Role of DNA-Dependent Protein Kinase in Mediating Cyst Growth in Autosomal Dominant Polycystic Kidney Disease

DNA-dependent protein kinase (DNA-PK) is a serine/threonine protein involved in DNA damage response (DDR) signaling that may mediate kidney cyst growth in autosomal dominant polycystic kidney disease (ADPKD) due to its pleiotropic effects on proliferation and survival. To test this hypothesis, the expression of DNA-PK in human ADPKD and the in vitro effects of DNA-PK inhibition in a three-dimensional model of Madin-Darby Canine Kidney (MDCK) cyst growth and human ADPKD cells were assessed. In human ADPKD, the mRNA expression for all three subunits of the DNA-PK complex was increased, and using immunohistochemistry, the catalytic subunit (DNA-PKcs) was detected in the cyst lining epithelia of human ADPKD, in a focal manner. In vitro, NU7441 (a DNA-PK kinase inhibitor) reduced MDCK cyst growth by up to 52% after long-term treatment over 6–12 days. Although human ADPKD cell lines (WT9-7/WT9-12) did not exhibit synthetic lethality in response to DNA-PK kinase inhibition compared to normal human kidney cells (HK-2), the combination of low-dose NU7441 enhanced the anti-proliferative effects of sirolimus in WT9-7 and WT9-12 cells by 17 ± 10% and 11 ± 7%, respectively. In conclusion, these preliminary data suggest that DNA-PK mediates kidney cyst growth in vivo without a synthetically lethal interaction, conferring cell-specificity in human ADPKD cells. NU7441 enhanced the anti-proliferative effects of rapamycin complex 1 inhibitors, but the effect was modest.


Introduction
Autosomal dominant polycystic kidney disease (ADPKD) is due to germ-line variants, predominantly in either PKD1 (85%) or PKD2 (15%), encoding the polytopic integral membrane, polycystin-1, and the calcium transient receptor channel, polycystin-2, respectively [1][2][3]. Both genotype as well as the total dose of ADPKD-causative genes, govern the total kidney cyst burden and severity of kidney disease [4][5][6]. The acquisition of somatic variants in the unaffected PKD allele has been hypothesized to cause a further reduction in gene dose and explain the focal nature of kidney cyst formation [7,8]. In previous studies, the genomic instability and DNA damage were increased in human and experimental models of PKD [9][10][11][12], suggesting that the DNA damage response (DDR) pathway is a potential therapeutic target in ADPKD [13].
Multiple DDR proteins work together, often with redundancy, to maintain genomic fidelity, despite exposure to numerous endogenous and exogenous genotoxic insults (~105 lesions each day) [14]. Inefficient repair of lesions by mutated or overexpressed genes results in DNA double strand breaks (DSBs), chromosomal rearrangement, and cancerous transformation [14]. In previous studies, pharmacological (but not genomic)

Focal Increase of DNA-PKcs in Cyst Lining Epithelial Cells of Human ADPKD
In the normal kidney, immunostaining for DNA-PKcs was weak with cytosolic staining (with negative nuclei) in the proximal tubular epithelial cells ( Figure 1A). In ADPKD, DNA-PKcs were absent in minimally cystic regions in either nuclei or cytosol ( Figure 1B). In contrast, focal and random immuno-positive nuclei were detected in the cyst lining epithelia ( Figure 1C). Stronger immunostaining for DNAPKcs was detected in the cytosol and apical membrane of epithelial cell lining dilated tubules and small cysts (50-200 µm) ( Figure 1D). The specificity of staining for DNA-PKcs was confirmed on antibody-negative controls (data not shown).

Focal Increase of DNA-PKcs in Cyst Lining Epithelial Cells of Human ADPKD
In the normal kidney, immunostaining for DNA-PKcs was weak with cytosolic staining (with negative nuclei) in the proximal tubular epithelial cells ( Figure 1A). In ADPKD, DNA-PKcs were absent in minimally cystic regions in either nuclei or cytosol ( Figure 1B). In contrast, focal and random immuno-positive nuclei were detected in the cyst lining epithelia ( Figure 1C). Stronger immunostaining for DNAPKcs was detected in the cytosol and apical membrane of epithelial cell lining dilated tubules and small cysts (50-200 µm) ( Figure 1D). The specificity of staining for DNA-PKcs was confirmed on antibody-negative controls (data not shown).

Pharmacological Inhibition of DNA-PK MDCK Cyst Growth In Vitro
Before assessing the in vitro effects of NU7441 on MDCK cyst growth, the effect on the number of viable cells was first assessed using an MTT assay. As shown in Figure 2, the number of viable cells was reduced at all time points following NU7441 treatment at 10 µM, whereas lower doses were similar to the vehicle.

Pharmacological Inhibition of DNA-PK MDCK Cyst Growth In Vitro
Before assessing the in vitro effects of NU7441 on MDCK cyst growth, the effect on the number of viable cells was first assessed using an MTT assay. As shown in Figure 2, the number of viable cells was reduced at all time points following NU7441 treatment at 10 µM, whereas lower doses were similar to the vehicle. Therefore, studies in MDCK cysts were performed using low (0.625 µM) and moderate (2.5 µM) doses. As shown in Figure 3, in forskolin-induced MDCK cysts, NU7441 reduced mean cyst diameter by 27% and 52% at 0.625 and 2.5 µM concentrations, respectively, on day 6 of treatment compared to the vehicle. This effect was sustained for 12 days of culture ( Figure 3). No cyst formation when cells were treated with sirolimus (50 nM) or 10 µM NU7441 (data not shown). Therefore, studies in MDCK cysts were performed using low (0.625 µM) and moderate (2.5 µM) doses. As shown in Figure 3, in forskolin-induced MDCK cysts, NU7441 reduced mean cyst diameter by 27% and 52% at 0.625 and 2.5 µM concentrations, respectively, on day 6 of treatment compared to the vehicle. This effect was sustained for 12 days of culture ( Figure 3). No cyst formation when cells were treated with sirolimus (50 nM) or 10 µM NU7441 (data not shown).

DNA-PK Inhibition Does Not Cause Synthetic Lethality of Human ADPKD Cells
To evaluate whether the inhibition of DNA-PK caused synthetic lethality in human ADPKD cells, the anti-proliferative of NU7441 was compared to normal human kidney cells. As shown in Figure 4, the number of viable cells at various concentrations of NU7441 (0.04, 0.156, 0.625, 2.5, and 10 µM) in human ADPKD cell lines (WT9-7, WT9-12) was similar to a normal human kidney cell line (HK-2). Cell viability decreased with prolonged exposure and had less than 75% viability by 96 h at doses greater than 0.625 µM.

DNA-PK Inhibition Enhances the Anti-Proliferative Effects of Sirolimus in Both Human ADPKD and Normal Kidney Cells
We next evaluated whether a non-toxic dose of NU7441 (0.04 µM; Figure 4) could enhance the anti-proliferative effects of TORC1 inhibition at sub-therapeutic doses (2.5 to 50 nM sirolimus) on ADPKD cells compared to normal kidney cells. As shown in Figure 5A, sirolimus alone, at the lowest dose (2.5 nM), reduced the number of viable cells by 31 ± 2% and 35 ± 3% in WT9-7 and WT9-12 cells, respectively (p < 0.05), and, interestingly, this level of suppression was similar at higher doses. In contrast, in HK-2 cells the number of viable cells was only reduced at 25nM of sirolimus (19% reduction in the number of viable cells; p < 0.05 compared to vehicle). There was a greater reduction in the number of viable cells in WT9-7/WT9-12 cells compared to HK-2 cells, suggesting that ADPKD cells are more sensitive to the anti-proliferative effects of sirolimus ( Figure 5A). Therefore, studies in MDCK cysts were performed using low (0.625 µM) and moderate (2.5 µM) doses. As shown in Figure 3, in forskolin-induced MDCK cysts, NU7441 reduced mean cyst diameter by 27% and 52% at 0.625 and 2.5 µM concentrations, respectively, on day 6 of treatment compared to the vehicle. This effect was sustained for 12 days of culture ( Figure 3). No cyst formation when cells were treated with sirolimus (50 nM) or 10 µM NU7441 (data not shown).

DNA-PK Inhibition Does Not Cause Synthetic Lethality of Human ADPKD Cells
To evaluate whether the inhibition of DNA-PK caused synthetic lethality in human ADPKD cells, the anti-proliferative of NU7441 was compared to normal human kidney cells. As shown in Figure 4, the number of viable cells at various concentrations of NU7441 (0.04, 0.156, 0.625, 2.5, and 10 µM) in human ADPKD cell lines (WT9-7, WT9-12) was similar to a normal human kidney cell line (HK-2). Cell viability decreased with prolonged exposure and had less than 75% viability by 96 h at doses greater than 0.625 µM.

DNA-PK Inhibition Enhances the Anti-Proliferative Effects of Sirolimus in Both Human ADPKD and Normal Kidney Cells
We next evaluated whether a non-toxic dose of NU7441 (0.04 µM; Figure 4) could enhance the anti-proliferative effects of TORC1 inhibition at sub-therapeutic doses (2.5 to 50 nM sirolimus) on ADPKD cells compared to normal kidney cells. As shown in Figure  Similarly, results are presented as mean cell viability (%) ± standard deviation (n = 12) after normalization to vehicle control and additional normalization to single therapy with 10nM sirolimus for comparison. Significance between treatments was tested by independent t-test for each cell line. ** p < 0.01, *** p < 0.001, **** p < 0.0001 between combined 10nM sirolimus and 40 nM NU7441 and 10 nM sirolimus treatment alone.
To determine whether DNA-PK inhibition enhanced the anti-proliferative effects of TORC1 inhibition, we compared the effects of NU7441+sirolimus with sirolimus alone in the cell lines. As shown in Figure 5, NU7441+sirolimus further reduced the number of viable cells by 17 ± 10% (p < 0.05) and 11 ± 7% (p < 0.05) in WT9-7 and WT9-12 cells, respectively, compared to sirolimus alone ( Figure 5C,D). Similarly, in HK-2 cells, the combined treatment reduced the number of viable cells by 13 ± 9% compared to sirolimus alone (p < 0.05).

Discussion
In this study, the hypothesis that DNA-PK expression is increased in human ADPKD and that inhibition of this DNA repair kinase reduces cyst growth in vitro was evaluated.
The key findings were the following: (i) the expression of genes of all three subunits of the DNA-PK complex was increased in human ADPKD transcriptome, as well as localized to cyst lining epithelial cells in human ADPKD; (ii) long-term inhibition of the catalytic activity using specific small molecule inhibitor NU7441 in MDCK 3D cysts reduced MDCK growth by up to 52%; (iii) human ADPKD cells do not exhibit synthetic lethality Results are presented as mean ± standard deviation (n = 12) after normalization to vehicle control. Significance was determined by Kruskal-Wallis followed by post hoc Dunn-Bonferroni test. One-way analysis of variance (ANOVA) and a post hoc Tukey honestly significant difference (HSD) test were carried out between cell lines. * p < 0.05 versus vehicle control of the same cell line; § p < 0.05 versus HK-2 cell line of the same dose and treatment. The color of symbols represent the line color of respective cell lines. In the same normal and ADPKD cell lines, sirolimus (10 nM) was tested in combination with low-dose NU7441 (40 nM) (B-D). Similarly, results are presented as mean cell viability (%) ± standard deviation (n = 12) after normalization to vehicle control and additional normalization to single therapy with 10nM sirolimus for comparison. Significance between treatments was tested by independent t-test for each cell line. ** p < 0.01, *** p < 0.001, **** p < 0.0001 between combined 10nM sirolimus and 40 nM NU7441 and 10 nM sirolimus treatment alone.
To determine whether DNA-PK inhibition enhanced the anti-proliferative effects of TORC1 inhibition, we compared the effects of NU7441+sirolimus with sirolimus alone in the cell lines. As shown in Figure 5, NU7441+sirolimus further reduced the number of viable cells by 17 ± 10% (p < 0.05) and 11 ± 7% (p < 0.05) in WT9-7 and WT9-12 cells, respectively, compared to sirolimus alone ( Figure 5C,D). Similarly, in HK-2 cells, the combined treatment reduced the number of viable cells by 13 ± 9% compared to sirolimus alone (p < 0.05).

Discussion
In this study, the hypothesis that DNA-PK expression is increased in human ADPKD and that inhibition of this DNA repair kinase reduces cyst growth in vitro was evaluated.
The key findings were the following: (i) the expression of genes of all three subunits of the DNA-PK complex was increased in human ADPKD transcriptome, as well as localized to cyst lining epithelial cells in human ADPKD; (ii) long-term inhibition of the catalytic activity using specific small molecule inhibitor NU7441 in MDCK 3D cysts reduced MDCK growth by up to 52%; (iii) human ADPKD cells do not exhibit synthetic lethality to DNA-PK inhibition by NU7441; and (iv) finally, the increased sensitivity of human ADPKD cells to TORC1 inhibition was enhanced by combination treatment with NU7441.
On analysis of public transcriptome datasets from ADPKD patients compared to the normal renal cortex, we found that the catalytic subunit of DNA-PK, essential for its function as a kinase and in NHEJ, is increased in a manner that is positively correlated to increased cyst size. Interestingly, the expression of catalytic subunits in the nuclei by immunohistochemistry was sporadic and did not affect all cysts in human ADPKD. The discrepancy between these findings could be due to the differences in technique used to evaluate DNA-PK, as well as our immunohistochemistry being limited to only one protein epitope.
To further evaluate the potential function of DNA-PK in ADPKD, the effects of a specific inhibitor, NU7741, were examined. In vitro, NU7441 reduced cyst diameter in a dose-dependent manner at doses that did not reduce the number of viable cells. Using an MTT assay, we determined that although the reduction in cyst size at 2.5 µM could be due to a reduction in cell numbers, the slower cyst growth at 0.625 µM could not be explained by this mechanism, and alternate mechanisms, such as apoptosis or alterations in intra-cystic fluid accumulation, are possible. Consistent with the results of the present study, PIK-75 (a dual inhibitor of DNA-PK and phosphoinositide 3-kinase (PI3K)), reduced cyst growth in a murine IMCD3 forskolin-induced cyst model [24]. However, this was characterized by a low viable cell count and not further examined. In contrast, the same study tested NU7441 in the IMCD3 cysts but did not identify any change in cyst size in the dose range tested in the MDCK cyst model in the current study [24], suggesting possible species and/or disease-specific effects. Unraveling the additional mechanisms by which DNA-PK inhibition influences MDCK cyst growth warrants further investigation.
Contrary to the main hypothesis, we did not find evidence that DNA-PK inhibition induces synthetic lethality in human ADPKD cells, as the number of viable cells on exposure to NU7441 was similar to normal kidney cells. Previous literature indicated variability in the toxicity of NU7441, where normal human fibroblasts tolerated up to 1 µM [30], whereas treatment of normal renal HK-2 cells with NU7441 and NU7026 did not demonstrate toxicity at doses of up to 5 µM [31]. In contrast, we found that concentrations above 0.156 µM affected both human ADPKD cell lines, WT9-7 and WT9-12, and normal renal HK-2 cells to a similar extent. These findings suggest that long-term use of NU7441 may have off-target toxicity to normal parenchyma if used continuously and may require a pulsed treatment approach. Further in vivo studies are needed to test this hypothesis.
Finally, we evaluated the effect of a combined treatment of sirolimus with NU7441 on the viability of human ADPKD and normal kidney cell lines. As NU7441 did not affect the number of viable cells at a concentration of 0.04 µM of NU7441 in any cell line, we postulated that this dose, combined with an equivalent sub-therapeutic dose of sirolimus, might be efficacious and potentially be a strategy to overcome the toxicity of using TORC1 inhibitors alone in human ADPKD [28,29]. Similar strategies have been harnessed for the treatment of various malignancies where DNA-PK inhibition renders malignant cells susceptible to chemo-and radio-therapies [32][33][34][35]. In addition, dual DNA-PK and TORC1 inhibition is under investigation in cancer but has not been evaluated as a potential therapy in ADPKD. Our results showed that low-dose NU7441 sensitized the anti-proliferative effects of sirolimus in human ADPKD cells, but the effect was modest. As these results suggest dual DNA-PK and TORC1 inhibition might reduce kidney cyst growth in vivo, further studies using a genetic ortholog of ADPKD would be an important next step.
In conclusion, the results of this study show, for the first time, that the expression of DNA-PK is increased in human ADPKD. However, our data indicate that monotherapy with DNA-PK inhibitors may not be suitable for long-term treatment of attenuating cyst growth in vivo, given that there is no evidence of the synthetic lethality of this pathway. On the other hand, it is possible that there may be an opportunity for combining DNA-PK inhibitors with TORC1 inhibitors to minimize the off-target and toxic effects of the latter through dose reduction. Further preclinical studies, testing the pharmacological inhibition in vivo, are warranted to fully evaluate this possibility and the potential of DNA-PK inhibition as a disease-modifying therapy in ADPKD.

Immunohistochemistry for DNA-PK in Human ADPKD
Paraffin-embedded kidney tissue from ADPKD patients (n = 6) and non-cancerous portions of renal tissue (n = 2) were obtained from archival samples from a previously published study [11]. Written informed consent was provided by all patients, and the study was approved by the Human Research Ethics Committee at Westmead Hospital

Effect of NU7441 Treatment on Viability of ADPKD Cells
Cell viability was assayed using an MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay, as per manufacturer protocol (11465007001, Roche Diagnostics, Mannheim, Germany). Briefly, MDCK, HK-2, WT9-7, and WT9-12 cells were seeded at 2 × 10 3 cells/well in 100 µL media in a 96-well plate and treated with NU7441, either alone at varying concentrations [31,40,41] or combined with sirolimus at concentrations below therapeutic blood levels (2.5-50 nM) [42][43][44][45]. NU7441 is a highly selective and potent ATP-competitive DNA-PK inhibitor, with an IC 50 of 14 nM [41]. NU7441 was added at four-fold serial dilutions from 0 to10 µM, with the 0 µM vehicle control containing dimethyl sulfoxide (DMSO) at a volume equal to the highest dose. Cells were incubated with the drug for 24, 48, and 96 h. To test sirolimus +/− NU7441, sirolimus doses were tested, including, and below, therapeutic blood plasma levels (2.5-50 nM) of the drug in combination with a non-toxic dose of NU7441 (40 nM) (media were changed every two days). The assay (Roche Diagnostics, Mannheim, Germany) was performed according to the manufacturer's specifications. Following drug treatment, MTT solution (0.5 mg/mL) was added to each well and incubated for 4 h to allow the development of formazan crystals by normal metabolic activity. Solubilization solution was then added followed by overnight incubation. Absorbance was measured the following day at 570 nm (720 nm reference). Percentage viability was calculated for each sample as 100 * (absorbance/average absorbance of vehicle control). Measurements were taken for n = 4 replicates per treatment group over three repeat experiments (total n = 12) at each time point.

Effect of NU7441 Treatment on Three-Dimensional (3D) MDCK Cyst Model
Three-dimensional cysts of MDCK cells was grown, as previously described [46]. Briefly, MDCK cells in collagen were incubated in the continuous presence of 10 µM forskolin to induce cyst growth and treated with either vehicle (DMSO), 0.625 µM, 2.5 µM, or 10 µM of DNA-PK inhibitors; NU7441 (IC 50 of 14 nM), LTURM34 (IC 50 of 34 nM), and NU7026 (IC 50 of 230 nM), or 0.05 µM sirolimus (positive control). All drugs were obtained from Selleck Chemicals (Houston, TX, USA). Media with treatment were refreshed every two days, and images were obtained at days 6 and 12. Cyst diameter was determined for 60 randomly selected cysts per treatment (20 cysts/well × triplicates) at each time point using Image J (v1.52a, U.S. National Institutes of Health, Bethesda, MD, USA). Experiments were repeated thrice for a total of n = 180 for vehicle and 0.625 µM, and twice for a total of n = 120 for 2.5 µM NU7441 at each time point.

Statistical Analyses
For statistical analysis of gene expression data, fold changes were transformed to a linear scale by log 2 transformation. All data were analyzed using the JMP ® Pro statistical software package (v14.2.0, SAS Institute, Cary, NC, USA) and graphed using GraphPad Prism (v8.2.1, San Diego, CA, USA). A Shapiro-Wilks test was performed to determine the normality of distribution of data. Variance of data was assessed by Bartlett's test for normally distributed data and Levene test for data that did not follow a normal distribution. An independent T-test was used to compare two normally distributed datasets. For multiple groups, Kruskal-Wallis and Dunn-Bonferroni post hoc analysis or one-way analysis of variance (ANOVA) and post hoc analysis by Tukey Kramer honestly significant difference (HSD) tests were carried out, depending on the distribution of data. p-values less than 0.05 was defined as statistically significant. For gene expression data, p-values were adjusted for false discovery rate (FDR) and presented as q-values.