Mechanisms of injury in APOL1-associated kidney disease

Abstract

Background:

An improved understanding of the pathogenesis in apolipoprotein L1 gene (APOL1)-gene associated chronic kidney disease (CKD) arose from observations in kidney transplantation. APOL1 genotyping could soon improve the safety of living kidney donation in individuals with recent African ancestry and alter the allocation of deceased donor kidneys.

Methods:

This manuscript reviews the potential mechanisms that underlie development of APOL1-associated nephropathy. Roles for circulating APOL1 protein versus intrinsic renal expression of APOL1 are discussed, as well as the requirement for modifying genetic and/or environmental factors.

Results:

Abundant evidence supports local kidney production of APOL1 renal-risk variant protein in the development of nephropathy; this is true in both native kidney disease and after renal transplantation. Only a minority of kidneys from individuals with APOL1 high-risk genotypes will develop CKD or manifest shorter renal allograft survival after transplantation. Therefore, modifying factors that explain why only a subset of kidneys develops nephropathy remain critical to identify. It appears likely that environmental exposures, as opposed to major APOL1-second gene interactions, will prove to be stronger modifiers of the risk for nephropathy.

Conclusions:

The evolving understanding of the pathogenesis in APOL1-associated nephropathy will identify biomarkers predicting nephropathy in individuals at high genetic risk and lead to novel therapies to prevent or slow native CKD progression and prolong survival of transplanted kidneys. In the interim, the National Institutes of Health-sponsored “APOL1 Long-term Kidney Transplantation Outcomes” (APOLLO) Network will determine whether APOL1 genotyping in individuals with recent African ancestry improves outcomes and safety in kidney transplantation.

Introduction

Discovery of the powerful apolipoprotein L1 gene (APOL1) association with several nondiabetic chronic kidney diseases (CKD) has altered our understanding of the epidemiology in end-stage renal disease (ESRD). The majority of individuals with recent African ancestry diagnosed with HIV-associated nephropathy (HIVAN), focal segmental glomerulosclerosis (FSGS) and “hypertension-attributed” ESRD have APOL1-associated kidney disease.¹ Essential hypertension, the second most commonly reported etiology of ESRD in African Americans, is an unlikely initiator of ESRD.² This manuscript discusses potential biologic mechanisms underlying the development and progression of CKD in those possessing APOL1 renal-risk genotypes.

Biologic mechanisms in APOL1-associated nephropathy

In isolation, genetic association analyses are unlikely to precisely identify disease-causing variants. Several fundamental questions need to be answered to determine whether the APOL1 G1 and G2 renal-risk variants directly cause CKD.³ Several have been addressed, including the strength of association, consistency of findings, temporal sequence and biological plausibility. Recent efforts now focus on the remaining questions which center on specificity of association, biological gradient, coherence and experiment. This review focuses on efforts to address 2 fundamental questions. Is APOL1 protein present in kidney tissue? Assuming that APOL1 protein is present, do G1 and G2 renal-risk variant proteins impact cellular function and via what mechanisms?

APOL1 is a trypanosome lytic factor (TLF).⁴ Hence, the presence of circulating APOL1 protein was widely known. Circulating factors are reported to be causes of idiopathic FSGS, including soluble urokinase-type plasminogen activating receptor (suPAR) and others yet to be identified in recurrent FSGS after kidney transplantation.^5,6 It was therefore possible that circulating APOL1 protein might contribute to the development of CKD. APOL1 is a small 42 kd protein that is present in the kidney, enriched in glomeruli.^7,8 Characterization of circulating APOL1 protein using fast protein liquid chromatography and nondenaturing gradient gel electrophoresis revealed that serum APOL1 was present almost exclusively in 20 nm and 12.2 nm diameter protein complexes.⁹ These complexes are consistent with TLF1 and TLF2.¹⁰ There is virtually no free APOL1 in the serum.⁹ This differs from circulating free apolipoprotein A, which is filtered in the glomerulus and reabsorbed in the proximal tubule.¹¹ An individual’s APOL1 genotype does not associate with circulating APOL1 protein concentrations in healthy African Americans⁹ and plasma levels of risk-variant APOL1 do not associate with the presence or the severity of CKD.^12,13

This lack of association between APOL1 genotype and serum APOL1 protein levels or presence of kidney disease support results from clinical studies in kidney transplantation. Poorer long-term outcomes after receipt of a kidney from African American, versus European American, deceased donors¹⁴ and higher risk for subsequent nephropathy in African American living kidney donors have long been appreciated.¹⁵ The causes of these phenomena had been unknown. However, shorter renal allograft survival is present predominantly in recipients of kidneys from African American deceased donors with APOL1 high-risk genotypes.^16–18 Renal histology in those with failed allografts typically showed APOL1-associated kidney diseases, not rejection or sequelae of BK virus infection.¹⁶ Recipients of kidneys from African American deceased donors without APOL1 renal-risk genotypes show similar allograft survival as those from European American donors lacking APOL1 renal-risk variants. In addition, healthy African American living kidney donors with APOL1 high-risk genotypes were significantly more likely to develop Stage 3 or higher CKD, including ESRD, 12 years postdonor nephrectomy compared to those without APOL1 renal-risk genotypes.¹⁹ In contrast, there is preliminary evidence that recipient APOL1 genotypes do not affect renal allograft survival.²⁰ In aggregate, these findings reveal that circulating APOL1 protein is highly unlikely to be the primary cause of CKD.

Subsequent efforts to understand the role of APOL1 in development of native CKD and posttransplant allograft failure have focused on locally synthesized protein in the kidney ^7,8. Table 1 summarizes the results of these reports. Ma et al first reported that APOL1 mRNA was present in renal podocytes, glomerular endothelial cells and renal tubule cells on kidney cryosections, but not mesangial cells ⁸. In addition, APOL1 mRNA and protein were present in cultured human differentiated podocyte and renal tubule cell lines, as well as a glomerular endothelial cell line, but not in a mesangial cell line. The APOL1 expression level was comparable to that in HepG2 cells, a human hepatocyte cell line ⁸. These data provided fundamental support for locally synthesized APOL1 in kidney disease.

Table 1.

Postulated mechanisms of APOL1 risk variant protein-mediated renal toxicity

Mechanism	Model	Overview	Ref
Altered endosomal trafficking with inhibition of autophagic flux	Glomerulosclerosis in APOL1 TG mice with podocyte expression	No renal phenotype in TG mice with renal tubule cell expression	21
Cell membrane injury with resultant K+ efflux leading to induction of stress-activated protein kinases	Tetracycline-inducible human embryonic kidney (HEK) cell line	Toxic downstream effects of p38, JNK and ERK MAPK activation	25
Mitochondrial dysfunction & upregulation of TGFβ expression	HEK293 Tet-on cells; human primary PTC lines	Mitochondrial dysfunction preceded intracellular K+ depletion (see above)	26 29
Altered intracellular binding of APOL1 risk variant protein to others	Computer model	Altered VAMP8 binding leading to lysosomal injury and autophagy	31
Defective autophagy from altered endolysosomal trafficking	Drosophila melanogaster; Saccharomyces cerevisiae	TG flies and yeast assessing conserved trafficking processes	37
Interaction with suPAR concentration	Human cohorts with kidney disease	Higher plasma suPAR concentration interacts with APOL1 genotype	67
Innate immune system activation eradicating urinary JC polyoma virus	PCR in human urine	Absence of JC viruria suggests immune system activation with higher risk of renal fibrosis	46 47
Environmental exposures mediate development of kidney disease	Genome-wide association study human in human	No APOL1-by-SNP interaction effect reached genome-wide evidence of statistical significance	63

RRV – renal-risk variant; TG – transgenic; JNK – c-Jun N-terminal kinase; ERK – extracellular signal-regulated kinase; MAPK – mitogen-activated protein kinase; PTC - proximal tubule cell; PCR – polymerase chain reaction

Beckerman et al subsequently reported that transgenic expression of human APOL1 renal-risk variants in podocytes induced kidney disease in mice.²¹ This milestone study demonstrated that APOL1 G1 and G2 renal-risk variants directly cause kidney disease; they are not simply in linkage disequilibrium with other disease causing variants. Another report using a different APOL1 transgenic mouse model demonstrated podocyte depletion (but no overt renal phenotype) with the G2 renal-risk variant; this may have reflected lower levels of gene expression.²² Although the kidney disease phenotype only developed in podocyte-specific transgenic mice, implicating podocytes in development of APOL1-associated nephropathy, this does not exclude additional effects of APOL1 renal-risk variant proteins on human renal tubule cell and glomerular endothelial cell injury with contributions to CKD progression.²¹ Mouse proximal renal tubule cells have a shorter life span and lower proliferative capacity than human proximal tubule cells.^23,24 Renal tubule cell cycles and environmental exposures in mice also differ from those in humans. Finally, mice lack the APOL1 gene. Therefore, they likely do not fully recapitulate human kidney disease or intracellular handling of APOL1 proteins. The precise mechanisms whereby APOL1 G1 and G2 renal-risk variants induce injury to cells in the kidney, compared to the reference G0 variant, remain unknown.

To examine the intracellular function of APOL1, Olabisi et al utilized a tetracycline (Tet) inducible human embryonic kidney cell line (T-REx-293) conditionally overexpressing reference APOL1 G0 and variant G1 and G2 proteins.²⁵ They speculated that the nephrotoxicity of APOL1 renal-risk variants may be mediated by risk variant protein-induced loss of potassium (K⁺) from cells with subsequent induction of stress-activated protein kinase (SAPK) pathways. APOL1 forms K⁺-permeable cation-selective pores in the plasma membrane. Pores formed by G1 or G2 mediate increased efflux of intracellular K+, resultant depletion of intracellular K+ and subsequent activation of p38, JNK, and ERK MAPKs. The aberrantly activated SAPKs (p38 and JNK) cause cell death likely via their downstream mediators.

Ma et al used a similar cell model, HEK293 Tet-on cells, conditionally over-expressing comparable low levels of APOL1 while maintaining cell viability. A unique feature was that APOL1-associated pathways were detected from unbiased global gene expression profiles generated with 2 independent array systems.²⁶ An unsupervised machine-learning algorithm identified unique expression patterns across cells expressing APOL1 G0, G1, and G2. The top hits identified were up-regulation of the TGF-β signaling pathway and mitochondrial dysfunction, seen with both the G1 and G2 renal-risk variants. TGF-β induces activation of ERK, JNK and p38 MAPK through non-SMAD dependent pathways.²⁷ Hence, K⁺ depletion-induced cellular stress may interact with TGF-β signaling, a known critical mediator of renal fibrosis.²⁸ To prove mitochondrial dysfunction was involved, oxygen consumption rates were subsequently assessed using the Seahorse assay to validate bioinformatic findings.²⁶ APOL1 G0 overexpression increased oxygen consumption rate and preserved respiratory capacity, while G1 and G2 overexpression decreased oxygen consumption rate and reduced respiratory capacity. Decreases in mitochondrial function were present prior to intracellular K⁺ depletion in cells overexpressing G1 and G2. Hence, mitochondrial dysfunction-induced adenosine triphosphate (ATP) deficiency could inhibit plasma membrane Na⁺/K⁺ ATPase, resulting in intracellular potassium depletion. These results suggest mitochondrial dysfunction precedes intracellular K⁺ depletion.

APOL1 renal-risk variant effects on mitochondrial function were independently replicated.²⁹ The presence of APOL1 protein on the mitochondrial membrane was detected by fluorescence microscopy.²⁹ APOL1 protein in the podocyte endoplasmic reticulum (ER) also supported bioinformatic results suggesting ER stress in APOL1 G1- and G2-induced cellular dysfunction.^29,30 Because APOL1 protein localizes in membranes²⁵, its involvement in ER trafficking was not surprising. Madhavan et al reported that APOL1 G1/G2 variants may alter computationally modeled C-terminal conformations and binding to vesicle-associated membrane protein 8 (VAMP8).³¹ Incidentally, the C-terminal conformational change in G1 and G2 proteins likely leads to the reduced binding avidities of HPR-α to the APOL1 G1 and G2 variant serum lipoprotein complexes.³² VAMP8 is a soluble N-ethylmaleimide-sensitive factor-attachment protein receptor (SNARE) involved in autophagy through the direct control of autophagosome membrane fusion with the lysosome membrane. APOL1 is reportedly associated with regulation of autophagy and can induce lysosomal injury in human cells including podocytes.^33–35 Contributions of autophagy and endosomal trafficking to the functional integrity of podocytes have been clearly documented.³⁶

The postmitotic podocyte is the likeliest glomerular target for any toxic, immune, metabolic, or oxidant stress to the kidneys. These are terminally differentiated cells and they do not normally proliferate. Autophagy is a major pathway delivering damaged proteins and organelles to the lysosome in order to maintain cellular homeostasis. Recent work by Kruzel-Davila et al suggests that APOL1 renal-risk variants may impair conserved core intracellular endosomal trafficking processes and lead to defective autophagy.³⁷ Autophagy is a protective mechanism against aging in podocytes and protecting glomerular architecture. It represents a putative target to ameliorate human glomerular disease and age-related loss of renal function.³⁶ Current data support APOL1 G1 and G2 renal-risk variants as inducers of cellular dysfunction, with the likeliest causes relating to altered intracellular trafficking, autophagy, ER stress, mitochondrial dysfunction and intracellular potassium depletion.

Mitochondria regulate cellular energy homeostasis and cell death. Removal of damaged mitochondria through autophagy or mitophagy, a process selectively removing damaged or excessive mitochondria, is critical to maintain proper cellular function. This is particularly true for podocytes, because they no longer proliferate and permanent loss of function or cell death can have irreversible consequences on kidney function. It is possible that a mechanism underlying APOL1 G1- and G2-induced mitochondrial dysfunction is the process of mitochondrial fragmentation/fission versus fusion. These “mitochondrial dynamics” play critical roles in maintaining functional mitochondria when cells experience a metabolic or environmental stress.³⁸ Fusion allows mitochondria to compensate for each other’s defects by sharing components and helping to maintain energy output in the face of stress. However, fragmented mitochondria will be eliminated by autophagy when a threshold of cellular damage develops. Fission may segregate the most severely damaged mitochondria to preserve the health of the mitochondrial network, as well as regulate morphology and facilitate mitochondrial trafficking.³⁸ APOL1 is a mitochondrial membrane protein.²⁹ If G1 and G2 variant-induced mitochondrial fission/fragmentation cannot be compensated by appropriate mitophagy, perhaps due to defective intracellular trafficking as has been proposed, cell death machinery may be activated.^31,37,39 This mechanism may also apply to ER-phagy induced by APOL1 G1- and G2-related ER stress, when intracellular trafficking for autophagy is impaired.^30,40 Mitochondrial fission creates smaller, more discrete mitochondria, which, depending on the context, may be more capable of generating reactive oxygen species (ROS).⁴¹ These can activate TGF-β1, and TGF-β1 induces redox imbalance by creating a vicious cycle of increasing ROS production and suppressing antioxidant defenses.⁴² Although APOL1 renal-risk variants may form pores in the plasma membrane leading to intracellular potassium depletion, chronic ATP deprivation from mitochondrial dysfunction likely leads to reductions in intracellular K⁺ with activation of cell death pathways.

The evidence for intracellular dysfunction due to APOL1 G1 and G2 variants is mounting. However, only 20% of African Americans with APOL1 renal-risk genotypes ultimately develop nephropathy. As such, additional environmental or genetic “modifiers” must play roles in increasing APOL1 gene expression levels or amplifying APOL1 protein effects by interactions in biological pathways. The strongest known modifier in APOL1-associated kidney disease is HIV infection; odds ratios in HIVAN range from 29 to 89.^43,44 HIV is a potent inducer of interferons and other innate antiviral immune components, and interferons and other inflammatory factors further enhance APOL1 expression.⁴⁵

To assess whether other, non-HIV, viral infections might influence nephropathy risk via interactions with APOL1 in African Americans, APOL1 genotypes and clinical phenotypes were assessed along with the presence of urine JC and BK polyoma virus and plasma Human Herpes Virus 6 (HHV6) and cytomegalovirus (CMV) replication.⁴⁶ These 4 viruses share properties with HIV, including maintaining renal reservoirs of infection (JC and BK polyomavirus) or lymphotrophism (HHV6 and CMV). The presence of JC virus replication in urine was associated with a significantly lower risk of kidney disease in those with 2 APOL1 renal-risk variants. This surprising result was replicated in African Americans with and without APOL1 high-risk genotypes.⁴⁷ JC polyomavirus does not typically cause kidney disease. Therefore, we suspect that those with upregulated innate immune systems more aggressively eradicate JC polyoma virus from the kidneys and urinary tract. As a consequence, they are more likely to develop renal inflammation with resultant APOL1-induced glomerulosclerosis given other stressors. In contrast, those with downregulated innate immune systems permit JC polyomavirus to replicate and are less likely to express APOL1 in their kidneys or develop nephropathy. JC virus could be an epiphenomenon, a marker of innate immune system activation. In this fashion, activity of the innate immune system may serve as a modifier of risk for APOL1-associated kidney disease.

Most circulating APOL1 protein is secreted by the liver.⁴⁸ Donor APOL1 genotypes have no effect on outcomes after liver transplantation.⁴⁹ A question that arises is why organ donor APOL1 renal-risk variants only influence outcomes in kidney transplantation but not liver?^18,50 One factor may be that the renal interstitium is more acidic than liver tissue.⁵¹ Because APOL1 forms pH-gated cation-selective channels in planar lipid bilayers and confers pH-switchable ion permeability to phospholipid vesicles, we speculate that the low pH in the kidney may interact with APOL1 G1 and G2 to alter intracellular function.^52,53 APOL1 is abundant in renal tubule cells and interstitial changes strongly influence kidney transplant outcomes.⁸ Hence, effects of APOL1*pH interactions remain plausible explanations for divergent observations in liver and kidney transplantation. One could argue that APOL1-associated native kidney disease is more likely a podocyte (glomerular) disorder. However, crosstalk between sirtuin 1 (Sirt1) in renal proximal tubule cells modulates the expression of podocyte Claudin-1, a key regulator of albuminuria and glomerular function.⁵⁴ With this crosstalk between tubular cells and podocytes, it may be possible to have an improved understanding of the nature of idiopathic FSGS in the native kidney.

Additional in vivo human data may be collected with initiation of the National Institutes of Health-sponsored APOL1 Long-Term Kidney Transplantation Outcomes Network (APOLLO).⁵⁵ These data may improve understanding of the pathways leading to APOL1 renal-risk genotype-related accelerated failure of transplanted kidneys.

Predicting nephropathy onset in those with APOL1 renal-risk genotypes

APOL1 has transformed our understanding of nephropathy in African Americans. HIVAN, idiopathic FSGS, “hypertension-attributed” nephropathy (a primary kidney disease manifesting as solidified glomerulosclerosis with low level proteinuria), severe lupus nephritis and sickle cell nephropathy contribute the majority of cases of nondiabetic ESRD in African Americans and have ORs for association with APOL1 renal-risk variants ranging from 5 to 89.^{43,44,56–58} However, inheriting 2 APOL1 risk variants is necessary, but insufficient for the development and/or progression of CKD. In fact, the kidney diseases above all require a modifier or second hit. This supports the existence of APOL1-by-second gene or APOL1-by-environment interaction effects in CKD.

APOL1-by-second gene interaction

Several groups have studied the APOL1-by-second gene interaction effect on risk for nondiabetic CKD. For example, we performed a candidate-gene approach and reported that variants in or near the podocin (NPHS2), serologically defined colon cancer antigen 8 (SDCCAG8), ecto-NOX disulfide-thiol exchanger 1 (ENOX) and bone morphogenetic protein 4 (BMP4) genes might interact with APOL1 to alter risk for CKD.^59,60 APOL1-by-second gene interaction effects have also been reported with a null variant in the apolipoprotein L3 (APOL3) and glutathione s-transferase mu 1 genes (GSTM1).^61,62

Langefeld et al recently conducted genome-wide association studies (GWAS) in 1749 nondiabetic ESRD cases and 1136 controls recruited at Wake Forest, and replicated results in 901 cases with lupus nephritis (LN)–ESRD and 520 controls with systemic lupus erythematosus lacking nephropathy.⁶³ No APOL1-by SNP interaction effect reached genome-wide evidence for statistical significance. It was concluded that strong gene-by-APOL1 interaction effects on ESRD were unlikely. Suggestive evidence of associations was akin to the results observed in GWAS of other complex traits. That is, a large number of variants with small effect sizes; ORs ranging between 1.1–1.6. We concluded from this GWAS that strong second hits in APOL1-associated nephropathy more likely reflect environmental effects.

APOL1-by-environment interaction

Many APOL1-by-environment interaction effects have been reported; including interactions with non-HIV viral infections (such as JC polyomavirus, JCPyV) and biomarkers of immune response and inflammation. We reported that infection/colonization of the kidneys and urinary tract with JCPyV had a paradoxical protective effect on risk for CKD and presence of APOL1 renal-risk genotypes modified association effects. Other studies extended this protective association with JCPyV to non-APOL1 associated etiologies of nephropathy.^46,47,64,65

Circulating factors can lead to rapid recurrence of FSGS after a kidney transplant.⁶ This is more common in younger patients whose native disease was refractory with heavy proteinuria and rapidly declining kidney function; this form of recurrent FSGS does not appear to be APOL1-associated. However, circulating soluble urokinase-type plasminogen activator receptor (suPAR), a marker of immune system activation and a signaling molecule, may mediate some cases of FSGS and other forms of CKD.^5,66 Interactions between suPAR concentration and APOL1 genotype have been reported, suggesting circulating suPAR interacts with APOL1 to activate ανβ integrin and produce CKD.⁶⁷ Additional work is required to assess the role of this relationship in transplantation.

Nichols et al demonstrated that trypanosome-derived molecules tended to act as innate immune Toll-like receptor (TLR) agonists and lead to overexpression of APOL1 protein and proinflammatory cytokines.⁴⁵ These analyses were inspired by the observation that all patients on therapeutic interferon who developed collapsing glomerulopathy, a variant of FSGS, possessed recent African ancestry and APOL1 renal-risk genotypes. Results were supported by data generated in cell cultures where interferons and TLR could increase APOL1 expression up to 200-fold. This manuscript overcame an important obstacle by using human cells. Results were likely more relevant than those obtained in rodent models since the APOL1 gene and renal-risk variants are only present in higher order primates. This study highlighted the complex interplay between immunity and APOL1. Nadkarni et al assessed interaction effects between inflammatory biomarkers and APOL1 for predicting CKD risk.⁶⁸ Plasma tumor necrosis factor receptors 1 and 2 (TNFR1, TNFR2) and kidney injury molecule (KIM1) concentrations were independently associated with renal outcomes. Although these markers may help refine the predictive power of APOL1, it remains uncertain whether elevated levels reflected the presence of ongoing kidney disease and additional replication studies are required.

Conclusions

Considerable progress has been made in understanding the pathogenesis of APOL1-associated native CKD and the effects of donor APOL1 genotype on allograft survival after kidney transplantation. Identification of disease modifiers will provide new targets for CKD prevention and prolonging allograft survival after kidney transplantation. We believe it is likely that effective treatments for APOL1-associated nephropathy will soon be developed. They will target the renal production of risk variant APOL1 protein and modifiers of APOL1 gene effect. These advances are important to all physicians and healthcare professionals who treat patients with (and at risk for) CKD. The earliest benefits of clinical APOL1 genotyping will likely take place in transplantation. Retrospective studies support that kidney donor APOL1 genotypes predict the likelihood of long-term renal allograft survival and the safety of living donation in individuals with recent African ancestry. The ongoing APOLLO study is prospectively assessing the outcomes in deceased donor kidney transplantation based on donor and recipient APOL1 genotypes (and their interaction), as well as longitudinally following living kidney donors for postdonation renal function and proteinuria. If retrospective results hold, APOL1 screening may make living kidney donation safer in African Americans and more accurately assess the quality of deceased donor kidneys. If the donor APOL1 genotype proves more informative than donor race in determining outcomes after deceased donor kidney transplantation, approximately 85% of kidneys donated by African Americans will be shown to be of better quality than currently thought. We believe this will reduce the discard of good quality kidneys, increase the numbers of kidneys available for all on the wait list and improve matching of kidney donors with recipients to permit longer renal allograft survival. It is apparent that the 2010 laboratory discovery of the role of APOL1 in nephropathy has reached the clinic.