Immunosuppressive calcineurin inhibitor cyclosporine A induces proapoptotic endoplasmic reticulum stress in renal tubular cells

Current immunosuppressive strategies in organ transplantation rely on calcineurin inhibitors cyclosporine A (CsA) or tacrolimus (Tac). Both drugs are nephrotoxic, but CsA has been associated with greater renal damage than Tac. CsA inhibits calcineurin by forming complexes with cyclophilins, whose chaperone function is essential for proteostasis. We hypothesized that stronger toxicity of CsA may be related to suppression of cyclophilins with ensuing endoplasmic reticulum (ER) stress and unfolded protein response (UPR) in kidney epithelia. Effects of CsA and Tac (10 µM for 6 h each) were compared in cultured human embryonic kidney 293 (HEK 293) cells, primary human renal proximal tubule (PT) cells, freshly isolated rat PTs, and knockout HEK 293 cell lines lacking the critical ER stress sensors, protein kinase RNA–like ER kinase or activating transcription factor 6 (ATF6). UPR was evaluated by detection of its key components. Compared with Tac treatment, CsA induced significantly stronger UPR in native cultured cells and isolated PTs. Evaluation of proapoptotic and antiapoptotic markers suggested an enhanced apoptotic rate in CsA-treated cells compared with Tac-treated cells as well. Similar to CsA treatment, knockdown of cyclophilin A or B by siRNA caused proapoptotic UPR, whereas application of the chemical chaperones tauroursodeoxycholic acid or 4-phenylbutyric acid alleviated CsA-induced UPR. Deletion of protein kinase RNA–like ER kinase or ATF6 blunted CsA-induced UPR as well. In summary, inhibition of cyclophilin chaperone function with ensuing ER stress and proapoptotic UPR aggravates CsA toxicity, whereas pharmacological modulation of UPR bears potential to alleviate renal side effects of CsA.

Current immunosuppressive strategies in organ transplantation rely on calcineurin inhibitors cyclosporine A (CsA) or tacrolimus (Tac). Both drugs are nephrotoxic, but CsA has been associated with greater renal damage than Tac. CsA inhibits calcineurin by forming complexes with cyclophilins, whose chaperone function is essential for proteostasis. We hypothesized that stronger toxicity of CsA may be related to suppression of cyclophilins with ensuing endoplasmic reticulum (ER) stress and unfolded protein response (UPR) in kidney epithelia. Effects of CsA and Tac (10 mM for 6 h each) were compared in cultured human embryonic kidney 293 (HEK 293) cells, primary human renal proximal tubule (PT) cells, freshly isolated rat PTs, and knockout HEK 293 cell lines lacking the critical ER stress sensors, protein kinase RNA-like ER kinase or activating transcription factor 6 (ATF6). UPR was evaluated by detection of its key components. Compared with Tac treatment, CsA induced significantly stronger UPR in native cultured cells and isolated PTs. Evaluation of proapoptotic and antiapoptotic markers suggested an enhanced apoptotic rate in CsA-treated cells compared with Tac-treated cells as well. Similar to CsA treatment, knockdown of cyclophilin A or B by siRNA caused proapoptotic UPR, whereas application of the chemical chaperones tauroursodeoxycholic acid or 4-phenylbutyric acid alleviated CsA-induced UPR. Deletion of protein kinase RNA-like ER kinase or ATF6 blunted CsA-induced UPR as well. In summary, inhibition of cyclophilin chaperone function with ensuing ER stress and proapoptotic UPR aggravates CsA toxicity, whereas pharmacological modulation of UPR bears potential to alleviate renal side effects of CsA.
Calcineurin (Cn) inhibitors (CNIs) are considered as the first-line immunosuppressive therapy in patients undergoing organ transplantation (1,2). Two drugs of this class, cyclosporine A (CsA) and tacrolimus (FK506, Tac), have been widely integrated into clinical practice. Cn is a holoenzyme consisting of a catalytic subunit (CnA), a regulatory subunit (CnB), and calmodulin, which functions as a Ca 2+ -dependent protein serine/threonine phosphatase. Alpha and beta isoforms of the catalytic subunit (CnAα and CnAβ) fulfill distinct and nonredundant tasks in various tissues (1,3). CnAβdependent dephosphorylation of nuclear factor of activated T-cell (NFAT) transcription factors is critical to T-lymphocyte activation and underlies the therapeutic action of CNI (4). Offtarget effects of CNI include renal complications because of suppression of CnAα activity in kidney tissue (5,6). A majority of organ-transplanted patients receiving CNI exhibit signs of kidney damage after several years of treatment (6). CNI-induced renal vasoconstriction, hyperactivity of the renin-angiotensin system (RAS), and dysregulation of major electrolyte transport systems cause hypertension and homeostasis disorders (6,7). At the cellular level, CNIs have been shown to induce endoplasmic reticulum (ER) stress and trigger unfolded protein response (UPR), which in turn may promote apoptosis of kidney epithelia (8,9).
Several retrospective studies suggested that CsA may be associated with stronger nephrotoxicity than Tac (10)(11)(12)(13)(14). Along the same line, a recent experimental study in cultured human cells demonstrated higher toxicity of CsA compared with Tac (15). The two CNIs inhibit Cn activity by forming complexes with distinct members of the immunophilin family: CsA binds to cyclophilins, whereas Tac interacts with 12 kDa FK506-binding protein (FKBP12). These individual interaction patterns may provide a reasonable explanation for distinct toxicity profiles of CsA versus Tac. Since the chaperone activity of cyclophilins is required for protein maturation, CsA may impair proteostasis ( Fig. 1) (16,17). Indeed, studies of CsA versus Tac in cell culture showed substantially stronger association of CsA with ER stress and UPR, but the role of cyclophilins herein remained to be elucidated (15,18). UPR is a complex signal transduction pathway that is triggered by pathophysiological alterations of proteostasis leading to accumulation of misfolded protein (19,20). Misfolded protein releases inositol-requiring protein 1 (IRE1α), protein kinase RNA-like ER kinase (PERK), and activating transcription factor 6 (ATF6) from their complexes with binding immunoglobulin protein (BiP) by competitive interactions with the latter. IRE1α, PERK, and ATF6 serve as ER-stress sensors and initiate a number of adaptive reactions promoting cell survival. PERK phosphorylates and deactivates the eukaryotic translation initiation factor 2α, thereby inhibiting protein translation. Parallel activation of IRE1α and ATF6 stimulates autophagy and accelerates protein folding via several signaling pathways, such as transcription and splicing of X-box binding protein 1 (XBP1) (19). Failure of these adaptive mechanisms triggers the proapoptotic UPR branch via stimulation of the C/EBP homologous protein (CHOP), which is a transcription factor that induces expression of proapoptotic genes (9,21,22). Interestingly, Cn itself has been implicated in adaptive UPR signaling, enabling long-term cell survival upon ER stress (23). Therefore, Cn inhibition may impair the ability of kidney epithelia to compensate for metabolic stress. Pharmacological targeting of UPR using chemical chaperones, inhibitors of translation, or autophagy enhancers has been considered as an emerging strategy to delay progression of chronic kidney disorders (9). In this context, improved understanding of UPR-related molecular pathways affected by CNI may deliver new options for managing their nephrotoxicity.
The present study addresses mechanisms underlying distinct toxicity profiles of CsA versus Tac with a focus on ER stress and UPR. Based on comparative analysis of CsA and Tac in cultured human embryonic kidney 293 (HEK 293) cells, primary human renal proximal tubular (PT) epithelial cells (HRPTEpCs), and freshly isolated rat PTs, we provide several lines of evidence in support of the hypothesis that the more pronounced toxic effects of CsA on proteostasis are rooted in the suppression of cyclophilin chaperone activity. Using genetic and pharmacological approaches, we show that targeting UPR bears therapeutic potential for alleviation of CsA nephrotoxicity.

Establishing CNI treatment protocols
First, we aimed to define the optimal CNI concentrations and treatment duration to set up a protocol permitting comparative analysis of their effects on ER and UPR in cultured cells. Earlier work reported ER stress and UPR in human cervical cancer (HeLa) and human lung carcinoma (A549) cells after application of 10 μM CsA for 6 h (15). Based on that protocol, we performed a dose-response analysis by treating native HEK 293 cells with increasing doses of CsA versus Tac (5-80 μM each) for 6 h to evaluate the effects on cell viability using the 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) assay (24). A substantial decrease of cell viability below 50% was observed starting with 20 μM CsA or 40 μM Tac ( Fig. 2A), which we defined as limitation precluding evaluation of ER stress and UPR because of the critical reduction of living cell mass. Since CsA has been reported to induce striking morphological changes such as cytoplasmic vacuolization (16,25), we also compared effects of CsA versus Tac (10 μM each) on cell morphology by light and electron microscopy. Ultrastructural analysis revealed no obvious morphological alterations after application of CsA or Tac for 6 h, whereas prolonged treatment for 24 h resulted in formation of cytoplasmic vacuoles in CsA-treated but not in Tac-treated cells (Fig. 2, B-G). Moreover, treatment of cells with CsA for 48 h resulted in a marked reduction of cell confluence compared with Tac or vehicle, as detected by light microscopy (Fig. S1). From these preliminary experiments, we concluded that the previously published treatment protocol for human cancer cells (10 μM CsA versus 10 μM Tac for 6 h) is applicable for our experimental settings (15), whereas increased CNI doses or treatment duration critically reduce the cell viability. Therefore, further experiments in HEK 293 cells, primary HRPTEpC, and microdissected rat PTs were performed using 10 μM CsA versus 10 μM Tac for 6 h.
To verify the equivalent efficacy of the chosen treatment protocols in terms of Cn inhibition, we evaluated phosphorylation of the established Cn substrate, NFAT (3). Immunoblotting using an antibody to phosphorylated NFAT showed comparable increases of phospho-NFAT levels in lysates from CsA-treated or Tac-treated HEK 293 cells suggesting that both CNIs attenuated Cn-dependent NFAT dephosphorylation to a comparable extent when applied at 10 μM concentrations for 6 h (Fig. S2).

CsA induces stronger ER stress and UPR than Tac
Next, we compared effects of CsA versus Tac on key UPR proteins using thapsigargin (Tg; 0.1 μM), an established . Schematic drawing illustrates common versus distinct pathogenetic mechanisms mediating nephrotoxic effects of CsA versus Tac. Both calcineurin inhibitors (CNIs) may lead to hyperactivity of renin-angiotensin-aldosterone system (RAAS), vasoconstriction, renal salt retention, and hypertension because of inhibition of the phosphatase activity of calcineurin (7,(54)(55)(56). In addition, CsA causes endoplasmic reticulum (ER) stress and proapoptotic unfolded protein response (UPR) because of suppression of chaperone function of cyclophilins (CYP; present study), whereas inhibition of FKBP12 by Tac may aggravate hypertension (57). Chemical chaperones (TUDCA or 4-PBA) alleviate the CsA-induced ER stress likely because of improved protein folding. Deletion of PERK (PERK KO) or ATF6 (ATF6 KO) blunted the CsA-induced proapoptotic UPR, suggesting that pharmacological modulation of these signaling pathways may improve cell survival. 4-PBA, 4-phenylbutyric acid; ATF6, activating transcription factor 6; TUDCA, tauroursodeoxycholic acid.
suppressed cyclophilin activity in the CsA-induced ER stress and UPR.
Since the absence of PERK or ATF6 may affect the proteostasis per se, we evaluated their downstream signaling components, namely CHOP and sXBP1. The baseline CHOP abundance was unaltered in PERK-KO cells but increased in ATF6-KO cells (+67%, p < 0.001) as compared with control unmodified cells (Fig. 8A). Evaluation of sXBP1 revealed similar baseline levels in control, PERK-KO, and ATF6-KO cells (Fig. 8B). Hence, we concluded that ATF6 deficiency is associated with altered baseline proteostasis resulting in activation of CHOP-inducing UPR branches such as PERK signaling (9).

Discussion
Today, there is still considerable clinical demand for CsA in patients well adjusted on this drug or those requiring conversion from Tac to CsA for nonrenal side effects of Tac such as diabetes (29,30). However, retrospective analysis of CNI nephrotoxicity suggested that Tac has more favorable pharmacological profile compared with CsA (6,13,14,31). Cytotoxicity of CsA has been associated with impaired proteostasis and integrated stress response in cultured nonrenal human cells (15). In line with this, the present results obtained in cultured cells of renal origin and isolated rat PTs demonstrate that CsA stimulates multiple UPR pathways, as evident from enhanced levels of sXBP1, p-IRE1α, and CHOP (8,9,32). Moreover, increased levels of CHOP, Bax, and cCas-3 along with decreased BCL-2 abundance in CsA-treated cells indicate failure of the adaptive UPR branches and switch to the proapoptotic mode (19). In contrast to CsA, Tac exerted only moderate or no effects on the UPR products in the present study. These results were obtained not only in littledifferentiated HEK 293 cells but also in more differentiated HRPTEpCs and dissected native rat PTs, which underlines their relevance in the pathophysiological context of CNI nephrotoxicity. Stronger cellular toxicity of CsA cannot be explained by different efficiency of CsA and Tac in terms of Cn inhibition, since our treatment protocols produced similar increases of NFAT phosphorylation levels. Therefore, we tested the possibility that toxic effects of CsA may be aggravated by the suppression of cyclophilins.
Cyclophilins assist in protein folding by catalysis of the cis-trans isomerization of proline imidic peptide bonds (17,33). Since prolyl isomerization is an intrinsically slow process, its catalysis by cyclophilins may be considered as a ratelimiting step in protein folding (33,34). Information on physiologic and pathophysiologic roles of different cyclophilin isoforms in the kidney is scarce and in part controversial. Earlier studies reported beneficial effects of pharmacologic cyclophilin inhibition or genetic silencing of cyclophilin D in animal models of acute kidney injury or chronic kidney disorders, including CsA nephrotoxicity (35)(36)(37). In contrast, other studies documented protective effects of CYPA and CYPB in mouse models of CsA nephrotoxicity or RASinduced kidney epithelial injury (38)(39)(40). It is tempting to speculate that interactions of CsA with CYPA or CYPB may impair the proteostasis because of suppression of peptidylprolyl cis-trans isomerase activity, whereas CsA-cyclophilin D complexes appear to hamper the mitochondrial function (35,38,39,41). Knockdown of CYPA or CYPB led to increased CHOP and cCas-3 levels in the present study, which supports a concept that CsA-induced suppression of cyclophilin-mediated chaperone function may cause ER stress A B Figure 8. Effects of inactivation of PERK and ATF6 on the unfolded protein response. A, representative immunoblot showing baseline levels of CHOP (approximately 27 kDa). B, sXBP1 (approximately 56 kDa) in lysates from PERK-deficient and ATF6-deficient HEK 293 cells; both genes were inactivated using CRISPR/Cas9 gene editing strategy. The graphs showing densitometric evaluation of CHOP and sXBP1 signals are placed below the respective immunoblots. N = three independent experiments. Data are the means ± SD. **p < 0.01, ***p < 0.001, ****p < 0.0001, NS. ATF6, activating transcription factor 6; CHOP, C/EBP homologous protein; HEK 293, human embryonic kidney 293 cell line; NS, not significant; PERK, protein kinase RNA-like ER kinase; sXBP1, spliced X-box binding protein 1. and proapoptotic UPR. According to their subcellular distribution, CYPA may be involved in folding and trafficking of cytosolic and membrane proteins, whereas CYPB may contribute to protein maturation and quality control in ER (40,42). The fact that knockdown of either cyclophilin isoform produced significant increases of CHOP and cCas-3 levels in the present study points to their nonredundant roles in proteostasis. Since Tac produced only moderate effects on UPR signaling in the present study, we conclude that FKBP12 chaperone function is less important for intact proteostasis. Indeed, FKBP12 has been implicated in modulation of the cell cycle rather than in protein folding and proteostasis (43). However, administration of Tac in vivo may produce metabolic stress in kidney epithelia indirectly, for example, because of RAS hyperactivity or constriction of renal arteries (7).
Targeting chaperone systems has been increasingly recognized as an emerging strategy for alleviation of kidney diseases (9). Previous studies in animal models suggest that ER stress contributes to progression of chronic kidney disorders to renal fibrosis, which can be retarded by use of chemical chaperones (9,44). The present results indicate that stimulation of protein folding using TUDCA or 4-PBA may protect cells from CsAinduced ER stress, as reflected by blunted increases of CHOP and cCas-3 levels in HEK 293 cells treated with CsA and either chemical chaperone. While chemical chaperones improve protein folding in a nonselective way, characterization of specific UPR pathways affected by CsA may provide further options to prevent or reduce its cell toxicity. To address this issue, we intervened in early UPR events by inactivating either PERK or ATF6 using CRISPR/Cas9 gene editing in the HEK 293 cell model.
PERK and ATF6 sense ER stress and recruit distinct mechanisms to restore protein folding capacity (19). Inactivation of either PERK or ATF6 produced no effects on baseline abundance of sXBP1, whereas abundance of CHOP was moderately increased in ATF6-KO, which may reflect stimulated proapoptotic UPR (19,45). Notably, our attempts to inactivate another ER-stress sensor, IRE1α, produced no viable cell line suggesting that intact IRE1α signaling is critical to cellular proteostasis and survival (32). Indeed, deletion of IRE1α has been shown to impair protein folding, autophagy, and mitochondrial function in podocytes (46). In the present study, PERK-KO cells exhibited milder increases of CHOP and p-IRE1α in response to CsA or Tg suggesting a blunted UPR. ATF6-KO cells showed blunted effects of CsA and Tg on CHOP and p-IRE1α as well, whereas stimulation of sXBP1 was more pronounced in this cell line. Tac produced little or no effect on UPR in the KO lines, apart from decreased sXBP1 levels in PERK-KO and moderately increased p-IRE1α levels in ATF6-KO cells. Although these results suggest that suppression of either PERK-mediated or ATF6-mediated signaling may interfere with the proapoptotic UPR branch and improve cell viability, further experiments including animal studies are mandatory to assess the individual impacts of these pathways in mediating the CsA toxicity.
In summary, the present comparative analysis of CsA versus Tac toxicity profiles suggests that the detrimental effects of CsA on proteostasis result from suppression of cyclophilins rather than Cn inhibition (Fig. 1). In addition, in vivo administration of CsA or Tac may indirectly cause or aggravate ER stress in kidney epithelia because of further local and systemic effects, such as sympathetic stimulation, rennin-angiotensinaldosterone system hyperactivity, vasoconstriction and hypoxia of renal tissue, and imbalanced workload of kidney epithelia (7,47,48). Therefore, we consider pharmacologic alleviation of ER stress using chemical chaperones or autophagy enhancers as an emerging therapeutic strategy to prevent or reduce CNI nephrotoxicity (8). In the clinical context, our data suggest that Tac possesses milder cytotoxicity compared with CsA and support the current trend of preferential Tac use in organ-transplanted patients (49).

Treatments
Cells were treated with CsA (Sigma-Aldrich) or Tac (FK-506 monohydrate; Sigma-Aldrich). Both drugs were dissolved in dimethyl sulfoxide (DMSO; Roth). Dose-response curves for CsA and Tac were established in native HEK 293 cells. Cell suspensions (100 μl containing approximately 10 4 cells) were inoculated to a 96-well plate and incubated overnight at 37 C with 5% CO 2 . Subsequently, cells were treated with CsA or Tac in concentrations of 0, 5, 10, 20, 40, and 80 μM, respectively, for 6 h each. Cell viability was evaluated using MTT (BioFroxx GmbH). MTT (5 mg/ml) was dissolved in PBS, and the solution was filtered for sterilization (0.22 μm; Whatman). Five microliters of the solution was added to each well at the end of the treatment. After ensuing incubation for 90 min at 37 C, the medium was discarded, and precipitated formazan crystals were dissolved in 100 μl DMSO and incubated with gentle agitation for further 15 min at 37 C. Absorbance was measured at 560 nm in a microplate reader (Biochrom ASYS, Expert 96), and the values obtained in vehicle-treated cells were set at 100%. Based on the results of this assay, 10 μM CsA and 10 μM Tac doses were chosen for further experiments. To define the optimal treatment duration, cells were grown to 80% confluence, treated with CsA or Tac for 6, 24, or 48 h, and evaluated by light and electron microscopy. The 6 h duration showed the best results by viability assay as well as morphological estimation of cell survival and was therefore chosen for the main experimental protocol. To assess effects of CNI on UPR, native HEK 293 cells, PERK-deficient (PERK-KO) or ATF6-deficient (ATF6-KO) cells were grown to 80% confluence and treated for 6 h with vehicle (DMSO), CsA (10 μM), Tac (10 μM), and Tg (Sigma-Aldrich; 0.1 μM) serving as positive control. Effects of concomitant application of the chemical chaperones, 4-PBA (Abcam) or TUDCA (Calbiochem), were studied in native HEK 293 cells in the presence of CsA. Cells were then harvested and processed for qPCR or immunoblotting. Alternatively, cells were fixed for immunofluorescence or ultrastructural evaluation.

Immunofluorescence
HEK 293 cells were grown on coverslips to 80% confluence, treated, and fixed in 4% paraformaldehyde for 10 min.
Permeabilization of the cells was performed using 0.5% Triton X-100 (Merck) in Tris-buffered saline (TBS) for 30 min. Unspecific protein binding was blocked with 5% bovine serum albumin in TBS for 30 min at room temperature. Primary antibodies against CHOP or cCas-3 (both Cell Signaling Technology) were diluted in the blocking medium (1:500) and applied overnight at +4 C in a humidified chamber. After washing steps, cells were incubated with Cy3-coupled secondary antibodies (1:250; Dianova) for 1 h at room temperature. Nuclei were counterstained with DAPI (Sigma-Aldrich), and the actin cytoskeleton was visualized using phalloidin (Alexa Fluor 488 phalloidin; Invitrogen), which was added to the secondary antibody solution. Coverslips were mounted (ProLong Glass Antifade Mountant; Invitrogen) on glass slides, and fluorescent signals were documented by confocal microscopy (LSM 5 Exciter; Carl Zeiss Microscopy GmbH). At least three independent experiments with five wells per treatment were quantified.

Ultrastructural analysis
Cells were fixed in 2.5% glutaraldehyde buffered with 0.1 M sodium cacodylate buffer, harvested with a cell scraper, and processed for EPON embedding as described previously (53). Ultrathin sections were cut (70 nm thickness; Ultracut S ultramicrotome; Leica), contrasted in 5% alcoholic uranyl acetate and 1% lead citrate, and imaged using an EM 906 transmission electron microscope (Zeiss).

Immunoblotting
Cells were harvested, dissolved in homogenization buffer (250 mM sucrose, 10 mM triethanolamine [AppliChem, Pan-Reac; ITW Reagents]), adjusted with a protease inhibitor cocktail (cOmplete; Roche), and sonicated five times for 1 s each. Cell debris was removed by centrifugation at 1000g at +4 C for 10 min, and the supernatant was collected. Protein concentrations were measured using Micro BCA protein assay kit (Thermo Scientific).
Protein lysates (30 μg per lane) were separated using SDS-PAGE (10% or 14%) and transferred to a nitrocellulose membrane (Macherey-Nagel). Membranes were then blocked Cyclosporine A induces ER stress in kidney epithelia with 5% bovine serum albumin in TBS for 30 min at room temperature, followed by overnight primary antibody incubation at +4 C. Antibodies used for immunoblotting are listed in Table 2. Signals were generated by incubation with the respective horseradish peroxidase-conjugated secondary antibodies (Dako; 1:2.000) for 1 h at room temperature, followed by incubation with Amersham ECL Western blotting detection reagent (GE Healthcare) and visualized in Intas ECL Chemo-Cam Imager (Intas Science Imaging). Densitometric evaluation was performed using ImageJ software (National Institutes of Health).

RNA extraction and real-time qPCR
Total RNA was isolated using PeqGOLD TriFast (VWR Life SCIENCE) according to the manufacturer's protocol. Reverse transcription was then performed with the help of Tetro Reverse Transcriptase kit (Bioline GmbH) using 1 μg total RNA. Specific primers used for evaluation of the expression of UPR-related genes are listed in Table 3 (27). Real-time qPCR was performed using HOT FIREPol EvaGreen qPCR Mix (Solis BioDyne) and 7500 Fast Real-Time PCR System (Applied Biosystems). Relative quantity of mRNA levels was normalized to endogenous GAPDH (forward: 5 0 -TGC ACC ACC AAC TGC TTA GC-3 0 ; reverse: 5 0 -GGC ATG GAC TGT GGT CAT GAG-3 0 ) expression.

siRNA-mediated knockdown
Specific CYPA and CYPB siRNA probes were generated by Ambion GmbH (Table 4). Transfection was performed at a cell confluence of 30% to 40% using INTERFERin (Polyplus-Transfection) according to the manufacturer's instructions.

Statistical analysis
Results were analyzed by routine parametric statistics for normal distribution as assumed from the experimental design. Comparative analysis between two groups was performed by unpaired t test. Comparative evaluation of multiple groups was performed using one-way ANOVA followed by Tukey's multiple comparisons test. GraphPad Prism 8 (GraphPad Software Inc) was used to analyze parameters. A probability level of p < 0.05 was accepted as significant.

Data availability
All data are contained within the article.
Supporting information-This article contains supporting information.
Acknowledgments-We thank Anette Drobbe for secretarial help and Katja Dörfel for excellent technical assistance. This work received financial support from the Deutsche Forschungsgemeinschaft (grants MU2924/2-2, BA700/22-2, and SFB 1365-C04/-S01). Conflict of interest-K. M. and S. B. report that financial support was provided by Deutsche Forschungsgemeinschaft. All other authors declare that they have no conflicts of interest with the contents of this article.