Within the glomerulus, podocytes are able to uptake LDL, APOB, and APOE containing lipoproteins. Their capacity to uptake APOB and APOE containing lipoproteins is increased when compared to LDL, and this process is associated with the suppression of cellular sterol synthesis and cholesteryl ester formation (50).

Lipoprotein glomerulopathy [LPG, Online Mendelian Inheritance in Man (OMIM) #611771] is a rare genetic disorder mainly affecting people of Japanese and Chinese origin that is caused by a heterozygous mutation in the APOE gene on chromosome 19q13. Several mutations in APOE were identified resulting in the expression of a dysfunctional APOE protein with impaired LDL receptor binding (34, 53, 54). Patients with LPG have proteinuria and nephrotic syndrome with elevated serum APOE levels, abnormal lipoprotein deposition in glomerular capillaries in form of laminated thrombi that contain APOB and APOE, a variable degree of mesangial proliferation, and dysbetalipoproteinemia (55, 56). Occasionally, a similar renal phenotype can be observed in type III hyperlipoproteinemia, a disease that is characterized by homozygous mutation in the ε2 allele of the APOE2 gene. The kidney phenotype includes proteinuria, glomerulosclerosis with mesangial and interstitial foam-cell accumulation, and the presence of intraglomerular thrombi (57, 58). Genetic polymorphisms in APOE were shown to differentially affect LDLR binding. Thus, the APOE2 variant binds poorly to LDLR, whereas the APOE4 variant binds with high affinity (59). The APOE2 variant was associated with persistent proteinuria and renal failure in patients with non-insulin-dependent diabetes mellitus (NIDDM) (60) and may contribute to the severity of the renal phenotype in IgA nephropathy (61). Conversely, the polymorphism in APOE4 is thought to constitute a protective factor in patients with NIDDM (62).

Constitutive ApoE^−/− mice generated through partial replacement of exon 3 and part of intron 3 of the ApoE gene by a neomycin cassette (63) are a model for experimental atherosclerosis and develop a renal phenotype similar to that observed in patients with type III hyperlipoproteinemia and LPG (64–66). Kidneys from ApoE^−/− mice showed glomerular macrophage infiltration with foam-cell formation, deposition of extracellular matrix, glomerular hyperplasia, mesangiolysis, and lipid deposits in glomerular capillaries at 24 weeks, which further progressed to a LPG-like phenotype at 36 weeks (66). In a different study, in which the APOE-Sendai mutation (Arg145Pro mutation found in patients with LPG) was introduced into ApoE^−/− mice showed that, although similar, the LPG-like lesions observed in the APOE-Sendai mice were different from the lesions in aged ApoE^−/− mice (65, 67) indicating that APOE variants may contribute to a differential phenotype in LPG. ApoE^−/− mice, which were fed a high fat diet developed a renal phenotype at 16 weeks of age with lipid droplet and cholesterol crystal accumulation within the glomerulus. This phenotype was associated with decreased basal plasma renin concentrations and low mean arterial blood pressure, indicating impaired renin-angiotensin system (RAS) function. After renal artery constriction, these mice were unable to increase renin secretion from the juxtaglomerular cells, demonstrating a lack of a normal response to renal hypoperfusion (68). Similar glomerular changes were observed in another study when male mice were fed a high cholesterol diet for 20 weeks (69) indicating that high fat diet accelerates the glomerular injury phenotype. Likewise, streptozotocin (STZ) injected into ApoE^−/− mice accelerated renal injury when compared to non-diabetic mice and was attenuated when advanced glycation end product (AGE) accumulation was prevented by the use of an AGE formation inhibitor. Renal changes observed in diabetic ApoE^−/− mice included increased albuminuria and structural changes in glomeruli and tubulointerstitium (70). Taken together, these studies suggest that ApoE deficiency and mutations that affect ApoE binding to LDLR and the resulting hyperlipidemia play a role in rendering renal cells more susceptible to glomerular injury.

In further support of this hypothesis, mice with a combined deficiency of inhibitor of differentiation 3 (Id3) and ApoE^−/− spontaneously develop glomerulonephritis characterized by lipid deposition specifically in glomeruli, the presence of enlarged hypercellular glomeruli, mesangial expansion, and increased extracellular matrix deposition (71). ApoE deficiency also accelerates the renal phenotype in MRL-Fas^lpr mice, a mouse model of human lupus (72). Further, decreased ApoE expression in glomerular podocytes is observed in Tg26 (human immunodeficiency virus, HIV) and Nef (negative regulatory factor) transgenic mice, two models of HIV-associated nephropathy (HIVAN) (73), whereas increased ApoE and ApoB expression was found in association with accumulation of oil red O positive lipid droplets in glomerular visceral epithelial cells and mesangial cells in rats with puromycin aminonucleoside or adriamycin induced nephrosis (74). Increased glomerular APOE expression was observed in patients with idiopathic nephrotic syndrome but it is rather a rare occurrence in focal segmental glomerulosclerosis (FSGS) (75).

APOM is a 26-kDa apolipoprotein and is a member of the lipocalin family expressed in the liver and in the kidney. In the plasma, APOM is associated with HDL particles (48); in kidney proximal tubular cells, APOM binds to megalin, thus preventing its excretion in the urine by megalin-mediated endocytosis (76). Mutations, polymorphisms, or allelic variants of the APOM gene have not been associated with any glomerular phenotype. Although APOM is strongly expressed in kidney tubular epithelial cells, it is not found in the urine of mice or human beings. However, megalin knockout mice were shown to be characterized by urinary loss of low molecular weight proteins underlining an important role of proximal tubules in resorption of proteins (77). Megalin is expressed in glomerular podocytes where it can bind to α-galactosidase A, suggesting a potential role in Fabry disease, and it may act as a pathogenic antigen in membranous glomerulonephritis (78, 79).

APOL1 is a secreted HDL-associated protein that binds to APOA1 and promotes cholesterol efflux. It is coded by a gene on human chromosome 22q12, and APOL1 mutations are associated with increased susceptibility to FSGS (OMIM #612551), HIVAN, and hypertensive nephropathy in patients of African ancestry (80–82). Using immunofluorescence staining of normal human kidneys, APOL1 was localized to podocytes of the glomerulus, the promixal tubules, and the extraglomerular arterial endothelium. In kidney biopsies from patients with FSGS and HIVAN, APOL1 expression levels in podocytes were found decreased and de novo appearance of APOL1 within cells of the arterial medial wall was observed (51). The observation that normal human podocytes express APOL1 opens up the possibility that APOL1 may be contributing to cellular cholesterol homeostasis under physiological conditions. Likewise, it is possible that decreased podocyte APOL1 expression under disease conditions leads to decreased cholesterol efflux from podocytes and cellular cholesterol accumulation, thus contributing to the pathogenesis of glomerular diseases such as FSGS and HIVAN. Future studies to investigate the role of the different APOL1 variants and their role in podocyte cholesterol homeostasis under physiological and disease conditions are needed to shed light on their contribution to the pathogenesis of glomerular diseases.

Familial LCAT deficiency (OMIM #245900) is a rare genetic disorder caused by a mutation in the LCAT gene. LCAT is a HDL-associated enzyme that converts cholesterol to cholesteryl esters by transfer of the SN2 fatty acid from a phosphatidylcholines (lecithins) with the formation of lyso-phosphatidylcholine on the surface of HDLs. It, therefore, plays an important role in reverse cholesterol transport from peripheral tissues to the liver. LCAT deficiency leads to cholesterol accumulation in many tissues (83, 84). Patients with familial LCAT deficiency are characterized by diffuse corneal opacities, target cell hemolytic anemia, and proteinuria with chronic progressive glomerulopathy resulting in renal failure. Massive lipid deposits in the glomerular basement membrane (GBM) and in the mesangial region can be detected (85, 86). The appearance of abnormal choleastic lipoprotein, LpX, in the plasma of the patients was described (84, 87, 88).

Lecithin-cholesterol acyltransferase was originally shown to be expressed only in liver (89) but later also in brain, testis, ileum, kidney, spleen, and adrenal tissue (84, 90). More detailed analysis of LCAT expression in kidneys from Mongolian gerbils by immunohistochemistry showed expression mainly in the proximal and distal convoluted tubules and in the collecting duct epithelial cells (91).

Lcat deficient mice are characterized by reduced total cholesterol, HDL cholesterol, and APOA1 plasma levels (92, 93). When fed a regular chow diet, Lcat deficient mice show no glomerular phenotype. However, when fed a high fat diet, Lcat deficient mice developed glomerular lesions including reduction of vascular space, mesangial expansion and sclerosis, and increased extracellular matrix with accumulation of lipid droplets and macrophages. Lipid droplets in glomeruli were composed of free cholesterol and polar lipids as shown by filipin and oil red O staining. Interestingly, the glomerular phenotype was only observed in mice that simultaneously accumulated LpX (94). In support of the observation that elevated plasma LpX levels are necessary to mediate the glomerular phenotype, double mutant mice with Lcat deficiency that constitutively overexpress a NH2-terminus segment of SREBP-1a, which leads to accumulation of predominantly LpX in plasma were shown to spontaneously develop glomerular lesions with glomerular and tubulointerstitial lipid deposits at 6 months (95).

Niemann–Pick disease type C (OMIM#607616) is a genetic, neurogenerative disorder caused by mutations in the genes NPC1 on chromosome 18q11 in the vast majority of the cases, and NPC2 on chromosome 14q24. Patients present with ataxia, vertical supranuclear gaze palsy (VSGP), and dementia. Mutations in this gene(s) lead to the inability to transport free cholesterol between intracellular compartments. As a consequence, free cholesterol and other lipids accumulate in late endosomes/lysosomes in all organs including the kidney simultaneously causing a delayed response in sterol-responsive pathways to exogenous cholesterol (96, 97). NPC1 and NPC2 are expressed in the kidney (98, 99). NPC2 expression was further localized to the distal and proximal convoluted tubules of the kidney (99). Although rather rare, Niemann–Pick disease-associated renal pathology was described in biopsies from patients and included foamy podocytes, vacuolated tubular epithelial cells, and collections of foam cells in the interstitium (100). Association with a phenotype resembling membranoproliferative glomerulonephritis type II was described as well (101). These observations indicate that cholesterol accumulation in glomerular or tubular cells contributes to the renal pathogenesis observed in Niemann–Pick disease. In Npc1 deficient mice, a single injection with hydroxy-propyl-beta cyclodextrin, a cholesterol-depleting agent, was shown to increase the life span and weekly injections normalized cholesterol metabolism in nearly every organ (102–104).

Tangier disease (OMIM #205400) is an autosomal recessive and familial hypoalphalipoproteinemia (FHA, OMIM #604091), an autosomal dominant disorder caused by a mutation in the ABCA1 gene on chromosome 9q31 leading to accumulation of esterified cholesterol in tissues due to impaired reverse cholesterol transport resulting in reduced levels of plasma HDL (105–108). Patients present with heterogeneous clinical symptoms including liver, spleen, lymph node, and tonsil enlargement and peripheral neuropathy in children and adolescents. Occasionally, cardiovascular disease is observed in adults but the presence of a renal phenotype in Tangier patients is extremely rare (109, 110). FHA is the more common disease and is, like Tangier disease, characterized by low plasma HDL but without the clinical manifestations of Tangier disease. Abca1 deficient mice as a model for Tangier disease were generated by partial (111) or complete (112) replacement of the exons encoding the first ATP-binding cassette of Abca1. When fed on a high fat diet, Abca1 deficient mice were characterized by significantly decreased plasma HDL levels and lipid accumulation in several tissues including the kidney. Phenotypical differences were observed within the two models of Abca1 deficiency, possibly due to differences in the high fat diet composition or to slight differences in the targeting strategies (111, 112). In the study by Christiansen-Weber, a glomerular phenotype was observed, which included the presence of a thickened and “split” GBM and mesangial cell proliferation. Furthermore, immunoglobulin and C3 complement complexes characteristic of menbranoproliferative glomerulonephritis were detected (111), whereas no renal phenotype was described in the other study. The observation that Abca1 deficient mice develop a glomerular phenotype on a high fat diet deficiency may indicate that Abca1 deficiency leads to increased susceptibility for the development of renal disease.

Lipid droplets in kidney biopsies from patients with DKD were first identified by Kimmelstiel and Wilson (42) and were more recently localized within podocyte foot processes of patients with DKD (41). We recently demonstrated that glomerular ABCA1 expression is decreased in patients with type 2 diabetes (T2D) and early DKD and in human podocytes treated with the sera from patients with type 1 diabetes (T1D) and DKD in the absence of changes in LDLR and HMGCR expression (52). These observations indicate that lipid accumulation in podocytes due to defective cellular cholesterol homeostasis may play an important role in glomerular injury in DKD.

In support of these observations, lipid droplets in various experimental models of DKD have been reported. In two models for DKD with type 1 diabetes, Akita, and OVE26 mice, renal triglycerides and cholesterol accumulation was described and was associated with increased renal cholesterol synthesis and decreased efflux. This coincided with increased expression of SREBP-2 and HMGCR and decreased expression of liver X receptor (LXR)-α, LXR-β, and ABCA1 in kidneys of these mice (113). In diabetic STZ-injected DBA2/J mice, the DKD phenotype, which includes albuminuria, mesangial expansion, fibrosis and lipid accumulation in glomeruli, and tubulointerstitium is associated with increased HMGCR expression in kidneys. Aliskiren (a renin inhibitor) and valsartan (an angiotensin II receptor antagonist) reduced albuminuria and glomerulosclerosis in these animals and was associated with decreased HMGCR expression and decreased lipid accumulation in glomeruli (114), suggesting a potential crosstalk between the RAS and lipid metabolism in the kidney. FvB-Lepr^db/db mice, a model for DKD with type 2 diabetes, spontaneously develop a renal DKD-like phenotype that includes glomerulosclerosis, tubulointerstitial fibrosis, GBM thickening, and proteinuria. In addition, renal triglyceride and cholesterol accumulation in glomeruli and tubules associated with increased activity of SREBP-1 and -2 was observed (115). Likewise, renal cholesterol accumulation in high fat/sucrose-fed (116) (type 2 diabetes model for DKD) and STZ-induced rats (type 1 diabetes model for DKD) was associated with increased expression of SCAP, SREBP-2, HMGCR, and LDLR. Atorvastatin, a lipid-lowering agent with antioxidative and anti-inflammatory effects reduced cholesterol synthesis in the kidneys improving renal function and kidney morphology in these models (116).

It should be noted that STZ has also been shown to impair renal function during the first week after injection in rodents affecting not only proteinuria but also renal microcirculation and oxidative pathways (117, 118). Studies to elucidate the contribution of hyperglycemia after STZ treatment or STZ-induced renal toxicity demonstrated that a majority of the renal damage was a result of the diabetic condition and only early proteinuria was caused by the nephrotoxic effects of STZ (119).

In the non-obese diabetic (NOD) mouse model of T1D, ABCA1 protein expression was decreased in kidneys and circulating macrophages, and associated with increased cholesterol content indicating that impaired reverse cholesterol transport may contribute to kidney phenotype observed in diabetic NOD mice (120). Diabetes-induced reduction of kidney ABCA1 expression in association with cholesterol accumulation was observed in cyclophosphamide-induced diabetic NOD and in LDLR-deficient-GP (glycoprotein) lymphocytic choriomeningitis virus (LCMV)-induced diabetic mice (121). In BTBR ob/ob mice, a type 2 diabetes model of DKD, cholesterol accumulation occurs in kidneys, a phenotype that can be prevented by cholesterol depletion with cyclodextrin, underlining the causative role of cholesterol accumulation in glomerular injury (52). Together these studies indicate that cholesterol accumulation due to impaired reverse cholesterol transport occurs in glomerular podocytes and contribute to the pathogenesis of DKD. As statins, which decrease cholesterol synthesis do not substantially affect the progression of kidney disease (122–124), other specific strategies that facilitate transport of sequestered cholesterol such as targeting reverse cholesterol transport are needed and may prove beneficial in treatment of DKD.

Chronic renal failure induced by 5/6 nephrectomy in rats is associated with neutral lipid accumulation in the remnant renal tissue, glomerulosclerosis and tubulointerstitial injury, and proteinuria. In remnant kidneys, increased expression of LXR, ABCA1, ABCG1, ACAT1, SR-B1, scavenger receptor A1 (SR-A1), lectin-type oxidized LDL receptor (LOX-1), carbohydrate-responsive element binding protein (ChREBP), fatty acid synthase (FAS), and acyl-CoA carboxylase (ACC) was detected, whereas expression of SREBP-1, SREBP-2, HMGCR, peroxisome proliferator-activated receptor alpha (PPAR-α), fatty acid binding protein (L-FABP), and carnitine palmitoyltransferase 1A (CPT1A) was decreased (125, 126). Cholesterol accumulation in proximal tubules is associated with HMGCR activation in acute kidney injury (AKI) (127, 128). The observation that cholesterol accumulation can result from glomerular injury suggested that renal cholesterol accumulation may represent a mechanism of stress response (129). In a model of experimental glomerulonephritis with concomitant proteinuria, cholesterol accumulation occurred in the renal cortex, i.e., the glomerulus and proximal tubular cells. While esterified cholesterol accumulation was present in glomerular and tubular cells, free cholesterol accumulation occurred in proximal tubular cells and was associated with increased ABCA1 and decreased SR-B1 expression (130).

Hypoxia-induced dyslipidemia is a phenomenon that was first described decades ago (131–133) and shown to play an important role in foam-cell formation and cytokine secretion in atherosclerosis and in non-alcoholic steatohepatitis (NASH) (134–137). Hypoxia is also a pathological feature observed in kidney disease and hypoxia-inducible factors (HIFs), such as HIF-1 and HIF-2 are thought to play an important role in AKI. HIFs initiate the transcription of a variety of genes involved in the regulation of tissue oxygen levels, glucose metabolism, apoptosis, lipid metabolism, and immune responses (138–143). HIF target genes include hemeoxygenase-1, vascular endothelial growth factor, plasminogen activator inhibitor-1, tissue-inhibitor of metalloproteinase-1, and connective tissue growth factor (139, 142, 144), all of which have been linked to the pathogenesis of a variety of glomerular diseases including DKD, FSGS, and HIVAN (145, 146). HIF-1 was shown to contribute to lipid accumulation by increasing lipid influx and synthesis in hepatocytes through increased LDL and very low-density lipoprotein (VLDL) uptake and increased levels and activity of HMGCR, respectively (147). More recently, it was demonstrated that activation of HIF-2 can also induce changes in hepatic lipid metabolism leading to the development of severe fatty liver disease in mice (148). Additionally, hypoxia-inducible protein 2 (HIG2) was identified as a novel lipid droplet protein and as a specific target gene of HIF-1 (149), further underlining a role for hypoxia in altered lipid metabolism in disease. In the kidney, HIF-1 was localized to the tubular epithelia, whereas HIF-2 was found predominantly in glomerular (podocytes), endothelial, and interstitial cells (150, 151). The induction of HIFs or inhibition of HIF degradation through hypoxia seems to have protective effects in AKI (152–154). Chronic hypoxia may also contribute to the pathogenesis of DKD (155) and genes induced by hypoxia can promote tubulointerstitial injury and renal fibrosis. Thus, novel therapeutic approaches targeting mechanisms of hypoxia-induced transcription may prove beneficial in targeting dyslipidemia in a variety of kidney diseases.

Inflammation, foam-cell formation, and lipid accumulation are characteristic occurrences associated with glomerulosclerosis. In non-alcoholic fatty liver disease (NAFLD), a two hit hypothesis proposed that a first hit such as obesity, T2D, or the initial lipid accumulation in the liver sensitizes it to a second hit such oxidative stress and proinflammatory cytokines ultimately causing hepatocellular injury and liver inflammation (156). In hepatocytes, tumor necrosis factor alpha (TNFα) or interleukin-1 beta (IL-1β) mediated inflammatory stress significantly reduced intracellular cholesterol efflux by inhibiting PPAR, LXR, and ABCA1 expression and increased LDLR and SREBP-2 expression and suggesting that inflammatory stress may exacerbates progression of fatty liver in NAFLD (157). We recently described an important role of circulating factors in the pathogenesis of DKD and showed that treatment of human podocytes with the sera from patients with DKD leads to cholesterol accumulation (52). Therefore, it seems possible that circulating inflammatory factors, such as TNFα or IL-1β, contribute to the progression of DKD. In support of this observation, TNFα is a major predictor of DKD progression in T1D and T2D (158–160) and inflammatory cytokines, such as TNFα or IL-1β, were shown to modify cholesterol-mediated LDL receptor regulation in mesangial cells. These studies suggest that inflammatory cytokines contribute to lipid-mediated renal damage (161, 162). Furthermore, IL-1β was shown to promote intracellular cholesterol accumulation in human mesangial cells via downregulation of ABCA1 (163). Interestingly, IL-1β treatment of hepatic and mesangial cells also interrupted LDLR feedback regulation, thus causing statin resistance (164), which might explain why statins are not effective in preventing the progression of DKD. More recently, Nod-like receptor protein 3 (Nlrp3) knockout mice were shown to be protected against obesity-induced renal fibrosis and microalbuminuria which, in wildtype mice, was associated with cholesterol and lipid accumulation in kidneys and increased SREBP-2, LDLR, and SREBP-1c expression (165). Whether cytokines contribute to lipid accumulation observed in podocytes and the exact mechanisms that lead to cytokine-induced cholesterol accumulation remain to be investigated.