APOL1 risk allele RNA contributes to renal toxicity by activating protein kinase R

APOL1 risk alleles associate with chronic kidney disease in African Americans, but the mechanisms remain to be fully understood. We show that APOL1 risk alleles activate protein kinase R (PKR) in cultured cells and transgenic mice. This effect is preserved when a premature stop codon is introduced to APOL1 risk alleles, suggesting that APOL1 RNA but not protein is required for the effect. Podocyte expression of APOL1 risk allele RNA, but not protein, in transgenic mice induces glomerular injury and proteinuria. Structural analysis of the APOL1 RNA shows that the risk variants possess secondary structure serving as a scaffold for tandem PKR binding and activation. These findings provide a mechanism by which APOL1 variants damage podocytes and suggest novel therapeutic strategies.


G0 vector
Shown are the constructs used In the experiments described herein, with the location of the two renal risk variants, the two G1 mutations (referred to in the text as upstream and downstream respectively) and G2 six-base pair deletion.    containing the G1 and G2 variants was sufficient to enhance PKR activation. RNAs transcribed in vitro using T7 sequence were incubated with PKR and ATP for 20 min. Poly(I:C) served as the positive control for PKR phosphorylation. Vehicle was used for negative control. Western blots identifying phospho-PKR are presented. The quantified intensity of phospho-PKR to each RNAs was added below the gel images. (e) Truncated G1 and G2 RNA (NM_001136540.1, 1180-1453) promote increased PKR activation in a time dependent manner, compared to truncated G0 RNA. Synthetic RNAs (0.10 μM) were incubated with 4 μM PKR and ATP for the indicated times. Poly(I:C) served as the positive control for PKR phosphorylation. Vehicle was used as the negative control. Western blots identifying phospho-PKR are presented. The quantified intensity of phospho-PKR to each RNA was added below the gel images. P values were calculated using a Student one-tailed t-test (b) or a Wilcoxon t-test (c & e). P < 0.05 is statistically significant.  Cell viabilit Ratio of phospho-PKR / total-PK (%Empty)

Supplementary Figure 4.
(a) Structural models for truncated APOL1 RNA variants were generated using RNAstructure software from SHAPE-derived reactivity profiles. The software converts RNA reactivity values into pseudo-free energy constraints that are then assimilated into the secondary structure prediction algorithm. In general, numerous structural models are generated for every data set and ranked by collective pseudo- Empty Stable dsRNA G0 free energy. Structures having relatively low pseudo-free energy values are considered more stable and likely to have a greater prevalence in structurally heterogeneous populations. Relative pseudo-free energies are given here for the three lowest energy conformers for each of the G0, G1 and G2 truncated APOL1 RNA variants. To allow for direct comparision of relative energies, the most stable predicted conformer in each of the respective groups is set to zero (0) kcal/mol. Whether or not a predicted conformer contains a PKR docking site is also indicated. The absolute pseudo-free energies determined for the lowest (1L) were G0 -175.6, G1 -159.3, G2 -164.1 kcal/mole, respectly. (b) Q tit ti f ll i bilit f t bl HEK293FT ll li i APOL1 Th ll l d (b) Quantitation of cell viability of stable HEK293FT cell line expressing APOL1. The cells were lysed and incubated with CellTiter-Glo to measure the amount of ATP in the cells. Cell viability of G1 and G2 variants decreased cell viability and recovered with synonymous mutations. (c) To make stable dsRNA form of APOL1 G0 RNA variant based on the lowest-energy secondary structural model, 5 nucleotides were changed (black ball) refer to complimentary sequence. (d)Stable HEK293FT cell lines expressing APOL1 variants were harvested. Constructs expressing stable dsRNA type APOL1 G0 variant increased phospho-PKR as well as G1 variant. All results are presented as ratio of controls (G0 or empty) was normalized to 100% and P values were All results are presented as ratio of controls (G0 or empty) was normalized to 100% and P values were calculated using a Student one-tailed t-test.    Figure 6 (a)APOL1 expression in transgenic mice with human APOL1 gene locus. Sections of mice kidney cortex stained with antibody against APOL1, demonstrating glomerular expression in all three APOL1-expressing mouse lines but not in wild-type mice. (b) Glomeruli were isolated from BAC-APOL1 transgenic mice using magnetic particle (upper panel: arrows indicated glomeruli, small dots are magnet beads). Glomeruli were treated with phosphatase inhibitor with/without R NPHS1-APOL1-G1-deltaRNA indicated glomeruli, small dots are magnet beads). Glomeruli were treated with phosphatase inhibitor with/without PKR inhibitor for 30 min and lysed for analysis. Phospho-PKR was increased in G1 and G2 glomeruli. Phospho-PKR signal reduced in the presence of PKR inhibitor (c) NPHS1-APOL1-deltaRNA mouse. Transgene schema is shown.
(d) Quantitative-PCR detected APOL1-deltaRNA mRNA in renal cortex from transgenic mice demonstrates transgene expression in G0 and G1 mice. (e) Urine albumin/creatinine ratio (mg/g) was measured in BAC-APOL1 transgenic mice. Risk variant transgenic mice did not manifest albuminuria, even at 8-12 weeks of age. , g (f) NPHS1-APOL1-G1-delta-RNA transgenic mice received a PKR inhibitor or vehicle. Western blot of kidney cortex lysates demonstrated reduced phospho-PKR in NPHS-APOL1-G1-deltaRNA mice with the PKR inhibitor compared to vehicle control. P values were calculated using a Student one-tailed t-test.  Results are presented as ratio of controls (G0/G0) was normalized to 100% and P values were calculated using a Student one-tailed t-test. (c) BAC-APOL1 transgenic mice manifested more proteinuria following podocyte injury, assessed as albumin/creatinine ratio (g/gCre), with higher levels in BAC-APOL1-G2 transgenic mice compared to BAC-APOL1-G0 transgenic mice. Urine protein was measured on days 7 and 10 after initiation of interferone γ, puromycin aminonucleoside and basic FGF exposure, which together induce podocyte injury. Each value represents data from one mouse. BAC-APOL1-G0 and G2 values are the same as main Figure 4d. Results are presented as means ± s.d. P values were calculated using a Student one-tailed t-test.

Supplementary Figure 9
Transiently-transfected HEK293 cell lines expressing APOL1. The G0 and downstream G1 rs60910145 G variant showed minimally increased phospho-PKR, while the upstream G1 rs73885319 G variant manifested greatly increased phospho-PKR comparable to the G1 allele containing both upstream and downstream variants. All results are presented as ratio of controls (empty) was normalized to 100% and P values were calculated using a Student one-tailed t-test. Non-denaturing PAGE analysis of partial RNA sequence. SybrGreen stain of the truncated APOL1 RNA folded in the absence or presence of 1 or 3 mM MgCl2 and r n on a 5% nati e gel Heterogeneit of folded in the absence or presence of 1 or 3 mM MgCl2 and run on a 5% native gel. Heterogeneity of APOL1 RNA gel mobility is apparent, increasing in the presence of higher Mg concentrations.  (b) In G1/G2 cultured human podocytes, PKR inhibitor restored the podocyte marker (WT-1) compared to vehicle treatment, as measured by qRT-PCR, while the PKR inhibitor had no effect on WT-1 expression.
Results are presented as ratio of controls (vehicle) was normalized to 100% and P values were calculated Results are presented as ratio of controls (vehicle) was normalized to 100% and P values were calculated using a Student one-tailed t-test. (a) The PKR inhibitor decreased mitochondrial respiration of G1/G2 cultured human podocytes. Cells were seeded and reached a final density of 20,000 cells/well. Effects of the PKR inhibitor were compared using a Seahorse XF-24 extracellular flux analyzer to measure oxygen consumption rate (OCR) in the presence of metabolic inhibitors and uncoupling agents. The first injection (shown as vertical line A), oligomycin, an inhibitor of ATP synthesis. This agent distinguishes the fraction of OCR attributed to ATP synthesis by blocking the oxygen consumption required to overcome proton leakage across the inner mitochondrial membrane; it also provides the basal respiration rate by blocking non-mitochondrial respiration The second injection also provides the basal respiration rate by blocking non mitochondrial respiration. The second injection (shown as vertical line B) was an uncoupling agent (FCCP) that disrupts ATP synthesis. This agent is used to calculate the spare respiration capacity of cells, defined as the quantitative difference between maximal and basal respiration rates. The third injection (shown as vertical line C) was a combination of a complex I inhibitor (rotenone) and a complex III inhibitor (antimycin A); this combination blocks mitochondrial respiration and enables calculation of mitochondrial and non-mitochondrial cellular respiration. Data were expressed as mean±SD at each time point. P values were calculated using a Student one-tailed t-test. P < 0.05 is statistically significant. (n = 3) y g ( ) (b) (c) PKR inhibitor decreased mitochondrial membrane potential of G1/G2 cultured human podocytes. cultured human podocytes were incubated for48 hours with a final concentration of 1000 U/ml interferone alfa with/without 1 μM PKR inhibitor. Cells were incubated for 20 min with a final concentration of 200 nM Mito tracker and 200 nM TMRE (tetramethylrhodamine ethyl ester, a live cell fluorescence marker of mitochondrial membrane potential). P values were calculated using a Student one-tailed t-test.
Primer pair for quantitative real-time PCR T tG F d R Target Gene  Forward  Reverse  Interferon α  GAC TCC ATC TTG GCT GTG A  TGA TTT CTG CTC TGA CAA CCT  Interferon β  TGG GAG GAT TCT GCA TTA CC  CAG CAT CTG CTG GTT GAA GA  β-Actin  CGC ACC ACT GGC ATT GTC  TTC TCC TTG ATG TCA CGC AC  APOL1  CAG CCT TGT ACT CTT GGA ACC  GCT TTT GAT GAC CAG GTC GTG  GAPDH  GAA ATC CCA TCA CCA TCT TCC AGG  GAG CCC CAG CCT TCT CCA TG