Differential regulation of basal and interferon-induced APOL1 gene expression in podocytes

Abstract

APOL1 genetic variants associated with kidney disease become pathogenic when APOL1 gene expression is induced to high levels, however, the mechanism of APOL1 gene regulation is not well characterized. Using human podocyte cell lines, the induction of APOL1 gene expression from various pathogen recognition receptors was universally blocked with a soluble receptor for type I interferons, indicating APOL1 induction from these different innate immune mechanisms was secondary to autocrine production of type I interferon. ATAC-seq and chromatin immunoprecipitations identified the APOL1 gene is regulated by an ISRE-containing promoter and an intronic enhancer containing an interferon-γ activation sequence (GAS) element which explains its responsiveness to both type I and type II interferons. RNAi-mediated gene silencing identified the transcription factor required for basal APOL1 expression levels (IRF1) was not required for interferon-induced expression, which utilized additional transcription factors known to mediate interferon receptor signaling (IRF9 and STATs). An inhibitor of Janus Kinases (JAK), which activate signaling from interferon receptors, was able to attenuate the interferon-induced, but not basal levels of APOL1 expression. These differential gene regulatory mechanism may be useful in optimizing treatment strategies for APOL1 mediated kidney disease.

NEW & NOTEWORTHY

APOL1 kidney disease is triggered by events that induce high levels of APOL1 expression. We show here basal and induced APOL1 expression are regulated by different mechanisms, with induced expression largely dependent on interferon signaling. Treatment strategies may be developed to selectively block the pathogenic induced levels of APOL1 while preserving basal levels that are known to provide beneficial functions against parasitic infections.

Graphical Abstract

INTRODUCTION

Human apolipoprotein L1 (APOL1, gene name APOL1) evolved as an innate immune defense against blood borne trypanosome infections that cause human African trypanosomiasis (sleeping sickness) (1). APOL1 variant alleles, known as G1 and G2, further evolved with superior parasite killing functions, but also with an increased risk for chronic kidney disease (CKD) (2). The mechanism of CKD is not fully understood but appears to be a two-step process, combining the inherited genetic susceptibility with a second hit stressor to trigger CKD (3). The best described second hit stressors are viral infections and other immune and inflammatory conditions. These immune second hits activate high levels of APOL1 gene expression. In individuals with APOL1 high risk genotypes, these high levels of APOL1 expression are linked with podocyte loss or cell death leading to kidney pathogenesis.

Considering the importance of induced APOL1 gene expression to APOL1 kidney disease pathogenesis, little is known about APOL1 gene regulation. In this study, we identified required transcription factors controlling basal versus interferon-induced APOL1 expression. We also determined the various immune events associated with high levels of APOL1 expression only induced APOL1 gene expression secondary to de novo production of type I interferon. Therefore, interferons are the principal, common mediator for the induced, high levels of APOL1 expression. These observations have implications on the use of targeted treatment strategies that may selectively block induced APOL1 expression that is deleterious to the kidney while preserving basal APOL1 expression to maintain protection against Trypanosome infection.

MATERIALS AND METHODS

Podocyte cell lines and RNAi-mediated gene silencing.

Human podocyte cell lines (ATCC CRL-3465 and CRL-3462) were derived from subjects with either an APOL1 low risk (G0/G0) or an APOL1 high risk (G1/G2) genotype and were immortalized with a temperature sensitive SV40 large T antigen. Cells were propagated at 33°C in RPMI with 10% fetal bovine serum supplemented with glutamine, antibiotics (pen/strep) and for all experiments were differentiated 5-10 days at 37°C. Podocytes were treated for various times with different innate immune agents as follows: 200units/ml interferon-β (InVivoGen), 100 ng/ml interferon-γ (GenScript), 1 μg/ml ssRNA40 (ssRNA InVivoGen), 1 μg/ml dsDNA-EC (dsDNA from E. coli, InVivoGen), polyinosinic:polycytidylic (dsRNA, InVivoGen) or lipopolysaccharides (LPS, Sigma). Nucleic acid treatments were transfected with either LyoVec (InVivoGen) or lipofectamine 3000 (ThermoFisher) following manufacturer recommended methods. Podocytes were also treated with inhibitors including 10 μM baricitinib (Sigma) or 100 ng/ml B18R a type I interferon soluble receptor or B8R a type II interferon soluble receptor (InVivoGen). After treatment, cells were harvested and processed for Western blotting using the following antibodies (supplier, catalog number, dilution): IRF1 (Cell Signaling, #8478, 1:1000), IRF9 (Cell Signaling, #76684 1:1000), STAT1 (Cell Signaling, #14995 1:500), T701 phospho-STAT1 (Cell Signaling, #9167, 1:1000), APOL1 (Proteintech, #66124, 1:2500), tubulin (Millipore, #T5168, 1:10,000). Western blot detection was by ECL chemiluminescence (Pierce) with images captured and quantified using iBright Imaging System and software (ThermoFisher).

For gene knockdown experiments, cells were differentiated for 5 days, transfected with ThermoFisher Silencer Select premade siRNA for IRF1 (ID S7501), IRF9 (ID s20290), STAT2 (ID S13528), or a negative control siRNA (4390843) as recommended using the Lipofectamine RNAimax reagent (ThermoFisher). After 48 hours, cells were treated with different stimulating agents with or without different inhibitors for 6 hrs. Cells were harvested directly in gel loading sample buffer, heated to 95°C for 10 min and processed for gel electrophoresis on 4-20% gradient precast denaturing polyacrylamide gels followed by immunoblotting using antibodies listed above.

Chromatin Immunoprecipitation (ChIP).

Podocytes were differentiated in 150mm dishes and stimulated with interferons (200 units/ml) for 1 hour. ChIP used cells at 50-75% confluence which was sufficient for 2-3 immunoprecipitations (50μg per immunoprecipitation). The SimpleChIP Enzymatic Kit (agarose beads) from Cell Signaling was used according to manufacturer–s instructions. The micrococcal nuclease was diluted 1:10 and 2μl were used for each 150mm plate harvest, and the recommended sonication step was added to lyse nuclei. Antibodies were qualified for use in ChIPs and used at recommended dilutions for IRF1 (Cell Signaling ,#8478, 1:50), IRF9 (Cell Signaling, #76684, 1:50), Phospho-STAT1 (T701, Cell Signaling, #9167, 1:50). Immunoprecipitated DNA was analyzed using standard PCR with reagents and recommended conditions with the Qiagen HotStarTaq Plus Master Mix. Primers for the promoter (forward: TTTGCTGTGCTTGGTCATGAGCTGC, reverse: AGACAGGTCCTCCAGTCCCCAAGATA) and enhancer (forward: ACCAAAGTCACTTCCCGGAATTGACC, reverse: ACCGTTAGCCTCAACTAGGATACAGC) were used for amplification. Thermocycle consisted of a hot start at 95°C for 1 min followed by 34 cycles of 95°C, 30 sec; 60°C, 30 sec, 68°C, 30 sec. PCR products were resolved on 1.8% agarose gels and imaged with SYBR green staining.

RNA quantification.

Gene expression was quantified from total cellular RNA extracts using a Qiagen RNeasy extraction kit and cDNA synthesis using First Strand cDNA synthesis kit (Roche) per manufacturer’s instructions. APOL1 RNA abundance was determined using quantitative PCR (Quant Studio 5, Applied Biosystems) with primers (forward: ATCTCAGCTGAAAGCGGTGAAC, reverse: TGACTTTGCCCCCTCATGTAAG) and a hot start thermocycle (30 cycles of 95°C for 15 sec, 60°C for 1 min), with the SYBR Green Master Mix (Applied Biosystems). IFNA1, IFNB1, IFNG RNA abundance was determined from 100ng of RNA with NanoString nCounter (Brueker, human inflammation V2 panel). NanoString data were analyzed using ROSALIND (RRID:SCR_006233, https://rosalind.bio ) using recommended parameters for NanoString data. All expression data were normalized to internal controls and log2 transformed.

Transposase-Accessible Chromatin sequence (ATAC-seq).

Existing ATAC-seq data (4) was used to extract peak profiles for the APOL1 gene in podocytes and kidney endothelia. Data and images were obtained from the open access Kidney Interactive Genomics website (https://humphreyslab.com/SingleCell/ ).

Statistics.

All experiments were repeated a minimum of 3 times, number of replicas for each experiment are given in figure legends. Data are reported as means with standard deviations. Comparisons between treated and untreated were by t test or ANOVA with p<0.05 considered significant after adjustment for multiple comparisons (Tukey’s).

RESULTS

Prior studies have shown APOL1 gene expression is inducible by a wide range of stimuli including viruses and viral mimetics such as single- and double-stranded RNA (ssRNA, dsRNA), or bacterial DNA and endotoxins such as LPS (5). Using co-treatments with these pathogen-derived stimuli with an inhibitor of type I interferon that functions as a soluble receptor (B18R), the induction of APOL1 protein expression was universally blocked but with no change to basal levels (Figure 1A,B). APOL1 gene expression was also induced with various stimuli, indicating changes in APOL1 protein expression were, in part, originating from new gene expression (Figure 1C). These treatments also induced type I interferon gene expression, specifically interferon-β (IFNB1), but not interferon-α (IFNA1) or type II interferon-γ (IFNG) (Figure 1D). This suggested all these stimuli were not directly inducing APOL1 expression, but instead were inducing the autocrine production of type I interferons, which secondarily induced APOL1 expression. Since interferons signal through receptors by activating Janus kinases (JAK), co-treatment with the JAK inhibitor baricitinib similarly suppressed the induced APOL1 expression. Importantly, baricitinib did not inhibit the basal level of APOL1 expression (Figure 1E,F). These observations indicate the APOL1 gene has different transcriptional controls for basal and induced expression.

Figure 1. Stimulation of APOL1 expression from pathogen recognition receptors is dependent on autocrine production of type I interferon.

Protein and gene expression of high risk (HR) and low risk (LR) APOL1 genotype podocytes treated for 24h with different interferons or innate immune stimuli, with and without the type I interferon inhibitor B18R or the JAK inhibitor baricitinib. A. Western blots of APOL1 protein showing induction by many innate immune stimuli which was concurrent with phosphorylation of STAT1. The B18R inhibitor is a soluble receptor that selectively binds type I interferons (β) but not type II interferon (γ). B18R inhibited both induced APOL1 expression and phosphorylation of STAT1. B. Quantification of bands in panel A normalized to tubulin (mean ± standard error, n=3; *adjusted p<0.05 compared to no stimulus, by ANOVA). There were no significant differences in the responses to stimuli or inhibitors when comparing HR and LR. C. APOL1 gene expression quantified by qPCR with and without immune stimuli (mean ± standard error, n=3; **adjusted p<0.001 by ANOVA). Induced and uninduced levels were not different between HR and LR. D. Interferon-β (IFNB1) gene expression quantified by NanoString. Interferon-α (IFNA1) and interferon-γ (IFNG) gene expression were not detected. Normalized log2 fold change data are shown (mean ± standard error, n=4-6; with no difference between HR and LR, **adjusted p<0.01). E. Western blots using the JAK inhibitor baricitinib to inhibit signaling from interferons through inhibiting STAT1 phosphorylation downstream of interferon binding to the receptor. Both B18R and baricitinib suppressed the induced but not basal APOL1 expression. F. Quantification of induced APOL1 expression by the various indicated stimuli and inhibition by baricitinib. Baricitinib treatment blocked the stimuli-induced APOL1 expression but did not affect the basal, uninduced APOL1 expression levels (quantification normalized to tubulin; mean ± standard deviation, n=3, *p<0.05, **p<0.01 by t test).

Two transcriptional regulatory regions were identified using existing ATAC-seq data (Figure 2A). One open chromatin region corresponded to a previously identified promoter that contains an interferon stimulated response element (ISRE) known to bind interferon response factors (IRFs) and signal transducers and activators of transcription (STATs) (5). A second open chromatin region, an enhancer, was located within intron 3 and contained a high fidelity consensus site for an interferon-γ activation sequence (GAS) also known to bind STATs. To confirm these are functional regulatory regions, ChIPs were performed with human podocyte cells lines treated with either type I (β) or type II (γ) interferons. Type I and II interferons signal through separate receptors (IFNAR and IFNGR) and activate different transcription factor complexes (interferon-stimulated gene factor 3 “ISGF3” and γ-activated factor “GAF”) via phosphorylation events (Figure 2B). In the absence of interferon stimulation, the promoter region bound IRF1 (Figure 2C,D). With interferon-β stimulation, the promoter region bound IRF1, IRF9 and phospho-STAT1 and the enhancer region bound phospho-STAT1 (Figure 2C). With interferon-γ stimulation, phospho-STAT1 bound the intragenic enhancer element and there was also binding of IRF1 and phospho-STAT1 at the promoter (Figure 2D). Although the unique complexes activated by the different interferon receptors were observed on the expected cognate binding sites (i.e. ISGF3 binding ISRE and phospho-STAT1 of GAF binding GAS), there were additional complexes identified including interferon-β-induced enhancer binding of phospho-STAT1 and interferon-γ-induced promoter binding of phospho-STAT1. These additional complexes were likely due to known cross-activation from the different interferon receptors since both interferon receptors are able to phosphorylate STAT1 (6).

Figure 2. APOL1 gene regulatory elements bind different transcription factor complexes depending on stimulus.

A.Schematic diagram of the APOL1 genomic region showing exon structure (exons 1-7 for transcript variant 2, ref seq NM_145343.3) with flanking and untranslated (UT) regions. ATAC-seq peak profiles of accessible chromatin identified two potential transcriptional regulatory sites for promoter (P) and enhancer (E) elements, which were similar in human podocytes and kidney endothelia. The promoter sequence contained a canonical ISRE element and the enhancer region contained a high fidelity GAS element. B. Schematic diagram of known signal transduction from interferon type I and type II receptors (IFNAR and IFNGR) which activate kinases (JAK1, JAK2, TYK2) that phosphorylate transcription factors forming alternative complexes. IFNAR signaling forms the tripartite interferon-stimulated gene factor 3 (ISGF3) which binds ISRE elements whereas IFNGR signaling forms STAT1 homodimers referred to as γ-activated factor (GAF) that binds GAS elements. There is the potential for cross activation due to the common phosphorylation of STAT1. C-D. Chromatin immunoprecipitations of human podocytes with low risk (LR) and high risk (HR) APOL1 genotypes stimulated with interferons for 1 hr. Shown are agarose gels of PCR products using different primers to detect DNA sequences of the promoter and enhancer regions. Immunoprecipitating antibodies (IP Ab) are shown above gel. C. Interferon-β treatment induced binding at the promoter region of IRF1, IRF9 and phospho-STAT1 and at the enhancer region of phospho-STAT1. D. Similar chromatin immunoprecipitations using interferon-γ stimulation resulted in binding at the enhancer region of phospho-STAT1. There also was some enhancement of binding of IRF1 and phospho-STAT1 at the promoter region.

With both interferon-β and -γ treatment, IRF1 binding was maintained or enhanced at the promoter region. Unlike IRF9 and the STATs, IRF1 is constitutively produced and resides in the nucleus, whereas IRF9 and the STATs reside in the cytoplasm and require post translational modifications to translocate to the nucleus to activate gene transcription (7). The abundance of the IRF1 protein is known to increase with interferon treatment through both new protein synthesis (the IRF1 promoter contains a GAS element) and through inhibition of protein degradation. To further define the role of IRF1, RNAi-mediated gene silencing was used to knockdown IRF1, IRF9 and STAT2 to determine the responses of the different interferons. In the absence of IRF1, basal APOL1 expression was abolished, however, the APOL1 gene could still be induced by dsRNA treatment (Figure 3A, B). However, the IRF1 knockdown decreased the effect of interferon-γ on inducing APOL1 expression (Figure 3C,D). As expected, IRF9 knockdown inhibited interferon-β induction, but had no effect on induction by interferon-γ (Figure 3C,D). The knockdown of STAT2 was incomplete, but like the knockdown of IRF9, interferon-β induction of APOL1 expression was attenuated whereas the reduction in STAT2 did not change interferon-γ increases in APOL1 expression (Figure 3C,D).

Figure 3. Regulatory transcription factors for basal and induced APOL1 gene expression are different.

A. Western blots of high risk (HR) and low risk (LR) APOL1 genotype podocytes treated for 24 hours with dsRNA. RNAi gene silencing (knockdown “KD”) of IRF1 resulted in the loss of basal APOL1 expression but still permitted induction of APOL1 expression by dsRNA. A non-specific siRNA was used as a negative control (nc) for the KD. B. Quantification of gel bands, comparing negative control (nc KD) with IRF1 KD (mean ± standard deviation, *p<0.05, by t test, n=4). C. Western blots of podocytes with knockdown of IRF1, IRF9 and partial knockdown of STAT2 in combination with inhibitors for type I (B18R) and type II interferons (B8R) to show specificity of the response from each receptor (monitored by phosphorylation of STAT1). IRF9 knockdown attenuated interferon-β induction (indicating ISGF3 was primarily responsible for induction) but had no effect on basal APOL1 expression. As expected, IRF9 and STAT2 knockdown had no effect on interferon-γ induced APOL1 expression, but IRF1 knockdown limited interferon-γ induction suggesting IRF1 was contributing to the maximal interferon-γ response. D. Quantification of bands in panel C (normalized to tubulin, mean ± standard error, n=3; *p<0.05 **p<0.01 by ANOVA).

In summary (Figure 4), IRF1 was a necessary transcription factor for basal APOL1 expression but not for interferon-mediated APOL1 induction. Although not required for interferon induction, IRF1 did appear to have a function in interferon-induced expression, since its absence attenuated APOL1 induction, particularly for interferon-γ. Type I interferon-β induced the expected ISGF3 complex acting at the promoter ISRE which appeared to be primarily responsible for interferon-β induction, although phospho-STAT1 (GAF) binding was also evident on the enhancer. Type II interferon-γ induced the expected phospho-STAT1 binding at the enhancer, but also binding of phospho-STAT1 with IRF1 at the promoter. An IRF-STAT1 complex is a known alternative complex that can form on ISRE elements (8), but another possibility is this IRF1-STAT1 complex may represent the formation of a cis-regulatory module (Figure 4). These cis-regulatory modules are common in interferon-stimulated genes that contain both ISRE and GAS elements (8) and may explain two observations from the interferon-γ treatments. First, IRF1 knockdown reduced the ability of interferon-γ to increase APOL1 expression indicating IRF1 may be supporting the phospho-STAT1 binding or function at the GAS element. And second, in the ChIPs there was low level detection of enhancer PCR products from transcription factors that only bind the promoter, suggesting the promoter and enhancer DNA fragments may have co-immunoprecipitated.

Figure 4. Summary of transcription factor binding to the APOL1 gene in podocytes.

Identified transcription factor binding at the promoter ISRE and enhancer GAS sites for basal expression or with stimulation by type I (β) or type II (γ) interferons. Two options are presented for interferon-γ stimulation with the possible formation of a cis-regulatory module of an enhancer-promoter loop, aggregating the ISRE and GAS elements into one functional complex.

DISCUSSION

This study has provided three novel observations. First, the induction of APOL1 gene expression in immune and inflammatory settings is an interferon-dependent event. Although it has been widely recognized APOL1 is inducible by interferons, reports have suggested a variety of immune and inflammatory stimuli can induce APOL1 gene expression. However, our studies indicate these stimuli only induce APOL1 expression as a secondary event from the autocrine production of type I interferons. This would make interferons the common, primary mechanism by which the APOL1 gene is induced. Second, we have shown the APOL1 gene contains an intragenic enhancer that contains a GAS element, which now explains its strong responsiveness to interferon-γ. Third, our studies show the transcription factors controlling the constitutive, basal level of APOL1 expression were different than the transcription factors controlling interferon-induced expression. Prior genome-wide studies characterizing promoters responsive to type I and type II interferons or examining IRF1-dependent promoters categorized APOL1 as an interferon stimulated gene and in a subgroup in which basal expression is dependent on IRF1 (9–11). Interferon stimulated genes with basal expression controlled by IRF1 are believed to provide constitutive innate immune functions, being able to combat low levels of infectious agents in the absence of signaling events that would subsequently induce their expression. Our studies confirm these findings in podocytes, in which the basal level of APOL1 expression was dependent on IRF1 and its induced expression was controlled by interferons.

Several preclinical and clinical trials are testing different anti-APOL1 strategies for the treatment of APOL1 kidney disease (12). One proposed treatment is repurposing the immunomodulatory agent baricitinib, a JAK1/2 inhibitor, clinically used for other inflammatory conditions (13). In our study, baricitinib was effective at blocking interferon-induced levels of APOL1 expression, as has been shown previously (14,15), but had no effect on reducing the constitutive basal level. Since the proposed anti-APOL1 therapies are systemic treatments, these drugs would block APOL1 expression or function in all cells, not just in the kidney. This would include the liver, the primary source of APOL1 in blood which provides protection against trypanosome infection (16). Although we did not conduct studies in hepatocytes here, two groups have published similar analyses of the APOL1 gene in hepatocyte cell lines. These reports similarly found basal APOL1 expression was dependent on IRF1 binding at the same promoter region (9, 17). Thus, it is possible the therapeutic use of JAK inhibitors may have an advantage in selectively inhibiting only the pathogenic, induced levels of APOL1 associated with CKD, while not inhibiting any beneficial immune functions of basal APOL1 expression.

DATA AVAILABILITY

All data are presented in the manuscript.