Pharmacological treatment with galectin-1 protects against renal ischaemia-reperfusion injury

Galectin-1 protein (GAL-1) has important anti-inflammatory properties, but related pharmacologic approaches to effectively treat or prevent renal ischaemia and reperfusion injury are highly limited. Here, we investigated the effect of GAL-1 in a renal ischaemia-reperfusion injury rat model and an in vitro hypoxia-reoxygenation model with a proximal renal tubular epithelial cell line. In vivo, pretreatment with GAL-1 attenuated the renal parameters changed by ischaemia-reperfusion/hypoxia-reoxygenation, with recovery of renal function, protecting against influx of leukocytes, cell death and oxidative stress. Ischaemia-reperfusion/hypoxia-reoxygenation was also associated with increased renal endogenous expression of GAL-1 and intercellular adhesion molecule 1 (ICAM-1) plus augmented levels of proinflammatory cytokines IL-1β, TNF-α and MCP-1 and decreased anti-inflammatory IL-10 in urine, all of which were abrogated by GAL-1 treatment. In vitro studies demonstrated renal tubular epithelial cells as an important source of GAL-1 during hypoxia-reoxygenation and confirmed the protective effects of exogenous GAL-1 through downregulation of proinflammatory cytokine release by proximal renal tubular epithelial cells. Collectively, our findings confirm the important anti-inflammatory role of GAL-1 in kidney ischaemia and reperfusion injury and indicate its promising use as a therapeutic approach.

GAL-1 reduces neutrophil and macrophage influx to the juxtamedullary kidney through the modulation of adhesion molecules. After IR, juxtamedullary region showed an intense influx of neutrophils and macrophages compared to the SHAM group ( Fig. 2A-B,E-F and Q-R). In contrast, GAL-1 and DEXA treatments reversed this effect, showing diminished inflammatory cell numbers compared to the non-treated IR group ( Fig. 2B-D,F-H and Q-R).
To better understand this anti-migratory effect of GAL-1 under IR, we investigated blood leukocyte recruitment under different experimental conditions via flow cytometry (Fig. 3). The analysis of IR dot plot graphs demonstrated a high incidence of CD62L/CD11b-positive cells (Fig. 3B) compared to the other experimental groups (Fig. 3A,C and D). Analysis of the percentages of CD62L/CD11b-positive cells confirmed this observation (Fig. 3G), and no difference was detected for only positive CD62L (Fig. 3E) or CD11b cells (Fig. 3F) among the studied groups.
Additionally, renal IR injury enhanced ICAM-1 and GAL-1 expression in the interstitial tissue of necrosed tubules of the juxtamedullary region compared to the SHAM group ( Fig. 2I-J,M-N and S-T). Pre-treatment with GAL-1 reduced its endogenous expression ( Fig. 2P and T) and ICAM-1 levels ( Fig. 2L and S) in the juxtamedullary region in comparison to the untreated IR and SHAM groups. No change in the interstitial expression of GAL-1 proteins was detected in the DEXA-treated group ( Fig. 2O and T), but ICAM-1 expression was reduced in this group ( Fig. 2K and S).

GAL-1 and DEXA treatments protect HK-2 cells against HR-induced cytotoxicity. The viability
of the cells was investigated in HK-2 control cells exposed to hypoxia and reoxygenation (HR) and HR incubated with GAL-1 (4 µg/mL) or DEXA (1 µM) at experimental times 4, 8, 16 and 24 h after HR. At all times assessed, the viability of the HK-2 cells submitted to HR was significantly lower than that of the control cells (Fig. 4) and was abrogated after GAL-1 or DEXA treatments. This effect could be observed by the presence of dead cells not adhered after 24 h of the HR procedure.

Release of pro-and anti-inflammatory cytokines is regulated by GAL-1 treatment in in vivo and
in vitro models of IR. In vivo, renal IR markedly increased the urine levels of IL-1β, MCP-1 and TNF-α ( Fig. 5A-C) compared to the levels in SHAM animals. Both pharmacological treatments, GAL-1 and DEXA, decreased these levels in the urine after the IR condition ( Fig. 5A-C). Animals treated with GAL-1 also showed an increase in the levels of the anti-inflammatory cytokine IL-10 ( Fig. 5D) compared to the levels in animals treated with IR. In vitro, HR induced an increase in MCP-1 levels in HK-2 cell supernatants at 4, 16 and 24 h (Fig. 5F), IL-1β levels at 16 h (Fig. 5E) and IL-8 levels at 8 and 16 h (Fig. 5G), as well as reductions in IL-10 levels at 16 h (Fig. 5H). Pre-treatments with GAL-1 and DEXA abolished these effects on the cytokine release in HK-2 cells. The addition of exogenous β-galactose did not alter the anti-inflammatory activity of GAL-1 in the production of these cytokines.
HR on HK-2 cells modulates GAL-1 endogenous expression and release. The release of GAL-1 was investigated in the supernatants of HK-2 cells under control, HR and DEXA-treated HR conditions at different time points. Increased levels of GAL-1 were exhibited in the HR condition at 4 and 16 h, as well as in treatment with DEXA, compared with control cells (Fig. 6A).
In accordance with these results and those obtained with cytokine release in HK-2 cell supernatants, the time course of 16 h was chosen for the analysis of endogenous GAL-1 expression through immunofluorescence. These studies were conducted in control, HR and DEXA or GAL-1-treated HR conditions (Fig. 6B). Control HK-2 cells exhibited GAL-1 expression predominantly in the cytoplasm, and after the HR assay, higher GAL-1 levels in the Data represent the mean ± SD, N = 6/group. + p < 0.05, IR and IR + DEXA vs SHAM and IR + GAL-1. *p < 0.05 and **p < 0.01 vs SHAM. One-way ANOVA or Kruskal-Wallis test.
SCIENtIFIC RePORtS | (2018) 8:9568 | DOI:10.1038/s41598-018-27907-y nucleus were evidenced. This pattern of endogenous GAL-1 immunostaining was reversed by GAL-1 treatment but not by DEXA. Densitometry confirmed these observations, showing that GAL-1 expression was increased in the nucleus and diminished in the cytoplasm of HK-2 cells after HR, treated or not treated with DEXA, compared to the expression in other groups (Fig. 6C). No significant alterations in the GAL-1 immunoreactivity were detected between control/HR and GAL-1-treated cells.

Discussion
IR-induced renal injury results from a complex process involving the release of reactive oxygen species and proinflammatory mediators, the upregulation of adhesion molecules and leukocyte recruitment that culminates with rapid kidney dysfunction and high mortality rates 2,4 . Here, an experimental in vivo rat model of renal IR and an in vitro HR model with a human proximal tubule cell line were used. We tested GAL-1 as a suitable therapeutic tool to diminish the side effects of IR-induced renal injury. (E-G) Percentage of CD62L + , CD11b + or double labelled to CD62L + /CD11b + . The percentage of CD62L + /CD11b + leukocytes increased in the IR group. These alterations were reversed with DEXA or GAL-1 treatments. (F) There was no significant difference in CD62L + and CD11b + expression among the studied groups. Data represent the median ± SD of the percentage of cells. **p < 0.01, IR vs SHAM. ++ p < 0.01 and +++ p < 0.001 vs IR, one-way ANOVA test, N = 6/group. The IR procedure used in our study is widely employed 3,16-18 , and right nephrectomy followed by left IR is a model designed with the objective of creating an underlying decrease of 50% of nephron mass and rat survival of at least a few days to study acute kidney injury. Considering that no polyuria and hyperfiltration was detected in the IR group at the 48 h time point, instead of reduced urine output and creatinine clearance, hypertrophy of the remaining kidney did not occur in our study.
Initially, we verified that treatment with GAL-1 30 min before the IR process, as well as DEXA, reversed the damage on renal function, as it reduced plasma creatinine and BUN and sodium/potassium excretion and increased the creatinine clearance and urine volume. In vitro studies confirmed this protective effect of GAL-1 and DEXA on tubular cells (HK-2 cells), which rescued their viability after the HR assay.
Our histological findings revealed no differences in the score of tubular necrosis between untreated or treated (GAL-1 or DEX) IR groups. Despite this, the untreated IR group exhibited elevated urine levels of KIM-1, a transmembrane glycoprotein of renal tubules that is considered a sensitive and specific marker for renal proximal injury monitoring [19][20][21][22] , an effect mitigated by GAL-1 treatment but not by DEXA. Our results are in accordance with a previous study using the IR model that also showed no differences in the KIM-1 urinary levels after 24 and 120 h of DEXA treatment compared to the untreated group 23 . We emphasize that KIM-1, undetectable in normal kidneys, has consistently been demonstrated as a biological marker of early renal damage 21,22,24 , which indicates a partial protective effect of GAL-1 in our study.
Corroborating the KIM-1 data and kidney injury, we found an elevation in cell death by apoptosis in IR animals, which was reversed by GAL-1 treatment but not by DEXA. In fact, the anti-apoptotic role of GAL-1 is possibly related to the activation of focal adhesion kinase-1 (FAK1) in hypoxic epithelial cells 25 . In this study, GAL-1 treatment increased the phosphorylation of FAK1, whereas inhibition of GAL-1 reduced hypoxia-induced pFAK1 levels in H441 cells (human lung adenocarcinoma cells). Furthermore, GAL-1 inhibition activated caspase-3 in the lungs of mice injured with bleomycin and hypoxia, indicating that GAL-1 inhibition can induce the apoptotic pathway in hyperactivated epithelial cells via reduced FAK1 activation.
Similar protective effects of GAL-1 on renal function have been observed in a rat model of glomerulopathy 26 . Alternate-day therapy with GAL-1 (i.p. administration; 1 mg/kg) during two weeks of reduced proteinuria and renal macrophage influx was induced by rabbit nephrotoxic serum. However, GAL-1 treatment failed to inhibit apoptosis and the influx of CD8 + cells and to reduce the anti-rabbit IgG production. In contrast, the injection of GAL-1 into allogeneic rat renal recipients can significantly prolong animal survival and decrease the CD8 + T cell subsets in a dose-dependent manner 27 . These studies suggest that the therapeutic effect of GAL-1 may be mediated by several different mechanisms, including the modulation of cell-cell or cell-matrix adhesion and proliferation.
In fact, the anti-migratory effect of GAL-1 on leukocytes has been associated with the expression of L-selectin and ß2-integrin [13][14][15]28,29 and described in different experimental models of inflammation 7,9,30 . As expected, in our study, the intravenous administration of GAL-1 or DEXA was also sufficient to reduce macrophage and The anti-migratory effect of GAL-1 on macrophages and neutrophils in the IR-injured kidneys probably resulted from the decreased expression of ICAM-1 in the juxtamedullary region. Accordingly, ICAM-1-null mice or wild-type mice treated with ICAM-1 antibody are protected from renal IR injury [30][31][32] . Intratubular leukocytes are known to adhere to ICAM-1-positive tubular cells in the renal biopsies of patients with progressive glomerulonephritis 33 , and post-ischaemic kidneys express increased levels of adhesion molecules on endothelial cells 34 . Therefore, increased expression of ICAM-1 contributes to leukocyte recruitment and infiltration into post-ischaemic tissue and then leads to enhanced leukocyte-interactions, which can promote kidney injury 2,34 .
Increased influx of leukocytes to tissue contributes to the enhancement of reactive oxygen species (ROS), which causes oxidative stress [35][36][37] . In addition, ROS and peroxynitrite generation contributes to nitrotyrosine formation, a marker of nitrosative stress [36][37][38] . Here we also demonstrated that IR induced elevated plasma TBARS and nitrotyrosine renal levels. GAL-1 treatment partly abolished these alterations once plasma TBARS levels were not reduced. Collectively, these data suggest an antioxidant protective effect of the GAL-1 protein through regulation of macrophage/neutrophil influx to the renal tissue.
In the next step of our study, we examined the expression of endogenous GAL-1 under the IR condition. The rat kidneys exhibited GAL-1 immunoreactivity in the cortical region, mainly in the glomeruli and walls of the blood vessels, confirming previous findings [39][40][41] . After IR, increased GAL-1 expression was detected in the interstitial tissue of necrosed tubules of the juxtamedullary region. Another study using a rat model of IR injury revealed an increased number of GAL-1-positive interstitial cells in the vicinity of altered tubules after 48 h, which declined at one week to reach control values 41 . GAL-1 overexpression was also detected in rodent kidneys with Meckel syndrome renal cystic and juvenile cystic kidney diseases relative to the normal kidney 42,43 . Additionally, the increased number of GAL-1-positive cells in patients with diffuse mesangial proliferation and glomerulosclerosis in comparison with control patients indicates that GAL-1 may be crucially involved in the pathogenesis of glomerulopathy in these subjects 39 . However, glomeruli from human nephritic syndrome had a significantly lower expression of GAL-1 than the control glomeruli from individuals with minor glomerular abnormalities. These observations were corroborated by a rat model of puromycin aminonucleoside-induced nephrosis that showed drastically reduced GAL-1 expression in the glomeruli at days 2 and 5 40 . Such differences in GAL-1 expression in kidney diseases reflect the pleiotropic functions of this lectin in renal tissues, which will be further evaluated to address innovative therapies.
In vitro, after the HR process, HK-2 cells markedly increased the nuclear expression of GAL-1 and decreased its cytoplasmic levels relative to control cells, which suggests nuclear protein translocation. Similar data have been observed in a tubular HKC cell line submitted to a high-glucose condition and TGF-β1 treatment 41 . Equally important, the HR condition augmented GAL-1 levels in HK-2 cell supernatants, detected at 4 and 16 h, showing this cellular type as a potential source of this lectin, as corroborated by previous data 39,44 .
Exogenous treatment with GAL-1 decreased its endogenous expression in the juxtamedullary interstitium in relation to the IR group and abrogated its high nuclear levels under the HR condition. Similar results were observed in rodent models of endotoxin-induced uveitis and allergic conjunctivitis 28,29 , where the inflammatory response increases the expression of endogenous GAL-1 in eye tissues, while treatment with exogenous GAL-1 reduces its endogenous expression, possibly through a process of inhibitory self-regulation. On the other hand, in a rat acute rejection model of allogeneic renal transplantation, injection of GAL-1 increased its endogenous expression in the recipient rat renal cortex in a dose-dependent fashion, but the exact mechanism of the increased GAL-1 expression was unclear. Furthermore, GAL-1 therapy plays an important role in inducing allogeneic rat immunotolerance after renal transplantation 27 .
Finally, to better understand the role of GAL-1 in the IR condition, we analysed the urine of animals and supernatants of HK-2 cells under different conditions. Again, GAL-1 and DEXA treatments exhibited an anti-inflammatory function in vivo by decreased urine levels of IL-1β, MCP-1, and TNF-α and increased IL-10 and in vitro by reducing the release of IL-1β, MCP-1 and IL-8 in an independent carbohydrate-recognition-domain manner (with addition of β-galactose). These data corroborate previous experimental models of acute inflammatory or autoimmune diseases in which exogenous GAL-1 reduces the levels of TNF-α, IL-1β, IL-2, IL-6, IL-8, IL-12 and IFN-γ and increases IL-5 and IL-10, consequently regulating the activity of leukocytes 6,7,9,13,27,[45][46][47][48] .
Despite the anti-inflammatory, antioxidant and antiapoptotic effects of GAL-1 on kidney IR model, no protective effect was evidenced in the acute tubular necrosis, an unexpected finding. A possible explanation could be the multifactorial aspect of renal injury models. In fact, studies involving protective handling on renal injury models have shown controversial results. The administration of hydroxytyrosol, a natural antioxidant of olive oil, completely counteracts cyclosporine A (CsA)-induced oxidative stress, the decreased lipid peroxidation level and haem oxygenase-1 (HO-1) gene expression in rats. However, the renal histology is partially protected 49 . Similarly, pharmacological treatment with annexin A1, a glucocorticoid-induced and anti-inflammatory protein, partially protected against CsA-induced histological changes in an acute CsA nephrotoxicity model 50 . On the other hand, the blockade of the renin-angiotensin system by the administration of losartan and enalapril prevents fibrosis induced by CsA, but does not improve renal function 51 . Thus, future studies using a chronic IR model will be necessary to address and confirm the protective effect of GAL-1 on the development of fibrosis, as well as in cytokine/ chemokine release and kidney blood flow.
In summary, the results showed the pleiotropic protective effects of GAL-1 in the early phase of IR-induced kidney injury by improving renal function, inhibiting the release of proinflammatory mediators and KIM-1, and downregulating the expression of the adhesion molecule, which culminates in reduced leukocyte transmigration. Our data indicate that GAL-1 is a new biological target for therapeutic intervention in inflammatory kidney conditions. In vivo study: renal IR injury. Male Wistar rats (200-250 g) were randomly distributed into the following four groups (N = 6/group): SHAM (control), IR, IR treated with GAL-1 or dexamethasone (DEXA). The animals were housed under a 12 h:12 h light-dark cycle and allowed access to food and water ad libitum.
The IR procedure consisted of right nephrectomy and occlusion of the left renal artery with a non-traumatic vascular clamp for 30 min 16,17 . This model was selected because it causes moderate to severe, reversible renal injury in a consistent and reproducible way 16,17 . Recombinant human GAL-1 (100 μg/animal; PeproTech, London, UK) or DEXA (100 μg/animal; Sigma-Aldrich, St. Louis, MO, USA) were administered intravenously 30 min before surgery; the untreated IR group received only phosphate-buffered saline (PBS). The doses of GAL-1 and DEXA were selected according to previous studies 26,27 and after preliminary experiments at our laboratory. Animals from the SHAM group were submitted to the same anaesthesia (20 and 50 mg/mL of xylazine and ketamine, respectively -1 mL/100 g rat) and surgical procedures described above but without the renal artery clamped. The animals were euthanized 48 h after reperfusion.
Renal function study. After 1 day of IR or SHAM procedure, the rats were placed in metabolic cages, with urine volume collected and measured for 24 h, and samples were taken at the end of this period. Furthermore, the animals were euthanized through an anaesthesia overdose, blood samples were drawn, and renal tissue samples were collected. Sodium and potassium levels were determined using an electrolyte analyser, and creatinine levels were ascertained with a colorimetric assay in the 24-h urine and in the blood samples. The blood urea nitrogen (BUN, urea/2.1428) was also ascertained with a colorimetric assay. Commercial kits were used for the measurements of creatinine and urea (Biotecnica, Varginha, MG, Brazil).

Determination of plasma thiobarbituric-acid-reactive substances (TBARS). Plasma TBARS was
determined by quantification of substances that react with thiobarbituric acid, generating colouring that might be detected in a spectrophotometer at 535 nm, compared to the reading of the absorbance of the standard malondialdehyde 20 µM/L 52,53 . The measurement of TBARS is a well-established method for monitoring lipid peroxidation.
The cytotoxic effect of the HR condition was determined by the trypan blue dye exclusion assay. The HK-2 cells were trypsinized, and an equal volume of cell suspension was mixed with trypan blue. Viable cells were counted as clear cells, and dead cells, as blue ones using a Countess Automated Cell Counter (Invitrogen, Frederick, MD, USA). The number of live cells per millilitre was calculated using the following formula: %viability (live cell count/total cell count) 100 = × .
Immunofluorescence. To detect the expression of GAL-1, we grew HK-2 cells on coverslips for 16 h after HR, fixed them in 4% paraformaldehyde for 24 h, washed them in PBS, Tween 20 (0.4%), blocked them with 1% BSA diluted in 3% normal goat serum, and incubated them with polyclonal rabbit anti-GAL-1 Ab (42-5900, Invitrogen, Camarillo, USA; 1:100 in normal goat serum 1.5%). After washing, the cells were incubated with a goat anti-rabbit Ab conjugated with FITC (Serotec, Oxford, UK; 1:100 in normal goat serum 1.5%) for 1 h. The slides were mounted with a solution containing glycerol and PBS (1:1). Negative control samples were incubated only with normal goat serum without primary Ab. The cells were analysed using a filter with a wavelength of 546 nm on an Axioskop 2-Mot Plus Zeiss microscope, and the levels of the protein were quantified by densitometry. Statistical analysis. After normality determination by Kolmogorov-Smirnov test, data with a normal distribution were subjected to analysis of variance (ANOVA), followed by the Newman-Keuls post hoc test for multiple comparisons. Samples with a non-normal distribution were subjected to the Kruskal-Wallis test followed by Dunn's test. A p value of <0.05 was considered significant.