Antagonism of scavenger receptor CD36 by 5A peptide prevents chronic kidney disease progression in mice independent of blood pressure regulation

Abstract

Scavenger receptor CD36 participates in lipid metabolism and inflammatory pathways important for cardiovascular disease and chronic kidney disease (CKD). Few pharmacological agents are available to slow the progression of CKD. However, apolipoprotein AI-mimetic peptide 5A antagonizes CD36 in vitro. To test the efficacy of 5A, and to test the role of CD36 during CKD, we compared wild type to CD36 knockout mice and wild type mice treated with 5A, in a progressive CKD model that resembles human disease. Knockout and 5A-treated wild type mice were protected from CKD progression without changes in blood pressure and had reductions in cardiovascular risk surrogate markers that are associated with CKD. Treatment with 5A did not further protect CD36 knockout mice from CKD progression, implicating CD36 as its main site of action. In a separate model of kidney fibrosis, 5A-treated wild type mice had less macrophage infiltration and interstitial fibrosis. Peptide 5A exerted anti-inflammatory effects in the kidney and decreases renal expression of inflammasome genes. Thus, CD36 is a new therapeutic target for CKD and its associated cardiovascular risk factors. Peptide 5A may be a promising new agent to slow CKD progression.

Introduction

Chronic kidney disease (CKD) affects about 11.5% of people in the United States [1, 2] and is associated with high morbidity and mortality. Increasing rates of diabetes and hypertension contribute to its increasing prevalence worldwide [3]. CKD patients have an increased prevalence of hypertension and dyslipidemia, are at higher risk for cardiovascular disease, and have higher adjusted all-cause hospitalization and mortality rates [1]. Few pharmacological interventions are available to slow CKD progression. Renin-angiotensin-aldosterone system (RAAS) inhibitors are standard-of-care in CKD, but progression to ESRD is common despite treatment [4]. As several recent approaches to slow CKD progression have failed [5–7], there is urgent need to discover novel therapeutics to slow CKD progression, which would benefit millions of individuals worldwide.

CD36 is a widely expressed cell surface class B scavenger receptor that recognizes a large variety of ligands, including apoA-1, apoL-1, thrombospondin-1, serum amyloid A (SAA), lipopolysaccharide (LPS), oxidized low-density lipoprotein (oxLDL), fatty acids [8], apoptotic cell surfaces, and microparticles [9–12]. CD36 contributes to foam cell formation and accumulation during atherosclerosis, and is upregulated by oxLDL [13]. CD36 binds with high affinity to a structurally conserved family of oxidized phosphor-choline glycerophospholipids (oxPC_CD36) that are enriched in atherosclerotic lesions [14,15]. It has been speculated that individuals with increased plasma oxPC_CD36 may be at an increased risk for atherosclerotic lesion development and progression [14,15]. Current studies suggest a role for CD36 in mediating cardiovascular disease associated with type II diabetes, insulin resistance, and obesity in rodent models [16] and in hypertensive humans [17]. Human genetic studies also support a role for CD36 in dyslipidemia, obesity and metabolic diseases [18–20], and CD36 protein expression is increased in kidney proximal convoluted tubule cells (PCT) in patients with proteinuria [21].

CD36 deficiency protects against atherosclerosis in ApoE-knockout (KO) mice fed a high-fat diet [22]. In this study we demonstrate that 5A, an ApoA-I mimetic peptide that has been used to attenuate atherosclerosis, acute vascular inflammation, and oxidative stress in experimental animal models [23–26], is a ligand for CD36 in vitro using transfected HeLa cells that overexpress CD36. We further tested 5A effects in vivo using an animal model of CKD that highly resembles the human disease. 5A-treated mice were protected from CKD progression, similar to CD36 KO mice. 5A also had beneficial effects in a second in vivo model of renal fibrosis. 5A-treated mice had decreased renal expression of inflammation and inflammasome-associated genes.

Results

Competition of 5A peptide and known CD36 ligands with Alexa-Fluor488 oxLDL in CD36-overexpressing HeLa cells

The potency of known CD36 ligands, including oxPC_CD361 and oxPC_CD362, cold unlabeled oxLDL, and synthetic peptide 5A, as inhibitors of Alexa-Fluor-488 oxLDL uptake was analyzed using CD36-overexpressing HeLa cells. Data in figure 1 show that unlabeled oxLDL competed with Alexa-Fluor-488 oxLDL in a dose-dependent manner. OxPC_CD361 and 2 also efficiently competed with labeled oxLDL. 5A peptide efficiently competed with labeled oxLDL, in contrast to control peptide L3D-37pA that did not have any inhibitory effect upon Alexa-Fluor-488 oxLDL uptake in CD36-overexpressing HeLa cells (Figure 1). Calculated IC₅₀ and _appK_i for 5A were 7.5 and 1.45 µM, respectively. In a small set of mice that received Alexa-Fluor488 labeled 5A peptide through osmotic mini-pump for five days (5mg/kg/day), 5A levels were not detectable in plasma, while it reached high levels in both urine and kidney tissue (Supplementary figure 1). Three hours after intravenous injection of Alexa-Fluor488-labeled 5A peptide, 5A is seen within the kidney using 2-photon microscopy, most likely within renal tubular epithelial cells (Supplementary figure 1).

Figure 1. Competition of CD36 ligands with Alexa-Fluor-488 oxLDL binding to CD36-overexpressing HeLa cells and CD36 antagonism by 5A peptide in vitro.

Cells were incubated with 2 nM Alexa-Fluor-488 oxLDL with indicated concentrations of unlabeled competitors (X-axis) for 2 h at 37 °C. Unlabeled oxLDL was used as a control. Cell-associated fluorescence was measured by a plate reader.

CD36KO mice, and WT mice treated with 5A, are protected against CKD progression after 5/6 nephrectomy with angiotensin II infusion

Because CD36KO mice are on C57BL/6 background and the C57BL/6 strain is resistant to CKD progression following 5/6Nx, we used the 5/6Nx plus continuous angiotensin II (AngII) infusion model to induce progressive CKD in this strain [27]. C57BL/6 WT mice subjected to 5/6Nx without AngII infusion did not develop CKD and were used as controls. Four weeks after 5/6Nx, WT mice that received continuous infusion of AngII had a substantial decline in kidney function, with significant elevation in BUN (Figure 2a), creatinine (Figure 2b), and a progressive rise in urinary albumin-to-creatinine-ratio (ACR) (Figure 4c), accompanied by renal histological damage [glomerulosclerosis (Figure 5a) and interstitial fibrosis (Figure 5b). CD36KO mice subjected to 5/6Nx+AngII were significantly protected from the decline in kidney function at 4 weeks, with BUN, serum creatinine, and histological injury similar to controls. This group also had significantly reduced levels of albuminuria compared to WT mice. WT mice subjected to 5/6Nx+AngII that received continuous infusion of CD36 antagonist peptide 5A by osmotic minipump (5 mg/kg/day) were also significantly protected from CKD progression, and had albuminuria levels similar to KO mice (Figure 4c).

Figure 2. Effects of CD36KO and 5A peptide therapy on CKD progression.

After 4 weeks, untreated WT mice subjected to 5/6Nx+AngII had increased levels of BUN (a) and serum creatinine (b), and (c) urinary albumin-to-creatinine ratio (ACR). These changes were reduced in KO and 5A-treated WT mice [(N=11–19/group; ANOVA with Dunnett's post-hoc test for a–d; N=11–19/group for e, mixed-model analysis (see Experimental Procedures)].

Figure 4. Effects of CD36KO and 5A peptide therapy on metabolic and electrolyte abnormalities associated with CKD.

Four weeks after 5/6Nx+AngII, WT mice had typical metabolic and electrolyte changes associated with CKD, which were prevented in CD36KO and 5A-treated WT mice. 5A did not cause liver toxicity. WT mice subjected to 5/6Nx+AngII had very high serum FGF-23 and intact PTH levels, which were significantly lower in CD36KO mice. (N=8–10/group; ANOVA with Dunnett's post-hoc test for a–i; N=6–13/group; ANOVA with Dunnett's post-hoc test for j and k).

Figure 5. Effects of CKD and CK36KO on telemetry blood pressure.

WT and CD36KO mice (N=6/group) were subjected to 5/6Nx+AngII and telemetry. Model-based weekly mean values: systolic (a), diastolic (b), and mean arterial blood (c) pressures (Y-axis) significantly increased over time (X-axis), from baseline to 4 weeks, p<0.001, in both groups, without statistical differences between the groups.

CD36KO mice, and WT mice treated with 5A, are protected from metabolic complications of CKD, including risk factors for CVD

At 4 weeks, WT mice subjected to 5/6Nx+AngII developed a metabolic profile typical of patients with CKD, including hypercalcemia (Figure 4a), hyperphosphatemia (Figure 4b), hypermagnesemia (Figure 4e), dyslipidemia (Figures 4f and 4g), and high serum fibroblast growth factor-23 (FGF-23) and intact PTH (Figures 4j and 4k). FGF-23 is a phosphate regulatory hormone involved in left ventricular hypertrophy, and associated with subclinical and clinical cardiac disease [28]. High serum phosphate and FGF-23 levels are independent risk factors for cardiovascular disease, particularly among CKD patients [29,30]. CD36KO and WT mice treated with 5A subjected to 5/6Nx+AngII had a more normal metabolic profile, similar to the control group that did not develop progressive CKD. Mice treated with 5A did not have liver toxicity, as assessed by serum transaminases (Figures 4h and 4i).

CD36KO mice are protected from CKD progression independent of blood pressure

Blood pressure was monitored weekly in conscious WT and KO mice subjected to 5/6Nx+AngII (N=6 /group) by radiotelemetry for 24 h periods. WT and KO mice subjected to 5/6Nx+AngII had similar baseline (before surgeries) blood pressure values. Systolic, diastolic, and mean blood pressure significantly (p<0.001) and gradually increased over time in both WT and KO mice without statistical differences between the two groups (Figure 5). Circadian patterns, where blood pressure was increased during nighttime and decreased during daytime (typical of any nocturnal animal), were not different between the two groups of mice (not shown).

Peptide 5A prevented up-regulation of renal inflammation- and inflammasome-associated genes in a progressive CKD model

A second set of experiments was performed to obtain remnant kidney tissue sufficient to measure renal mRNA expression of genes associated with inflammation and inflammasome. This was necessary because histology consumed the entire kidney tissue in the first experiment. As before, WT mice subjected to 5/6Nx+AngII developed CKD while 5A-treated WT mice were protected (Supplementary figure 2). Four weeks after 5/6Nx+AngII, renal gene expression of pro-inflammatory genes IL-6 and CXCL-1 were decreased in WT mice subjected to 5/6Nx+AngII and treated with 5A (Figure 6). 5A-treated WT mice subjected to 5/6Nx+AngII also had significantly lower expression of 1L-1β and NLRP3 genes in the remnant kidney (Figure 6).

Figure 6. Effects of 5a on renal mRNA expression of cytokines and NLRP3 on the progressive CKD model.

Four weeks after 5/6Nx+AngII renal mRNA expression of IL-6 (a), CXCL-1 (b), Il-1β (e) and NLRP3 (f) in WT mice treated with 5A was significantly decreased in comparison to non-treated mice. (N=4–6 /group; ANOVA with Dunnett's post-hoc test).

5A peptide prevents CKD progression in vivo via the CD36 receptor

To better understand the relationship between CD36 and 5A peptide in vivo, twenty CD36KO mice were subjected to 5/6Nx+AngII. Ten mice received concomitant 5A infusion through osmotic mini-pump, while ten mice did not. After 4 weeks CD36KO mice treated with 5A did not have any further improvement on renal outcomes compared to CD36KO mice that did not receive 5A Figure 7).

Figure 7. Relationship between CD36 receptor and 5A in the progressive CKD model.

Four weeks after 5/6Nx+AngII, CD36 KO mice that received 5A (5mg/kg/day) did not have better renal outcomes as analyzed by serum BUN and creatinine, renal histology, and albumin-to-creatinine ratio in urine in comparison to CD36 KO subjected to the same model without 5A infusion. (N=4–6 /group; ANOVA with Dunnett's post-hoc test).

5A treatment prevents renal macrophage infiltration and interstitial fibrosis in the unilateral ureteral obstruction (UUO) model

Renal interstitial fibrosis is a histological hallmark of CKD, and a strong predictor of renal functional loss in patients [31]. We tested the effects of targeting CD36 with 5A peptide in a renal fibrosis model (UUO) that does not require AngII infusion using a different strain (CD-1 mice). While mice with progressive CKD develop hypercholesterolemia on a regular diet, mice do not develop dyslipidemia with the UUO model (Figure 9). Because CD36 can scavenge oxLDL and oxHDL [32], alleviating lipid/cholesterol imbalance by CD36 inhibition cannot account for the anti-fibrotic effect in the UUO model. After UUO, renal vascular resistance increases, reducing renal blood flow to the obstructed kidney [31], and distribution and metabolism of therapeutic agents become difficult. Therefore, we started 5A infusion 24 h before surgery using an osmotic minipump, at two doses: 5 and 15 mg/kg/day. 5A-treated mice were protected from interstitial fibrosis at the higher dose (15 mg/kg/day) (Figures 8b), and had less kidney damage, as measured by cortical thickness (Figure 8a). 5A treatment prevented F4/80⁺ macrophage infiltration in the obstructed kidney, even at the lower 5A dose (Figure 8c). Ten days after UUO, the obstructed kidney had a higher gene expression of pro-inflammatory cytokines, and 5A peptide treatment decreased this elevation, but not significantly (Figure 9). 5A-treatment at a higher dose (15 mg/kg/day) significantly prevented increases in gene expression of TGF-β1 and IL-1β (Figure 9).

Figure 9. Effects of 5A therapy on renal mRNA expression of cytokines and NLRP3 in the UUO model.

Renal mRNA expression of cytokines and inflammasome-associated genes. TGF-β and IL-1β gene expression are decreased in the 5A 15 group. (N=4 /group; ANOVA with Dunnett’s post-hoc test).

Figure 8. Effects of 5A therapy on kidney fibrosis, macrophage infiltration (F4/80⁺ cells) in the UUO model.

Ten days after UUO, untreated WT CD-1 mice had substantial cortical thinning (a), interstitial fibrosis (b), and macrophage infiltration of the obstructed kidney (c); 5A-treated mice (15 mg/kg/day) were protected. (d and e) Representative pictures of obstructed kidney sections stained with Masson`s-trichrome from all groups demonstrating increased cortical thinning (d; 1.5×), and interstitial fibrosis (e; 400×) in the untreated UUO group. (N=8 /group; ANOVA with Dunnett’s post-hoc test).

CD36 receptors are localized in proximal convoluted tubule cells

CD36 receptor is expressed in PCT [21, 33] and in renal interstitial macrophages [33, 34]. Using kidney sections from a KO mouse as a negative control, we confirmed by IHC that CD36 is expressed in proximal convoluted epithelial cells (Supplementary figure 4).

Discussion

In this study we demonstrate that 5A peptide binds to and competes for CD36 in CD36-overexpressing HeLa cells in vitro. In vivo, both genetic and pharmacologic inhibition of CD36reduce CKD progression and its associated cardiovascular risk factors without changes in blood pressure. 5A exhibits anti-inflammatory effects in the kidney, in part, via decreased renal expression of inflammasome genes NLRP3 and IL-1β. Our data suggest that the impact of CD36 on CKD progression is independent of blood pressure and likely involves more direct anti-inflammatory and/or anti-fibrotic effects.

WT mice subjected to 5/6Nx+AngII infusion developed progressive CKD with increasing levels of albuminuria and blood pressure over time. After 4 weeks these mice had a substantial decline in kidney function with renal histological damage. All these changes were prevented in CD36KO and 5A-treated mice. 5A treatment did not further protect CD36KO mice subjected to 5/6Nx+AngII, suggesting that 5A acts mainly through CD36 receptor during CKD progression. In the unilateral ureteral obstruction (UUO) model, 5A treatment prevented interstitial fibrosis and cortical thinning of the obstructed kidney. Therefore, 5A had similar renal protective effects in two different models of chronic kidney damage and fibrosis.

Other groups have also demonstrated that CD36 has a substantial role in kidney fibrosis. In mice fed a high-fat diet and subjected to UUO, strong CD36 tubular expression was induced in the obstructed kidney [28]. Further, CD36KO mice fed high-fat diet for 8 weeks and then subjected to UUO have significantly less renal fibrosis compared with WT mice [35]. Now we demonstrate in a progressive model of CKD that targeting CD36–-either genetically by using a KO mouse line, or pharmacologically, by using the 5A peptide–-reduces functional deterioration. As in patients [36], CKD also leads to lipid metabolism changes in mice [37, 38]. Unlike previous work, dyslipidemia gradually developed over time in association with CKD, despite standard chow, and was lessened in CD36KO and 5A-treated mice. We also tested 5A in a UUO model using mice fed standard diet. Ten days after sham-surgery or UUO, all CD-1 mice (treated or non-treated with 5A) had similar cholesterol levels (Supplementary figure 3). 5A-treated (15 mg/kg/day) mice subjected to UUO had significantly less kidney fibrosis than non-treated mice, without altering cholesterol levels; thus, CD36 affects CKD beyond its role as a lipid scavenger receptor. While the role of CD36 during renal fibrogenesis may not depend on circulating cholesterol levels, it is possible that targeting CD36 during CKD progression in the 5/6Nx+AngII model may impact lipid metabolism, or that treated mice had less dyslipidemia because they had less CKD. Targeting CD36 in the progressive CKD model not only prevented changes in lipid metabolism but also in bone-mineral disorders that occur in parallel to CKD.

Renin-angiotensin-aldosterone system (RAAS) inhibitors, standard-of-care in CKD, decrease both albuminuria and renal fibrosis in parallel [27]. Despite the reduction in fibrosis in both models, in the 5/6Nx+AngII model there was only a partial decrease in ACR in both CD36KO and 5A-treated mice. This partial reduction in albuminuria may be explained by CD36 participating, together with megalin and cubilin [21], in albumin binding and endocytosis in renal proximal tubular cells. Blocking CD36 may reduce the uptake of filtered albumin by the proximal tubule; thus, targeting CD36 prevents CKD progression without dramatic changes in albuminuria. CD36 mediates albumin-induced cellular fibrosis in cultured proximal tubule cells [39]. Thus, inhibiting CD36 might have two competing effects on albuminuria: a) a decrease caused by ameliorating progressive CKD and b) an increase from impaired albumin reuptake. The net change, a partial decrease in albuminuria, appears to underestimate the benefit of CD36 inhibition on CKD, as CD36 inhibition prevents albumin-induced damage to the proximal tubule. Targeting CD36 may be an adjuvant therapy to RAAS inhibitors that could be more effective than monotherapy because of non-redundant mechanisms.

We have previously demonstrated that olmesartan (an AngII type-I receptor antagonist) reduces blood pressure, albuminuria, and CKD progression in CD-1 mice subjected to 5/6Nx, whereas mice treated with hydralazine, which similarly reduced blood pressure, did not improve CKD progression [27]. We again dissociate blood pressure and CKD progression: inhibiting CD36 genetically decreases CKD progression without changing blood pressure. Despite similar mean, systolic, or diastolic hypertension in both WT and CD36KO mice subjected to 5/6Nx+AngII, KO mice were protected from CKD whereas WT were not. Thus, we conclude that CD36 inhibitory effects on CKD progression are independent of systemic blood pressure. In addition to hypertension and dyslipidemia, CKD mice also manifested other risk factors for cardiovascular disease and vascular injury in the progressive CKD model, such as high serum FGF-23 levels and hyperphosphatemia, which were prevented in 5A-treated mice and in CD36KO mice.

Chronic inflammation is important in the progression of CKD, and CD36 activation is associated with downstream activation of inflammation [40, 41] and inflammasome [13, 42, 43, 44]. We found that 5A decreases renal cytokine mRNA levels in both models, and decreases macrophage infiltration 10 days after UUO. Pennathur et al. recently demonstrated that chimeric mice with CD36-deficient bone marrow-derived cells, including macrophages, are protected from renal fibrosis following unilateral ureteral obstruction or ischemia reperfusion injury. Along with in vitro work showing profibrotic, CD36-dependent signaling in macrophages isolated after UUO, the authors suggest that macrophage CD36 is a critical regulator of renal fibrogenic signaling [45]. Follow-up studies are necessary to further understand the direct and/or indirect impact of CD36 on renal epithelial cells during fibrosis / CKD progression. Besides renal effects, CD36 inhibition has been shown to decrease inflammation and fibrosis in other organs: CD36 antagonist peptide 5A exerts anti-inflammatory effects in lungs [24, 46], and reduces pulmonary macrophage infiltration in a model of dust mite-induced asthma [46]. Further, silencing the CD36 gene results in suppression of silica-induced lung fibrosis in rats [17]. While there was some evidence that 5A impacts inflammation and fibrosis on lungs, our study is the first to demonstrate that 5A effectively ameliorates the deterioration of kidney function in a clinically relevant model of CKD that replicates several functional and metabolic changes that occur in CKD patients.

In summary, our study suggests that CD36 is a new therapeutic target and 5A is a new potential therapeutic agent to slow CKD progression, however, we did not identify the ligand(s) for CD36 in CKD and cannot rule out actions of 5A through other receptors. CD36 inhibition might complement RAAS inhibition through non-overlapping effects, possibly allowing synergistic combination therapy. Targeting CD36 should be examined as a possible alternative or second line therapy for CKD, especially when RAAS inhibitors are contra-indicated.

Methods

Cell Cultures

HeLa (Tet-Off) cells were transfected with FuGENE 6 (Roche Diagnostics) using the expression plasmid pTRE2 (Clontech), encoding a human CD36 protein (pTRE2-CD36). Cells were co-transfected with pTRE2-hCD36 and pTK-Pur (Clontech), using a 1:20 ratio, and selected with 400 µg/ml puromycin. Puromycin-resistant cells were screened for the expression of the human CD36 protein utilizing mouse anti-hCD36 antibody (Abcam, Cambridge, MA) in Western blotting. Cells were cultured in Dulbecco's modified Eagle's medium, supplemented with 10% fetal calf serum, 2 mm glutamine, 100 units/ml penicillin, 100 µg/ml streptomycin, and 20mM HEPES at 37 °C in a 5% CO₂ humidified atmosphere.

Synthesis of peptides

5A is a 37-residue amphipathic peptide (DWLKAFYDKVAEKLKEAF-P-DWAKAAYDKAAEKAKEAA) that contains a proline between the two amphipathic helices. The peptide contains a high lipid-affinity helix paired with a low lipid-affinity helix with 5 alanine substitutions. 5A peptide was synthesized by a solid-phase procedure using Fmoc/DIC/HOBt chemistry and purified to > 99% by reverse-phase high-performance liquid chromatography (HPLC) as previously described [47]. Purity was assessed by MALDI-TOF-MS (Bruker Ultraflex) [48]. Control L3D-37pA peptide (DWLKAFYDKVAEKLKEAFPDWLKAFYDKVAEKLKEAF) was synthesized by the same solid-phase procedure and used for in vitro experiments.

Alexa 488-labeled ligand uptake and competition experiments

Human LDL was purchased from EMD Biosciences (San Diego); oxidized LDL (oxLDL) was prepared by incubation with 5 µm CuSO₄ at 37 °C for 24 h as described previously [49]. Oxidized LDL was conjugated with Alexa Fluor® 488, using a protein labeling kit (Molecular Probes, Eugene, OR) following the vendor's instructions. The Alexa-labeled preparations were analyzed by SDS-PAGE, using 10–20% Tricine pre-cast gels (Invitrogen). Gels were scanned using a Fluoroscan (model A, Hitachi). All incubations were performed in Dulbecco's modified Eagle's medium (HeLa cells) containing 0.1% bovine serum albumin at 37 °C. Uptake experiments with HeLa cells were performed by using Alexa 488-oxLDL at concentrations between 0.7 and 50 µg/ml, in quadruplicate, in a 96 wells plate. After 2 h of incubation, the cells were rinsed with ice-cold PBS, and read in a fluorescence plate reader (Wallac Victor 1420 Multilabel Counter, PerkinElmer Life Sciences). Competition experiments were performed using 6 µg/ml fluorescence-labeled oxLDL and unlabeled ligands ranging in concentration from 5 to 100 µg/ml (oxPC_CD361 and oxPC_CD362) and from 5 to 500 µg/ml (unlabeled oxLDL and 5A). OxPC_CD361 and oxPC_CD362 were purchased from Cayman Chemical (Ann Arbor, MI), KOdiA-PC catalog numbers 62945 and 62935, respectively. Following a 2 h incubation, HeLa cells were treated as described above. After washing with ice-cold PBS, cell-associated fluorescence was analyzed by a plate reader (Wallac Victor 1420 Multilabel Counter, PerkinElmer Life Sciences).

Animals

The National Institutes of Health (NIH) criteria for laboratory animal care were used in this study. CD36KO mice backcrossed into a C57BL/6 background (over 10 generations) were kindly provided by Dr. Kathryn Moore [40, 42]. CD36KO colony was maintained at a NIH animal facility. WT (CD36^+/+) C57BL/6 mice were obtained from the NCI-DCT Laboratory, Frederick, MD. Sixteen week-old male C57BL/6 mice were used for experiments with the progressive CKD model. Nine week-old male CD-1 mice were obtained from Charles River Laboratory, Wilmington, MA and used for kidney fibrosis experiments. All mice had free access to water and regular chow.

Surgeries

1) Progressive CKD model (5/6 nephrectomy plus angiotensin II infusion) - 5/6 nephrectomy (5/6Nx) was performed in C57BL/6 mice in two stages under isoflurane anesthesia. Via left flank incision, the left kidney was decapsulated to avoid ureter and adrenal damage, and the upper and lower poles were resected. Bleeding was controlled with microfibrillar collagen (Avitene, Davol, Warwick, RI, USA). The upper and lower poles were weighed. After one week, the entire right kidney was decapsulated and removed via right flank incision. Immediately after each surgical intervention, mice received a single dose of buprenorphine (0.1 mg/kg), followed by buprenorphine 0.05 mg/kg 18 h after the procedure. Animals with kidney mass resection [as determined by (removed left kidney weight at week −1) / (removed right kidney weight at week 0) between 0.5–0.65] were used for the study; and others, euthanized [27]. As previously described, C57BL/6 mice are resistant to CKD progression after 5/6Nx, and Angiotensin II (AngII) overcomes this strain-specific resistance [27]. Therefore, mice subjected to 5/6Nx without AngII infusion did not develop progressive CKD after 4 weeks and were considered controls. Mice were divided into 4 groups (N= 11–19 /group): 1. WT 5/6Nx without AngII (control); 2. WT 5/6Nx+AngII (WT); 3. CD36KO 5/6Nx+AngII (KO); 4. WT 5/6Nx+AngII+5A (WT+5A). AngII and 5A were infused through separate osmotic minipumps (see below) starting at the time of right nephrectomy in the progressive CKD model. Mice were followed for 4 weeks, with weekly urine collection, then euthanized. A second set of experiments with 2 groups was performed to obtain kidney tissue for mRNA expression analysis [(N=5–6/group): 1. WT 5/6Nx+AngII (WT); 2. WT 5/6Nx+AngII+5A (WT+5A). To better understand the relationship between CD36 receptor and 5A in vivo, a third set of experiments was performed in CD36KO mice subjected to the same model (5/6Nx+AngII), with one group receiving 5A, while the other group did not: 1. CD36KO 5/6Nx+AngII (KO); 2. CD36KO 5/6Nx+AngII+5A (KO+5A); N=10/group. 2) Kidney fibrosis model (unilateral ureteral obstruction, UUO) - Under isoflurane anesthesia CD-1 mice were subjected to right flank incision and the right ureter was ligated with a double suture below the pelvis and at the end of inferior kidney pole. Immediately post-surgery, mice received a single dose of buprenorphine (0.1 mg/kg), followed by buprenorphine 0.05 mg/kg 18 h after the procedure. Mice were divided in 4 groups (N=8/group): 1. Sham; 2. UUO; 3. UUO+5A 5 mg/kg/day (5A 5); 4. UUO+5A 15 mg/kg/day (5A 15), and euthanized 10 days after sham or UUO. All mice in the UUO study were subjected to implantation of osmotic mini-pump, under isoflurane anesthesia, with sterile water (vehicle) or 5A, twenty-four hours prior to sham or UUO procedures.

Drug administration

AngII (Val5-AngII) 0.75 µg/kg/min (Sigma-Aldrich, St Louis, MO, USA), diluted in sterile water, or vehicle (sterile water) was continuously infused by subcutaneous osmotic mini-pump (Alzet model 1004, Cupertino, CA). 5A was continuously administered through a second osmotic mini-pump, at the dose of 5 mg/kg/day in the progressive CKD model. Osmotic mini-pumps were inserted at the time of right nephrectomy. In the UUO model, osmotic minipumps with 5A were started 24h before UUO (5 or 15 mg/kg/day). To evaluate 5A levels within the kidney after subcutaneous administration via osmotic mini-pump, mice not subjected to any surgical procedure received Alexa 488-labeled 5A through osmotic mini-pump for 5 days (5mg/kg/day) and were compared with mice that received sterile water through osmotic mini-pumps (N= 2–3/group). Five days after mini-pumps implantation mice were euthanized under anesthesia to collect plasma, urine, and kidneys. Samples were read directly, or kidneys were homogenized in 10 mls/g of TPER (Pierce, Rockford, IL, USA) and read using a Fluoroscan model A (Hitachi).

Fluor-labeled peptide uptake

2-photon microscopy was performed on kidneys of a WT mouse 3h after an intravenous bolus injection of Alexa-Fluor-488-labeled-5A peptide (5mg/kg).

Blood and urine measurements

Spot urine samples were collected before (baseline, −2 weeks), and at 1, 2, 3, and 4 weeks after completion of 5/6Nx. At 4 weeks, mice were euthanized under isoflurane anesthesia, and blood was collected by cardiac puncture. The remnant kidney was harvested and fixed in 10% formalin. Urine samples were not collected in the UUO model. Serum creatinine was measured by HPLC [47], and blood urea nitrogen (BUN) by colorimetric assay (QuantiChrom Urea assay kit DIUR-500, Hayward, CA, USA). Urine albumin-to-creatinine-ratio (ACR) was determined from albumin ELISA (Albuwell M; Exocell, Philadelphia, PA) and creatinine by Jaffe method. A biochemistry panel was measured using a Siemens Advia 1800 automated chemistry analyzer (Siemens Healthcare Diagnostics, Flanders, NJ). Serum FGF-23 was measured by ELISA (Immunotopics, Inc, cat#60-6300).

Morphologic evaluation of the kidney

Kidney specimens were fixed in 10% formalin, paraffin embedded (FFPE), and stained with Masson’s trichrome or periodic acid-Schiff (PAS) reagent (Sigma-Aldrich). Histological changes were assessed semi-quantitatively. The degree of glomerular damage was assessed in 10 randomly selected fields at 400× magnification from the degree of mesangial expansion in PAS-stained tissue and scored as follows: 1, <25%; 2, 25–50%; 3, 50–75%; 4, >75%; 5, completely sclerotic glomeruli. Interstitial fibrosis was assessed at 200× magnification on Masson’s trichrome-stained sections using 10 randomly selected fields for each animal and scored by the following criteria: 1, area of damage <10%; 2, 10–25%; 3, 25–50%; 4, 50–75%; and 5, 75–100%.

Immunohistochemistry (IHC) for CD36 and F4/80⁺

FFPE sections were treated for 1 h in an antigen retrieval solution and incubated with a protein blocking solution for 20 min before incubation with the rabbit polyclonal anti-CD36 antibody (Novus, NB400-144) for 1h at 1:1,600. An isotype rabbit IgG Ab was used as a control. Sections were then incubated with a goat anti-rabbit HRP-conjugated Ab for 30 min at 1:200 followed by DAB Chromogen for 5 min. Sections were counterstained, coverslipped and digitally captured. IHC for CD36 was performed on kidney sections from a KO mouse and a WT mouse, both subjected to 5/6Nx+AngII. IHC for F4/80⁺ cells (Ab from AbD Serotec, cat#MCA497GA) was performed in FFPE kidney sections from all UUO groups (N=4/group) and the number of positive cells counted in 10 randomly selected high-power fields (400×).

Radiotelemetry

In 6 mice from each group: [1. WT 5/6Nx+AngII (WT); 2. CD36KO 5/6Nx+AngII (KO)] in the progressive CKD model, a telemeter transmitter (Data Sciences International, St Paul, MN) was implanted in a subcutaneous pocket on the left flank, with the tip of the catheter inserted into the aortic arch (via the carotid artery). The animals were allowed to recover for about 1 week before 5/6Nx. Systolic, diastolic, and mean arterial pressure were measured and recorded for 10 sec every 30 min for 24h at baseline (−2 weeks), and once a week for 4 weeks after completion of 5/6Nx [27].

mRNA isolation and RT-qPCR

Because there was no remnant kidney tissue remaining after histology, a new experiment was performed for mRNA expression in the progressive CKD model with 2 groups (N=4–6 /group): 1. WT 5/6Nx+AngII (WT), and 2. WT 5/6Nx+AngII+5A (WT+5A). In the UUO model mRNA was isolated from mice of all groups (N=4 /group). For RNA isolation, tissue samples preserved in RNAlater (Life Technologies, Grand Island, NY) were homogenized in TRIzol Reagent using a Precellys 24 homogenizer (Bertin Technologies, France). All reagents used for RNA isolation, reverse transcription and real-time PCR were obtained from Life Technologies. RNA was isolated with the PureLink RNA Mini Kit after DNase treatment. RNA (2µg) was reverse-transcribed using a TaqMan Reverse Transcriptase Reagents Kit. Real-time qPCR assays were performed with a StepOne Real-Time PCR System (Applied Biosystems), with 40ng of cDNA per reaction. A list of TaqMan Gene Expression assays used in the study is as follows: (IL-6) Mm00446190_m1, (CXCL-1) Mm04207460_m1, (IL-1β) Mm00434228_m1, (TGF-β1) Mm01178820_m1, (TNF-α) Mm00443258_m1, and (NLRP3) Mm00840904_m1. The relative levels of gene expression were measured by the comparative C_T (ΔΔC_T) method [50] with mouse GAPDH used as a reference gene. All gene expression results were analyzed using 2^−ΔΔC_T formula.

Statistical analysis

Differences in parameter values between the groups at fixed time points were examined for statistical significance by ANOVA, with subsequent post-hoc analysis using Dunnett's adjustment for multiple comparisons to a reference group; nonparametric Kruskal-Wallis tests were examined when ANOVA assumptions were not met. Student`s t-test was used for comparisons between 2 groups. Differences between the groups in weekly mean/sinusoid parameter values (ACR and telemetry data) were analyzed using mixed-effects models that extend repeated measures ANOVA to imbalanced data settings. Analyses were conducted using SAS v9.3 (SAS Institute, Cary, NC), GraphPad Prism version 5.0c (GraphPad Software Inc., San Diego, CA, USA), or R v3.0 (r-project.org). Data were expressed as mean ± SEM, except telemetry data was expressed as mean ± upper and lower confidence intervals.

Supplementary Material

1. Supplementary figure 1. 5A peptide is detected in the kidneys and urine after infusion.

(a) 5A concentrations in plasma, urine, and kidneys after infusion via subcutaneous osmotic minipump. (b) Localization of fluorescently-labeled 5A within the kidney after intravenous injection.

2. Supplementary figure 2. Effects of 5A peptide treatment on biochemistry profiles of mice subjected to progressive CKD model (2^nd experiment for mRNA isolation).

Experiments with 5A-treated and untreated WT mice subjected to 5/6Nx+AngII were performed for renal mRNA expression of cytokines and NLRP3. On this set of experiments, mice that did not receive 5A also had loss of kidney function accompanied by an altered metabolic profile. 5A-treated mice were protected. (N=4–6 /group; ANOVA with Dunnett's post-hoc test).

3. Supplementary figure 3. Effects of 5A peptide treatment on biochemistry profile of mice subjected to UUO.

Ten 10 days after UUO, mice did not have important metabolic or electrolyte changes. Kidney function is not an outcome for the UUO model because the unobstructed kidney compensates for the damaged kidney. (N=8 /group; ANOVA with Dunnett's post-hoc test).

4. Supplementary figure 4. Localization of CD36 in the kidney.

Immunohistochemistry shows CD36 expression on proximal tubular cells of WT mice but not in CD36KO mice (400×).

Figure 3. Effects of CD36KO and 5A peptide therapy on renal histology.

After 4 weeks, untreated WT mice subjected to 5/6Nx+AngII had increased glomerulosclerosis (a) and interstitial fibrosis (b) scores. These changes were reduced in KO and 5A-treated WT mice. On the right, representative pictures at high-power field HPF (400×). [(N=11–19/group; ANOVA with Dunnett's post-hoc test for a–d; N=11–19/group).