Genetic variants in the APOL1 gene associated with kidney disease in Black individuals were first reported in 2010.1 There is now little doubt that the two coding risk alleles, G1 and G2, are the causal drivers of disease in a spectrum of clinical phenotypes, including FSGS, hypertension-associated kidney failure, HIV nephropathy, and several others. Despite an abundance of evidence that G1 and G2 act as toxic, gain-of-function variants, there is no consensus about how APOL1 injures kidney cells, the range of cell types affected, and the reason for the very different clinical phenotypes we observe in patients.

Long before any connection between APOL1 and kidney disease was established, APOL1 was known as an important trypanolytic factor in human serum, protecting humans and some other primates from the common African trypanosome Trypanosoma brucei. The G1 and G2 risk variants increased in frequency because they have enhanced activity against chronic and acute African sleeping sickness caused by trypanosomes Trypanosoma brucei gambiense and Trypanosoma brucei rhodesiense, respectively. Because trypanolytic activity is believed to rely on pore/channel function, an obvious candidate mechanism is deleterious channel activation of the risk variants in the kidney. Although many alternative mechanisms have been proposed, evidence supporting APOL1 risk variant toxicity through aberrant channel function, presumably in the podocyte, is the most developed hypothesis.

The fact that APOL1 can and does operate as an ion channel is clear. Elegant studies by Thomson and Finkelstein in a lipid bilayer system worked out many of the important characteristics of APOL1 channel activity. They showed that the protein enters the bilayer from a low-pH environment but requires neutralization of the pH to switch from low to high cation conductance.2 Subsequent bilayer experiments indicate that it is the C-terminus of APOL1 that functions, likely in at least a dimeric state, as the channel rather than the N-terminus as was originally proposed due to its purported structural relationship to bacterial colicin domains.3,4 Experiments in liposomes suggest that APOL1 risk variants may flux more current than non-risk G0 APOL1 haplotypes, providing a potential explanation for differences in toxicity to the kidney.5

Many investigators have extended the examination of channel function into kidney cells. Whole-cell patch and x-ray fluorimetry techniques in HEK293 cells demonstrated that overexpression of APOL1 drives cation currents and that those currents were larger with APOL1 risk variants than with G0.6,7 Trypanosomal virulence factor serum resistance-associated (in the case of G0) or anti-APOL1 antibodies (in the case of both G0 and G1) could block the currents, showing that APOL1 was expressed on the cell surface and acting as an ion channel.7 Several of these findings were extended to primary human podocytes in a more physiologic model where APOL1 expression was induced with interferon instead of through artificial overexpression systems.7 Deleting APOL1 using gene editing in these podocytes eliminated the current induced by interferon.7

Clear connections were established between APOL1 risk variant overexpression and an array of cell phenotypes, especially cell death. HEK293 cells proved exquisitely sensitive to the toxic effects of APOL1 risk variants. This easy-to-measure readout became a major workhorse for examining the APOL1 mechanism of injury. Many pathways from APOL1 risk variant overexpression to cell death have been delineated through careful and compelling experimentation, leading to a rapidly expanding list of cellular dysfunctions initiated by APOL1. Among the important events along this pathway are cell surface ion flux, mitochondrial dysfunction, elevation of intracellular calcium, and activation of specific cell death pathways. Blocking any of these events can rescue cell death. Important facets of this dysfunction can be recapitulated in primary mouse and human podocytes.8

These data provide ample evidence to support the APOL1 ion channel model of kidney injury. However, there are nontrivial gaps in the model that need to be filled in before the channel theory can be adopted with confidence and other theories discarded. It is premature to conclude that all APOL1 nephropathy results from ion channel activation of risk variants in the plasma membrane of podocytes.

First, a causal chain that starts with APOL1 risk variant channel activity and leads to cell death is at present circumstantial. Documenting the order in which events can be detected is suggestive of their chronology, but can be confounded by the different sensitivity of the assays for different cellular functions. Localizing the site of APOL1 toxic activity is equally problematic. APOL1 likely accumulates where it is non-toxic, whereas even a tiny fraction of the cellular APOL1 that is hard to detect by conventional methods may be toxic to the cell. Only a few percent of total APOL1 is found in the cell membrane, which certainly does not imply that cell membrane APOL1 is not important, but neither does it suggest that APOL1 cannot cause major dysfunction in cellular compartments with even smaller fractions of the protein mass. A compound shown to bind APOL1, block its channel function, and reduce proteinuria in early human studies of APOL1-mediated FSGS may also bind and block the activity of APOL1 everywhere in the cell.9 The field awaits a demonstration that specifically blocking APOL1 membrane current without affecting other pools of APOL1 in the cell prevents cell death or other readouts.

Second, APOL1 cell models are useful but limited. HEK293 cells, transformed podocytes, and even primary podocytes have little in common with the complex, highly differentiated podocyte in vivo. Cell death is a useful readout primarily because it is easy to quantify, but early pathology in mouse models seem to point toward podocyte foot process effacement rather than cell death.10 Experiments in cell models can at best provide hypotheses for further testing in more complex systems such as mouse models. Most current mouse models, while recapitulating many important features of human disease, currently reflect hyperacute, severe disease that represents only a small fraction of human APOL1 nephropathy.

Third, any claims about the mechanism of APOL1 kidney disease will have to grapple with the problem of recessive, toxic gain-of-function. In other words, why are two risk variants required for the large increase in risk of APOL1 nephropathy? Arguments fall into two major categories, neither of which is entirely satisfactory. The G0-rescue model posits that G0 may be able to neutralize the toxic effect of G1 or G2, possibly by direct binding.11 This model has intuitive appeal, but to date, the model has not been supported by experimental data.10 The other main hypothesis is the threshold model, which argues that one risk variant is not enough to cause disease, but with two risk alleles, some disease-causing threshold of APOL1 expression is crossed. While possible, it is surprising that the dosage threshold can be transferred intact to human APOL1 transgenic mice, where risk variant homozygotes but not hemizygotes develop kidney disease in response to interferon.10 A compelling channel model will need to address this recessive gain-of-function paradox by elucidating the properties of mixed genotype APOL1 multimeric channels or by identifying the cellular factors that explain the extraordinary sensitivity of the cell to two-fold difference in protein quantity from a highly inducible gene.

If you know the normal cellular function of a protein, testing the mechanism of a loss-of-function variant is straightforward in principle if not practice. Understanding gain-of-function variants seems much trickier, especially in a gene without a known function in the cell of interest. APOL1 is an innate immunity gene that does not yet appear to have a role in the basic operation of the kidney. A loss-of-function variant fails to do its essential job, but a gain-of-function protein can in theory misbehave in any number of ways, in any number of cell types, and in many different organelles. The growing list of phenotype differences between G1 and G2 risk variants is evidence in itself that APOL1 gain-of-function properties may be more diverse than currently recognized. Incorporating the study of human variants that modify APOL1 toxicity (for example N264K) could help illuminate risk variant mechanism of injury.12

Most likely, our model for causality needs updating. APOL1 nephropathy can present as explosive proteinuria with rapid glomerular injury and kidney functional impairment or can progress over years with slow, insidious decline in kidney function with vascular injury and low-grade proteinuria. It seems conceivable, and even likely, that more than one mechanism is possible in more than one cell type. We would not be surprised if collapsing glomerulopathy was caused by APOL1 channel activity in the podocyte plasma membrane, whereas slowly progressive kidney disease involved mitochondrial injury in the endothelium. The exact opposite is possible too. We would all be better served by talking about the mechanisms rather than the mechanism of APOL1 nephropathy. At the least, we should resist premature closure and keep an open mind. Anyone who has spent time studying APOL1 knows that it is a gene full of surprises.

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at https://links.lww.com/JSN/E635.

Funding

D.J. Friedman and M.R. Pollak: US Department of Defense (W81XWH2010826), NIDDK (DK138503), and NIMHD (MD014726).

Acknowledgments

The content of this article reflects the personal experience and views of the authors and should not be considered medical advice or recommendation. The content does not reflect the views or opinions of the American Society of Nephrology (ASN) or JASN. Responsibility for the information and views expressed herein lies entirely with the authors.

Author Contributions

Conceptualization: David J. Friedman.

Visualization: David J. Friedman.

Writing – original draft: David J. Friedman, Martin R. Pollak.

Writing – review & editing: David J. Friedman, Martin R. Pollak.

References 1. Genovese G, Friedman DJ, Ross MD, et al. Association of trypanolytic ApoL1 variants with kidney disease in African Americans. Science. 2010;329(5993):841–845. doi:10.1126/science.1193032 2. Thomson R, Finkelstein A. Human trypanolytic factor APOL1 forms pH-gated cation-selective channels in planar lipid bilayers: relevance to trypanosome lysis. Proc Natl Acad Sci U S A. 2015;112(9):2894–2899. doi:10.1073/pnas.1421953112 3. Schaub C, Lee P, Racho-Jansen A, et al. Coiled-coil binding of the leucine zipper domains of APOL1 is necessary for the open cation channel conformation. J Biol Chem. 2021;297(3):101009. doi:10.1016/j.jbc.2021.101009 4. Schaub C, Verdi J, Lee P, et al. Cation channel conductance and pH gating of the innate immunity factor APOL1 are governed by pore-lining residues within the C-terminal domain. J Biol Chem. 2020;295(38):13138–13149. doi:10.1074/jbc.RA120.014201 5. Bruno J, Edwards JC. Kidney-disease-associated variants of Apolipoprotein L1 show gain of function in cation channel activity. J Biol Chem. 2021;296:100238. doi:10.1074/jbc.RA120.013943 6. Olabisi OA, Zhang JY, VerPlank L, et al. APOL1 kidney disease risk variants cause cytotoxicity by depleting cellular potassium and inducing stress-activated protein kinases. Proc Natl Acad Sci U S A. 2016;113(4):830–837. doi:10.1073/pnas.1522913113 7. Vandorpe DH, Heneghan JF, Waitzman JS, et al. Apolipoprotein L1 (APOL1) cation current in HEK-293 cells and in human podocytes. Pflugers Arch. 2023;475(3):323–341. doi:10.1007/s00424-022-02767-8 8. Datta S, Antonio BM, Zahler NH, et al. APOL1-mediated monovalent cation transport contributes to APOL1-mediated podocytopathy in kidney disease. J Clin Invest. 2024;134(5):e172262. doi:10.1172/JCI172262 9. Egbuna O, Chertow GM. Inaxaplin for proteinuric kidney disease in persons with two APOL1 variants. Reply. N Engl J Med. 2023;388(26):2491. doi:10.1056/NEJMc2304780 10. McCarthy GM, Blasio A, Donovan OG, et al. Recessive, gain-of-function toxicity in an APOL1 BAC transgenic mouse model mirrors human APOL1 kidney disease. Dis Model Mech. 2021;14(8):dmm048952. doi:10.1242/dmm.048952 11. Limou S, Dummer PD, Nelson GW, Kopp JB, Winkler CA. APOL1 toxin, innate immunity, and kidney injury. Kidney Int. 2015;88(1):28–34. doi:10.1038/ki.2015.109 12. Gbadegesin R, Martinelli E, Gupta Y, et al. APOL1 genotyping is incomplete without testing for the protective M1 modifier p.N264K variant. Glomerular Dis. 2024;4(1):43–48. doi:10.1159/000537948