The impact of SGLT2 inhibitors on αKlotho in renal MDCK and HK-2 cells

αKlotho is a transmembrane protein predominantly expressed in the kidney serving as a co-receptor for phosphate homeostasis-regulating hormone FGF23 and has an extracellular domain that can be cleaved off and is a hormone. αKlotho deficiency results in accelerated aging and early onset of aging-associated diseases while its overexpression strongly expands the lifespan of mice. Moreover, αKlotho exerts health-beneficial anti-inflammatory, anti-neoplastic, anti-fibrotic, and anti-oxidant effects. Higher αKlotho levels are associated with better outcomes in renal and cardiovascular diseases. SGLT2 inhibitors are novel drugs in the treatment of diabetes by inhibiting renal glucose transport and have additional nephro- and cardioprotective effects. We explored whether SGLT2 inhibitors affect αKlotho gene expression and protein secretion. Experiments were performed in renal MDCK and HK-2 cells, and αKlotho transcripts were determined by qRT-PCR and Klotho protein by ELISA. SGLT2 inhibitors canagliflozin, sotagliflozin, and dapagliflozin enhanced whereas empagliflozin reduced αKlotho gene expression in MDCK cells. By the same token, canagliflozin, sotagliflozin, dapagliflozin, but not empagliflozin down-regulated p65 subunit of pro-inflammatory NFκB. In HK-2 cells, all SGLT2 inhibitors reduced αKlotho transcripts. Canagliflozin and sotagliflozin, however, increased Klotho protein concentration in the cell culture supernatant, an effect paralleled by up-regulation of ADAM17. Taken together, our investigations demonstrate complex effects of different SGLT2 inhibitors on αKlotho gene expression and protein secretion in renal MDCK and HK-2 cells.


Introduction
aKlotho is a renal protein the lack of which results in a phenotype recapitulating human aging: aKlotho-deficient mice die at an age of a few weeks only whilst exhibiting aging-associated diseases affecting almost all organs including emphysema, hearing loss, infertility, or hypogonadism amongst others (1)(2)(3). The animals exhibit massive hyperphosphatemia and excess of active vitamin D, 1,25(OH) 2 D 3 , with a low-phosphate diet or low-vitamin D diet normalizing their phenotype (4,5). Hence, further research has demonstrated a pivotal role of aKlotho in maintaining phosphate homeostasis: aKlotho is a transmembrane protein that enhances binding affinity of fibroblast growth factor 23 (FGF23) for its renal receptor (6,7). FGF23 is produced in bone as a hormone and reduces production of 1,25(OH) 2 D 3 and reabsorption of phosphate, thereby lowering serum levels of 1,25(OH) 2 D 3 and phosphate (8)(9)(10)(11). The joint action of FGF23 and aKlotho in the regulation of phosphate and 1,25(OH) 2 D 3 explains why FGF23 deficiency is comparable to aKlotho deficiency in mice (12)(13)(14). In addition to being the co-receptor of FGF23, a soluble form of aKlotho exists called sKL which results from the cleavage of transmembrane aKlotho by proteases ADAM10 and ADAM17 (15, 16). SKL has hormone-like properties and influences various cellular effects including signaling or membrane transport (17).
SGLT2 inhibitors such as dapagliflozin, canagliflozin, empagliflozin, or sotagliflozin were introduced as new antidiabetics in the early 2010s (32). They efficiently lower blood glucose levels by inhibiting renal SGLT2, a Na + -dependent glucose transporter which normally accomplishes complete glucose reabsorption along with SGLT1, another isoform that is not inhibited by specific SGLT2 inhibitors (33)(34)(35). As a consequence, the patients excrete more glucose in their urine (36,37). Recent large clinical studies of high quality have unequivocally proven that SGLT2 inhibition has further significant and diabetes-independent cardio-and nephroprotective effects that are, in many aspects, regarded as a milestone in cardiology and nephrology (38)(39)(40). SGLT2 inhibition is the only approved drug therapy in heart failure with preserved ejection fraction (HFpEF), and in CKD, SGLT2 inhibitor therapy is the first major breakthrough since the introduction of angiotensin-converting enzyme (ACE) inhibitors in the 1990s, since the drugs convincingly delay loss of kidney function (41)(42)(43)). The precise underlying mechanisms of SGLT2 inhibitor-dependent cardio-and nephroprotection have, hitherto, remained enigmatic although many beneficial effects have been characterized including the prevention of hyperfiltration in the kidney or anti-inflammatory and anti-fibrotic effects (44, 45).
Given the strong cardio-and nephroprotection provided by aKlotho, which is also, at least in part, dependent on its antiinflammatory and anti-fibrotic effects, we hypothesized that benefits of therapy with SGLT2 inhibitors may be related to up-regulation of cellular aKlotho levels. This hypothesis has already been formulated in a recent review (46). To verify it, we performed experiments in canine MDCK and human HK-2 kidney cell lines.
For experiments, cells (150,000 MDCK or 120,000 HK-2 cells per well) were seeded into 6-well plates, grown for 24 h, and then treated for 24 h with or without canagliflozin, sotagliflozin, empagliflozin (all from Selleck Chemicals, Planegg, Germany), dapagliflozin, or phlorizin dihydrate (both from Sigma-Aldrich, Schnelldorf, Germany) at the indicated concentrations. Control cells were treated with the appropriate amount of solvent dimethyl sulfoxide. For the measurement of soluble Klotho protein in the cell culture medium, supernatants were collected after 24 h of treatment and stored at -70°C until further use.

RNA isolation and qualitative expression analysis
Total RNA from MDCK, HK-2, and UMR-106 cells was isolated using a phenol-chloroform extraction technique (RNA-Solv Reagent, Omega Bio-Tek, Norcross, GA, USA or TriFast Reagent, VWR, Bruchsal, Germany) according to the manufacturer's protocol. For qualitative detection of SLC5A1 and SLC5A2 transcripts and quantitative gene expression analysis in MDCK cells, total RNA was extracted by means of phenol-chloroformextraction, subjected to DNase treatment, and RNA purification (NucleoSpin RNA, Macherey-Nagel, Düren, Germany). Synthesis of complementary DNA (cDNA) was performed using 1.2 µg of total RNA, random primers, and GoScript Reverse Transcription System (Promega, Mannheim, Germany).
Polymerase chain reaction (PCR) was conducted on a Biometra TAdvanced thermocycler (Analytik Jena, Jena, Germany) using 2 µL cDNA, 0.25 µM canine forward and reverse primers or 0.5 µM human primers, respectively, 10 µL GoTaq Green Master Mix (Promega), and sterile water up to 20 µL. Conditions for PCR included 3 min at 94°C; followed by 40   Relative mRNA transcript levels were calculated by the 2 -DDCt method using TBP as internal reference gene and normalized to the control group. For quantitative analysis of SGLT1 and SGLT2 expression, relative mRNA transcripts were calculated using the 2 -DCt formula and TBP as internal reference.
Klotho protein determination in cell culture supernatant HK-2 supernatants were subjected to enzyme-linked immunosorbent assay (ELISA) for measurement of soluble Klotho protein (IBL, Hamburg, Germany).

Statistics
Data are reported as means ± standard error of mean (SEM) in bar graphs with scatter plots, and n is the number of independent experiments. Data were tested for normality using Shapiro-Wilk test. Differences between control and treatment group were assessed by two-tailed paired Student's t test, Wilcoxon matched-pairs signed rank test (for data not passing normality), one-sample t test or Wilcoxon signed rank test (for data not passing normality), as indicated in the figure legends. P < 0.05 was considered statistically significant. SPSS version 27.0 (IBM, Armonk, NY, USA) or GraphPad Prism 6.0 (GraphPad Software, San Diego, California, USA) were used for statistical analysis.

Results
We utilized canine MDCK and human HK-2 kidney cells to study the regulation of aKlotho. At first, we investigated the expression of SGLT1 (encoded by SLC5A1) and SGLT2 (encoded by SLC5A2) in the two cell lines by PCR. According to qualitative analysis, mRNA specific for both glucose transporters was detectable in MDCK cells seemingly to a small extent ( Figure 1A). For HK-2 cells, mRNA for both transporters could be verified with a markedly stronger band for SGLT2 ( Figure 1B). In addition, we performed quantitative analysis and examined relative expression (n=3 for all analyses) of SGLT1 and SGLT2 in both cell lines by qRT-PCR. As a result, cycle threshold (Ct) value for canine SLC5A1 (SGLT1) in MDCK cells was 37.46 ± 1.499, whereas a positive signal for SLC5A2 (SGLT2) could not be detected in these cells (Ct > 40 for SLC5A2). In HK-2 cells, Ct value for SLC5A1 (SGLT1) was 37.39 ± 0.661 resulting in 3.93 x 10 -5 ± 2.055 x 10 -5 arbitrary units of relative SLC5A1 mRNA transcription normalized to the housekeeping gene TBP. For SLC5A2 (SGLT2), Ct value was 30.35 ± 0.128 resulting in 4.04 x 10 -3 ± 0.447 x 10 -3 arbitrary units of relative SLC5A2 mRNA transcription normalized to TBP.
Next, we treated MDCK cells with increasing concentrations of the specific SGLT2 inhibitor canagliflozin, which is widely used in the treatment of patients, for 24 h and utilized qRT-PCR to assess the effect on aKlotho gene expression. As detailed in Figure 2A, canagliflozin up-regulated aKlotho gene expression in a dosedependent manner. As a further step, we investigated whether also sotagliflozin, a dual SGLT1/SGLT2 inhibitor approved for use in patients, up-regulates aKlotho. To this end, we treated MDCK cells with different concentrations of sotagliflozin for 24 h, and measured aKlotho transcripts by qRT-PCR. Similar to canagliflozin, sotagliflozin enhanced aKlotho mRNA abundance ( Figure 2B). Further experiments explored the effect of a 24 hincubation with or without dapagliflozin ( Figure 2C) or empagliflozin ( Figure 2D) on aKlotho in MDCK cells. Whereas dapagliflozin ( Figure 2C) up-regulated aKlotho gene expression to an extent similar to canagliflozin (Figure 2A) or sotagliflozin ( Figure 2B), empagliflozin surprisingly reduced aKlotho transcripts in MDCK cells ( Figure 2D).
Finally, we tested phlorizin, a non-specific glucose transporter inhibitor inhibiting both, SGLT1 and SGLT2, which is not approved for use in patients. As illustrated in Figure 3, treatment of MDCK cells with 30 µM phlorizin (24 h) resulted in a small, but statistically significant up-regulation of aKlotho transcripts whereas higher concentrations had no significant effect on aKlotho (Figure 3).
Owing to its anti-inflammatory properties (45), we investigated whether SGLT2 inhibitors impact on the expression of p65 subunit (encoded by RELA) of pro-inflammatory transcription factor complex NFkB. As illustrated in Figure 4, the three SGLT2 inhibitors which enhanced aKlotho down-regulated RELA expression in MDCK cells (Figures 4A-C). Empagliflozin, which did not significantly affect aKlotho, and phlorizin did not significantly modify RELA in MDCK cells (Figures 4D, E).
Our results thus far indicate that canagliflozin, sotagliflozin, and dapagliflozin, but not empagliflozin, up-regulated aKlotho in MDCK cells. We wondered whether this effect translates into enhanced Klotho protein secretion. In order to answer this question, we carried out a new series of experiments in human HK-2 cells since ELISA-based Klotho protein quantification is not feasible in canine MDCK cells. First, we examined the effect of the SGLT2 inhibitors on aKlotho transcripts in HK-2 cells. As demonstrated in Figure 5A, canagliflozin, sotagliflozin, dapagliflozin as well as empagliflozin down-regulated aKlotho gene expression in HK-2 cells in sharp contrast to MDCK cells, whereas phlorizin had no significant effect on aKlotho mRNA levels in HK-2 cells. As a second step, we employed ELISA to measure Klotho protein in the cell culture supernatant. According to Figure 5B, Klotho protein abundance was slightly but significantly higher in the supernatant of HK-2 cells treated with canagliflozin or sotagliflozin compared to control cells ( Figure 5B, upper panel). Dapagliflozin or empagliflozin did, however, not significantly modify the Klotho protein amount in the cell culture supernatant ( Figure 5B, lower panel).
ADAM10 and ADAM17 are proteases that are involved in aKlotho shedding, a process yielding soluble Klotho in the blood or cell culture supernatant (15). To investigate whether ADAM10 or ADAM17 regulation accounts for the different effects of gliflozins on Klotho protein in the cell culture supernatant, we tested the effect of SGLT2 inhibitors on ADAM10 and ADAM17 expression in HK-2 cells. QRT-PCR analyses revealed that none of the SGLT2 inhibitors had a significant effect on ADAM10 transcripts ( Figure 6A) in HK-2 cells. However, a significant increase in ADAM17 transcript levels was observed in response to canagliflozin or sotagliflozin but not dapagliflozin or empagliflozin treatment ( Figure 6B), an effect likely to contribute to the different actions of SGLT2 inhibitors on Klotho protein in the cell culture supernatant of HK-2 cells.
Finally, we investigated whether the SGLT2 inhibitors also impact on gene expression of Fgf23, a bone-derived hormone which requires renal aKlotho as a co-receptor. To this end, UMR-106 osteoblast-like cells were treated without or with FIGURE 3 Effect of phlorizin on KL transcript level in MDCK cells. Arithmetic means ± SEM of relative KL mRNA abundance normalized to TBP (n=7) in MDCK cells treated without (Ctr) or with phlorizin for 24 h. *p < 0.05 indicates significant difference from control-treated cells. (One sample t test).

Discussion
Our investigations revealed that common SGLT2 inhibitors used in the treatment of patients with diabetes, CKD, or heart failure modify aKlotho gene expression and Klotho protein secretion in renal MDCK and HK-2 cells.
In canine MDCK cells, canagliflozin, sotagliflozin, and dapagliflozin up-regulated aKlotho gene expression. In contrast, empagliflozin down-regulated aKlotho transcripts in MDCK cells. Importantly, empagliflozin has the highest selectivity for SGLT2 over SGLT1 among the four gliflozins tested (47). According to our quantitative expression analysis, MDCK exhibited low SGLT1 and virtually no SGLT2 expression. Therefore, it can be assumed that exposure to empagliflozin resulted in higher remaining SGLT1 activity compared to exposure to the other SGLT2 inhibitors in MDCK cells. It is tempting to speculate that this difference in remaining SGLT1 activity may play a role in the gliflozin effect on aKlotho in MDCK cells. Certainly, aKlotho up-regulation in MDCK cells was independent of SGLT2 since this transporter was not expressed in MDCK cells. In line with this, cardioprotective effects of SGLT2 inhibitors are not necessarily related to SGLT2, since expression of SGLT2 in the heart and cardiomyocytes is controversial (48-50). Therefore, upregulation of aKlotho expression in MDCK cells by canagliflozin, sotagliflozin, and dapagliflozin is probably also an SGLT2-independent effect.
Therefore, other SGLT2-independent effects of gliflozins must account for their effect on aKlotho in MDCK cells. As a matter of fact, such effects exist and play a role in nephro-and cardioprotection by SGLT2 inhibitors: Among them are antiinflammatory properties (51,52). Accordingly, we observed a small, but statistically significant down-regulation of p65 subunit (encoded by RELA) of pro-inflammatory NFkB transcription factor (53, 54) by canagliflozin, sotagliflozin, and dapagliflozin. This effect can clearly be expected to be anti-inflammatory and may be one of the SGLT2-independent effects contributing to the up-regulation of aKlotho. Interestingly, in line with its failure to significantly modify aKlotho expression, empagliflozin did not alter RELA expression, either. Therefore, suppression of inflammation may be relevant for the up-regulation of aKlotho by SGLT2 inhibitors in MDCK cells in the absence of SGLT2. Alternatively, RELA suppression may be secondary to enhancement of aKlotho in canagliflozin-, sotagliflozin-, or dapagliflozin-treated MDCK cells as aKlotho has anti-inflammatory properties (20,22).
Given the cellular responses of HK-2 cells to SGLT2 inhibitors, the situation even becomes more complex. All SGLT2 inhibitors reduced aKlotho gene expression in HK-2 cells to a small, but statistically significant extent, an observation in contrast to MDCK cells. Our expression analysis suggested markedly higher SGLT2 than SGLT1 expression in HK-2 cells. Since the strong upregulation of aKlotho in MDCK cells in the absence of SGLT2 expression suggests that this up-regulation is SGLT2-independent, the different expression pattern of SGLT1 and SGLT2 in MDCK cells and HK-2 cells may contribute to the contrasting effects on aKlotho transcripts in the two cell lines. Therefore, inhibition of SGLT2 in HK-2 cells may be accompanied by down-regulation of aKlotho expression. In the kidney, highest aKlotho expression is observed in the distal tubule, but also the proximal tubule expresses it (55,56). Whereas MDCK cells are derived from distal tubule, HK-2 cells are of proximal tubule origin (57, 58). The different origin of the two cell lines may also help explain the different cellular response to gliflozin treatment. Since SGLT2 is expressed in proximal tubular cells, the gliflozin effect on SGLT2 and aKlotho may be related to actions in different cell types in the kidney.
Despite the small but significant suppressive action of all four SGLT2 inhibitors on aKlotho transcripts in HK-2 cells, canagliflozin and sotagliflozin enhanced Klotho protein secretion. Hence, both SGLT2 inhibitors ultimately increased Klotho protein levels in the supernatant of HK-2 cells, an effect that can be expected to be in line with up-regulation of aKlotho gene expression by SGLT2 inhibitors in MDCK cells. In contrast, dapagliflozin and empagliflozin did not significantly affect Klotho protein levels in the cell culture supernatant. In the kidney, soluble Klotho is generated due to cleavage of transmembrane aKlotho by ADAM10 and ADAM17 proteinases (15). Our expression analysis of ADAM10 and ADAM17 indeed revealed that canagliflozin and sotagliflozin, but not dapagliflozin or empagliflozin up-regulated ADAM17 expression in HK-2 cells. This effect is likely to explain why canagliflozin and sotagliflozin, but not dapagliflozin or empagliflozin increased Klotho protein concentration in the cell culture supernatant of HK-2 cells although all four gliflozin suppressed aKlotho gene expression in these cells according to qRT-PCR analysis.
Phlorizin, a dual inhibitor SGLT1/SGLT2 not in clinical use since inhibition of both Na + -dependent glucose transporters, can be expected to cause serious adverse effects, only had moderate (MDCK cells) or no significant effect (HK-2 cells) on aKlotho although its glucosuric, hence anti-diabetic effect may even be stronger than that of gliflozins. This finding corroborates the notion that the health benefits of SGLT2 inhibitors for the heart or kidneys, which may include up-regulation of aKlotho as unraveled by this study, are independent of the glucose-lowering effects of this class of drugs.
It is a major finding of this investigation that the four SGLT2 inhibitors studied yielded different results in our cell culture experiments: In MDCK cells, empagliflozin failed to up-regulate aKlotho in contrast to canagliflozin, sotagliflozin, and dapagliflozin, and in HK-2 cells, canagliflozin and sotagliflozin, but not dapagliflozin and empagliflozin increased Klotho protein in the cell culture supernatant. It is tempting to speculate whether these findings indeed translate into a different capacity of the SGLT2 inhibitors to upregulate aKlotho levels in vivo and whether this would have a clinical significance. Thus far, most clinical studies revealed similar health benefits of different gliflozins. However, it must be kept in mind that the clinical studies rarely include direct comparisons of different gliflozins. In a recent study of acute kidney injury in rats, differences between empagliflozin and sotagliflozin became apparent (59). Whether or not the differences of the four gliflozins with regard to aKlotho matter in vivo, must be investigated in the future. A next step could be an ex vivo study addressing the effect of the four gliflozins on aKlotho in proximal tubule cells from diabetic rats to investigate whether aKlotho regulation is different in disease.
Our study brings together two novel concepts of nephro-and cardioprotection, i.e. therapy with SGLT2 inhibitors and increasing the availability of aKlotho (Figure 7). Given the well-established beneficial effects of gliflozins on the kidney and heart and promising results of Klotho delivery particularly in kidney disease in conjunction with association studies showing better outcome with higher Klotho levels in renal and cardiovascular diseases, our study sheds new light on the relationship between SGLT2 inhibitors and aKlotho levels (46).
Our experiments with UMR-106 cells revealed that the four different SGLT2 inhibitors studied here markedly differed in their effects on FGF23. In humans, treatment with dapagliflozin or canagliflozin elevates serum FGF23 levels, effects at least in part secondary to changes in phosphate metabolism (60, 61). Why canagliflozin and sotagliflozin directly decreased and dapagliflozin and empagliflozin directly increased Fgf23 gene expression in cell culture experiments remains enigmatic and must be clarified in future studies.
It is a major limitation of our study that it is solely cell culturebased. Clearly, additional animal experiments and human studies B A FIGURE 6 Effect of SGLT2 inhibition on ADAM10 and ADAM17 mRNA expression in HK-2 cells. Arithmetic means ± SEM of relative ADAM10 (A) and ADAM17 (B) mRNA expression normalized to TBP in HK-2 cells either treated with vehicle control (Ctr), canagliflozin (60 µM; n=10), sotagliflozin (60 µM; n=10), dapagliflozin (100 µM; n=8), or empagliflozin (100 µM; n=8) for 24 h. **p < 0.01 indicates significant difference from control cells. (One sample t test or Wilcoxon signed rank test). Proposed interdependence of SGLT2 inhibitors and aKlotho with regard to nephro-and cardioprotection. are warranted to decipher the effect of SGLT2 inhibitors on aKlotho in vivo. Furthermore it would be interesting to study whether other antidiabetics also impact on aKlotho.
Taken together our study revealed distinct actions of four different SGLT2 inhibitors on aKlotho and ADAM17 gene expression and Klotho protein secretion in two different renal cell lines, distal tubular MDCK cells and proximal tubular HK-2 cells.

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions
MFe and MFö designed the research. MFe, MFö, and LW interpreted data. MFe and MFö wrote the manuscript. MFe and LW performed the research and analyses. All authors contributed to the article and approved the submitted version.