ApoL1 Overexpression Drives Variant-Independent Cytotoxicity

Abstract

Coding variants in the APOL1 gene are associated with kidney diseases in African ancestral populations; yet, the underlying biologic mechanisms remain uncertain. Variant-dependent autophagic and cytotoxic cell death have been proposed as pathogenic pathways mediating kidney injury. To examine this possibility, we conditionally expressed APOL1-G0 (reference), -G1, and -G2 (variants) using a tetracycline-regulated system in HEK293 cells. Autophagy was monitored biochemically and cell death was measured using multiple assays. We measured intracellular Na⁺ and K⁺ content with atomic absorption spectroscopy and APOL1-dependent currents with whole-cell patch clamping. Neither reference nor variant APOL1s induced autophagy. At high expression levels, APOL1-G0, -G1, and -G2 inserted into the plasma membrane and formed pH-sensitive cation channels, causing collapse of cellular Na⁺ and K⁺ gradients, phosphorylation of p38 mitogen-activated protein kinase, and cell death, without variant-dependent differences. APOL1-G0 and -G2 exhibited similar channel properties in whole-cell patch clamp experiments. At low expression levels, neither reference nor variant APOL1s localized on the plasma membrane, Na⁺ and K⁺ gradients were maintained, and cells remained viable. Our results indicate that APOL1-mediated pore formation is critical for the trypanolytic activity of APOL1 and drives APOL1-mediated cytotoxicity in overexpression systems. The absence of cytotoxicity at physiologic expression levels suggests variant-dependent intracellular K⁺ loss and cytotoxicity does not drive kidney disease progression.

Common genetic variations in the APOL1 gene have been linked to progressive kidney diseases in populations with African ancestry.^1–9 APOL1-G0 is the reference sequence and the two kidney disease–associated haplotypes are APOL1-G1, consisting of two nonsynonymous sequence variants in high linkage disequilibrium (rs73885319, p.S342G and rs60910145, p.I384M), and APOL1-G2, a 6-bp in-frame deletion resulting in loss of two amino acids (rs71785313, p.N388_Y389del). The APOL1-G1 and -G2 variants are rare in non-African populations, but are common in populations with African ancestry, such as blacks, where the allele frequencies have been reported as 18%–21% and 13%–15%, respectively.¹

The APOL1 gene is restricted to humans and a few nonhuman primate species and has a role in innate immunity. Trypanosoma brucei is a blood-borne parasite endemic to sub-Saharan Africa, which infects livestock, leading to nagana, but humans resist infection caused by this parasite. Circulating APOL1, contained in a subclass of HDL particles, mediates this resistance by lysing the trypanosome (trypanolysis).¹⁰ Its ability to form channels in lipid bilayers is central to its trypanolytic activity.^11–14 The APOL1-G2 variant confers enhanced protection from infection with the trypanosomal subspecies, Trypanosoma brucei rhodesiense, and the APOL1-G1 variant is associated with less severe clinical disease after infection with another trypanosomal subspecies, Trypanosoma brucei gambiense.¹⁵ These trypanosomal subspecies cause African sleeping sickness in humans and a single APOL1-G1 or -G2 allele confers these selective advantages, which have likely driven their high frequencies in black populations.

APOL1 has also been localized in arteriolar endothelial cells and podocytes within the normal human kidney as well as in kidney biopsy samples of individuals with APOL1-associated diseases.^16,17 The lack of an association between circulating levels of normal and variant APOL1 with kidney phenotypes^18,19 and the positive association of kidney allograft failure in kidney transplant recipients that received organs with the risk genotypes^20–22 point to APOL1 within the kidney as the driver of APOL1-associated kidney diseases. However, the functional role of APOL1 within the cells of the kidney is currently unknown.

Overexpression of APOL1 in mammalian cell lines or model organisms has been used to model the mechanisms mediating APOL1-associated kidney diseases. Overexpression of APOL1-G0 and/or the APOL1-G1 and -G2 variants causes cell death.^23–30 Autophagic cell death was the first mechanism proposed for APOL1 toxicity in mammalian cells,²³ but not all subsequent reports replicated this result.^26,28,30 Lysosomal damage, a proposed mechanism of trypanolysis,¹² has also been implicated in APOL1-mediated mammalian cytotoxicity.²⁶ Finally, similar to other trypanolytic mechanisms,^11–14 APOL1-mediated cell death can result from its channel activity when it inserts into the plasma membrane, which allows loss of cellular ion gradients, osmotic swelling, and ultimately cell death.^28,30 A key question that remains outstanding is whether the APOL1-G1 and -G2 cause differential effects on cytotoxicity compared with APOL1-G0, and whether cytotoxicity is the mechanism driving the pathogenesis of APOL1-associated kidney diseases. Therefore, we have designed the following experiments in order to address the relative toxicity of APOL1-G0 compared with APOL1-G1 or -G2, the mechanism of APOL1-mediated cell death, and the effect of APOL1 expression on autophagic processes.

Results

We first generated cell lines that conditionally expressed with tetracycline treatment the reference sequence of APOL1 (APOL1-G0) and the APOL1 variants that have been associated with progressive kidney diseases, APOL1-G1 and -G2. APOL1 levels were measured after the addition of tetracycline (1 μg/ml) in three clones for each APOL1 genotype (Figure 1). Tetracycline-inducible APOL1 expression was variable across clones (Figure 1, A and B), detectable after 4 hours of tetracycline treatment, and continued to increase at the 8-hour time point (Figure 1B). Using tetracycline (at 1 μg/ml) minimized differences in the APOL1 induction rate and steady state abundance compared with lower tetracycline concentrations. One clone from each genotype with matched levels of APOL1 expression was examined for APOL1 expression using confocal immunofluorescent microscopy (Figure 1C). The APOL1 expression level across a population of cells and its subcellular distribution appeared qualitatively similar across all three genotypes.

Figure 1.

Comparable expression of APOL1-G0, -G1, and -G2 can be achieved in stable cell lines. APOL1 expression was assayed by immunoblotting 293 cell lines that conditionally express APOL1-G0, -G1, and -G2 (G0, G1, and G2) with tetracycline induction. Three clones of each genotype were studied (labeled A, B, C for each genotype). (A) Band density was quantitated using ImageJ software, and the expression of APOL1 normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) is shown for each clone tested before tetracycline treatment (0 hour) and 4 and 8 hours after the addition of tetracycline (1 µg/ml). (B) APOL1 expression, determined by immunoblotting, after 0, 4, and 8 hours of tetracycline induction for three G0, G1, and G2 clones each. Blots were stripped and reprobed for GAPDH expression as loading controls. (C) One clone for each genotype (G0 C, G1 C, G2 B) judged to have comparable levels of APOL1 expression by immunoblotting was examined with confocal immunofluorescent microscopy for APOL1 expression before and 5 hours after the addition of tetracycline (1 μg/ml). Scale bars=25 µm.

Because APOL1 expression has been associated with autophagic cell death,^23,26 we evaluated the effect of APOL1 expression on autophagy (Figure 2). We examined three clones for each genotype (Figure 1) to account for clone-dependent differences. The normalized expression of LC3II in each of the nine clones examined is shown (Figure 2, A and B) after 8 and 16 hours of induction, respectively, graphically illustrating the degree of variability between clones. The remaining graphs (Figure 2, C–F) depict average normalized LC3II or p62 abundance of three independent clones for each genotype in a single bar. APOL1 did not affect autophagy at 8 hours (Figure 2C, Supplemental Figure 1A) or 16 hours (Figure 2D, Supplemental Figure 1A). Consistent with these data, APOL1 did not significantly change the normalized p62 expression at 8 hours (Figure 2E) and 16 hours (Figure 2F). Autophagic flux was increased with serum starvation, but was not altered by the expression of APOL1-G0, -G1, or -G2 (Supplemental Figure 1B). Although p62 levels tend to be lower at 16 hours, the abundance of p62 did not increase with bafilomycin A treatment consistent with the conclusion that neither reference nor variant APOL1s activate autophagic flux. Tetracycline treatment alone did not alter autophagic flux in parental 293 cell lines or affect the expression of GAPDH used for normalization (Supplemental Figure 1C).

Figure 2.

Stable expression of APOL1 does not induce autophagy. The protein abundance of LC3II, p62, and GAPDH was assessed by immunoblotting in three clones of each genotype (G0, G1, and G2) after tetracycline induction (1 μg/ml). The bar graphs show the normalized expression of (A–D) LC3II or (E and F) p62, two biochemical markers of autophagy, 8 or 16 hours after tetracycline addition. (A and B) The results for three clones of each genotype, demonstrating genotype-independent, clonal variability in LC3II abundance. (C–F) Bar graphs of normalized LC3II and p62 abundance stratified by APOL1 genotype (mean±SD). Conditions include no treatment (Cont), treatment with bafilomycin A (Baf, 100 nM) alone to induce an autophagic block, treatment with tetracycline alone (Tet, 1 μg/ml), and treatment with tetracycline and bafilomycin A (Tet+Baf). Cells were treated with tetracycline for 8 or 16 hours, as indicated. Bafilomycin was added for the last 3 hours. Within-genotype treatment differences were not statistically significant (single-factor ANOVA).

Increased cytotoxicity of the APOL1-G1 and -G2 variants has been proposed as a mechanism of APOL1-associated kidney diseases; therefore, we examined our stable cell lines for variant-dependent death. In contrast to published data, cytotoxicity/viability ratios determined by a fluorogenic assay were similar in all genotypes compared with uninduced control cells (Figure 3A). The average cytotoxicity/viability ratio of the clones expressing APOL1-G0, the reference sequence, was higher than those expressing either the APOL1-G1 (P value 0.14) or APOL1-G2 (P value <0.001) kidney disease risk variants. An alternative assay of cell viability uses methyl-thiazolyl-tetrazolium (MTT), which is reduced in metabolically active cells to produce a blue precipitate, formazan.³¹ Eight hours after APOL1-G0, -G1, or -G2 induction, cell viability was equivalent to uninduced control cells (Figure 3B). However, after 16 hours of APOL1 expression cytotoxicity increased and was indistinguishable between genotypes (Figure 3C).

Figure 3.

APOL1-induced cytotoxicity is variant-independent. (A) Fluorescent assay demonstrating no difference in the average cytotoxicity-to-viability ratio, normalized to control untreated cells, in 293 cells stimulated with tetracycline to express APOL1-G0, -G1, or G2 (G0, G1, G2). Each bar represents mean±SD of three independent experiments with one to two independent clones, with two to four technical replicates (n=12–14). (B) MTT assays after 8 hours of tetracycline (1 μg/ml) induction demonstrated no loss of viability in G0, G1, or G2 cells compared with uninduced cells (mean±SD, n=2–3 independent experiments, each with three technical replicates, for each APOL1 genotype). (C) In contrast, at 16 hours of tetracycline (1 μg/ml) induction, cytotoxicity is marked in G0, G1, and G2 cells compared with uninduced control cells (mean±SD, n=2 independent experiments, each with two to three technical replicates, for each APOL1 genotype). MTT assays in control cells of each genotype were normalized to 100% viability in each independent experiment.

Many cell death assays use biochemical, morphologic, or physiologic surrogate measures for cell death, which quantify early kinetic processes that regulate but do not always result in cell death.³² To address this possibility we performed a clonogenic survival assay, which is generally regarded as the gold standard for the determination of cellular cytotoxicity/survival after a cytotoxic exposure such as APOL1. Two hours of induction (tetracycline, 1 μg/ml) of reference or variant APOL1s resulted in almost complete loss of cell viability (Figure 4). Treatment with glycine to nonspecifically block ion-fluxes in pyroptosis,³³ digoxin to inhibit autosis,³⁴ or wortmanin to inhibit autophagy³⁵ failed to rescue the reference or variant APOL1 cell lines from cell death in the clonogenic assay (Figure 4).

Figure 4.

Clonogenic survival assays demonstrate variant-independent cell death after APOL1 expression. Bar graph shows the fraction of surviving cells in the test condition normalized to untreated control cells (mean+SD, n=3). The vertical axis indicates the APOL1 genotype and treatments. The normalized surviving fraction of 1.0 is highlighted in red. Respective APOL1 genotypes are abbreviated G0, G1, or G2. C-Del338+Tet, 16 hours after plating the cells tetracycline 1 µg/ml was added to media to induce expression in a stable 293 cell line of an APOL1 transgene with a deletion of carboxy terminal amino acids, 339–398; Dig, digoxin (1 nM, 6 hours); Gly, glycine (5 mM glycine added to culture media throughout assay); 2HrTet, tetracycline (1 μg/ml, 2 hours); LowTet, 16 hours after plating the cells tetracycline was added at 5 ng/ml for APOL1-G0 and APOL1-G1 and 10 ng/ml for APOL1-G2; NoTx, no additions; Tet, tetracycline (1 μg/ml, 16 hours); v., versus; Wort, wortmanin (50 nM, 16 hours).

APOL1 has been reported to have channel properties when expressed in oocytes or when purified and reconstituted into lipid bilayers.^11–13,28 APOL1 appears to form monovalent cation channels permeable to both Na⁺ and K⁺. In order to determine the relationship between the appearance of channel activity and the onset of cell death markers, we used atomic absorption spectroscopy to measure cell-associated Na⁺ and K⁺ after 0, 4, 5, 6, 7, 8, 9, and 10 hours of tetracycline (1 μg/ml)-induced expression of APOL1-G0, -G1, and -G2. Cell-associated K⁺ and Na⁺ were similar in cell lines expressing the three APOL1 genotypes in the absence (time 0) and presence of tetracycline for up to 4 hours (Figure 5, A and C). However, continued treatment with tetracycline for 5–10 hours resulted in a progressive loss of cellular K⁺ and gain of cellular Na⁺ in clones from each APOL1 genotype, reaching maximal effect by 10 hours (Figure 5, B and D). Expression of APOL1 with a deletion of the carboxy-terminal amino acids 339–398, for up to 24 hours, failed to induce any change in cellular Na⁺ or K⁺, nor was there evidence of cell death using the clonogenic assay (Figure 4, Supplemental Figure 2B). Cytosolic free Ca²⁺ concentration ([Ca²⁺]_i) measured using fura-2, was similar with or without 8 hours of induction of any of the APOL1 genotypes. Furthermore, expression of the APOL1 genotypes had no effect on the transient increase in [Ca²⁺]_i observed after stimulation of endogenous purinergic receptors (Supplemental Figure 3). Together, these results demonstrate that APOL1 expression correlates with a variant-independent loss of K⁺ and a gain of Na⁺ by the cell, without a change in Ca²⁺ homeostasis or signaling, which precedes loss of cell viability.

Figure 5.

APOL1-G0, -G1, -G2 cause a time-dependent loss of cellular K⁺ and a gain of cellular Na⁺. Cell-associated (A) K⁺ and (C) Na⁺ were measured as described in the Concise Methods in G0, G1, or G2 cells without or with tetracycline (1 µg/ml) for 4 hours (mean±SD, n=5–6). Cellular (B) K⁺ and (D) Na⁺ content as a function of time after tetracycline addition is shown, normalized to time 0 (mean±SD, n=3). (B) Hatches between the time 0 and the 5-hour time measurement added for clarity, illustrating this line connects these two time points and does not represent additional interval measurements. Tet, tetracycline.

Previous reconstitution studies showed that the channels formed by APOL1 are cation selective and inhibited by a reduction in bath pH.¹³ If in fact APOL1 channels are responsible for the loss of K⁺ and gain of Na⁺, we reasoned that APOL1 channel activity should be measurable at induction times >4 hours. In order to test this hypothesis, whole-cell patch-clamp experiments were performed on APOL1-G2 cells. Preliminary studies showed the channels induced by APOL1-G2 expression were permeable to Na⁺, K⁺, and Cs⁺. In subsequent experiments, we used a normal bath solution, and a Cs⁺ pipette solution to eliminate endogenous K⁺ currents. Under these recording conditions, whole-cell currents were small in APOL1 cells in the absence (not shown) or presence of tetracycline for 4–5 hours (Figure 6B). When measured 6–7 hours after induction, large currents were observed immediately upon establishment of whole-cell recording mode (Figure 6, A and B). The currents had a reversal potential near 0 mV and showed a slight outwardly rectifying I-V relationship (Figure 6C).

Figure 6.

APOL1 expression is associated with the appearance of cation-selective and pH-sensitive whole-cell membrane currents. (A) Whole-cell membrane currents were recorded in an APOL1-G2–expressing 293 cell 6 hours after addition of tetracycline (1 µg/ml). Voltage ramps were applied every 6 seconds and the outward current at +60 mV (red) and inward current at −80 mV (black) during each ramp are plotted as a function of time after rupture of the patch for whole-cell recording. At the time indicated by the horizontal bar (top), the bath solution was changed from pH 7.2 to pH 5.1 and subsequently returned to pH 7.2. (B) Summary of inward (−80) and outward (+60) current amplitudes recorded at pH 7.2 at 4–5 or 6–7 hours after tetracycline induction. Values are mean±SD, n=3–5. (C) Representative current-voltage relationships (I-V) of the cell shown in (A) obtained in bath solution of the indicated pH. Note that the blue and black traces are virtually superimposable. (D) Summary of inward (−80) and outward (+60) current amplitudes at the different pH values; mean±SD, n=3. (E) Representative I-Vs showing the inhibition by reduced pH and the shift in reversal potential upon replacement of extracellular Na⁺ with NMDG (n=3). (F) Experiments were performed as in (A). The I-Vs show inhibition by reduced pH in APOL1-G0–expressing 293 cells (representative of n=3). Note that the blue and black traces are superimposed. G0, APOL1-G0; G2, APOL1-G2; TET, tetracycline.

To determine if these currents in cells have characteristics previously reported for APOL1 channels reconstituted in lipid bilayers, the extracellular bath solution was rapidly changed from pH 7.2 to pH 5.1. Within 12 seconds, both inward and outward currents were reduced to near background levels (Figure 6, A, C, and D). The inhibition was rapidly reversed upon return of the extracellular pH to 7.2. Inhibition by reduced pH and the subsequent recovery could be repeated multiple times in the same cell. Another characteristic of reconstituted APOL1 channels is their cation selectivity. Under our recording conditions, a reversal potential near 0 mV could be explained by activity of either a cation channel or a Cl⁻ channel. To evaluate selectivity, we first confirmed that the recorded cell expressed a large reversible pH-sensitive current with a reversal potential near zero (Figure 6E). Subsequent superfusion of the cell with bath solution in which the Na⁺ was replaced with the impermeable cation, N-methyl-D-glucamine (NMDG), resulted in a shift of the reversal potential to approximately −50 mV and a pronounced outward rectification of the I-V relationship (Figure 6E), consistent with a channel permeable to both Na⁺ and Cs⁺, but not Cl⁻. A similar negative shift in reversal potential with NMDG was observed during recording of APOL1 currents using pipette solutions in which the Cl⁻ was replaced by aspartate (not shown), again consistent with a lack of anion permeability of the APOL1 channels. Lastly, similar pH-sensitive currents were observed in cells expressing the G0 variant of APOL1 (Figure 6F). Taken together, our results are consistent with the conclusion that both overexpressed APOL1-G0 and -G2 form constitutively active cation channels in the plasma membrane, which leads to the dissipation of Na⁺ and K⁺ concentration gradients without a change in [Ca²⁺]_i and before cell death.

Using a similar stable expression system, others have reported variant-dependent cytotoxicity after insertion of APOL1 into the plasma membrane.³⁰ We also examined lower APOL1 expression for variant-dependent differences in cytotoxicity (Figure 7). We first established the dose-response relationship between low tetracycline concentrations and APOL1 expression (Supplemental Figure 4). As expected, lower tetracycline concentrations reduced APOL1 abundance in all stable lines (Figure 7A); in addition, the rate and steady state levels (Supplemental Figure 4) of APOL1 were more variable than at high-tetracycline concentrations. Within this lower dose range, we chose tetracycline concentrations to induce comparable APOL1 levels in G0, G1, and G2 clones (Figure 7A, Supplemental Figure 4). Because cytotoxicity is preceded by plasma membrane channel activity, we assessed the presence of APOL1 at the plasma membrane using biotinylation and precipitation of surface proteins in cells in the absence of tetracycline and after 8 and 24 hours of incubation in the presence of high and low concentrations of tetracycline, respectively (Figure 7A). Biotinylated reference and variant APOL1s were only precipitated from plasma membranes isolated from cells exposed to a high concentration of tetracycline. Consistent with the absence of plasma membrane APOL1 in cells exposed to no or matched low dose tetracycline concentrations, cytotoxicity was markedly attenuated but increased as the tetracycline concentration was increased from the matched low-dose concentrations across genotypes (Figure 7B). There was little evidence of cell death using clonogenic survival assays in APOL1-G0, -G1, or -G2 cells induced with matched low-dose tetracycline concentrations and APOL1 expression levels (Figure 4). MTT assays of APOL1-G0, -G1, and -G2 cells treated for 24 and 48 hours with the matched low-dose tetracycline concentration also demonstrated no cytotoxicity (Figure 7C). We were unable to find a level of APOL1 expression that resulted in variant-dependent cell death. Phosphorylation of p38 MAPK has been reported to be a downstream effector of APOL1-mediated loss of intracellular potassium.³⁰ Therefore, we assessed phosphorylation of p38 MAPK in cells expressing APOL1 at low and high levels. p38 MAPK was phosphorylated in the presence of APOL1-G0, -G1, and -G2 and phosphorylated p38 MAPK amounts were similar across genotypes and increased with higher APOL1 expression independent of genotype (Figure 7D).

Figure 7.

APOL1-G0, -G1, and -G2 proteins, induced with low doses of tetracycline, do not localize to plasma membrane, cause cell death, or stimulate p38 MAPK phosphorylation. (A) APOL1-G0, -G1, or -G2 clones were cultured without tetracycline (0); with low (L) concentrations of tetracycline adjusted to induce equivalent amounts of G0, G1, and G2 protein (G0, 5 ng/ml; G1, 5 ng/ml; G2, 10 ng/ml); or with high (H) tetracycline concentrations (1 μg/ml, all genotypes). Cell surface proteins were biotinylated and cells were collected in lysis buffer. APOL1 was detectable in the input lysates of tetracycline-stimulated but not untreated cells (left panel). Biotinylated APOL1 was only identified in membranes from cells treated with high concentrations of tetracycline (right panel, labeled IP). (B) Fluorogenic cytotoxicity/viability assay demonstrates dose-dependent cytotoxicity with increasing tetracycline doses across APOL1 genotypes. An increased cytotoxicity/viability ratio was observed in APOL1-G0 and APOL1-G1 clones at the “1×” tetracycline dose compared with the no tetracycline control, despite using doses chosen to provide matched APOL1 expression. Cells were treated for 24 hours with no tetracycline (0), or increasing doses of tetracycline (1×–5×); these doses correspond to 5, 10, 15, 20, and 25 ng/ml for APOL1-G0 and -G1; and, 10, 20, 30, 40, and 50 ng/ml for APOL1-G2, respectively. (C) MTT assays also show no significant cytotoxicity after treatment with low concentrations of tetracycline. Bars represent mean±SD (n=2). (D) Immunoblots of APOL1, phosphorylated p38 MAPK, total p38 MAPK, and GAPDH from one of three experiments yielding similar results. Lanes 1–3, lysates from 293 cells used to generate the stable cell lines (293 No. Tet); lanes 4–6, lysates from G0 (0), G1 (1), and G2 (2) cells treated for 24 hours with low concentrations of tetracycline, as above; lanes 7–9, lysates from G0 (0), G1 (1), and G2 (2) cells treated for 8 hours with tetracycline, 1 μg/ml. IB, immunoblot; IP, immunoprecipitate; M.W., molecular weight; Phospho-p38, phosphorylated-p38 mitogen-activated protein kinase; Tet, tetracycline.

Discussion

We generated and characterized 293 cells that conditionally expressed APOL1-G0, -G1, or -G2 to study biologic mechanisms underpinning the genetic association of APOL1 variants with kidney disease. Building on APOL1’s known trypanolytic activity, published in vitro evidence has suggested that variant-dependent, autophagic, or cytotoxic cell death is responsible for progressive kidney disease. We did not detect variant-dependent, differential cytotoxicity, findings which differ from several prior reports using HEK293 cells,^25,30,36,37 HEK293T cells,³⁸ or human podocytes.²⁶ The reasons that our data differ from earlier publications are unclear. Established cell lines such as 293 cells can exhibit heterogeneous phenotypes between laboratories.³⁹ We also found significant variability in the tetracycline induction kinetics of APOL1 expression across independent clones expressing all three genotypes. These results suggest that APOL1 induction kinetics are as important as APOL1 expression level in determining the phenotypic response to a particular tetracycline dose. Variations in APOL1 induction kinetics could confound interpretation of short-term, surrogate assays for cell death used in prior studies. We confirmed variant-independent cytotoxicity by measuring cell survival after 7–10 days.³² We found that APOL1-G0, -G1, and -G2 can all be cytotoxic, an outcome that depends on expression levels and not upon nonsynonymous changes in the amino acid sequences of the variant APOL1s.

The notion that cytotoxicity drives variant APOL1-associated kidney disease is on the basis of data generated using model systems ectopically expressing APOL1 from heterologous promoters. However, no studies have reported APOL1-mediated cytotoxicity when APOL1 is expressed under its endogenous promoter. Importantly, we were able to induce reference and variant APOL1 expression without causing cytotoxicity, better recapitulating human kidney phenotypes. APOL1 is normally expressed in healthy human kidney cells^16,17,40 and most blacks that carry the high risk APOL1 genotype do not develop kidney diseases. Therefore a “second hit” seems to be required for kidney disease initiation and progression in individuals with high-risk APOL1 genotypes.

Reference and variant APOL1s target to the plasma membrane when sufficiently overexpressed in cell culture. However, APOL1 does not localize to podocyte plasma membranes nor colocalize with the podocyte plasma membrane receptor phosphatase, GLEPP1, in human kidneys.¹⁶ When associated with plasma membranes in cells, some report variant APOL1s permit greater intracellular potassium depletion than reference APOL1.^30,38 However, we found no variant-dependent differences in the intracellular potassium loss and found comparable channel activity in cells expressing APOL1-G0 compared with those expressing APOL1-G2. We characterized the mechanism of K⁺ efflux for the first time in mammalian cells and found APOL1-G0 and APOL1-G2 both had pH-sensitive, nonselective, cation channel activity, consistent with previous reports.^11,13 Channel characteristics of cells expressing APOL1-G0 or -G2 were identical. In contrast to these data with full length APOL1 proteins, chloride channel activity has been recorded when the pore-forming domain of APOL1 (amino acids 60–325) was reconstituted in lipid bilayers.¹² A recent report has provided evidence that may reconcile these observations, reporting that APOL1 channel conducts chloride at low pH and potassium when the pH is neutralized.⁴¹ Taken together, our data would suggest a model where increasing the overexpression of APOL1 leads to variant-independent APOL1 localization to the plasma membrane, ion channel opening, loss of transcellular Na⁺ and K⁺ gradients, and cell death.

We have also previously generated transgenic mice expressing APOL1-G0 or APOL1-G2 under control of the Nephrin promoter for podocyte-specific expression.⁴² Similar to the APOL1 expression levels in cells that were not cytotoxic, these mice did not develop kidney damage as assessed by serum creatinine levels, proteinuria, and histology even when aged to 300 days, and we could not demonstrate necrotic, apoptotic, or autophagic podocyte cell death. Interestingly, the APOL1-G2 mice did have a postnatal reduction of glomerular podocyte density by 200 days when compared with wild-type or APOL1-G0 transgenic mice, suggesting transgenic expression of APOL1-G2 in podocytes may promote accelerated but clinically silent podocyte detachment. Another murine model of inducible, podocyte-specific APOL1-G0, -G1, or -G2 expression also reported variant-dependent podocyte depletion.³⁷ Unlike mice expressing APOL1 under the Nephrin promoter, these animals consistently demonstrated albuminuria, azotemia, and histologic kidney damage with the expression of variant but not reference APOL1s, which were attributed to cytotoxicity. These data are not entirely consistent with the epidemiology of common human kidney disease, where the majority of individuals with APOL1 risk genotypes do not develop overt kidney disease in the absence of an additional stressor. A third mouse model utilizing hydrodynamic gene delivery of APOL1-G0, -G1, or -G2 also found variant-dependent cytotoxicity and proteinuria, which was dependent upon the soluble urokinase plasminogen activator receptor, suPAR.⁴³ In vitro APOL1-G1 and -G2 augmented suPAR-dependent α_vβ₃ activation, permitting podocyte detachment. Transgenic expression of human APOL-G1 and -G2 in zebrafish glomeruli caused histologic abnormalities but not proteinuria or edema.⁴⁴ Two reports of ectopic expression of variant but not reference APOL1s in the Drosophila nephrocyte, a cell type with some similarities to the podocyte, increased endocytosis and loss of nephrocytes with aging.^45,46 Experiments in Saccharomyces cerevisiae also found a genetic interaction between APOL1 expression and deletion of components in the endocytic pathway.⁴⁶ Together, these results suggest that processes such as podocyte adhesion and endocytic trafficking, in addition to cytotoxicity, may also contribute to renal pathology associated with APOL1 risk variants.

In summary, our data fail to demonstrate variant-dependent cytotoxic or autophagic cell death, suggesting neither of these pathways mediate progressive APOL1-associated kidney disease. APOL1-G0, -G1, and -G2 can all localize to the plasma membrane at high expression levels, which leads to channel formation, and may cause cell death. The channel activity characteristics of APOL1-G0 were similar to those of APOL1-G2, suggesting that kidney diseases do not result from variant-dependent changes in APOL1 channel activity. Importantly, this report demonstrates that APOL1-G0, -G1, and -G2 can all be expressed without causing cytotoxicity, and this system can be used to identify alternative cell processes that mediate APOL1-associated kidney diseases.

Concise Methods

Cell Line Creation and Culture

Tetracycline-regulated expression (T-Rex)–293 cells (Life Technologies) with the tetracycline repressor were used to generate stable transfectants for the expression of APOL1-G0, -G1, or -G2. Clones for each genotype were derived from independent single-cell clones. Cells were maintained in culture according to manufacturer instructions in DMEM+Glutamax (Life Technologies), 10% tetracycline-free, FBS (GE Healthcare Hyclone), 2 mM L-glutamine, 1% Pen-Strep, 5 µg/ml blasticidin, and 400 μg/ml zeocin. APOL1 expression was induced by the addition of a stock solution of tetracycline in sterile water to achieve the indicated final concentrations for the specified time periods.

Antibodies and Reagents

Antibodies for immunoblotting include a rabbit polyclonal anti-APOL1 (HPA01885; Sigma) at 1:4000 or rabbit monoclonal anti-APOL1 (ab108315, lot#GR8593–2; AbCam) at 1:4000, rabbit monoclonal anti-LC3II (3868; Cell Signaling) at 1:1000, rabbit polyclonal anti-p62 (P0067, lot#108K4764; Sigma) at 1:2000, rabbit monoclonal anti-GAPDH (2118; Cell Signaling) at 1:2000, mouse monoclonal anti–Na/K-ATPase (Cat#sc-21712, Lot#G1615; Santa Cruz Biotech) at 1:200, and rabbit monoclonal anti-tubulin (2128; Cell Signaling) at 1:1000. For secondary detection of primary antibodies in immunoblotting HRP-conjugated protein A (P8651, lot#110M6029; Sigma) was used at 1:15,000.

Assays of Cell Death and Cytotoxicity

The MTT assay was performed using the thiazolyl blue tetrazolium bromide reagent (M2128; Sigma). The fluorogenic MultiTox-Fluor Multiplex Cytotoxicity Assay (Promega) was used according to manufacturer instructions. Clonogenic survival assays were done as described with the following minor differences.⁴⁷ Briefly, cells from each APOL1 genotype were grown in culture without APOL1 induction, trypsinized, and harvested in a single-cell suspension, and counted using a hemocytometer. Between 50 and 200 cells were plated in each well of a six-well culture dish; cells were allowed to attach for 8–16 hours, before addition of tetracycline for APOL1 induction or the experimental drug treatment. Cells were maintained in culture for 7–10 days, changing media every 48 hours, until single clonogenic colonies were visible with light microscopy; wells were then washed with PBS, and cells fixed and stained with 6% gluteraldehyde and 0.5% crystal violet in water for 20 minutes. Fixative was removed and wells thoroughly rinsed and colonies counted by two independent observers. Three wells on each plate were used as control to calculate plating efficiency and three wells were treated as the experimental wells and used to calculate surviving fraction. Cells were then exposed to specified treatment conditions and maintained in culture for 10 days. Media with treatment supplements were changed every 48 hours and at least three replicates were done for each condition.

Biotinylation of Cell Surface Proteins

After tetracycline induction of cultured cells, they were washed with ice-cold PBS, and incubated with EZ-Link Sulfo-NHS-LC-biotin (Thermo Fisher Scientific) in PBS (1 mg/ml), 4 ml total volume per 10-cm dish for 5 minutes on ice. Plates were washed once with PBS and residual NHS quenched with 4 ml of 0.1 M glycine in PBS, then washed with ice-cold PBS and lysed with the addition of 300 μl of lysis buffer (10 mM Tris, pH=7.5, 1 mM EDTA, 175 mM NaCl, 1% Triton X-100, PMSF 0.7 mM, 1× protease inhibitor cocktail [P8340; Sigma]), incubated on ice for 10 minutes, and centrifuged at 13,000 × g for 10 minutes. Next, each sample was quantitated with DC protein assay kit (Biorad) and 500–1000 μg of total protein was transferred to a fresh microfuge, transferred to a fresh microcentrifuge tube containing 60 µl of streptavidin resin (Thermo Fisher Scientific), and incubated overnight at 4°C. Beads were collected with centrifugation at 13,000 rpm for 1 minute at 4°C, washed four times with 500 μl of lysis buffer, resuspended in Laemmli sample buffer for separation of proteins with SDS-PAGE, and transferred to PVDF membranes for immunoblotting.

Immunoblotting

Immunoblots were done as previously described,¹⁶ images were captured as TIFF files, and band densities were determined using ImageJ software available from the National Institutes of Health (http://rsb.info.nih.gov/ij/).

Confocal Microscopy

Cells were grown on glass coverslips, fixed in PBS containing 4% paraformaldehyde for 20 minutes at room temperature, permeabilized with 0.2% Triton X-100 for 10 minutes at room temperature, blocked with 10% normal goat serum (NGS) for 1 hour at room temperature, then incubated with anti-APOL1 antibody diluted 1:50 in PBS with 1% NGS overnight at 4°C, then washed and incubated with goat anti-rabbit-FITC secondary antibody diluted 1:400 in PBS with 1% NGS for 1 hour in the dark. Coverslips were washed and mounted using Vectashield with DAPI. Images were captured on a Leica SPE RGB+405 one Spectral Channel System with ACS Objectives.

Cellular Na⁺ and K⁺ Content

The cellular content of Na⁺ and K⁺ was measured as previously described with minor modifications.⁴⁸ Briefly, cells were grown in 30-mm culture dishes in the absence or presence of tetracycline to induce the expression of APOL1-G0, -G1, or -G2 variants. At selected time points, the culture medium was aspirated from the dish and the cells were rapidly washed three times with 2 ml of ice-cold solution containing 110 mM MgCl₂ and 5 mM Tris-Cl, pH 7.4 with HCl. After aspiration of the final wash, the cells were solubilized in 2 ml of 10% nitric acid. Cellular particulate matter was removed by centrifugation and the resultant supernatants were diluted (generally 1:5 or 1:10) in 10% nitric acid for the subsequent measurement of Na⁺ and K⁺ by atomic absorption spectroscopy (Agilent 55B AA Spectrometer). Concentrations of Na⁺ and K⁺ in the diluted samples were compared with standard curves for Na⁺ and K⁺, and are expressed as micrograms per liter.

Whole-Cell Patch Clamping

The giga-seal technique for current recording was utilized in the whole-cell mode as previously described.⁴⁹ Briefly, cells, grown on 35-mm plastic cell culture dishes, were placed on the stage of a Nikon inverted microscope immediately before use. The normal extracellular solution contained (in millimolar) 137 NaCl, 5.4 KCl, 2 CaCl₂, 1 MgCl₂, 10 glucose, and 10 HEPES (pH 7.2). The normal pipette solution contained (in millimolar) 124 CsCl₂, 2 MgCl₂, 1 CaCl₂, 11 EGTA, and 10 HEPES (pH 7.2 and pCa 8). In some experiments, the Na⁺ in the bath solution was isosmotically replaced with NMDG. To change extracellular pH, the cells were rapidly superfused (via a perfusion pipette placed in close proximity to the cell) with normal bath solution or with bath solution in which the Hepes buffer was replaced with MES buffer, pH 5.2. Data were obtained using an Axopatch 1C amplifier (Axon Instruments) and sampled on-line using pCLAMP 9.0 software. The ground electrode was an Ag-AgCl pellet connected to the bath via an agar bridge containing 150 mM KCl. All recordings were made at room temperature (approximately 22°C). To generate current-voltage (I-V) relations, voltage ramps from −80 to +60 mV over 1 second were repetitively applied at 6-second intervals. Unless otherwise indicated, the holding potential between ramps was −60 mV. All figures show representative current traces normalized to cell capacitance (pA/pF).

Statistical Analyses

To correct for multiple comparisons, the single-factor ANOVA test was used to assess for significant differences across multiple groups. Where significance was detected, the Tukey honestly significant difference test was used to assess for significance between specific groups. A significance threshold of <0.05 was used for all experiments.

Disclosures

None.