Impact of cytotoxic agents or apoptosis stimulants on αklotho in MDCK, NRK-52E and HK2 kidney cells

αKlotho is a transmembrane protein acting as a co-receptor for FGF23, a bone hormone regulating renal phosphate and vitamin D metabolism. αKlotho expression is controlled by PPARγ. Soluble αklotho (sKL) regulates cellular signaling impacting stress resistance and death. αKlotho deficiency causes early onset of aging-associated diseases while its overexpression markedly increases lifespan. Cellular stress due to cytotoxic therapeutics or apoptosis induction through caspase activation or serum deficiency may result in cell death. Owing to αklotho’s role in cellular stress and aging, this study explored the effect of cytotoxic agents or apoptosis stimulants on cellular αklotho expression. Experiments were performed in renal MDCK, NRK-52E and HK-2 cells. Gene expression was determined by qRT-PCR, sKL by ELISA, apoptosis and necrosis by annexin V binding and a fluorescent DNA dye, and cell viability by MTT assay. Cytostatic drugs cisplatin, paclitaxel, and doxorubicin as well as apoptosis induction with caspase 3 activator PAC-1 and serum deprivation induced αklotho and PPARG gene expression while decreasing viability and proliferation and inducing apoptosis of MDCK and NRK-52E cells to a variable extent. PPARγ antagonism attenuated up-regulation of αklotho in MDCK cells. In HK-2 cells, αklotho gene expression and sKL protein were down-regulated by chemotherapeutics. SKL serum levels in patients following chemotherapy were not significantly changed. In summary, potentially fatal stress results in up-regulation of αKlotho gene expression in MDCK and NRK-52E cells and down-regulation in HK-2 cells. These results indicate that different renal cell lines may exhibit completely different regulation of αklotho.


INTRODUCTION
The αklotho gene product was discovered in mice in 1997 as a protein with strong anti-aging properties [1,2]. Mice almost completely lacking αklotho exhibit a dramatically shortened life span of a few weeks only whilst suffering from a broad range of diseases and symptoms mimicking human aging [1]. Observed abnormalities affect nearly every organ and tissue [1] and include frequent aging-associated diseases including fibrosis [3,4], lung emphysema [5], multiple organ atrophy [1], or hearing loss [6,7]. The accelerated aging of αklotho-deficient mice is paralleled by massive calcification in most tissues [1,8]. Importantly, the reduction of dietary phosphate or vitamin D intake of the animals almost completely normalizes their phenotype pointing to a dominant role of phosphate and vitamin D excess in their rapid aging [9,10]. Indeed, αklotho protein has important functions in the homeostasis of these nutrients [11]: It is a transmembrane protein predominantly expressed in the kidney that enhances the binding affinity of fibroblast growth factor 23 (FGF23) for its membrane receptor [12,13]. FGF23 is a proteohormone released by bone cells that inhibits phosphate reabsorption and 1,25(OH)2D3 (biologically active vitamin D) synthesis in the kidney [14,15] and has gained attention as a marker indicating disease [16,17]. Hence, the lack of αklotho or FGF23 results in abnormally high serum phosphate and 1,25(OH)2D3 levels that account for enhanced calcification and contribute to rapid aging and early death to a large extent [18].
Chemotherapy with platinum derivative cisplatin, anthracycline doxorubicin, or paclitaxel is standard of care in many forms of cancer. Although the three compounds differ in their cellular targets, they have in common that they exert cytotoxic effects which compromise proliferation and may ultimately result in apoptotic cell death [41][42][43]. Apoptosis of cultured cells without prior cell damage may be induced by activation of executioner caspase 3 with PAC-1 or by growth factor deprivation through serum depletion [44,45].
In view of the versatile effects of αklotho on cell survival and death [46,47], this study aimed to investigate whether cytotoxic drugs or initiation of apoptosis affect αklotho gene expression in three different renal cell lines and in patients receiving chemotherapy.

RESULTS AND DISCUSSION
As a first step, MDCK and NRK-52E cells were used to study αklotho gene expression. MDCK cells were treated with antineoplastic platinum derivative cisplatin for 24 h, and αklotho mRNA levels were analyzed by qRT-PCR. As illustrated in Figure 1A, cisplatin up-regulated αklotho gene expression in MDCK cells, an effect reaching significance at 3 µM cisplatin. The effect was not paralleled by decreased viability of MDCK cells even at 10 µM cisplatin ( Figure 1B), but by reduced cell proliferation ( Figure  1C). We determined the rate of apoptosis and necrosis by means of an assay analyzing annexin V binding and a DNA-binding dye which is impermeable to the membrane of intact cells. As illustrated in Figure 1D, cisplatin induced apoptosis without significantly influencing necrosis of MDCK cells. In another series of experiments, NRK-52E cells were treated without or with cisplatin for 24 h, and αklotho gene expression, viability, proliferation, and apoptosis/ necrosis were assessed. Again, cisplatin (10 µM) significantly enhanced αklotho expression ( Figure 1E), an effect paralleled by decreased cell viability ( Figure  1F) and proliferation ( Figure 1G). Again, cisplatin induced apoptosis without significantly stimulating necrosis of NRK-52E cells ( Figure 1H).
Further experiments were performed to elucidate whether cytostatic compound paclitaxel also affects αklotho. To this end, MDCK cells were incubated with different concentrations of paclitaxel for 24 h or with vehicle control, respectively. It is shown in Figure 2A that 120 nM paclitaxel significantly stimulated the abundance of αklotho mRNA. By the same token, 120 nM paclitaxel significantly lowered the viability ( Figure  2B) and proliferation ( Figure 2C) of MDCK cells. These effects were paralleled by enhanced apoptosis and necrosis ( Figure 2D). We also studied the effect of 120 nM paclitaxel in NRK-52E cells. This concentration of the antimitotic agent significantly upregulated αklotho gene expression within 24 h ( Figure  2E), too, whilst down-regulating viability ( Figure 2F) and proliferation ( Figure 2G) of NRK-52E cells. Similar to MDCK cells, paclitaxel induced apoptosis and necrosis in NRK-52E cells ( Figure 2H).
As a third common antineoplastic drug, we tested anthracycline doxorubicin. A 24 h-exposure to 100 nM or 300 nM doxorubicin led to a significant increase in the abundance of αklotho transcripts in MDCK cells ( Figure 3A). Doxorubicin treatment (300 nM) did not significantly affect viability ( Figure 3B) but reduced proliferation ( Figure 3C) of MDCK cells. Doxorubicin induced apoptosis while slightly reducing the number of necrotic cells ( Figure 3D). In NRK-52E cells, 300 nM doxorubicin readily stimulated αklotho expression within 24 h ( Figure 3E) and compromised cell viability ( Figure 3F) as well as proliferation ( Figure 3G).  . (E) Arithmetic mean ± SEM of αklotho transcript levels relative to TBP in NRK-52E cells incubated without or with 120 nM paclitaxel for 24 h (n = 5, paired t-test). (F, G) Arithmetic mean ± SEM of NRK-52E cell viability (F) or number (G) upon treatment without or with 120 nM paclitaxel for 24 h (F: n = 5, one-sample t-test; G: n = 5, one-sample t-test). (H) Rate of apoptosis and necrosis of NRK-52E cells treated with or without 120 µM paclitaxel for 24 h (n = 5, one-sample t test). * p < 0.05, ** p < 0.01, *** p < 0.001 indicate significant difference from vehicle control; Abbreviations: a. u.: arbitrary units; ctr: control. AGING Apoptosis and necrosis were enhanced by doxorubicin in NRK52-E cells ( Figure 3H).
Since different classes of cytostatic drugs with proapoptotic properties similarly enhanced αklotho expression in MDCK and NRK-52E cells within 24 h, we sought to explore whether direct apoptosis induction also affects αklotho. To this end, we treated the cells with and without caspase 3 activator PAC-1 for 24 h. As demonstrated in Figure 4A, 10 µM PAC-1 induced αklotho expression in MDCK cells, an effect paralleled by decreased cell viability ( Figure 4B) and proliferation ( Figure 4C). PAC-1 enhanced apoptosis without significantly modifying necrosis ( Figure 4D). Also in NRK-52E cells, PAC-1 treatment (10 µM) resulted in a significant surge in αklotho transcripts within 24 h ( Figure 4E) and decreased their viability ( Figure 4F) and proliferation ( Figure 4G). The rates of apoptosis and necrosis were significantly higher in NRK-52E cells upon exposure to PAC-1 ( Figure 4H).
Depriving cells of growth factors through serum depletion similarly favors apoptosis [45]. We therefore aimed to test whether αklotho expression is affected by serum depletion. As depicted in Figure 5A, a 24 hincubation of MDCK cells in the absence of serum significantly up-regulated αklotho gene expression without significantly impacting on cell viability ( Figure  5B) and proliferation ( Figure 5C). Serum depletion upregulated apoptosis whereas necrosis-dependent fluorescence was reduced in serum-starved cells ( Figure  5D). In NRK-52E cells, serum depletion did not significantly affect αklotho mRNA levels within 24 h ( Figure 5E). However, viability and proliferation were moderately but significantly lower in NRK-52E cells incubated in the absence of serum compared to control cells ( Figure 5F, 5G). Serum depletion induced apoptosis and did not significantly affect necrosis in NRK-52E cells ( Figure 5H).
We performed further experiments to identify the mechanism underlying enhanced αklotho expression in
In order to confirm that PPARγ is indeed required for cisplatin to up-regulate αklotho expression, we exposed MDCK cells to cisplatin in the presence and absence of PPARγ antagonist SR202. As illustrated in Figure 8, SR-202 significantly blunted cisplatin-dependent upregulation of αklotho. Hence, PPARγ contributes to enhancement of αklotho expression, but may not fully explain it.
Transmembrane αklotho forms a complex with FGFR1 to serve as a receptor for FGF23. A further series of experiments sought to clarify whether the effect of   chemotherapeutics and apoptosis stimulants also affect FGFR1 and/or FGF23 expression in MDCK cells. As demonstrated in Figure 9A, 9B, cisplatin up-regulated FGFR1 expression and protein. Similar effects on FGFR1 expression were observed following incubation with doxorubicin ( Figure 9C), PAC-1 ( Figure 9D), and upon incubation in serum-free medium ( Figure 9E). The expression of FGF23, which is mainly expressed in bone, could not be detected in unstimulated (Ct value: > 40, n = 5) MDCK. Cisplatin-treated MDCK cells exhibited lower Ct values for FGF23, however expression was still very low (Ct value: 37.1 ± 1.85, n = 5).
ELISA-based quantification of αklotho protein is particularly feasible in human cells. Therefore, we performed further experiments in human proximal tubular cell line HK-2. We treated these cells with the cytotoxic agents and apoptosis stimulants in a way similar to MDCK and NRK-52E cells and measured αklotho transcripts as well as sKL protein by ELISA.  Figure 10E), and serum-free incubation ( Figure 10G) significantly down-regulated αklotho gene expression. In line with this, sKL protein concentration was lower in the cell culture supernatant of HK-2 cells upon incubation with cisplatin ( Figure  10B), doxorubicin ( Figure 10F), and in the absence of serum ( Figure 10H) and virtually unchanged upon exposure to paclitaxel ( Figure 10D).
As a last step, we analyzed sKL in serum samples from patients before and after chemotherapy ( Table 1) and found that the serum sKL concentration was not significantly different after chemotherapy compared to samples obtained before therapy (Figure 11).
According to our study, αklotho expression was upregulated by antineoplastic cytostatic agents cisplatin, paclitaxel, and doxorubicin in MDCK and NRK-52E cells within 24 h. Moreover, caspase 3 activator PAC-1 enhanced αklotho expression in both cell lines, whereas serum depletion was only effective in MDCK cells. Caspase 3 activation and serum depletion can be expected to induce apoptosis [44,45]. In sharp contrast, the same treatment resulted in down-regulation of both, αklotho transcripts and sKL protein, in HK-2 cells. The serum concentration of sKL was not significantly affected by chemotherapy.
Treatment with antineoplastic agents induces cellular stress through different mechanisms: Cisplatin impairs DNA replication by enabling inter-and intrastrand crosslink adducts [41], anthracycline derivative doxorubicin is a topoisomerase II inhibitor and DNA intercalator [42], and paclitaxel is an antimitotic agent that prevents spindle assembly by interacting with  tubulin [50]. Ultimately, the cellular impairments induced by these drugs may result in apoptotic cell death, a consequence intended in therapeutic use of these agents in the treatment of different types of cancer [51]. In line with this, cisplatin, doxorubicin, and paclitaxel reduced viability and proliferation of MDCK and NRK-52E cells, albeit to a variable extent. Moreover, the treatment was followed by induction of apoptosis and partially by secondary necrosis. In these two cell lines, apoptosis was paralleled by a marked upregulation of αklotho gene expression. In addition, expression of pro-apoptotic genes BAD, BAX, and BAX/BCL-2 ratio was induced by the chemotherapeutic agents, albeit to a variable extent. Also, direct induction of apoptotic cell death in the absence of cytotoxic drugs up-regulated αklotho mRNA levels in MDCK and Figure 11. The serum concentration of soluble klotho (sKL) in patients with cancer before and after administration of a cycle of chemotherapy. Arithmetic mean ± SEM of sKL serum concentration (n = 9; paired t-test) in patients 24 ± 4 h before and after administration of a cycle of chemotherapy. AGING NRK-52E cells. According to these results, αklotho expression was upregulated in injured and potentially moribund MDCK and NRK-52E cells prior to their putative death.
Several of the effects of αklotho on major intracellular signaling pathways can be expected to be pro-apoptotic: Inhibition of IGF-1 and insulin signaling [52] as well as Wnt signaling [53]. Also αklotho's role as a tumor suppressor fits to the concept of αklotho being proapoptotic [52]. Accordingly, our findings, i.e., upregulation of αklotho in MDCK and NRK-52E cells prone to death, may be a novel aspect of the cellular machinery which is part of the initiation and/or execution of apoptosis. Other effects of αklotho including increased anti-oxidant resistance [54], further anti-apoptotic properties [55], or reduced inflammation [29] may rather be associated with being pro-survival.
In view of the latter aspect of αklotho signaling, upregulation of αklotho in damaged and/or dying cells as revealed by our study could therefore be interpreted as an attempt to enhance cellular stress resistance and possibly overcome the injury. In line with this, αklotho has been shown to counteract another form of cell death, necroptosis [46]. Definitely, further studies are necessary to decipher the precise role of increased αklotho expression in cells exposed to potentially deadly noxae.
In an attempt to identify the mechanism underlying αklotho up-regulation in MDCK and NRK-52E following exposure to cytotoxic agents or other apoptosis stimulants, we uncovered a role for transcription factor PPARγ. PPARγ has been demonstrated to be relevant for αklotho expression [48] and is upregulated itself by cisplatin [49]. In line with this, we could confirm that the chemotherapeutics upregulate PPARG in both, MDCK and NRK-52E cells. Moreover, using PPARγ antagonist SR-202 we demonstrated that the cisplatin effect on αklotho in MDCK cells is indeed dependent on PPARγ albeit other factors are likely to be involved, too.
In the kidney, transmembrane αklotho forms a complex with FGFR1, yielding the receptor for bonederived hormone FGF23 [12]. In line with stimulation of αklotho expression, the cytotoxic agents and apoptosis inducers also up-regulated FGFR1 in MDCK cells.
While our experiments clearly demonstrated upregulation of αklotho in apoptotic MDCK and NRK-52E cells and uncovered PPARγ as a factor explaining, at least in part, this effect, a completely different response was found in HK-2 cells: The same treatment down-regulated both, αklotho gene expression and sKL concentration in the cell culture supernatant. Several factors may contribute to this discrepancy: Firstly, HK-2 is a human proximal tubule cell line from normal kidney that has been immortalized with human papilloma virus (HPV 16) E6/E7 genes, and these two genes are part of its genome [56]. In contrast, MDCK and also NRK-52E cells are spontaneously immortalized cells [57]. As a matter of fact, E6 and E7 genes used to immortalize HK-2 cells render them more resistant to apoptotic stimuli [58], an effect that may help explain the different response of HK-2 cells observed in our study. Secondly, it also appears possible that the origin of the cells (MDCK cells: dog, NRK-52E: rat, HK-2: human) also contributes to the different response [59]. Thirdly, the renal localization of αklotho may play a role: It is expressed in proximal and, at a higher level, in distal tubule. Renal phosphate handling mainly occurs in the proximal tubule, but its regulation is more dependent on αklotho in the distal tubule [60,61]. MDCK cells are from distal tubule [62], whereas NRK-52E cells are from proximal tubule [63] as are HK-2 cells [64]. Therefore, the different origin of the cell lines may also contribute to the contrasting results. Moreover, it has to be kept in mind that renal cell lines are only models that do not reflect all aspects of kidney physiology [65]. Therefore, our diverging results using the three different kidney cell lines also underscores that care must be taken when studying αklotho in cell culture.
In a pilot human study, we studied the impact of one cycle of chemotherapy on serum sKL in patients suffering from different types of cancer. We did not observe a significant change of sKL after chemotherapy. It is a major limitation of this small pilot study that patients with different forms of cancer, different chemotherapeutic regimens and different treatment cycles were included. Hence, several aspects may be relevant for our finding: Different forms of cancer themselves impact on αklotho [66]. Moreover, the disease stage and also the number of chemotherapy cycles may influence the effect on αklotho. Although distal tubule is thought to be the main source of sKL [61], also proximal tubule may produce sKL. Given the different response of distal tubular MDCK and proximal tubular HK-2 cells to chemotherapeutics, it appears to be possible that divergent effects also play a role in the human kidney. Definitely, further human studies are warranted to define possible effects of cytotoxic agents on sKL.
Since αklotho plays a particular role in patients with severe disease (e.g., CKD patients [67]), it would of course be of high clinical interest to know whether different responses of αklotho to chemotherapeutics are of clinical relevance and may reflect a different In conclusion, our study shows that the expression of αklotho gene is stimulated in MDCK or NRK-52E cells exposed to cytotoxic chemotherapeutics cisplatin, doxorubicin or paclitaxel or treated with apoptosis inducers PAC-1 or serum depletion. The effect is, at least in part, dependent on PPARγ. In contrast, the same treatment down-regulates αklotho gene expression and sKL protein in HK-2 cells.  The primers used in qPCR analysis are provided in Table  2. αklotho, PPARG, FGFR1, BAD, BAX, and BCL-2 mRNA levels were normalized to TBP mRNA.

Viability assay (MTT assay)
Cells were seeded into 96-well plates and treated as described for 24 h and for another hour with 3-[4,5dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide (MTT; Sigma-Aldrich, Schnelldorf, Germany). Thereafter, the MTT solution was replaced by dimethyl sulfoxide (DMSO; AppliChem, Darmstadt, Germany), and absorption was measured at 550 nm and 690 nm (reference) on a FluoStar Omega plate reader (BMG Labtech, Ortenberg, Germany). Results were normalized to vehicle-treated cells and are given as percentage of viable cells.

ELISA
HK-2 supernatants and patients' serum samples were subjected to ELISA for measurement of soluble αklotho protein according to the manufacturer's protocol (IBL, Hamburg, Germany).

Apoptosis and necrosis assay
The rate of apoptosis and necrosis was measured using the RealTime-Glo Annexin V Apoptosis and Necrosis Assay (Promega) according to the manufacturer's protocol.

Western blotting
MDCK cells were cultured in T25 cell culture flasks (Greiner Bio-One) for 24 h under standard conditions, then incubated with or without 10 µM cisplatin for another 24 h. After cell lysis using RIPA buffer (Cell Signaling Technology, Frankfurt, Germany) supplemented with protease and phosphatase inhibitor cocktail and EDTA (Halt, Thermo Scientific), total protein concentration was measured by Bradford assay (Bio-Rad). Thirty µg of total protein were subjected to standard 10% SDS-PAGE and Western Blotting. The following antibodies were used: anti-FGF receptor 1 (D8E4), anti-β-actin (8H10D10), anti-rabbit IgG HRPlinked (all from Cell Signaling Technology), and antimouse IgG HRP-linked antibody (Abcam, Cambridge, UK). For visualization, membranes were incubated for 2 min with Westar Nova 2.0 (β-actin) or Westar Supernova (FGFR1) ECL substrate (both from Cyanagen, Bologna, Italy). Densitometrical analysis was performed on a C-Digit ® Blot scanner (Li-Cor, Lincoln, NE, USA) and FGFR1 bands were normalized to β-actin bands using the Image Studio ™ software (Li-Cor).

Patients
Serum samples were collected from cancer patients of the Department of Oncology, University Hospital of Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany. The study was approved by the ethics committee of Martin-Luther-University (approval no. 2014-75). Blood samples were collected 20 ± 4 h before and after chemotherapy, centrifuged and frozen at −70°C until analysis. Patient characteristics are depicted in Table 1.

Statistics
Data represent arithmetic mean ± standard error of the mean (SEM) with n denoting the number of independent experiments. Groups were tested for normal distribution using Shapiro-Wilk test. The cell number and viability experiments were analyzed with one-sample t-test or alternatively with one-sample Wilcoxon signed rank test, as appropriate. Data with more than two groups were analyzed with repeated measures analysis of variance (ANOVA) followed by Dunnett's multiple comparison test or with nonparametric Friedman ANOVA and Dunn-Bonferroni post-hoc test. If p < 0.05, differences were considered significant. SPSS software was used for statistical data evaluation (IBM Version 27.0; Armonk, NY, USA).