lymphocytes (HUT78T) (Fig. 2C).
Next, we addressed the impact of (phospho)-YB-1 on renal Il10 expression in mice. We compared binding of YB-1 ( Fig. 2D and E) and p-YB-1 S102 (Fig. 2D and F) to the Il10 fourth intron in kidneys of three mice challenged with LPS or with PBS as control. Particularly, binding of p-YB-1 S102 to the fourth intron was enhanced 12 hrs following LPS treatment (Fig. 2F) along with elevated renal IL-10 mRNA expression (Fig. 2G). The importance of YB-1 for Il10 gene expression during LPS-triggered inflammation was further underlined by a reduced mRNA expression (Fig. 2H) in kidneys of Yb1 +/À mice that display half-maximal expression of YB-1 at both the transcript and protein levels [4,23,30] compared to their wild-type (WT) littermates. Vice versa, an increased YB-1 content enhanced IL10 fourth intron activity in renal mesangial cells (Fig. 2I) and in T lymphocytes (Fig. 2J) following LPS activation.

Half-maximal YB-1 expression results in decreased tubular damage in the early-phase of I/R but hinders renal regeneration at later stages
To further investigate the regulatory role of YB-1 not only in driving acute inflammation but also to monitor progression and resolution of renal inflammation/regeneration, we used a second model of kidney damage associated with sterile inflammation, the unilateral renal ischaemia-reperfusion injury model (I/R) in the early (day 1 after I/ R), established (day 5) and advanced disease stage (day 21). In this model, treatment with IL-10 has been previously shown to ameliorate both renal and systemic inflammation [31]. During the time course of I/R, YB-1 mRNA was significantly up-regulated in kidney cortex on day 1 (Fig. 3A), returned to normal levels on day 5 and was significantly reduced on day 21. Western blot analyses confirmed transient upregulation of renal YB-1 ( Fig. 3B/C).
In order to elucidate YB-1's function in this model, we used Yb1 +/À mice and compared these to WT littermates. Histological examination revealed tubular damage in all I/R groups, indicated by loss of the brush border, cast formation (arrow), acute tubular dilatation (d) and necrosis (arrowhead) (Fig. 3D). Compared to WT animals, Yb1 +/À mice exhibited less pronounced damage on days 1 and 5. However, 21 days after I/R, mice with half-maximal YB-1 expression exhibited significantly stronger damage compared to WT mice, demonstrated by the tubular damage score (Fig. 3D/E) and by areas of exceeding inflammation/fibrosis especially in Yb1 +/À mice (Fig. 3D, boxed region with a higher power image; marked by asterisk).
Our findings were further demonstrated by expression of the tubular damage markers neutrophil gelatinase-associated lipocalin (NGAL) (Fig. 3F/G). In WT mice, a transiently elevated expression of NGAL protein (Fig. 3F) and NGAL mRNA (Fig. 3G) was observed that peaked on day 1 and continuously decreased in the subsequent observation period until day 21. In contrast, in Yb1 +/À animals, NGAL mRNA expression significantly decreased on day 5, but significantly increased again on day 21 compared to WT mice (Fig. 3G, white bars). Elevated NGAL protein was detectable on day 21 only in Yb1 +/À mice (Fig 3F, far right). Augmented tubular damage on day 21 in Yb1 +/À mice was additionally confirmed by immunofluorescence of kidney injury molecule (KIM)-1 (Fig. 4A).
Temporal mRNA expression of CCL5, a key chemokine in tissue inflammation, was lower in Yb1 +/À compared to WT animals at earlier time-points in I/R, and however was significantly elevated on day 21 (Fig. 4B). This shifted immune response was reinforced by the analysis of infiltration of Ly6G + and F4/80 + immune cells. In I/R, early inflammatory events start from the medulla and further spreads to the renal cortex and this is triggered among others by neutrophils. In the medulla of Yb1 +/À mice, markedly fewer Ly6G + cells were detected on day 1; however at later time-points, increased numbers of medullary Ly6G + cells were detected in none of the two genotypes (data not shown). However, significant more renal cortical Ly6G + cells were detected on day 21 in Yb1 +/À compared to WT mice (Fig. 4C). On day 5, the number of cortical F4/80 + monocytes/macrophages was significantly higher in cortices of WT compared to Yb1 +/À mice whereas on day 21 no statistical significant difference was observed between the two genotypes (Fig. 4D).
Following I/R, renal cortical deposition of collagen type 1A (Col1A), a major extracellular matrix (ECM) component in renal fibrotic tissue, was significantly higher in WT compared to Yb1 +/À animals on day 5, however significantly lower on day 21 (Fig. 4E/F). In line with this, expression of PCNA in tubules as an index of tubular proliferation but also of tubular repair and renal regeneration was significantly higher in Yb1 +/À than in WT animals on day 1 but lower on day 21 (Fig. 4G/H).
To further explore differences during the regeneration processes in both genotypes, ischaemia time was shortened from 30 min. to 25 min. and expectedly kidneys analysed on day 21 exhibited considerably less tubular damage (Fig. 4I compared to Fig 3E). Advanced regeneration was also reflected in PCNA + tubular nuclei (Fig. 4J) and of F4/80 + cells (Fig. 4K) that was already lowered again in WT, however still significantly elevated in Yb1 +/À mice. Furthermore, expression of NGAL tended to be higher in Yb1 +/À mice but did not reach significance (Fig. 4L).
Taken together, Yb1 +/À mice exhibit an ameliorated inflammatory response and kidney damage at the earlier stages following I/R but exhibit aggravated kidney injury, inflammation, fibrosis and less regeneration on day 21.

YB-1 modulates renal IL-10 expression following I/R
To directly address anti-inflammatory properties of YB-1 in the different phases after I/R, we monitored presence of p-YB-1 S102 and its binding to the Il10 fourth intron region on days 1, 5 and 21 by ex vivo ChIP analyses of murine kidneys. Analyses of kidney extracts from WT animals exhibited phosphorylation of YB-1 especially at the two early time-points following I/R (Fig. 5A/B). On day 1, enhanced binding of p-YB-1 S102 (about fourfold) to the fourth intron of Il10 was observed that decreased only slightly on day 5 but clearly dropped until day 21 (Fig. 5C, broken line). In contrast to p-YB-1 S102 , binding of total YB-1 to the fourth intron was already strongly diminished on day 5 (Fig. 5C, solid line). During the time course of I/R, a transiently elevated renal expression of IL-10 mRNA was observed, peaking on day 5 (Fig. 5D, black bars). Thus, at the time of high Il10 gene activation, the ratio between p-YB-1 S102 and total YB-1 shifted towards p-YB-1 S102 . In Yb1 +/À compared to WT mice, a strongly and significantly reduced expression of renal IL-10 mRNA was observed in the acute phase of I/R, especially on day 5 (Fig. 5D, white versus black bars). However, on day 21, Yb1 +/À mice exhibited significantly enhanced IL-10 expression compared to their WT littermates that was paralleled by highest renal p-YB-1 S102 content on day 21 ( Fig. 5E/F) and binding of YB-1/p-YB-1 S102 to the Il10 fourth intron (Fig. 5G).
Taken together, IL-10 expression is transiently elevated during the time course of I/R and is regulated by binding of p-YB-1 S102 to the fourth intron region within the Il10 gene. In Yb1 +/À mice, reduced expression of IL-10 at early time-points occurs that might be responsible for the renewed inflammatory response on day 21.

Discussion
In this study, we identified YB-1 as an important regulator of the immune regulatory cytokine IL-10 via its fourth intron region in endotoxemia and in I/R-induced sterile inflammation and demonstrated that YB-1 exerts both pro-and anti-inflammatory properties in the course of inflammation.
Over the last years, YB-1 turned out to be an important player in modulating inflammatory events [3, 5-8, 22, 24, 33, 34]. Transient upregulation of YB-1 occurs during the acute phase of inflammation in experimental inflammation models in kidneys [4] and in hearts recovering from I/R injury [3], where the expression of YB-1 in the myocardium was markedly increased in the earlyphase after reperfusion. This is in line with our observation in the renal I/R model, where YB-1 mRNA and protein expression increased during the onset of inflammation, albeit to a lower extent as compared to other models. Recently, a decrease of YB-1 protein content 24 hrs after I/R in kidneys was described [35]. Conceivably, these opposing observations relate to differences in the experimental setting, for example bilateral [35] versus unilateral clamping that can cause different systemic effects with impact on kidney function.
Opposite effects of YB-1 during inflammatory processes have been described for CCL5 gene regulation in monocytes and in differentiated macrophages in the course of inflammatory responses [6][7][8] and also for IL-6 generation following LPS challenge. In the latter, YB-1 acts by differential control of IL-6 mRNA levels as an inductor in dendritic cells and as a repressor in infiltrating macrophages [5]. Obviously, the phosphorylation status of YB-1 decides on its compartment-specific functions, as this influences its subcellular location [6,28]. Akt/PKB-mediated phosphorylation at serine 102 (pYB-1 S102 ) [28] promotes nuclear YB-1 translocation [6,36,37] and affects YB-1-dependent gene activity, as demonstrated for the promoter of epidermal growth factor receptor [29] as well as for CCL5 [6] and-as we demonstrate now-for the fourth intron region within the IL10 gene, an important anti-inflammatory mediator.
Progression of chronic renal inflammation/kidney diseases critically depends on the decision between resolution and persistence of inflammation. Recently, we demonstrated that Yb1 +/À mice are protected in a renal inflammation/fibrosis model with persistent a pathological condition, namely unilateral ureteral obstruction (UUO) [24]. On the contrary, in I/R, the insult is initial and transient, which allows to analyse not only the onset and the progression of inflammation but also repair mechanisms that are more efficient at shorter ischaemia times. Countervailing effects of YB-1 were observed during the course of this model and the perception that YB-1 is involved not only in the onset but also in the resolution of inflammation is supported by the different outcome in WT and Yb1 +/À mice following I/R. While the expression of markers of inflammation (CCL5) and of fibrosis (Col1A1) was alleviated in the early-phase of inflammation (days 1 and 5) in Yb1 +/À mice compared to their WT littermates, they were increased in the later phase (day 21). Thus, factors induced at early time-points might be responsible for the limited regeneration in the subsequent course of the disease. This applies, for NGAL, a well-established marker of both renal injury and cell regeneration, which was expressed to a lesser extent at day 5 in the established inflammatory phase after I/R in Yb1 +/À mice. Furthermore, a key factor for a rapid and effective response to hypoxia includes HIF-1a that is regulated by YB-1 at the gene and/or mRNA level [38,39]. HIF-1a protein expression is enhanced by direct YB-1 binding to and subsequent activation of the translation of HIF-1a mRNA [38].
Another important feature to suppress the immune response is the generation of IL-10. Expression of IL-10 is achieved by various cells, is of high complexity and tightly regulated. Besides gene expression, additional post-translational regulators are involved that include for example miR-21 as demonstrated in macrophages [40]. In human Tregs and in murine natural killer (NK) cells, exogenous IL-2 mediates Stat5 and Stat4 binding to the fourth intron of IL10 [41,42].
Recently, Hedrich and colleagues demonstrated that the recruitment of Stat3 and Stat5 to the promoter and to the fourth intron of the IL10 gene in T cells varies following stimulation with different cytokines and that the imbalance between pStat3 and pStat5 in T cells from systemic lupus erythematosus (SLE) patients contributes to an increased IL-10 expression [21].
In renal I/R, IL-10 expression is achieved among others by regulatory T cells (Tregs). An adoptive transfer of these cells in renal I/R exerts protective effects that were lost when IL-10-deficient Tregs were transferred [17]. However, Tregs are not the only cell population that produces IL-10 in kidneys. Thus, in Treg-depleted mice, renal IL-10 expression was still markedly increased in the late renal recovery  phase following I/R [15]. Along these lines, observations from others [19,20] and our group proved that mesangial and tubular cells also account for IL-10 production upon an (inflammatory) insult. The regulation of IL-10 expression through YB-1 in both resident renal and immune cells via the fourth intron was demonstrated in the present study via cell culture assays and two in vivo models of renal inflammation, namely in LPS-triggered and sterile inflammation following renal I/R.
In summary, our study provides further evidence for the indispensability of YB-1 as an adaptor for the immune response and its decisive role for IL-10 expression. Thus, fine-tuning of the cellular YB-1 content and its subcellular localization is crucial in systemic inflammatory processes. (C) Ex vivo ChIP assay revealed binding of YB-1 (diamonds) and p-YB-1 S102 (circles) to the Ybox within the Il10 fourth intron in kidneys following I/R that was especially pronounced on day 1 but in case of p-YB-1 S102 still present on day 5. (D) IL-10 mRNA expression in WT (black bars) and Yb1 +/À (white bars) mice in the time course of I/R (n = 5-11). (E/F) Western blot analyses of renal protein extracts of Yb1 +/mice with anti-p-YB-1 S102 Ab (E) and quantification thereof (F). (G) Enhanced binding of p-YB-1 S102 to the Il10 fourth intron in kidneys following I/R was demonstrated by ex vivo ChIP assay. (H/I) OTA (20 lM) that induces damage in tubular cells enhanced IL10 fourth intron luciferase activity (H) and IL-10 mRNA expression (I) in renal tubular (HK-2) cells following YB-1 overexpression. Band intensities were quantified by densitometry, and values were normalized against GAPDH. Relative band intensities are depicted in histograms. Data are expressed as mean values AE S.D. *P < 0.05; *** P < 0.001; ****<0.0001. Fig. 4 (A) Immunofluorescence of tubular damage marker KIM-1 in WT and in Yb1 +/À mice. (B) Transcript numbers of chemokine CCL5 were determined by qRT-PCR in kidneys of WT and Yb1 +/À mice following I/R at different time-points (n = 5-10). (C/D) Numbers of renal infiltrating immune cells positive for Ly6G in renal medulla on day 1 and in renal cortex on days 1, 5 and 21 (C) and immune cells positive for F4/80 in cortex (D) in Yb1 +/À compared to WT mice following I/R (n = 5-14). (E/F) Representative images of Col1 immunostaining (E) and quantification thereof by computer-based morphometric analyses of the positively stained cortical area (%) (n = 5-7) (F). In Yb1 +/À mice, cortical Col1A1 deposition was significantly less on day 5 but more on day 21 compared to their WT littermates. (G/H) PCNA + tubular nuclei in post-ischaemic Yb1 +/À mice in comparison with WT mice on different days (G) and representative images on day 21 (n = 5-11) (H). (I-L) Tubular damage in PAS-stained cortical tissue in I/R groups (I) and (J/K) quantification of tubular PCNA + nuclei (J) and F4/80 + cells (K) and NGAL mRNA expression (L) in kidneys following shortened ischaemia (25 min.) time (n = 4-5). Scale bars, 50 lm. Data are expressed as mean values AE S.D. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. JW, SD, LG, VL, DMB, DH, IVM and PB performed the research; UR, TO, JF, TR and GSB designed the research study, analysed the data and wrote the manuscript.