A triglyceride‐rich lipoprotein environment exacerbates renal injury in the accelerated nephrotoxic nephritis model

Summary Hyperlipidaemia accompanies chronic renal disease either as a consequence of the renal dysfunction or as part of generalized metabolic derangements. Under both situations, the lipid profile is characterized by accumulation of triglyceride‐rich lipoproteins (TGRLs). This lipid profile is recognized as a risk factor for cardiovascular complications. Whether it may pose a risk for renal injury as well remains unclear. A hyper‐TGRL state was generated in C57BL/6 mice using poloxamer‐407 (P‐407) and immune complex‐mediated renal injury was triggered using the accelerated nephrotoxic nephritis (ANTN) model. The hyper‐TGRL animals were hypersensitive to ANTN demonstrated by greater haematuria and glomerular cellularity. These changes were accompanied by increased glomerular accumulation of CD68+ macrophages. The hypersensitive response to ANTN was not seen in low‐density lipoprotein receptor knock‐out mice fed with a high fat diet, where triglyceride levels were lower but cholesterol levels comparable to those obtained using P‐407. These data indicate that a hyper‐TGRL state might be more detrimental to the kidneys than low‐density lipoprotein‐driven hypercholesterolaemia during immune complex‐mediated nephritis. We speculate that the hyper‐TGRL environment primes the kidney to exacerbated renal damage following an inflammatory insult with increased accumulation of macrophages that may play a key role in mediating the injurious effects.


Introduction
The detrimental effects of abnormal lipid and lipoprotein profiles on the kidney are well documented [1]. Dyslipidaemia frequently complicates the course of renal diseases and contributes to their progression [1,2]. Dyslipidaemia may develop as part of a more global metabolic derangement, as occurs in type 2 diabetes (T2D) and the metabolic syndrome (MS), or as a consequence of an immune-mediated inflammatory state, as seen in patients with systemic lupus erythematosus (SLE), where its occurrence is known to increase the risk of renal complications [2][3][4][5]. Whether or not dyslipidaemia complicates or precedes the course of renal disease, disturbances in lipid and lipoprotein metabolism are thought to promote, if not incite, renal injury [6].
A variety of animal studies demonstrate accelerated renal injury upon high fat (HF) feeding [7][8][9]. The lupus-prone New Zealand black and white mice [(NZB3W)F 1 ] develop worse renal disease and increased mortality on a HF diet [7]. Guinea pigs fed a high cholesterol (CHOL) diet exhibited enhanced glomerular cellularity with mesangial matrix expansion [8], while rats fed a high CHOL diet develop significant proteinuria [9]. These studies have focused primarily on the effects of changes in CHOL levels on experimental renal disease, but the lipid profile in metabolic diseases such as T2D and MS is characterized by hypertriglyceridaemia secondary to accumulation of triglyceride-rich lipoproteins (TGRLs), i.e. very lowdensity lipoproteins (VLDL) [10,11]. This hyper-TGRL lipid profile also occurs in other non-metabolic inflammatory conditions such as systemic lupus erythematosus (SLE), where it is associated with enhanced morbidity and mortality [5,12]. Triglycerides (TGs) constitute the majority of dietary fat and there is an influx of TGRLs following each meal [13]; however, the role of hypertriglyceridaemia and a hyper-TGRL state in the pathogenesis of renal disease remain poorly investigated [8]. This is because inducing isolated hypertriglyceridaemia in laboratory animals is extremely difficult. Dietary induction of hypertriglyceridaemia in animals on high sugar diets is associated with hyperglycaemia, itself a risk factor for renal disease [14]. Genetically modified hypertriglyceridaemic mice display complex metabolic derangements, making it very difficult to distinguish the contribution of TGs to the general metabolic picture [15,16]. To determine the effect of TGRLs on experimental renal injury we adopted the poloxamer-407 (P-407)-induced model of hyperlipidaemia [17,18]. P-407 is a surfactant that induces dose-dependent hyperlipidaemia in rodents through inhibition of lipoprotein lipase (LpL) [18]. The hyper-TGRL state is characterized by the accumulation of VLDL leading to gross hypertriglyceridaemia accompanied by low HDL levels, a lipid profile comparable to that observed in T2D, MS and SLE [5,10]. We have shown [19] that the P-407-induced hyper-TGRL environment alters the distribution of circulating monocyte subsets, specifically causing a drop in Gr1 low monocyte numbers [20,21]. This was associated with extensive accumulation of CD68 1 macrophages within tissues, including the kidneys. We interpreted these changes as the result of the promotion of Gr1 low monocyte migration from blood to surrounding tissues driven by the hyper-TGRL environment [19]. Unexpectedly, the accumulation of CD68 1 macrophages in the kidneys during hyper-TGRL conditions did not result in renal damage [19]. However, we hypothesized that the presence of increased macrophage numbers in the kidneys under hyper-TGRL conditions would result in exacerbated renal damage following an inflammatory trigger. We tested this hypothesis by subjecting C57BL/6 (B6) mice to accelerated nephrotoxic nephritis (ANTN) during P-407-induced hyper-TGRL.

Materials and methods
Mice C57BL/6 (B6) mice were purchased from Charles River Laboratories (Wilmington, MA, USA). C57BL/6.LDLRdeficient mice (Ldlr -/-) were purchased from Jackson Laboratory (Bar Harbor, ME, USA). Mice aged 8-12 weeks (weighing between 18 and 20 g) were used for all experiments. The HF diet used consisted of a cholate-free western-type HF diet with the following composition: 15% cocoa butter, 1% corn oil, 0Á25% cholesterol, 40Á5% sucrose, 10% cornstarch, 20% casein, total fat content 16% (Arieblok Diet W, cat. 4021.06) purchased from (Hope Farms, Woerden, the Netherlands). All animals were housed in individually ventilated cages. All procedures were carried out according to the Institutional guidelines and the use of experimental animals was performed following the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. Animal studies were approved by the UK Home Office. P-407 administration and lipid levels P-407 (Pluronic F-127, Cat. no. P2443; Sigma-Aldrich, Dorset, UK) was dissolved overnight in cooled sterile phosphate-buffered saline (PBS). Mice were injected intraperitoneally (i.p.) with 200 ll of P-407 solution (10, 5 or 2Á5 mg equivalent to 0Á5, 0Á25 and 0Á125 g/kg, respectively) or PBS every other day. According to the experimental design blood samples were collected into heparin tubes at different time-points. In P-407-treated animals, blood collection was performed 24 h following drug administration. CHOL and TG levels were measured by colorimetric assay using the cholesterol and triglyceride infinity reagent (TR13421 and TR22421, respectively; Thermo-Scientific, Middletown, VA, USA).

Accelerated nephrotoxic nephritis (ANTN)
B6 mice were treated with P-407 injections for 9 days and then immunized i.p. with 200 lg sheep immunoglobulin IgG (I-5131; Sigma) in complete Freund's adjuvant (CFA; Sigma). Five days later, 200 ml 10% sheep nephrotoxic serum (NTS) was administered intravenously (i.v.) [22]. During the course of the experiment proteinuria and haematuria levels were determined using Hema-Combistix (Bayer, Reading, UK). Blood was collected prior to P-407 injection, before NTS administration and at the end of the experiment. Serum urea was measured using an ultraviolet (UV) method kit (R-Biopharm Rhone, Flasgow, UK), according to the manufacturer's instructions. Thirteen days after the NTS injection, overnight urine collection was performed and the animals were then killed. Paraffin sections were stained with period acid-Schiff (PAS) for glomerular cellularity scoring. PAS-stained kidney sections were scored blindly, as described previously [22].

Injection of nephrotoxic serum
B6 mice were treated with either 10 mg P-407 or PBS i.p. for 2 weeks and then allocated randomly to either receiving 200 ml of 10% NTS intravenously or not (n 5 5/group). Urine samples were collected 24 h after the NTS administration and proteinuria measured using the sulphosalicylic acid method [23].
For mouse C3 and IgG glomerular staining, snap-frozen kidney tissues were cut at a thickness of 5 lm and fixed in acetone. Sections were then blocked with 20% normal goat serum and stained with either fluorescein isothiocyanate (FITC)-labelled polyclonal goat anti-mouse C3 (MP Biomedical, Cambridge, UK) or FITC-conjugated polyclonal goat anti-mouse IgG Fcg chain-specific antibody (Sigma-Aldrich). Sections were then washed and mounted using Vectashield Hard Set mounting medium (Vector Laboratories, Peterborough, UK). Murine C3 and IgG-stained sections were visualized using an Olympus fluorescent microscope with digital camera (Olympus, Southend, UK) and images analysed using Image-Pro Plus software (Media Cybernetics, Silver Spring, MD, USA). Ten glomeruli were examined per section, with mean fluorescence intensity expressed in arbitrary fluorescence units (AFU). Neutrophils were assessed based on polymorphic nuclei from PAS sections and counted in 50 glomeruli per section.

Plasma C3 levels
Plasma C3 levels were measured using a sandwich enzymelinked immunosorbent assay (ELISA). NUNC Maxisorp 96-well polystyrene plates were coated with polyclonal goat anti-mouse C3 capture antibody (MP Biomedical, Cambridge, UK) and incubated overnight at 48C. Plates were then washed and blocked with 2% bovine serum albumin (BSA) in washing buffer. After blocking, plates were washed again and plasma samples were added (1 : 12000 dilution) and incubated for 1 h at room temperature. Standards of mouse C3 were obtained using an acute phase serum (Calbiochem, Hertfordshire, UK). Following incubation, plates were washed and the detection antibody [horseradish peroxidase (HRP)-conjugated form of capture antibody] (MP Biomedical) was added at a dilution of 1 : 25000 for 1 h at room temperature. Plates were developed using HRP substrate tetramethylbenzidine (TMB) substrate and the reaction was stopped with 20 ml stop solution, and the plates read at 450 and 540 nm.

Plasma cytokine and chemokine levels
Plasma cytokine and chemokine levels were measured using Biolegend LEGENDplex (Biolegend, San Diego, CA, USA), according to the manufacturer's instructions.

Statistical analysis
Parametric results were expressed as mean 6 standard error (s.e.), while the median was used for non-parametric data. Differences between two experimental groups were calculated using the unpaired t-test for parametric data and Mann-Whitney U-test for non-parametric data. Comparison between multiple experimental groups was achieved using analysis of variance (ANOVA) followed by Bonferroni's multiple comparison test if data were parametric, and the Kruskal-Wallis test followed by Dunn's multiple comparison test if data were non-parametric. Comparison between experimental groups in the percentage of mice showing haematuria and proteinuria at each time-point was performed using Fisher's exact test. Cut-off for significance was set at P < 0Á05.

A hyper-TGRL environment enhances renal damage during ANTN
To define the effect(s) of high TGRL levels on the progression of renal disease and end-organ damage we induced ANTN, a well-established experimental model of immune complex (IC)-mediated renal injury [22], in B6 mice that had been rendered hyperlipidaemic using P-407. B6 mice, treated with PBS or P-407, were pre-sensitized with sheep IgG after 9 days and received NTS at day 14 ( Fig. 1a). At this time-point (day 14) all the animals exhibited very elevated levels of TG (6497 6 848.6 mg/dl, mean 6 s.e.) and a less pronounced increase in their CHOL levels (911Á5 6 85Á27 mg/dl, mean 6 s.e.) when compared to the PBS-injected group (70Á21 6 8Á77 mg/dl and 101Á4 6 5Á91 mg/dl, mean 6 s.e., respectively; P < 0Á001, unpaired t-test) (Fig. 1b,c). However, despite the changes in the lipid profile, the animals showed no evidence of renal impairment (normal urea levels and absence of haematuria and proteinuria on urinalysis) ( Fig. 1d and data not shown). The P-407 treatment, however, can trigger an accumulation of CD68 1 cells in the kidneys, especially in the glomeruli [19].
Two weeks after the administration of NTS (day 28 of P-407 treatment), haematuria and uraemia were increased significantly in the P-407-treated mice compared to controls, while proteinuria did not differ ( Fig. 1d-f). However, the development of proteinuria occurred earlier in the P-407-injected mice, i.e. 2 days following NTS administration (Supporting information, Fig. S1). In the hyper-TGRL mice there was enhanced glomerular cellularity (Fig. 1g), and immunohistochemical staining for CD68 revealed a significant glomerular accumulation of CD68 1 cells (Fig.  1h,i), but not of neutrophils (Fig. 1j).
We next examined glomerular C3 deposition. Unexpectedly, glomerular C3 deposition was significantly lower in the P-407-treated mice when compared to the PBS-injected controls (P < 0Á001) (Fig. 2a,b). The difference in glomerular C3 deposition was not due to differences in circulating C3 levels, as these were comparable between the experimental groups throughout the course of the ANTN model (Fig. 2c). The reduced C3 staining was associated with markedly decreased glomerular IgG deposition in the P-407-treated   mice compared to the controls (P < 0Á01) (Fig. 2a,d). To explore whether the increased renal damage in the hyper-TGRL mice was due to an abnormal innate immune response to NTS, we administered NTS to PBS-and P-407treated mice without prior immunization with sheep IgG. Both groups showed enhanced proteinuria 24 h after the administration of NTS compared to their respective non-NTS-injected controls. However, the amount of proteinuria was not different between the hyper-TGRL and the normolipidaemic mice (Supporting information, Fig. S2). We then tested lower doses of P-407 (5 mg and 2Á5 mg), which resulted in lower levels of TG (3298 6 157Á2 and 965Á7 6 117Á8 mg/dl, respectively, mean 6 s.e.) and CHOL (395Á3 6 20Á01 and 265 6 15Á35 mg/dl, mean 6 s.e.) (Supporting information, Fig. S3a,b). At these lower levels of TG we found no increased glomerular damage (Supporting information, Fig. S3d-f), indicating that TG levels greater than $4000 mg/dl were required to aggravate the renal damage triggered by the ANTN model.
To investigate the mechanisms driving the hypersensitive response to ANTN seen with the P-407-treated mice, we measured a panel of cytokines and chemokines. As shown in Fig. 3, at day 28 the plasma levels of some inflammatory cytokines such as interleukin (IL)-1b, IL-6 and tumour necrosis factor (TNF)-a were higher in P-407-treated animals compared to PBS-injected controls at the end-point. There was also an increase in CCL2, but not in other chemokines. Together these data demonstrate that a TGRL rich environment primes the kidney for enhanced injury in response to ANTN. This effect seems to be unrelated to IC deposition, complement activation or neutrophil infiltration, but is accompanied by an increased CD68 1 cell infiltration and an enhanced inflammatory response.

LDL-driven hypercholesterolaemia does not promote renal injury following ANTN
The P-407-induced hyper-TGRL state was accompanied by increased CHOL levels; thus, using this model we could not rule out a potential contribution of CHOL to the kidney insult. To explore the role of CHOL in this model we induced ANTN in a model of dyslipidaemia characterized by predominant hypercholesterolaemia : low-density lipoprotein receptor-deficient (Ldlr -/-) mice on a HF diet. Ldlr -/and B6 mice were fed a HF diet for a period of 28 days, during which ANTN was induced (Fig. 4a). At baseline, the plasma levels of both CHOL and TG were generally higher in the Ldlr -/animals than in the B6 mice prior to any dietary intervention (Fig. 4b,c). Fourteen days after the commencement of HF diet, the HF-fed Ldlr -/group showed significantly higher plasma levels of TG and CHOL (450Á7 6 39Á43 mg/dl and 873Á0 6 34Á03 mg/dl, respectively) compared to the HF-fed B6 group (75Á12 6 7Á028 mg/dl and 145Á3 6 16Á46 mg/dl) (Fig.  4b,c). Notably, the hypercholesterolaemia achieved by diet modification in the Ldlr -/mice was comparable to that observed using the 10 mg P-407 dose (Fig. 1c), whereas the TG levels were much lower (Fig. 1b). At the end of the experiment the Ldlr -/-HF group had higher urea levels, but the uraemia was still within the normal range (below 20 mmol/l) (Fig. 4d). Haematuria, proteinuria, glomerular cellularity and glomerular CD68 1 cell numbers did not differ between the Ldlr -/-HF and B6 HF mice (Fig. 4e-i). Of note, in the HF-fed Ldlr -/mice, as in the P-407-treated animals, we detected an enhanced cytokine response (Fig. 5). Thus, the Ldlr -/-HF mice, despite cholesterol levels comparable to the P-407 treated animals, did not display a hypersensitive response to ANTN, suggesting a more detrimental role for hypertriglyceridaemia than hypercholesterolaemia in the renal injury.

Discussion
Elevated TGRL levels are recognized as a risk factor for cardiovascular disease [13,25,26], while their contribution to renal disease remains poorly understood [27,28]. Here we show that a hyper-TGRL state can be detrimental to the kidney. Animals with high TGRL levels displayed a hypersensitive response to renal injury triggered by the ANTN model with increased haematuria and glomerular cellularity. The increased renal damage in the hyper-TGRL mice was accompanied by marked accumulation of CD68 1 cells in the glomeruli implicating a role for these cells in mediating the injurious effects. By comparing animals with different TG levels but comparable CHOL levels, our data indicate that a hyper-TGRL status is more detrimental to the kidneys than LDL-driven hypercholesterolaemic conditions.
Dyslipidaemia and renal disease are commonly coencountered in clinical practice. Dyslipidaemia may occur as a complication of the primary renal disease itself, as seen in the nephrotic syndrome [1], or may precede its development as part of metabolic (e.g. MS and T2D) or inflammatory (e.g. SLE) conditions increasing the risk for renal complications [5,10]. The lipid profile seen in these conditions is dominated by hypertriglyceridaemia due to an accumulation of TGRLs. Several human studies have linked such a lipid profile to renal disease [29][30][31]. However, establishing the precise role of TGRLs in renal disease has been hindered by other confounding factors. The P-407 compound can induce a hyper-TGRL state in the absence of hyperinsulinaemia and/or hyperglycaemia, the most common confounding factors complicating other hypertriglyceridaemic animal models [32]. In addition, P-407 treatment has been shown to have no hepatotoxicity [33,34] nor to cause overt inflammation [19]. Here we show that the P-407-induced hyper-TGRL environment can prime the kidney for renal injury following a local insult. Although the detrimental effects of a hyper-TGRL state were observed mainly under extremely high TGRL levels, our findings cannot exclude harmful effects by lower TGRL levels. In our study the mice were exposed to a hyper-TGRL environment only for a short period (28 days), whereas patients with MS and/or T2D suffer from a hyper-TGRL state for years. Thus one could speculate that a mild hyper-TGRL environment could still be detrimental, but a longer period may be required for the consequences to be manifest.
Macrophages appear to play a role in mediating the enhanced inflammatory response to ANTN under hyper-TGRL conditions, as their numbers were greatly increased in the glomeruli of P-407-treated mice compared to their normolipidaemic controls. We have reported migration of the 'non-classical' Gr1 low monocytes into tissue with enhanced accumulation of CD68 1 macrophages in different organs, including the kidneys, after 2-week treatment with P-407, a time-point that corresponds to the time of NTS administration in the current study, but their presence was not associated with detectable renal injury [19]. However, the data reported here using the ANTN model indicate that the tissue CD68 1 macrophages might amplify the inflammatory response triggered by an insult. Therefore, what may have started as a beneficial physiological reaction aiming to maintain tissue homeostasis by clearing intrarenal lipids during hyper-TGRL conditions may become detrimental once its physiological limits are exceeded, especially during inflammation.
A role for lipids in promoting renal damage has been reported. The lupus-prone (NZB3W)F 1 mice developed accelerated renal disease with higher proteinuria and increased mortality when fed a HF diet [7]. Sprague-   Dawley (SD) rats fed a high CHOL diet developed proteinuria 10 weeks after diet modifications, and this was also associated with enhanced macrophage accumulation [9]. In both studies, the hyperlipidaemia was dominated by hypercholesterolaemia rather than hypertriglyceridaemia [7,9]. Here we show that not only hypercholesterolaemia, but also hypertriglyceridaemia and a hyper-TGRL environment, are harmful to the kidneys. This is consistent with observations in the hypertriglyceridaemic obese Zucker rats that develop focal and segmental glomerulosclerosis with proteinuria and impaired creatinine clearance associated with renal macrophage accumulation as they age [35,36]. However, the obese Zucker rat model is also known to have insulin resistance with or without hyperglycaemia, both risk factors for renal disease [37]. In contrast, the P-407 model generates a hyper-TGRL state in the absence of these confounding factors [32]. This enabled us to separate the effects of TGRLs from those of insulin resistance and hyperglycaemia. Although the findings in the P-407-treated mice and in the obese Zucker rats are consistent, hypertriglyceridaemic Nagase analbuminaemic rats (NAR) did not display signs of renal injury and macrophages were not detected in their kidneys at 8 months of age [9]. The discrepancy with our data may be because the increase in TGs in the NAR model is much lower than that achieved by P-407 administration, which corroborates the notion of a threshold effect. Moreover, although the hypertriglyceridaemia seen in the NAR is secondary to VLDL accumulation, the lipid profile seen in NAR is also characterized by high HDL levels [38]. As the HDL lipoprotein fraction is known to infer protection against atherosclerosis, we speculate that a similar protective effect may have occurred in the kidney in this model [39].
Feeding HF diet to Ldlr -/mice results in hyperlipidaemia, driven largely by hypercholesterolaemia secondary to LDL accumulation with mild hypertriglyceridaemia. Despite the increased CHOL levels we could not detect any significant renal damage or accumulation of CD68 1 macrophage in the kidneys of the HF Ldlr -/mice, despite a comparable cytokine/chemokine profile with the P-407treated mice. The similar systemic inflammatory response, however, does not exclude a different effect of TGRLs and CHOL on renal injury. It is known that some fatty acids, by-products of TGRL degradation and cholesterol bind to different nuclear receptors, which can modulate distinct downstream effector functions [40]. Thus, the divergence of the glomerular response in these two dyslipidaemic models may be due to different lipid metabolites or other signalling pathways. Our findings in the HF Ldlr -/mice also differ from those reported in high CHOL diet-fed SD rats where both renal damage and macrophage accumulation were seen [9]. In addition to the difference in the duration of hyperlipidaemia between the two studies, the hypercholesterolaemia observed in the Ldlr -/mice is secondary to LDL accumulation, whereas in the high CHOL-fed SD rats it is due primarily to VLDL and IDL accumulation [9]. The differences in the lipid profile, along with the duration of hypercholesterolaemia between the two models, may explain their contrasting renal phenotype.
In summary, our data demonstrate that a hyper-TGRL environment primes the kidney to enhanced damage in the ANTN model. These observations may have important clinical implications. The P-407-induced hyper-TGRL lipid profile recapitulates the profiles seen in a variety of metabolic and inflammatory conditions such as MS, T2D and SLE, where renal involvement is well documented. Moreover, as Zilversmit has long advocated a role for the postprandial hyper-TGRL state in the development of atherosclerosis [13], our study indicates that this role may extend to inflammatory kidney conditions.

Supporting information
Additional Supporting information may be found in the online version of this article at the publisher's web-site: Fig. S1. Haematuria and proteinuria in hyper-triglyceride-rich lipoprotein (TGRL) and normolipidaemic mice during the accelerated nephrotoxic nephritis (ANTN) model. Fig. S2. Proteinuria 24 h after the administration of the nephrotoxic serum.