The AGE receptor, OST48 drives podocyte foot process effacement and basement membrane expansion (alters structural composition)

Abstract Aims The accumulation of advanced glycation end products is implicated in the development and progression of diabetic kidney disease. No study has examined whether stimulating advanced glycation clearance via receptor manipulation is reno‐protective in diabetes. Podocytes, which are early contributors to diabetic kidney disease and could be a target for reno‐protection. Materials and methods To examine the effects of increased podocyte oligosaccharyltransferase‐48 on kidney function, glomerular sclerosis, tubulointerstitial fibrosis and proteome (PXD011434), we generated a mouse with increased oligosaccharyltransferase‐48kDa subunit abundance in podocytes driven by the podocin promoter. Results Despite increased urinary clearance of advanced glycation end products, we observed a decline in renal function, significant glomerular damage including glomerulosclerosis, collagen IV deposition, glomerular basement membrane thickening and foot process effacement and tubulointerstitial fibrosis. Analysis of isolated glomeruli identified enrichment in proteins associated with collagen deposition, endoplasmic reticulum stress and oxidative stress. Ultra‐resolution microscopy of podocytes revealed denudation of foot processes where there was co‐localization of oligosaccharyltransferase‐48kDa subunit and advanced glycation end‐products. Conclusions These studies indicate that increased podocyte expression of oligosaccharyltransferase‐48 kDa subunit results in glomerular endoplasmic reticulum stress and a decline in kidney function.


| INTRODUC TI ON
There is a rising global pandemic of diabetes, 1,2 defined by persistent hyperglycaemia, a principal risk factor for the development of concomitant chronic complications. 3,4 Hence, the numbers of individuals with diabetic kidney disease (DKD), a major complication, are burgeoning. DKD is an important risk factor for both end-stage kidney disease and cardiovascular disease. 3,5 While renin-angiotensin system blockade including angiotensin-converting-enzyme inhibition is first-line therapy for DKD, in general these agents only achieve a ~ 20% reduction in end-stage kidney disease. 6 Advanced glycation end products (AGEs) are a heterogeneous and complex group of non-enzymatic, post-translational modifications to amino acids and proteins, which include haemoglobin A 1C (HbA 1C ) a clinical marker used for the diagnosis of diabetes. Their endogenous formation can be exacerbated by chronic hyperglycaemia and oxidative stress 7 and therefore the accumulation of AGEs occurs at an accelerated rate in diabetes. 8,9 Increases in AGE formation on skin collagen [10][11][12] and within the circulation 13,14 predict poor prognosis for patients with diabetes including increased risk for kidney and cardiovascular disease. The kidneys are a major site of AGE detoxification through the filtration of circulating AGEs and their subsequent active uptake and excretion. 15,16 Therefore, AGE accumulation is seen in diabetic patients with chronic kidney disease. 17,18 Therapies which lower AGE accumulation have shown benefit in Phase II clinical trials for the treatment DKD in individuals with Type 2 diabetes. 19 Other AGE lowering therapies tested in clinical trials in DKD include pimagedine 20 which was withdrawn due to safety concerns, alagebrium chloride (ALT-711) which reached Phase II trials that were not completed due to financial constraints and benfotiamine which has shown mixed results. 21 OST48, an evolutionarily conserved type 1 transmembrane protein, 22 is encoded by the DDOST (Dolichyldiphosphooligosaccharide-protein glycosyltransferase 48 kDa subunit) gene and has been postulated to function as a receptor for AGE turnover and clearance. 23,24 OST48 is expressed in most cells and tissues, 25 including macrophages 26 and mononuclear cells. 27 In the kidney, previous studies have shown OST48 expression in glomerular cells including podocytes 25,28 and mesangial cells 26 where in the podocytes it is postulated to mediate the uptake and secretion of AGEs. 29 The process whereby AGEs are cleared by OST48 in the kidney is not well understood, 30 but is thought to involve degradation of AGEs and then excretion via the urine. 31 A strong link exists between impaired podocyte function and albuminuria, increased urinary AGE excretion and glomerulosclerosis. [32][33][34][35][36][37] Early damage to the glomeruli, specifically podocyte structural damage, is characteristic of DKD. [38][39][40] AGEs can induce podocyte cell-cycle arrest, 41 hypertrophy 41 and apoptosis, 42 while a systemic reduction in AGEs has been shown to improve podocyte and kidney function. 43  and showed no differences in any anthropometric parameters or long-term glycaemic control (Table 1). We have previously confirmed that this genetic modification does not affect N-glycosylation machinery. 44 Mice with experimentally induced diabetes were characterized by elevated fasting blood glucose and glycated haemoglobin (GHb) ( Table 1). Irrespective of genotype, diabetic mice exhibited a lower body weight, renal hypertrophy, increased consumption of food and water and a greater urinary output when compared with non-diabetic mice (Table 1). Interestingly, it appeared that there was also a significant interactive effect of diabetes on the abundancy of OST48 in glomeruli  Figure 1C). A podocyte-specific increase in OST48 expression was confirmed using ultra-resolution three-dimensional structured illumination microscopy (3D-SIM) ( Figure 1D). This included increased localization of OST48 to damaged podocyte foot processes, as indicated by nephrin loss (Figure 1D), which was particularly evident in 3D reconstruction videos ( Figure S2A and Video S1). This localization of OST48 to areas of damaged foot processes was unexpected, since the prevailing model is that increasing OST48 abundance could improve declining kidney function by sequestering AGEs. 45 These initial findings of damage to podocyte foot processes warranted further investigation of kidney function in DDOST+/− Pod−Cre mice, to determine whether increasing podocyte expression of OST48 affected glomerular filtration. for DDOST+/− Pod−Cre and 123% increase; p = .0047 for wild-type).

| Increases in podocyte OST48 expression decrease GFR, exacerbating DKD
Diabetes increased serum creatinine in wild-type mice ( Figure 2C) and decreased creatinine clearance ( Figure 2D), which tended to further decrease in DDOST+/− Pod−Cre mice, although this did not reach statistical significance ( Figure 2D). Compared to wild-type mice, DDOST+/− Pod−Cre mice averaged a 72% reduction in creatinine clearance following simultaneous blood and 24-hour urine collection ( Figure 2E; p = .0052) in agreement with FITC-sinistrin based GFR assessment (Figure 2A). Diabetes increased 24-hour urinary albumin excretion rate, but this did not differ between wild-type and DDOST+/− Pod−Cre mice ( Figure 2F and G).

| DDOST+/− Pod−Cre mice have damaged podocyte architecture and tubulointerstitial damage
Diabetes resulted in glomerulosclerosis in both genotypes ( Figure 3A). In the absence of diabetes, DDOST+/− Pod−Cre mice also had glomerulosclerosis ( Figure 3A) and greater glomerular collagen IV accumulation ( Figure 3B) than wild-type mice which was not further elevated by diabetes. Glomerular collagen IV accumulation was also present in all diabetic mice ( Figure 3B). By electron microscopy, glomerular basement membrane thickening and podocyte foot process effacement were seen in DDOST+/− Pod−Cre , but not in wild-type mice ( Figure 3C). In glomerular fractions from mouse kidney cortices, SWATH-MS proteomics identified significant increases in the abundance of collagen proteins ( Figure 3D). Specifically, both diabetes and increased podocyte OST48 expression caused an increased abundance of collagen 1, collagen 6 and other structural collagen proteins ( Figure 3D). Given that all DDOST+/− Pod−Cre mice exhibited a decline in kidney function, we suspected that in addition to damaged glomeruli there would be other structural evidence of progressive kidney damage in the tubules of these mice. As expected, diabetes increased tubulointerstitial fibrosis in WT mice ( Figure 4A  Note: Values are mean ± SD (n = 5-11).
Bold values indicate a significant (P < 0.05) effect.

| Tissue AGEs were localized to damaged podocytes
Podocyte AGE accumulation and OST48 appeared to be colocalized ( Figure 5A) and were prominent in areas where concomitant deterioration of podocyte foot processes was seen, as defined by loss of both nephrin ( Figure 5A) and synaptopodin ( Figure S3A). 3D-SIM confirmed the appearance of nodules/vacuoles containing both OST48 and AGEs in DDOST+/− Pod−Cre mice in areas of podocyte foot process deterioration ( Figure 5B, Figure 5C and Video S2). These changes were not observed in wild-type diabetic mice. Overall, there was a significant reduction in AGE con- (E) Change in creatinine clearance, which was determined from matched creatinine clearance values from week 0 of the study (6-8 weeks of age) to week 12 of the study. (F-G) Albumin was measured spectrophotometrically at 620nm in a biochemical analyser and the albumin excretion rate (AER) was determined based on the 24-hour urine flow rate. (F) AER measured at week 12 of the study. (G) Change in albuminuria over the study, which was determined from matched albuminuria values from week 0 (6-8 weeks of age) of the study to week 12 of the study. Results are expressed as mean ± SD with either two-way ANOVA or paired t-test analysis (n = 3-8) *p < .05, **p < .01, ***p < .001

| Podocyte OST48-mediated AGE accumulation resulted in ER stress
ER stress at sites of AGE accumulation within damaged podocytes was examined. DDOST+/− Pod−Cre mice exhibited an increase in the ER stress markers GRP-78 and spliced XBP-1 when compared to wildtype mice ( Figure 6A). Furthermore, these ER stress markers were further increased by diabetes ( Figure 6A). This was confirmed using Results are expressed as mean ±SD with either two-way ANOVA or unpaired t-test analysis (n = 5-9) *p < .05, **p < .01, ***p < .001, ****p < .0001. For proteomics, MSstatsV2.6.4 determined significant (p < .05) log fold changes in the protein intensities between the selected experimental group the wild-type nondiabetic group. Heatmap representation allow for compact visualization of complex data comparisons. Full detail of the quantitative data is available in the supplementary information synthesis, particularly biosynthesis of complex I components (NADH dehydrogenase:ubiquinone) ( Figure 6C).

| DISCUSS ION
Lowering AGE burden, including by facilitating renal AGE clearance, has been suggested as a possible treatment for DKD. However, the assertion that increasing renal AGE clearance receptors such as OST48 may protect against DKD has not been previously investigated. Here, we have shown for the first time that increasing OST48 in a podocyte-specific manner decreased glomerular filtration and caused podocyte structural damage including foot process effacement, leading to glomerulosclerosis and tubulointerstitial fibrosis, and some parameters were further exacerbated by diabetes.
Surprisingly, in non-diabetic mice the decline in GFR and increased Results are expressed as mean ± SD with either two-way ANOVA or unpaired t test analysis (n = 5-9) *p < .05, **p < .01. For proteomics, MSstatsV2.6.4 determined significant (p < .05) log fold changes in the protein intensities between the selected experimental group the wild-type non-diabetic group. Heatmap representation allow for compact visualization of complex data comparisons. Full detail of the quantitative data is available in the supplementary information | 9 of 15 ZHUANG et Al.
It was particularly interesting that albuminuria was not seen concomitantly with the loss of GFR in non-diabetic DDOST+/− Pod−Cre mice despite podocyte foot effacement, thickening of the glomerular basement membrane and glomerulosclerosis. Although surprising, this is consistent with clinical observations. For example, the prospective longitudinal Joslin Proteinuria Cohort studies 46,47 identified that the overall prevalence of normo-albuminuria in patients with progressive kidney disease was around 10% and that regression from micro-and macro-albuminuria to normo-albuminuria was repeatedly observed. We observed podocyte foot effacement and glomerular basement membrane thickening, which are commonly attributed as causes of micro-and macro-albuminuria. This finding is consistent with other postulated models of the origins of albuminuria, such as the high glomerular sieving coefficient (GSC) hypothesis. 48,49 This hypothesis suggests that albuminuria is driven by proximal tubule damage which prevents the reabsorption/ processing of albumin by these cells. In support of this model, tubulointerstitial fibrosis was more pronounced in diabetic mice regardless of their genotype, and they also had concomitant increases in urinary albumin excretion which not seen in non-diabetic mice.
Alternatively, the absence of macro-albuminuria seen in mice with increased podocyte expression of OST48 could be attributed to the significant decline in GFR, as this would limit the flux of albumin into the urine.
Although we showed that increases in podocyte OST48 expression decreased renal AGE accumulation and increased urinary AGE excretion, this was not associated with reno-protection. In fact, this resulted in significant renal disease. We have shown for the first time that OST48 and AGEs colocalize within podocytes, in areas of foot process denudation and damage. We observed that OST48 facilitation of podocyte AGE accumulation resulted in a cascade of ER stress and mitochondrial abnormalities, culminating in podocyte foot process effacement, GBM expansion and renal functional decline. Increasing the urinary flux of AGEs has indeed been previously shown to impair renal function in both healthy humans 43,50 and early in the development of diabetic kidney disease in rodent models. 37,51 This is interesting given that although AGE accumulation is facilitated and accelerated by the hyperglycaemia of diabetes, there is substantial evidence that AGE accumulation is a pathological mediator of many chronic kidney diseases, where AGE formation is driven by oxidative stress, dyslipidaemia, uraemia and insulin resistance independent to diabetes. 43,52 In those instances, AGEs have a major contribution to pathology including podocyte loss, where that chronic AGE administration can mimic chronic kidney disease in the absence of diabetes. Therefore, we believe that AGE accumulation in the podocytes is sufficient to drive kidney disease in the absence Taken together, these results suggest that specifically increasing podocyte exposure and uptake of AGEs thereby facilitating their flux into the urine is sufficient to induce significant kidney damage.
Moreover, it should be noted that modulation of surface AGE receptors on podocytes to facilitate excretion may not be an optimal target to improve DKD due to the consequential effects of podocyte effacement.

| Animals
All animal studies were performed in accordance with guidelines

| Glomerular and tubule enrichment
Kidneys were decapsulated and the medulla discarded.

| Glomerular filtration rate
At 0 and 12 weeks of the study, GFR was estimated in conscious mice using the transcutaneous decay of retro-orbitally injected FITC-sinistrin (10 mg/100 g body weight dissolved in 0.9% NaCl), as previously described. 54 GFR was calculated from the rate constant (α2) of the single exponential excretion phase of the curve and a semi-empirical factor.

| Liquid chromatography-tandem mass spectrometry proteomics
Fractions enriched for either glomeruli or tubules were processed as previously described. 55 Peptides were desalted and analysed by Information Dependent Acquisition Liquid chromatography-tandem mass spectrometry (LC-MS/MS) as described 56  Confocal images were visualized on an Olympus FV1200 confocal microscope (Olympus) and viewed in the supplied program (FV10, F I G U R E 6 Podocyte OST48 increased ER stress markers. (A) Confocal photomicrographs of either ER stress marker GRP-78 (green), or XBP-1 (green) and podocyte foot process marker, synaptopodin (red). (B-C) Heat map representation of SWATH-MS proteomics data for enzymatic pathways involved in (B) endoplasmic reticulum (ER) stress and oxidative stress (OS) enzymatic pathways, and (C) ubiquinone biosynthesis pathways. Significant proteins are represented as bolded cells, where red indicates an increase and blue indicates a decrease in protein concentrations. Data represented as means ±SD (n = 4-7/group). Scale bars from representative images of confocal microscopy were 30 μm. Data represented as means ±SD (n = 5/group). For proteomics, MSstatsV2.6.4 determined significant (p < .05) log fold changes in the protein intensities between the selected experimental group the wild-type non-diabetic group. Heatmap representation allow for compact visualization of complex data comparisons. Full detail of the quantitative data is available in the supplementary information Olympus). For ultra-resolution 3D microscopy, slides were visualized on an OMX Blaze deconvolution structured illumination (SIM) superresolution microscope (GE Healthcare Life Science).

| Glomerulosclerotic index
Glomerulosclerotic index (GSI) as a measure of glomerular fibrosis was evaluated in a blinded manner by a semi-quantitative method. 60 Severity of glomerular damage was assessed on the following parameters; mesangial matrix expansion and/or hyalinosis of focal adhesions, true glomerular tuft occlusion, sclerosis and capillary dilation.

| Fixation, tissue processing and acquisition of data from electron microscopy
Renal cortex was processed as previously described. 61 Sections were cut at 60nm on a Leica UC6 ultramicrotome (Leica) and were imaged at 80 kV on a Jeol JSM 1011 transmission electron microscope (Jeol) equipped with an Olympus Morada (Olympus, Japan) digital camera.
Quantification of foot process width and GBM expansion was assessed as previously described. 62

| Statistics
Results are expressed as mean ± SD (standard deviation) and as- The primary endpoint kidney function. Six mice/group will be required to observe a 10% change in eGFR (α = .05, power = 0.9 and SD = 0.1). Animal studies and statistical analysis were blinded under the ARRIVE guidelines.

ACK N OWLED G EM ENTS
The authors acknowledge the Translational Research Institute (TRI) for providing the excellent research environment and core facilities that enabled this research. We particularly thank Sandrine Roy and Ali Ju from the TRI Microscopy Core Facility. Furthermore, the au-

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets generated and analysed during the current study are available from the corresponding author upon reasonable request.
The mass spectrometry proteomics data have been deposited to the