APOL1-G2 accelerates nephrocyte cell death by inhibiting the autophagy pathway

Apolipoprotein L1 is encoded by the APOL1 gene, which carries well-established risk variants associated with a wide spectrum of kidney-related diseases, including focal segmental glomerulosclerosis (FSGS), HIV-associated nephropathy (HIVAN) and hypertension-associated end-stage renal disease (ESRD) (Kopp et al., 2011; Tzur et al., 2012; Ulasi et al., 2013). It is unique to humans and some primates (Thomson et al., 2014). Circulating APOL1 protein forms part of the trypanolytic complexes that protect humans against infections with the Trypanosoma brucei (T. brucei) that cause African sleeping sickness. However, subspecies of T. brucei (including T. brucei rhodesiense and T. brucei gambiense) have developed resistance to the ancestral G0 allele of APOL1, giving rise to the expansion of G1 and G2 alleles in the African population as they confer greater protection against the new subspecies.

APOL1-G0 is generally considered the reference allele; it encompasses all APOL1 variants except APOL1-G1 or APOL1-G2, both of which are known as the risk alleles. APOL1-G1 is denoted by the presence of two amino acid substitutions (S352G and I384M), while APOL1-G2 is marked by deletion of two amino acids (del388N389Y) (Genovese et al., 2010; Tzur et al., 2010). African Americans show an allele frequency for APOL1-G1 and APOL1-G2 of 23% and 15%, respectively, making these alleles the most common and powerful risk variants identified to date (Friedman and Pollak, 2020). Overall, ∼13% of African Americans carry the APOL1 high-risk genotype (two risk alleles) that confers a 3–30-fold increased risk of developing different types of kidney disease (Friedman and Pollak, 2020; Limou et al., 2014).

The APOL1-G2 allele appears to be the evolutionary older variant, which arose in response to the emergence of the serum resistance-associated’ protein (SRA) of T. brucei (Vanhamme et al., 2003; Molina-Portela et al., 2008), which binds and neutralizes APOL1-G0. In vitro studies have shown that APOL1-G2 exhibits almost no binding to SRA (Thomson et al., 2014); therefore, APOL1-G2 can render the parasite inactive, ensuring that humans remain safeguarded against infection (Beckerman and Susztak, 2018; Cooper et al., 2017). In contrast, APOL1-G1, has reduced binding to APOL1 compared to that of APOL1-G0, and appeared later in response to infections with T. brucei gambiense, which also evolved other mechanisms to block the trypanosomal effects of APOL1-G0 (Cooper et al., 2017; Friedman and Pollak, 2020; Thomson et al., 2014). The APOL1-G1 variant does not prevent infection with T. brucei gambiense, but reduces the symptoms of African sleeping sickness. In contrast, APOL1-G2 increases the risk of severe disease from a T. brucei gambiense infection (Cooper et al., 2017; Friedman and Pollak, 2020). These significant protective differences between APOL1-G1 and APOL1-G2 raise the question whether the two gene variants may also play different roles in the pathogenesis of chronic kidney disease. This notion is supported by findings in a cohort of sickle cell disease patients of Sub-Saharan African ancestry, which found that people carrying the APOL1-G2 risk allele were more likely to have more-severe nephropathy than those carrying the APOL1-G1 risk allele (Kormann et al., 2017). In addition, other differences were observed between the G1 and G2 APOL1 kidney risk alleles. In particular, APOL1-G2 does not appear to increase the risk of heterozygous individuals to develop ESRD) or HIVAN, or the age in which the kidney diseases appeared but, unlike the G1 allele, is present in all documented individuals with interferon-associated APOL1 nephropathy (Friedman and Pollak, 2020; Nichols et al., 2015). Whether the APOL1 risk variants G1 and G2 differ in their biologic behavior to trigger APOL1 kidney diseases remains unclear.

A recent study in fly garland cells – which carry out a filtration function similar to that of podocytes – has shown that both G1 and G2 alleles induce cytotoxicity when expressed, and implicated ER stress as the causal mechanism (Gerstner et al., 2022). However, technical limitations did neither allow the detection of acidification or functional uptake deficits, nor the distinction between APOL1-G1 and APOL1-G2 toxicity. Therefore, to determine whether the APOL1-G1 and -G2 alleles play similar or different roles precipitating APOL1 kidney diseases, we expanded the scope of our previous fly study (Fu et al., 2017a) by generating flies that selectively express the APOL1 nephropathy risk alleles G2 or G1G2 (the latter comprising both mutations in one allele), and comparing the results with those generated in control (hereafter referred to as Dot-Gal4), APOL1-G0 and APOL1-G1 transgenic flies. Although no patients are known to carry the APOL1-G1 and APOL1-G2 mutations in the same allele, we included the APOL1-G1G2 variant to study whether the APOL1-G1 and APOL1-G2 risk variants act synergistically or independently, or are able to reduce detrimental effects. We used flies, a low cost and a highly efficient screening system, because our previous findings by using this model have since been validated and expanded by other groups in yeast, cultured human podocytes and APOL1 transgenic mice (Beckerman et al., 2017; Fu et al., 2017a; Kruzel-Davila et al., 2017; McCarthy et al., 2021). We found that overexpression of APOL1-G1 and APOL1-G2 risk alleles – specifically in the fly's nephrocytes, the equivalent of mammalian podocytes – led to deficits in nephrocyte function, decreased cell numbers, impaired organelle acidification, and a worsening phenotype with age. In agreement with previous studies done in APOL1 transgenic mice (McCarthy et al., 2021), we found that the APOL1-G2 risk allele induced a more severe phenotype when compared to APOL1-G1. We further gained mechanistic insights by showing that G1 and G2 risk allele phenotypes are marked by disrupted nephrocyte slit diaphragm structures as well as altered endocytic membrane trafficking and autophagy pathways.

Since fly does not carry a gene homologous to APOL1, we produced transgenic Drosophila melanogaster lines that carry the human APOL1-G0 reference allele (obtained from OriGene, with M228 and R255 mutagenized to I228 and K255 to match APOL1-G1) or an APOL1-G1 risk allele that had been derived previously from a patient diagnosed with HIVAN (Fu et al., 2017a). Here, we expanded on this fly model by generating APOL1-G2 and APOL1-G1G2 transgenic fly lines. The APOL1-G2 risk allele carries two amino acid deletion points (at N388del and Y389del) and the APOL1-G1G2 risk allele carries both the APOL1-G1 point mutations (S342G and I384M) and the APOL1-G2 deletion mutations (N388del and Y389del). To ensure equal expression of all constructs we carried out western blot for APOL1 in flies that expressed the APOL1-G0 or APOL1-G1 [previous constructs; P element insertion (Fu et al., 2017a)], or G2 or APOL1-G1G2 (fixed docking site). We used the ubiquitous tubulin (tub; alphaTub84B) to drive expression (tub-Gal4; BDSC 30029) for the western blots as nephrocytes by themselves (Dot-Gal4 driver) did not yield sufficient material. Data showed equal expression for all APOL1 alleles in the fly larvae (Fig. 1A,B). These newly generated fly models for APOL1-dependent risk for susceptibility to renal disease were used in all assays described in this manuscript.

We have previously shown that APOL1 risk allele-induced perturbation of nephrocyte function affected adult fly longevity (Fu et al., 2017a). That study was carried out at 29°C, the temperature at which Gal4 is most stable. In our current study, we reduced the temperature to 25°C to maintain the flies, under which Gal4 is less stable. This reduced APOL1 expression and, thus, APOL1-associated toxicity (including that of APOL1-G0). By reducing baseline toxicity, we aimed to increase the sensitivity to assays to detect differences between APOL1-G0, APOL1-G1 and APOL1-G2. Under these new conditions APOL1-G0 flies survived ∼10 days longer. However, they still had slightly reduced lifespans compared to control flies (Dot-Gal4), which survived ∼60 days (50%) (Fig. 1C). Flies with nephrocyte-specific expression of human APOL1-G1, APOL1-G2 or APOL1-G1G2 alleles displayed reduced lifespans, such that most flies (50%) survived fewer than 40 days and virtually none made it past 45 days (Fig. 1C). These data support that APOL1 risk allele-induced disruption of nephrocyte function affects longevity in flies.

To measure functional changes induced by nephrocyte-specific APOL1 expression, we exploited the ability of nephrocytes to take up and sequester toxins. Ingested silver nitrate (AgNO₃) showed normal levels of sequestration in nephrocytes of 1-day-old adult control flies (Fig. 1D,E). The amounts of AgNO₃ sequestered in nephrocytes of APOL1-G1-, APOL1-G2- and APOL1-G1G2-expressing 1-day-old adult flies were significantly increased (∼35%) compared to that in nephrocytes of control flies (i.e. control nephrocytes) (Fig. 1D,E). Nephrocytes expressing APOL1-G0 showed only moderately increased AgNO₃ uptake compared to control nephrocytes (∼10%; P=0.004) (Fig. 1D,E). This finding suggests that expression of APOL1 risk alleles in nephrocytes, leads to increased endocytosis.

Since AgNO₃ uptake by nephrocytes is cumulative, it might mask any subtle differences conferred by the different risk alleles. To further assess the effects of APOL1 on nephrocyte function, we employed a complementary assay that measures uptake of hemolymph proteins by the nephrocytes. A myosin heavy chain (Mhc) promoter-driven atrial natriuretic factor (ANF)–red fluorescent protein (RFP) (Mhc-ANF-RFP) transgene was introduced to the model flies expressing human APOL1. The Mhc promoter directs expression of the ANF-RFP fusion protein in muscle cells, which is secreted into the hemolymph (fly blood system). ANF-RFP is then removed from circulation through endocytosis by the nephrocytes. Quantification is based on RFP signal intensity in the nephrocytes, as captured by imaging. We found that nephrocytes carrying APOL1-G2 or APOL1-G1G2 alleles showed significantly increased RFP levels compared to control nephrocytes (∼40%), and displayed homogenous RFP distribution throughout the nephrocyte cytoplasm (Fig. 1F,G). Nephrocytes expressing APOL1-G0 or APOL1-G1 showed only a moderate increase in RFP levels compared to those of control nephrocytes (∼20%) (Fig. 1F,G). These data confirmed the increased endocytosis seen in APOL1 risk allele-expressing nephrocytes in the AgNO₃ uptake assay and suggest this phenotype is stronger in nephrocytes carrying the APOL1-G2 and APOL1-G1G2 alleles compared to those carrying APOL1-G1. We postulate, the increased ANF-RFP signal is the result of reduced protein degradation in addition to increased endocytosis activity. If so, this could indicate altered organelle acidification as another consequence of APOL1 risk allele expression.

APOL1 risk allele-induced kidney function presents as age-related decline in human patients (Hoy et al., 2015). Therefore, we next tested if the nephrocyte functional phenotype in flies that express APOL1 risk alleles shows a similar age-related pattern. Unfortunately, Mhc, which drives ANF-RFP expression in our functional assay, is not active in the skeletal muscle cells of adult flies. Therefore, we used an ex vivo functional assay instead, in which fly nephrocytes were dissected and then assayed for their capacity to filter and endocytose 10 kD dextran fluorescent particles from artificial hemolymph. Consistent with data from the AgNO₃ and ANF-RFP assays, the 10 kD dextran ex vivo assay showed increased function of nephrocytes obtained from 1-day-old adult flies expressing an APOL1 risk allele (Fig. 2A,B). Nephrocytes from APOL1-G0 and APOL1-G1 flies showed moderate fluorescent dextran accumulation compared to those from control animals (∼25% and 30%, respectively), while APOL1-G2- and APOL1-G1G2-expressing nephrocytes showed significantly greater accumulation of fluorescent dextran (∼40%) (Fig. 2A,B). Notably, nephrocytes obtained from 10- or 20-day-old flies expressing APOL1 risk alleles showed a significant reduction in fluorescent dextran, indicating decreased nephrocyte function as opposed to the increased activity seen in nephrocytes obtained from 1-day-old adult flies. Nephrocytes from 10-day-old flies that expressed APOL1-G1, APOL1-G2 or APOL1-G1G2 showed a significant reduction in fluorescent dextran compared to those from control flies (∼40%) (Fig. 2A,B). Nephrocytes from 20-day-old flies displayed an even more severe phenotype, with those expressing APOL1-G2 or APOL1-G1G2 displaying ∼75% reduction in function compared to those of control flies. By contrast, nephrocytes expressing APOL1-G0 or APOL1-G1 showed a more moderate reduction (∼25% and 50%, respectively) (Fig. 2A,B). Together, these findings showed that flies carrying an APOL1 risk allele exhibit age-related changes in nephrocyte function.

Next, we wanted to delve deeper into the changes that contribute to the nephrocyte dysfunctional uptake induced by the APOL1 risk alleles. Therefore, we examined whether APOL1 expression affects endocytosis by studying the localization of early endosome marker Rab5, late endosome marker Rab7 and recycling endosome marker Rab11. We chose these Rab proteins, as we had previously identified them as most important among the Rab GTPases expressed in and required for nephrocyte function (Fu et al., 2017b). We found expression of APOL1-G0, APOL1-G1 or APOL1-G2 in nephrocyte did not cause any changes in early endosomes (Rab5) but was associated with a significant reduction in recycling endosomes (Rab11) when compared to 1-day-old control flies. This phenotype was even more pronounced in nephrocytes expressing APOL1-G1 or APOL1-G2 (∼50%) (Fig. 3A,B,D). Notably, we did observe a significant increase (∼50%) in late endosomes (Rab7) within nephrocytes of 1-day-old flies expressing APOL1-G1 compared to those of control flies (∼50%). This change was even greater (∼150%) in nephrocytes expressing APOL1-G2 (Fig. 3A,C).

Furthermore, we studied the expression of autophagy- and degradation-related markers, including autophagosome marker autophagy-related 8a (Atg8a, known as GABARAP in human) and autophagy receptor refractory to sigma P [Ref(2)P, also known as p62, and SQSTM1 in human] as well as ubiquitinylated proteins marked for degradation (by using the antibody against ubiquitinylated proteins antibody clone FK2). We found a dramatic increase of autophagosome (∼250%, Atg8a) and autophagy receptor [Ref(2)P; ∼100%] in nephrocytes of 1-day-old flies expressing APOL1-G1 or APOL1-G2, which was also significantly greater than that observed in APOL1-G0 expressing nephrocytes (Fig. 3A,E,F). We also observed an ∼50% increase in the accumulation of ubiquitinylated proteins in nephrocytes expressing APOL1-G0 or APOL1-G1 compared to that in 1-day-old control flies, and this change was even greater (∼70%) in nephrocytes expressing APOL1-G2 (Fig. 3A,G). These findings demonstrated that disrupted endocytosis mechanisms affecting the fusion of lysosomes and autophagosomes to inhibit autophagy pathways might contribute to the aberrant uptake function observed in nephrocytes expressing APOL1, especially its risk alleles.

We have previously shown that ubiquitous APOL1 transgene (G0 or G1 allele) expression leads to impaired organelle acidification in fly nephrocytes (Fu et al., 2017a). Thus, next we used fluorescent LysoTracker dye to examine the status of acidic vacuoles in flies with nephrocyte-specific expression of APOL1 allelic variants. We found that, in nephrocytes of 1-day-old flies, expression of APOL1-G0 or APOL1-G1 was associated with a significant reduction in LysoTracker fluorescence (∼50%) compared to that in control flies. This reduction was even more pronounced (∼70%) in fly nephrocytes expressing APOL1-G2 or APOL1-G1G2 (Fig. 4A,B). The reduction in LysoTracker fluorescence –indicative of aberrant organelle acidification – declined further with age, as evident in 10- and 20-day-old flies displaying dramatically reduced levels in their nephrocytes (i.e. ∼75%, ∼80% or ∼80% in flies expressing APOL1-G1, APOL1-G2 or APOL1-G1G2, respectively) compared to levels in nephrocytes of control flies (Fig. 4A,B). Nephrocytes expressing APOL1-G0 showed a significant reduction in LysoTracker fluorescence compared to control nephrocytes at day 20 (∼65%) but not at day 10 (∼45%), when compared to levels at day 1 (∼45% reduction compared to control at this timepoint) (Fig. 4A,B). These data show that nephrocyte-specific expression of any APOL1 allele resulted in impaired organelle acidification, and that this phenotype is most pronounced in flies expressing the risk alleles, especially APOL1-G2 and APOL1-G1G2.

Previously, we have shown that endocytosis-driven recycling is important for nephrocyte function and to maintain slit diaphragm structural integrity (Fu et al., 2017b; Wang et al., 2021; Wen et al., 2020). Since endocytosis is impaired in nephrocytes expressing an APOL1 risk allele (Fig. 3), we had a closer look at the slit diaphragm, i.e. the filtration unit essential to nephrocyte functioning. To examine its structural integrity in nephrocytes expressing APOL1, we carried out immunochemistry for the slit diaphragm protein polychaetoid (Pyd) of Drosophila. Localization of Pyd in the medial optical section of nephrocytes was found to be a fine and continuously delineated circumferential ring in control flies (Fig. 5A). On the surface of control nephrocytes Pyd presented as a uniform and smoothly distributed finger-print-like localization pattern (Fig. 5B). Nephrocyte-specific expression of APOL1 disrupted Pyd localization, such that Pyd protein was not longer at the surface but internalized. The vast majority of nephrocytes of 1-day-old flies appeared normal (∼10% show Pyd mislocalization) (Fig. 5D). However, with age, flies carrying any – i.e. non-risk or risk – APOL1 allele showed increasingly disrupted Pyd surface localization pattern and increased Pyd internalization (Fig. 5A,C). This APOL1-induced phenotype is significantly more severe in nephrocytes of 10-day-old flies expressing alleles APOL1-G2 or APOL1-G1G2 (Fig. 5A-C), with signs of disrupted Pyd localization in APOL1-G0 (∼30% of cells), APOL1-G1 (∼40% of cells), APOL1-G2 (∼50% of cells) and APOL1-G1G2 (∼50% of cells) (Fig. 5D). This phenotype increasingly worsened with age, yet was comparable across all APOL1 transgenic flies at 20-days-old (Fig. 5C), the age at which the vast majority of cells (∼80%) showed signs of a severely disrupted slit diaphragm structure (Fig. 5C,D). Taken together, these findings show that APOL1-induced impairment of nephrocyte function is probably due to structural disruption of the slit diaphragm, which is essential for filtration by the nephrocytes.

We previously have shown that expression of APOL1 transgenes (APOL1-G0, APOL1-G1) leads to an initial increase in cell function, followed by hypertrophy and, ultimately, cell death (Fu et al., 2017a). As this process is likely to relate to the age-related phenotypic decline observed in the APOL1 risk allele flies in this current study, we examined the size and number of nephrocytes. Data showed that, as the flies aged, expression of an APOL1 risk allele led to a significant increase in nephrocyte size (Fig. 6A,B). The 1-day-old flies did not show a difference in nephrocyte cell size regardless of their genotype. However, as the flies aged, those expressing APOL1 in their nephrocytes displayed significantly increased cell sizes. In 10-day-old flies, nephrocytes expressing APOL1-G0 displayed a moderate (∼20%) increase in cell size compared to control nephrocytes (Fig. 6B); however, those expressing any of the risk alleles (i.e. APOL1-G1, APOL1-G2 or APOL1-G1G2) showed cell sizes significantly larger (∼50%) than control nephrocytes or APOL1-G0-expressing cells (Fig. 6B). Nephrocytes from 20-day-old flies showed an even greater increase in size, with cells expressing APOL1-G1, APOL1-G2 or APOL1-G1G2 being ∼80% larger than control nephrocytes (Fig. 6B).

In addition, expression of APOL1 was associated with an age-related reduction in the number of nephrocytes (Fig. 6A,C). Moreover, we observed accumulation of cell debris in the vicinity of any remaining nephrocytes in flies expressing the APOL1-G1, APOL1-G2 or APOL1-G1G2 alleles (Fig. 6A, arrows). Overall, we found no difference in nephrocyte numbers in 1-day-old flies. However, in 10-day-old flies, expression of APOL1-G0 or APOL1-G1 led to a significant reduction (∼10%) in nephrocyte numbers (Fig. 6C), whereas flies expressing the APOL1-G2 or APOL1-G1G2 risk allele showed an even greater reduction (∼20%) in nephrocyte numbers compared to age-matched control flies (Fig. 6C). At 20 days, APOL1-G0 flies did not show a significant further reduction. However, nephrocyte numbers were significantly and more severely reduced in flies expressing APOL1-G1 (∼40%) and, even more so, in those expressing APOL1-G2 or APOL1-G1G2 (both ∼60%) (Fig. 6C).

A previous study using 3rd instar larval garland cells, another type of nephrocyte in the fly, had revealed that APOL1 induces an endoplasmic reticulum (ER)-stress response (Gerstner et al., 2022). Therefore, we checked the level of protein disulfide-isomerase (PDI), a marker for ER stress. Compared with control flies, PDI levels in nephrocytes expressing APOL1-G1, APOL1-G2 or APOL1-G1G2 were increased in 10-day-old flies (Fig. 6D,E), the timepoint we first observed nephrocyte loss (Fig. 6C). In 10-day-old flies, expression of APOL1-G1 led to a similar, significant increase in ER stress (∼50%) in the nephrocytes when compared to age-matched control flies (Fig. 6E). This increase in PDI levels was even greater in nephrocytes of flies expressing the APOL1-G2 or APOL1-G1G2 risk allele, showing ∼100% increase compared to control (Fig. 6E). Therefore, like in the larval garland cells used by Gerstner et al. (2022), APOL1 induced ER-stress in pericardial nephrocytes of adult flies.

In our current study we developed new transgenic Drosophila lines expressing human APOL1-G2 and APOL1-G1G2 exclusively in nephrocytes. The APOL1-G2 and APOL1-G1G2 transgenic lines, as well as the APOL1-G1 and APOL1-G0 lines had been made previously (Fu et al., 2017a), and all used the same nephrocyte-specific Gal4 driver (Dot-Gal4). Despite different insertion sites, all four constructs yielded equal APOL1 expression, which is important as even APOL1-G0 reference allele overexpression is known to induce toxicity. Given that both unnatural and different haplotype backgrounds can alter APOL1 cytotoxicity (Lannon et al., 2019), our transgenic flies are based on APOL1-G1 cDNA derived from cultured podocytes that had been obtained from a child diagnosed with HIVAN (Xie et al., 2014) as described before (Fu et al., 2017a). The APOL1-G1 cDNA contains the haplotypes E150, I228 and K255, and differs from the APOL1-G0 cDNA only at S342 and I384 (APOL1-G0 was obtained from OriGene and mutagenized to match APOL1-G1; M228I and R255K). The APOL1-G2 cDNA contains the APOL1-G1 haplotypes E150, I228 and K255, as well as dual deletion of amino acids N388 and Y389, characteristic of APOL1-G2. Therefore, all our transgenic constructs are physiologically relevant and carry common haplotypes seen in people of African ancestry.

All fly lines expressing APOL1, including APOL1-G0, in their nephrocytes showed – as the flies aged – a decline of nephrocyte function, increase in cell size, reduction in acidification of organelles and premature death of nephrocyte. Whereas we had previously kept the flies at 29°C, the temperature at which Gal4 is most stable (Fu et al., 2017a), we reduced the temperature to 25°C for this current study. This renders Gal4 less stable, thus reducing APOL1 expression. By doing so, we aimed to reduce baseline toxicity and increase assay ability to detect changes between the different APOL1-alelles. APOL1-G0 flies showed longer survival curves at 25°C compared to our previous study. When comparing the nephrocyte phenotypes among the different risk alleles, the data predominantly demonstrated that APOL1-G2- and APOL1-G1G2-expressing nephrocytes displayed more severe structural and functional changes, relative to nephrocyte-specific transgenic flies expressing the APOL1-G1 allele obtained from a child diagnosed with HIVAN and the minimized effects observed in APOL1-G0 nephrocytes. Although improved, our assays still show residual toxicity in APOL1-G0 nephrocytes and, therefore, we were unable to detect differences between APOL1-G1 and APOL1-G2 in all assessments, including autophagy markers. Therefore, additional assays that directly assess autophagy or a means to further reduce APOL1 baseline toxicity, are needed to tease out in which cases APOL1-G2 is more toxic than APOL1-G1, distinctions that could inform pathomechanistic insights. Additional informing would be to understand how APOL1-G2 induces a more severe phenotype than APOL1-G1. One possibility is a different intracellular localization of each. Unfortunately, technical limitations currently prevented us from providing direct evidence for this hypothesis.

Although, in humans, APOL1-G1 and APOL1-G2 are not expressed on the same allele, we still included this artificial variant APOL1-G1G2 to determine whether both risk variants act in a synergistic or independent manner, or whether one variant could minimize the detrimental effects of the other. We found that both variants act in a similar manner by disrupting key cell trafficking pathways and confirmed that these changes were driven initially by the increased endocytic activity of nephrocytes, leading to the accumulation of proteins. These changes were followed by the premature death of nephrocytes. Overall, the changes induced by the APOL1-G1 and APOL1-G2 alleles in fly nephrocytes show that the endocytic function is initially increased in association with perturbations in endosomal trafficking pathways and acidification processes that impair the autophagic flux and degradation mechanisms of these cells.

Fly nephrocytes are highly active in endocytosis, since they need to take up material from the hemolymph (the blood of the fly) and then sort the endocytic cargo for either degradation in the lysosome or recycling back to the hemolymph (Fu et al., 2017b; Wang et al., 2021; Wen et al., 2020). Rab GTPases play a critical role in endocytosis and cell trafficking, acting as switches that recruit effector molecules that regulate these processes. Previously, we have shown that Rab5, Rab7 and Rab11 play key roles in maintaining the normal function of fly nephrocytes (Fu et al., 2017b; Wang et al., 2021; Wen et al., 2020). Rab5 is localized to early endosomes and plays a critical role regulating the early steps of endocytosis in nephrocytes. However, we did not detect significant changes in the expression levels of Rab5 despite the enhanced endocytic function of nephrocytes detected in 1-day-old transgenic flies expressing the APOL1 risk alleles (evident in increased AgNO₃ and ANF-RFP uptake). Thus, the methods we used to detect changes in Rab5 expression might not be sensitive enough or, alternatively, other GTPases might regulate this process as well. However, we did find a significant upregulation of Rab7 expression in nephrocytes expressing the APOL1 risk alleles. We have previously reported that silencing Rab7 in nephrocytes results in the accumulation of clear vacuoles and a reduced number of lysosomes, in association with disrupted cell trafficking and degradation processes (Fu et al., 2017b). Therefore, we speculate that the changes in Rab7 expression levels (Fig. 3) reflect the activation of compensatory mechanism to re-establish the normal function of late endosomes as well as protein degradation processes. By contrast, expression of Rab11 was significantly reduced in nephrocytes expressing the APOL1 risk alleles. Rab11 regulates the recycling of endosomal compartments, as well as the exocytosis of recycling vesicles at the plasma membrane in mammalian cells (Takahashi et al., 2012). Our findings support the notion that the function of recycling endosomal compartments is affected as well. Finally, we found high levels of the autophagosome marker Atg8a and autophagy receptor Ref(2)P in nephrocytes expressing APOL1 risk alleles. These findings support that autophagosomes accumulate whenever the autophagic flux is impaired. Taken together, our findings suggest that APOL1 risk alleles impair the function of nephrocytes mainly by disrupting late endocytic pathways that regulate the fusion of late endosomes with autophagosomes or lysosomes, therefore impairing the autophagic flux as well as protein degradation processes. Interestingly, similar findings have been reported in cultured human podocytes and in transgenic mice expressing APOL1 exclusively in podocytes. (Beckerman et al., 2017). Briefly, in these studies, the APOL1 risk alleles reduced the autophagic flux of podocytes and increased the steady-state level of autophagosomes by impairing the fusion of lysosomes as well as degradation of autophagic cargo.

The podocyte slit diaphragm is an extracellular structure that bridges the filtration slits between neighboring podocytes and is an essential component of the glomerular filtration barrier. Two of the three layers of the human glomerular filtration barrier, the slit diaphragm and the basement membrane, are present in fly nephrocytes. Here, we showed that the failure to maintain appropriate cell trafficking pathways affects the structure of the slit diaphragm and disrupts the process of endocytosis in nephrocytes expressing APOL1 risk alleles (Fig. 7). Moreover, we reported that the presence of Pyd, a core component of the slit diaphragm, is greatly diminished and its characteristic fingerprint-like expression pattern is no longer visible (Fig. 5), indicating disrupted slit diaphragms as a plausible explanation for the reduced nephrocyte uptake function in older flies (Fig. 2). Disrupted slit diaphragm structures in humans are often indicative of dramatic podocyte morphological changes that lead to foot-process effacement associated with proteinuria, podocyte loss and, ultimately, kidney failure (Garg, 2018). Mice expressing the APOL1-G1 or APOL1-G2 risk allele in their podocytes display molecular, functional and structural podocyte changes, including foot-process effacement, that mimic the clinical features observed in patients with APOL1-induced kidney disease (Beckerman et al., 2017). Notably, the slit diaphragm proteins and their interactors, and the endocytosis and recycling pathway components, are highly conserved between fly and human (Wang et al., 2021). Overall, our findings are in line with previous studies carried out in yeast (Kruzel-Davila et al., 2017), cultured human podocytes, APOL1 transgenic mice (Beckerman et al., 2017; McCarthy et al., 2021) and the age-related changes observed in patients with APOL1-associated kidney diseases (Abid et al., 2020; Hoy et al., 2015). However, the pathological changes induced by APOL1 risk alleles in transgenic mice and cultured podocytes, appear to be regulated by the expression levels of these alleles and, in some transgenic mouse models – as it is the case in humans – a second ‘hit’ is needed, e.g. through high levels of interferon-γ (Aghajan et al., 2019; McCarthy et al., 2021) to induce the development of proteinuria and/or chronic kidney disease (Beckerman et al., 2017; Bruggeman et al., 2019; Okamoto et al., 2018; Ryu et al., 2019; Wakashin et al., 2020). One limitation of our fly model is that, given the high expression levels of APOL1 in the nephrocytes, a second hit is not needed to injure these cells.

The recent study in garland cells of APOL1 transgenic flies, attributed the pathogenic effects of APOL risk alleles almost exclusively to ER stress (Gerstner et al., 2022). Gerstner and colleagues did not detect significant changes in cell trafficking pathways, organelle acidification or even the structure of the slit diaphragm. However, like us, they report that APOL1 risk alleles increase the endocytic activity of nephrocytes and induce premature cell death (Gerstner et al., 2022); similar to their study, we detected a significant ER stress response. The different findings might be due to several factors: (i) garland cells and pericardial nephrocytes originate from different cell lineages, (ii) different overexpression constructs (e.g. drivers), (iii) different APOL1 background sequence and, (iv) different timepoints examined (3rd instar larva versus 1-, 10- and 20-day-old adult flies). Sufficient ER stress typically affects cell trafficking and degradation pathways; indeed, a study that used cultured human HEK293T cells found ER stress, endolysosomal disturbances and increased permeability of the cell membrane induced by APOL risk alleles (Granado et al., 2017). The authors suggested these to be secondary consequences of mitochondrial damage induced by APOL1 risk alleles. Since the garland cells showed ER stress in response to APOL1 and underwent premature cell death (Gerstner et al., 2022), perturbation of trafficking and degradation pathways would be expected, and additional studies are warranted.

In conclusion, we found that the APOL1 risk alleles G1 and G2 affect similar key endocytic and cell trafficking pathways, and that the G2 risk allele appears to induce more severe cytotoxic effects on nephrocytes than the G1 risk allele. Drosophila pericardial nephrocytes provide a unique opportunity to assess the effects of APOL1 risk alleles on different endocytic pathways in vivo, and to perform cost effective and large-scale genetic interactive screening to detect interacting molecules that affect endocytic activity and other cell trafficking pathways, and might be therapeutic targets to treat APOL1 kidney diseases. Previous studies in cultured human podocytes and APOL1 transgenic mice support the main findings of our study. Nonetheless, it remains to be determined how these changes mimic those induced by APOL1 risk alleles in patients with APOL1-associated kidney diseases.

Flies were crossed, reared and kept on standard food (Meidi LLC) at 25°C. The following Drosophila strains used in this study had been previously generated by us (Z.H.’s lab): Hand-GFP (expressing green fluorescent protein in the nuclei of nephrocytes and cardiomyocytes; Han and Olson, 2005; Huang et al., 2023), and Mhc-ANF-RFP (expressing ANF red fluorescent protein in muscle myosin heavy chain promoter-driven atrial natriuretic peptide-red fluorescent protein in muscle cells; Zhang et al., 2013). The following lines were obtained from the Bloomington Drosophila Stock Center (BDSC; Indiana University, IN): Dot-Gal4 (ID 6903; Ugt36A1 driver), tub-Gal4 (ID 30029; alphaTub84B driver), UAS-YFP-Rab5 (ID 24616), UAS-YFP-Rab7 (ID 23270), UAS-YFP-Rab11 (ID 9790), and UAS-mCherry-Atg8a (ID 37750). As control, w¹¹¹⁸ (BDSC; ID 3605) flies were used in the crosses.

The APOL1-G0 and APOL1-G1 Drosophila strains had been generated by us (Z.H.’s lab) and are described elsewhere (Fu et al., 2017a). cDNAs of APOL1-G0 and APOL1-G1 alleles were cloned into the pUAST vector and introduced into the germ cells of flies by standard P element-mediated germ line transformation. The APOL1-G2 and G1G2 cDNAs were made using point mutagenesis and are based on the APOL1-G1 cDNA (provided by P.E.R.) used previously to generate the APOL1-G0 and APOL1-G1 lines. This APOL1-G1 cDNA was obtained from a patient of recent African ancestry. To generate UAS-APOL1-G2 and UAS-APOL1-G1G2 constructs, cDNAs of APOL1-G2 and APOL–G1G2 alleles were cloned into the pUASTattB vector, and the transgenes were introduced into a fixed chromosomal docking site by germ line transformation. To match the previous APOL1-G0 and APOL1-G1 constructs, we incorporated a FLAG epitope tag at the APOL1 protein C-terminus encoded by the APOL1-G2 and APOL1-G1G2 expression constructs, even though we did not need the FLAG-tag for the assays in this current study.

3rd instar larvae from the tub-Gal4 transgenic lines were crossed to UAS-APOL1-G0, UAS-APOL1-G1, UAS-APOL1-G2 or UAS-APOL1-G1G2 transgenic lines. For each sample, five 3rd instar larvae were smashed in a 1.5-ml tube using a pestle. Immediately after, 500 µl RIPA lysis buffer with protease inhibitor was added, and tissue suspensions were sonicated, followed by centrifugation. The cleared tissue lysates were subjected to immunoblotting using anti-APOL1 (66124-1-Ig, Proteintech). Horseradish peroxidase (HRP)-conjugated secondary antibody (anti-mouse IgG-peroxidase antibody, Sigma #A4416; used at 1:2000 dilution) was used for detection, and the HRP signal was detected by using the enhanced chemiluminescence method (ECL) and recorded by using a G:BOX Chemi XRQ (Syngene).

Female adult flies from the Hand-GFP and Dot-Gal4 transgenic lines were crossed to UAS-APOL1-G0, UAS-APOL1-G1, UAS-APOL1-G2 or UAS-APOL1-G1G2 transgenic lines. Eggs from these crosses were laid on standard food containing 0.002% AgNO₃ (Sigma-Aldrich). AgNO₃ uptake by nephrocytes was assessed ex vivo in 1-day-old adult flies by dissecting heart tissue into artificial hemolymph and examining the cells by phase-contrast microscopy after 10 min fixation in 4% paraformaldehyde in 1×phosphate buffered saline (1×PBS) (Thermo Fisher Scientific). For quantification, 40 nephrocytes were analyzed from six female flies in a single cross per genotype. The results are presented as the mean±standard deviation (s.d.). Statistical significance: *P<0.05, **P<0.01, ***P<0.001.

Female adult flies from the Hand-GFP, Mhc-ANF-RFP and Dot-Gal4 transgenic lines were crossed with flies from the UAS-APOL1-G0, UAS-APOL1-G1, UAS-APOL1-G2 or UAS-APOL1-G1G2 transgenic lines. Progeny from these crosses display nephrocyte-specific expression of one of the APOL1 allelic variants with nephrocytes labeled by GFP, as well as ANF-RFP expression driven by Mhc. RFP uptake by nephrocytes was assessed ex vivo in 1-day-old adult flies by dissecting heart tissue into artificial hemolymph and by examining cells by using fluorescence microscopy after 10 min fixation in 4% paraformaldehyde in 1×PBS (Thermo Fisher Scientific). For quantification, 40 nephrocytes from six female flies in a single cross per genotype were analyzed. The results are presented as mean±s.d. Statistical significance: *P<0.05, **P<0.01, ***P<0.001.

Following egg laying, Drosophila larvae were kept at 25°C under standard conditions and at optimal temperature for UAS-transgene expression. Adult male flies were maintained in vials, in groups of 20 or fewer. The number of live flies in each group was recorded every second day. The assay was ended when no survivors were left for any of the lines. In total, 100 flies total from a single cross were assayed per genotype.

Female adult flies from the Hand-GFP and Dot-Gal4 transgenic lines were crossed to UAS-APOL1-G0, UAS-APOL1-G1, UAS-APOL1-G2 or UAS-APOL1-G1G2 transgenic fly lines. Dextran uptake by nephrocytes was assessed ex vivo in adult flies by dissecting heart tissue into artificial hemolymph and examining cells by fluorescence microscopy following a 20 min incubation with Texas Red-dextran (10 kD, 0.05 mg ml⁻¹, Invitrogen) and 10 min fixation in 4% paraformaldehyde in 1×PBS (Thermo Fisher Scientific). For quantification, 40 nephrocytes were analyzed from six female flies in a single cross per genotype. The results are presented as the mean±s.d. Statistical significance: *P<0.05, **P<0.01, ***P<0.001.

Adult flies from the Hand-GFP and Dot-Gal4 transgenic lines were crossed to UAS-APOL1-G0, UAS-APOL1-G1, UAS-APOL1-G2 or UAS-APOL1-G1G2 transgenic fly lines. LysoTracker intensity within nephrocytes was assessed ex vivo in adult flies by dissecting heart tissue into artificial hemolymph and examining cells by using fluorescence microscopy following 20 min incubation with LysoTracker (Red DND-99, Thermo Fisher Scientific) used according to the manufacturer’s instructions and 10 min fixation in 4% paraformaldehyde in 1×PBS (Thermo Fisher Scientific). For quantification, 40 nephrocytes were analyzed from six female flies in a single cross per genotype. The results are presented as the mean±s.d. Statistical significance: *P<0.05, **P<0.01, ***P<0.001.

Adult female flies were dissected and heat-fixed for 20 s at 100°C in artificial hemolymph, i.e. 70 mmol/l NaCl (Carolina), 5 mmol/l KCl (Sigma), 1.5 mmol/l CaCl₂·2H₂O (Sigma-Alrich), 4 mmol/l MgCl₂ (Sigma-Aldrich), 10 mmol/l NaHCO₃ (Sigma-Aldrich), 5 mmol/l trehalose (Sigma), 115 mmol/l sucrose (Sigma-Aldrich), and 5 mmol/l HEPES (Sigma-Aldrich), in H₂O. Primary mouse monoclonal anti-Pyd antibody (PYD2) was obtained from DSHB and used at a 1:100 dilution in 1×PBS with 0.1% Triton X-100 (Sigma) (PBST). Secondary antibody Alexa Fluor 555 (A-21422, Thermo Fisher Scientific) was used at a 1:1000 dilution in PBST. The nephrocytes were washed with PBST three times, blocked in PBST+2% bovine serum albumin (BSA; Sigma-Aldrich) for 40 min, incubated with primary antibodies at 4°C overnight, washed with PBST three times, incubated with secondary antibodies at room temperature for 2 h, washed with 1×PBST three times and mounted with Vectashield mounting medium (H-1000, Vector Laboratories).

Adult female flies were dissected and fixated for 10 min in 4% paraformaldehyde in 1×PBS (Thermo Fisher Scientific). Primary monoclonal mouse antibody ubiquitin FK2, raised against human polyubiquitin-B and polyubiquitin-C (BML-PW8810, Enzo Life Sciences) and used at a1:100 dilution in 1×PBS with 0.1% Triton X-100 (PBST, Sigma). Primary rabbit polyclonal antibody against Ref(2)P (ab178440, Abcam) was used at a 1:100 dilution in PBST. Primary rabbit polyclonal antibody against protein disulfide isomerase (PDI; P7122, Sigma) was used at a 1:100 dilution in PBST. Secondary antibody Alexa Fluor 555 (A-21422, Thermo Fisher Scientific) was used at a 1:1000 dilution in PBST. The nephrocytes were washed with PBST three times, blocked in PBST+2% BSA for 40 min, incubated with primary antibodies at 4°C overnight, washed with PBST three times, incubated with secondary antibodies at room temperature for 2 h, washed with 1×PBST three times and mounted with Vectashield mounting medium.

Imaging for AgNO₃ uptake was carried out with a Zeiss ApoTome.2 microscope using a 20× Plan-Apochromat 0.8 NA air objective. Confocal imaging for Dextran uptake, ANF-RFP, LysoTracker assays and anti-PDI was carried out using a Zeiss LSM900 microscope with a 20× Plan-Apochromat 0.8 NA air objective. Immunocytochemistry for anti-Pyd and anti-Ref(2)P were all carried out using confocal technology; a Zeiss LSM900 microscope with a 63× Plan-Apochromat 1.4 NA oil objective under Airyscan SR mode. Live-cell fluorescence imaging for the UAS-YFP-Rab5, UAS-YFP-Rab7, UAS-YFP-Rab11, and UAS-mCherry-Atg8a fly strains were all carried out using confocal technology; a Zeiss LSM900 microscope with a 63X Plan-Apochromat 1.4 NA oil objective under Airyscan SR mode. For quantitative comparison of intensities, common settings were chosen to avoid oversaturation. The results are presented as the mean±s.d. Statistical significance: *P<0.05, **P<0.01, ***P<0.001. ImageJ Software Version 1.52a was used for image processing.

Statistical analysis was performed using PAST software (Natural History Museum, Norway). Mean values are provided as the mean±s.d. The Kruskal–Wallis H-test followed by a Dunn's test was used for comparisons between multiple groups. Statistical significance: *P<0.05, **P<0.01, ***P<0.001. Details for sample size and replicates used for quantification and the statistical tests applied to determine significance are provided in the figure legends.

We thank the Bloomington Drosophila Stock Center at Indiana University (IN) for the fly stocks, and Developmental Studies Hybridoma Bank based at University of Iowa (IA) for the anti-Pyd antibody.

This article has an associated First Person interview with the first author of the paper.