Hypercholesterolemia-induced increase in plasma oxidized LDL abrogated pro angiogenic response in kidney grafts

Background Renal transplantation is increasingly associated with the presence of comorbidity factors such as dyslipidemia which could influence the graft outcome. We hypothesized that hypercholesterolemia could affect vascular repair processes and promote post-transplant renal vascular remodeling through the over-expression of the anti-angiogenic thrombospondin-1 interacting with vascular endothelial growth factor-A levels. Methods We tested this hypothesis in vitro, in vivo and in a human cohort using (1) endothelial cells; (2) kidney auto-transplanted pig subjected (n = 5) or not (n = 6) to a diet enriched in cholesterol and (3) a renal transplanted patient cohort (16 patients). Results Cells exposed to oxidized LDL showed reduced proliferation and an increased expression of thrombospondin-1. In pigs, 3 months after transplantation of kidney grafts, we observed a deregulation of the hypoxia inducible factor 1a—vascular endothelial growth factor-A axis induced in cholesterol-enriched diet animals concomitant with an overexpression of thrombospondin-1 and a decrease in cortical microvessel density promoting vascular remodeling. In patients, hypercholesterolemia was associated with decreased vascular endothelial growth factor-A plasma levels during early follow up after renal transplantation and increased chronic graft dysfunction. Conclusions These results support a potential mechanism through which a high fat-diet impedes vascular repair in kidney graft and suggest the value of controlling cholesterolemia in recipient even at the early stage of renal transplantation. Electronic supplementary material The online version of this article (10.1186/s12967-018-1764-4) contains supplementary material, which is available to authorized users.

recipient is of greater importance related to calcineurin or mTOR inhibitor exposure given the subsequent exposure length of the graft to the recipient milieu. In general, hypercholesterolemia is associated with increased systemic levels of oxidized low density lipoproteins (OxLDL) well known to be involved in endothelial cell dysfunction [4,5]. Renal microcirculation, one of the critical targets of I/R injury, plays a crucial role in the early regeneration phase but also in the extension of chronic renal disease through a complex interplay between proliferation, regeneration and capillary loss [6][7][8]. In healthy organs, vascular endothelial cells are protected from moderate injury by the release of autocrine signaling molecules such as vascular endothelial growth factor-A (VEGF-A) [9]. However, endothelial damage induced by I/R affect this regenerative response [10,11].
Previously, Cui et al. reported the role of thrombospondin-1 (TSP-1) as an important mediator for high fat-diet induced kidney dysfunction in mice characterized by macrophage infiltration and fibrosis development [12]. TSP-1 is also involved in numerous detrimental processes induced by kidney I/R, but little is known about this effect in kidney graft outcome and the link between TSP-1 and OxLDL [13]. TSP-1 is a matricellular glycoprotein that interacts with different receptors and has multiple functions including promotion of clot formation or fibrosis and is also well-known as an inhibitor of angiogenesis by different mechanisms such as interacting with VEGF-A levels or modulating the NO availability which may prevent kidney repair [14].
This study is aimed at determining the role of hypercholesterolemia and particularly oxidized LDL (OxLDL) in early vascular regenerative processes occurring after I/R injury in renal transplantation that are distinct from the chronic injury of atherosclerosis induced by dyslipidemia. We hypothesized that OxLDL could impede protective angiogenesis mechanism induced by I/R through the stimulation of TSP-1 expression and its effects on VEGF-A production.

Animal model and surgical procedures
Male Large White pigs (Sus Scrofa) were fed with a standard or a high-fat diet (standard diet + 20% lard and 2% cholesterol) immediately after weaning and maintained until euthanasia, i.e. 3 months after transplantation as previously described in previous studies [1,15]. Renal auto-transplantation model was performed 2 months after weaning as previously described in accordance with the institutional committee for the use and care of laboratory animals (CEEA Poitou-Charentes CE2012-4) [16]. Briefly, the left kidney was removed, flushed with 300 ml of UW preservation solution and preserved at 4 °C in the same solution in static conditions for 24 h. On the day of transplantation, the right kidney was removed and the left kidney grafted mimicking the nephron mass in transplanted situation. Two experimental groups were studied: Normal diet: transplanted kidneys collected 3 months after surgery from animals fed a standard diet (n = 6), Highfat diet: transplanted kidneys collected 3 months after surgery from animals fed a high-fat diet (n = 5). One transplanted high-fat diet pig died before completion of the study due to surgical complications and was not included in data analysis. Urinary and plasma creatinine or proteinuria were measured using an automatic analyzer (Modular, Roche Diagnostic, Meylan, France). Creatinine clearance was calculated by the formula: (Urinary volume × Urinary creatinine level)/plasma creatinine level. Peripheral blood was collected before kidney transplantation, at 3, 7 days and 1, 3 months after reperfusion. Urines were collected at 3, 7 days and 1, 3 months after reperfusion.

Apoptotic signals
Apoptotic renal cells were characterized by the TUNEL method using the DeadEnd Fluorometric TUNEL system (Promega, Fitchburg, Massachusetts, USA) as previously shown [20].

Real-time quantitative PCR
We used RNA extraction kit (Qiagen, Courtaboeuf, France). Genomic DNA was removed using DNA-free kit (Applied Biosystems, Foster City, California, USA) and first-strand reverse transcription (Applied) was performed. Real-Time PCR assays were performed on an ABI Prism 7300 (Applied) with porcine primers adapted for VEGF and TSP-1 mRNA expressions (Additional file 1: Table S1).

Tissue preparation, acquisition and analysis of images by high resolution micro-computed tomography
Three months post-transplant at the time of sacrifice, the kidney graft was removed and perfused with a saline solution containing 5000 IU/L of heparin. Then, the saline solution was replaced by an intravascular radioopaque silicone polymer (Microfil MV122; Flow Tech, Carver, Massachusetts, USA). The kidney was immersed in formalin solution at 4 °C and a cylinder biopsy (diameter: 1 cm; depth: 1.5 cm) of the polymer-filled kidney was performed and encased in paraffin. High resolution micro-computed tomography scans were performed, followed by image reconstructions using specific algorithm [21]. The average diameter and spatial density of cortical vascular segments of microvessels were calculated as previously described [17,22] and the spatial density classified according to diameter as small (< 40 µm), medium (40-80 µm), large (80-120 µm) or very large (> 120 µm) vessels.

Patients
A prospective cohort study was conducted on 16 kidney transplant patients at the Transplant Unit of the Poitiers University Hospital between January 2010 and June 2010. Adult recipients of a first or second kidney transplant were eligible for enrollment and patients with preemptive graft, infectious complications and early acute graft rejection have been excluded. The blood samples were taken during the normal follow-up of the patient, anonymized and since the study did not require additional blood sampling, an approval from an ethics committee was not required under French law according to the article L.1121-1 of the public health code. Written informed consents were obtained from each patient according to the Declaration of Helsinki. We classified patients in 2 groups related to the plasma cholesterol levels using a threshold of 1.80 g/L. Renal function was assessed by urinary protein excretion and eGFR according to the simplified modification of diet in renal disease formula MDRD. Peripheral blood was collected by venipuncture at day 0 (D0) before kidney transplantation, 1, 3, 7 and 14 days (respectively D1, D3, D7 and D14), 1, 3 and 12 months after surgery. Plasma samples were stored at − 80 °C prior to protein quantification by ELISA for OxLDL (Mercodia, France), and VEGF-A (RD System).

Statistics
Results are shown as means ± SD. We used a student t-test for two-group comparisons or a Mann-Whitney

The inhibition of endothelial cells proliferation induced by OxLDL is mediated by TSP-1
OxLDL-enriched culture medium induced the expression of ADAMTS-1 in HAEC and promoted the increase of TSP-1 secretion in supernatant without effect on phosphorylation of VEGF-R2 (Fig. 1a, c). Alpha5beta3 integrin expression on HAEC, marker of endothelial cell proliferation, was decreased in both conditions: in culture medium enriched with OxLDL or supplemented with TSP-1 human recombinant protein (Fig. 1b). Interestingly, SiRNA targeted TSP-1 (Additional file 2: Figure  S1) limited the inhibition of cell proliferation induced by OxLDL in presence of VEGF underlining the role of TSP-1 in these inhibited conditions ( Fig. 1d).

High-fat diet reduces pro-angiogenic pathways in renal auto-transplanted pig
In a clinical relevant porcine model, as previously described, hyperlipidic diet induced an increase of plasma cholesterol level in the last part of the follow up even if OxLDL plasmatic levels were significantly elevated at the day of surgical procedure and at 3 months post-transplantation [15]. As previously reported, only high-fat diet was associated with an increased proteinuria at 1 and 3 months while both experimental conditions, with or without high-fat diet, induced similar kinetics for creatinine clearance (Table 1) [15]. The effects of OxLDL on regenerative vascular processes suggested by our in vitro results were further assessed by measuring the expressions of proteins involved in these pathways at 3 months postreperfusion. The hypoxic pathway mediated by HIF1a well known to promote VEGF production was abrogated by the high-fat diet. Dyslipidemic pigs exhibited an increased expression of HIF1a in kidney grafts 3 months after while VEGF-A expression remained stable or even decreased when assessed by western blotting and immunohistochemistry respectively (Figs. 2, 3, 4a, b, Table 2, Additional file 3: Figure S2, Additional file 4: Figure S3). In addition, high-fat diet increased the expression of ADAMTS-1 (Fig. 4a) known as an activator of the TSP-1 pathway and decreased the expression of proteins involved in endothelial cell proliferation and migration such as SDF-1 (Fig. 4b). A trend was observed for MMP-9 and alpha5beta3 integrin (Fig. 4b).

TSP-1 pathway is activated by high-fat diet in kidney graft
To investigate the role of TSP-1 in the inhibition of the HIF1a pro-angiogenic pathways induced by hyperlipidic diet, we assessed TSP-1 expression in kidney graft. We observed in the high-fat diet group an increased protein expression of TSP-1 ( Fig. 5; Table 2, Additional file 5: Figure S4). These results observed by immunohistological studies showed an elevated staining of TSP-1 in the interstitium, tubules and glomerulus in high-fat diet group but not significant difference in glomerulus parietal cell ( Table 2). High-fat diet did not affect VEGF-A and TSP-1 mRNA expressions indicating rather a posttranscriptional regulation of these pathways (Additional file 6: Figure S5).

High-fat diet induces vascular and tissue remodeling in kidney graft
The increase of macrovascular media-to-lumen ratio indicated an intensive vascular remodeling in high-fat diet group compared to normal group (Fig. 6A). In addition, tubular atrophy was increased suggesting deleterious mechanism induced by hyperlipidic diet resulting in non functional fibrous tissue development assessed by Masson trichrome staining (Fig. 6B, C; Table 2) as  previously suggested [15]. High resolution micro-computed tomography showed a decrease of the density of vascular segments with a diameter inferior to 40 µm in the hyperlipidic diet kidney grafts, indicating a high microvessel injury (Fig. 6D). These results were supported by aminopeptidase staining indicating a decrease of capillary density in hyperlipidic diet group (Additional file 7; Additional file 8: Figure S6). In the cortex, this experimental group exhibited a reduction of the average vascular segment diameter supporting a vascular remodeling stimulated by high-fat diet (Fig. 6D). These injuries were associated with a pro-inflammatory milieu supported by ED1-positive cells infiltration significantly elevated in dyslipidemic pigs as well as apoptosis stimulation (Fig. 7A, B).

Hypercholesterolemia the day of surgery affects VEGF-A secretion and promotes chronic graft dysfunction in patients
This single-center study enrolled a total of 16 Caucasian renal transplant recipients. Characteristics of recipients and renal transplant allograft donors are listed in Table 3. All patients received a kidney graft from deceased donors. Using a threshold of 1.80 g/L of plasma cholesterol level, the day before renal transplantation, we classified patients in hyper or normocholesterolemic groups (Fig. 8a). There was no patient death and two grafts loss during the follow up (1 in each group). Plasma OxLDL measurement in the blood samples collected before surgery showed a high level in Table 2 Summary table of

semi-quantification of kidney stainings in transplanted animals fed either a normal or a high-fat diet maintained for 3 months after surgery (M3)
Values are mean ± SD  the hypercholesterolemia group but not reached statistical difference (p = 0.061; Fig. 8b). Interestingly, we observed a decrease of plasmatic VEGF-A level, 1 day after surgery in the hypercholesterolemia group which reached statistical significance at D3 and returned to basal level from D7 (Fig. 8c). The levels of plasma TSP-1 before surgery were increased in the hypercholesterolemia group (Additional file 9: Figure S7). Estimated glomerular filtration rate (eGFR) was similar in both groups up to 3 months post-surgery. However, 1 year after transplantation, the eGFR was significantly reduced in the hypercholesterolemia group (Table 3) and urinary protein excretion increased (Table 3).

Discussion
Organ shortage in renal transplantation has two major consequences: (1) it pushes transplantation centers to use organs obtained from older donors; (2) it increases the time spent on the waiting list, therefore increasing the average recipient age. This type of donors/recipients is linked with age-related co-morbidity factors such as hypertension or hypercholesterolemia, negatively affecting the transplantation outcome and leading back patients to dialysis [23].  TSP-1 is upregulated in the kidney from diet-induced obese mice with marked accumulation of TSP-1 in the glomeruli mesangium and tubular system [12]. They also demonstrated that TSP-1 deletion protects mice from obesity-induced renal fibrosis [12]. This detrimental effect was also observed in a syngeneic rat renal transplantation model. Briefly, a CD47 monoclonal antibody, which blocks the interaction between the ligand TSP-1 and its receptor CD47, used during the donor kidney perfusion, induced a marked improvement in post transplant survival [24]. These results were also observed in a renal ischemia reperfusion injury model [25]. In this study, we investigated the early effects of high-fat diet on pro-angiogenic pathway inhibition and TSP-1 expression in a porcine model of renal auto-transplantation with high clinical relevance. This allowed us to focus on I/R injury in the absence of any allogeneic response and easily extrapolate the results to human as porcine and human kidneys have the same multilobular architecture and similar cortical microcirculation [26]. Hypercholesterolemia is associated with increased circulating levels of OxLDL [4]. These oxidized lipoproteins are involved in endothelial cell dysfunction [5], the first cell type subjected to I/R injury in organ transplantation. However, vascular endothelial cells are protected by the release of autocrine signaling molecules such as the pivotal VEGF-A and are well known to proliferate, migrate and regenerate the injured tissue [9]. In a previous study, Sunitinib, a tyrosine kinase inhibitor which inhibits both VEGF and PDGF receptors, prevents chronic rejection changes in experimental kidney transplantation and preserved significantly renal graft function after transplantation highlighting the role of VEGF in early kidney graft function [27].
In the first part of this study, we investigated in endothelial cell model, the effects of OxLDL incubation on TSP-1 expression. We reported, for the first time, that OxLDL promotes TSP-1 secretion in endothelial artery cells according to the increase expression of ADAMTS-1 and decreases the expression of alpha5beta3 integrin known as an integrin required for angiogenesis and endothelial proliferation [28]. TSP-1 uses multiple pathways to modulate endothelial cell migration [29]. TSP-1 is known as a natural antagonist of alpha5beta3 integrin through alpha5beta3 ligation and it has been shown to negatively regulate focal adhesion formation on fibronectin, suggesting that integrin function may be impaired [29]. These results suggest a repressive role of OxLDL production stimulated by hyperlipidic diet on angiogenesis through a TSP-1-dependent pathway underlining by the SiRNA experiment. Renal I/R induces regenerative processes such as angiogenesis which could be affected by OxLDL permanent production. Since our in vitro model relies on the culture on one cell type and is far to mimic the complexity of the kidney physiology, we decided to further test our hypothesis in our porcine a b Normal Diet High-fat Diet In previous studies, it has been shown that high-fat diet affects renal VEGF-A protein expression and impairs microvascular structures in native porcine kidney without ischemic insults [30,31]. This aspect is well documented and suggests that high-fat diet in transplant condition could affect pro-angiogenic processes mediated by VEGF-A. We exclusively focused this study on the role of high-fat diet in early vascular regenerative processes after an I/R injury in renal transplantation and excluded the effect of high-fat diet in native kidney since it was previously reported [30,31]. In pigs fed with a high-fat diet beginning from weaning up, we previously reported an increase in plasma levels of OxLDL at the day of surgery and at 3 months after transplantation [15]. This later increase was concomitant with a rise of total cholesterol levels. The discrepancy between plasma cholesterol and OxLDL levels the day of surgery could be explained by the fast required before surgery and underlined the role of high levels of plasma OxLDL during the early step of reperfusion. During follow up, renal function assessed by creatinine clearance was similar in both groups. This monitoring window for 3 months might be too short to observe significant differences in graft function. However, we observed a significant elevated proteinuria in high-fat diet group at 3 months post-transplantation which is described as a pivotal time point to investigate tissue and vascular remodeling and pro-angiogenic pathways [10,17].
Renal tissue ischemia is one crucial factor in the development and progression of chronic kidney disease in general. It is known that HIF1a expression is increased in area of severe fibrosis as observed in high-fat diet group indicated by Masson trichrome staining [9]. Decreased oxygen tension activates the stabilization of HIF1a, inducing a large range of pro-survival responses such as VEGF-A expression [32]. However, in this study, high-fat diet promotes a HIF1a stabilization without observed transcriptional effects on VEGF-A expression. In addition, high-fat diet reduced the pro-angiogenic capacity of kidney graft as shown by the repressed expressions of SDF-1 and VEGF-A proteins [9,28,33]. Taken together, these results suggest an anti-angiogenic effect of hyperlipidic diet that could be mediated by TSP-1. Neovascularization induced by angiogenic factors is counterbalanced by the effects of angiogenic inhibitors, such as TSP-1, a glycoprotein expressed at very low levels in the healthy renal cortex and upregulated during some renal diseases [12]. Thakar et al. previously reported that TSP-1 null mice exhibited a significant protection against renal ischemia supporting the fact that TSP-1 acts as a regulator of ischemic damages in the kidney [34]. TSP-1 inhibits angiogenesis through its direct effects on endothelial cells and indirect effects on growth-factor mobilization and activation [9]. To investigate the role of TSP-1 in our conditions, we evaluated its expression in grafts. Hyperlipidic diet induced an over-expression of TSP-1 protein, 3 months after transplantation, in glomerulus, intensively marked in tubules and interstitium without mRNA expression change suggesting a post-transcriptional regulation of TSP-1. This hypothesis was supported by ADAMTS-1 overexpression. Indeed, ADAMTS-1 is a soluble matrix metalloprotease molecule that inhibits angiogenesis by mechanisms that may involve direct sequestration of VEGF or release activation of anti-angiogenic thrombospondin-derived peptides such as TSP-1 [35,36]. ADAMTS-1 was already described to be enhanced in rat proximal tubules after I/R [37] and also in atherosclerosis with macrophage invading the tissue [38]. We observed an increase of ED1-positive macrophages infiltrating interstitium in high-fat diet group, in accordance with the elevated level expression of ADAMTS-1. The significant increase of inflammatory ED-1 positive macrophages in the interstitium of kidney grafts exposed to high OxLDL levels was linked with the profibrotic milieu associated with tubular atrophy. In addition, in this group, TSP-1 could stimulate apoptotic process as suggested by the TUNEL staining increase [39]. In order to investigate the consequences of these pro-inflammatory and regenerative pathways on vascular remodeling, we characterized the cortical microvascularization damages. High-fat diet increased the microvascular rarefaction induced by I/R injury in our renal transplantation model particularly for small vascular segments with diameter inferior to 40 µm according to the decrease of capillary density and the inhibition of pro-angiogenic pathways observed. In addition, the decrease of vascular segment diameter supported a vascular remodeling according to the high ratio of media-to-lumen and fibrosis development.
In order to test the relevance of our results in human transplantation, we investigated the effect of hypercholesterolemia on the systemic levels of the pivotal angiogenic vascular growth factor: VEGF-A in the short-term follow up after renal transplantation. We observed that recipients with high levels of cholesterol the day of surgical procedure, exhibited a decrease of plasmatic levels of VEGF-A during the first 3 days after transplantation suggesting a reduction of the regenerative processes in the graft. In addition, in this preliminary study, we underlined that the level of cholesterolemia the day of surgical procedure could affect graft outcome as indicated by the decrease of filtration function (eGFR) and by the increase of urinary protein excretion in hypercholesterolemic patient, 1 year after reperfusion. Although patient characteristics were similar in both groups, we should notice that other confounding factors could be involved in these differences such as immunosuppressive therapies. Further analyses will be necessary to support the correlation between the increase of TSP-1 levels observed in plasma recipient before surgery and its role on plasma VEGF-A levels early after transplantation.

Conclusions
Taken together, these results support that: (1) OxLDL are involved in abrogation of regenerative capacity of kidney grafts (2) TSP-1 is overexpressed in porcine kidney grafts exposed to high-fat diet and associated to a repression of pro angiogenic pathways, (3) high-fat diet induces microvascular rarefaction and vascular remodeling and (4) hypercholesterolemic patients have an early decrease of VEGF-A plasma levels after transplantation surgery.
In conclusion, our study underlines the impact of a short term exposition to high-fat diet on the early stage of regenerative processes after renal transplantation, distinct from the chronic injury of atherosclerosis induced by dyslipidemia. In addition, high-fat diet or dyslipidemia in recipient of kidney graft could affect organ outcome identifying TSP-1 as a therapeutic target of interest and emphasizing the need to better control either cholesterol or OxLDL plasma levels in recipient at the early stage of renal transplantation.