Cellular responses to silencing of PDIA3 (protein disulphide-isomerase A3): Effects on proliferation, migration, and genes in control of active vitamin D

.


Introduction
Vitamin D 3 (cholecalciferol) is a seco-steroid that can be naturally synthesised from 7-dehydrocholesterol in the skin or obtained from the diet.This inactive precursor requires further catalysis towards its active form, 1,25-dihydroxyvitamin D 3 (1,25(OH) 2 D 3 ), via two sequential hydroxylation reactions catalysed by cytochrome P450 (CYP) enzymes [10,21,6].1,25(OH) 2 D 3 , which is also called calcitriol, is important for several physiological processes, including the regulation of uptake and secretion of calcium, magnesium and phosphorus in the intestine and kidneys, respectively.Other roles for 1,25(OH) 2 D 3 involve e.g.effects on osteogenesis and cell proliferation and differentiation [6,20].The concentration of 1,25(OH) 2 D 3 is under tight regulatory control [6,10].Of particular importance for the 1,25(OH) 2 D 3 levels are the actions by CYP27B1, which forms 1,25(OH) 2 D 3 from its immediate precursor 25-hydroxyvitamin D 3 , and CYP24A1, which converts 1,25(OH) 2 D 3 into more easily eliminated metabolites.The physiological level of 1,25 (OH) 2 D 3 is regulated in a negative feed-back loop via its effect on CYP27B1, by virtue of a negative vitamin D response element (nVDRE) in the CYP27B1 promoter [42].Expression of the catabolic CYP24A1 is stimulated by an increased level of 1,25(OH) 2 D 3 in order to avoid excess 1,25(OH) 2 D 3 concentrations , which may negatively affect cellular function [10,4,6].In addition to the well-studied bioactivation pathway that produces 1,25(OH) 2 D 3 , alternative vitamin D pathways, forming other hydroxymetabolites with different biological effects, have also been reported [21,37].
The most well-known receptor mediating 1,25(OH) 2 D 3 signalling is VDR (vitamin D receptor), that functions as a transcription factor to induce genomic responses [21,25,48,6].VDR is the most recognised modus operandi through which 1,25(OH) 2 D 3 exerts its effects.Upon ligand binding, VDR translocates to the nucleus.The existence of a membrane-bound pool of VDR (mVDR) has also been reported [2,32,48].There is growing evidence that VDR is not the sole receptor able to mediate 1,25(OH) 2 D 3 signalling.A number of non-genomic effects by 1, 25(OH) 2 D 3 and related compounds have been reported, including e.g.altered expression of genes in the immune system, activation of calcium channels, and changes in the growth of chondrocytes and in the second messenger molecules [2,7,23,32,48].
PDIA3 (protein disulphide-isomerase A3), also known as 1,25D 3 -MARRS (1,25D 3 -membrane-associated, rapid response steroid-binding receptor) or ERp57, is a thiol reductase with protein disulphide isomerase activity that has been associated with some of the classic functions of 1,25(OH) 2 D 3 and is considered to be a mediator of the rapid cellular response to 1,25(OH) 2 D 3 [18,22,48].In the endoplasmic reticulum (ER) lumen, PDIA3 acts as a molecular chaperone and redox-catalyst, in order to modulate the folding of glycoproteins [24].It is also involved in unfolded protein response (UPR), following ER stress, and the assembly of major histocompatibility complex (MHC) class I [13,15].Various reports place PDIA3 in different subcellular compartments, to match the diverse roles attributed to it [8,41].The physiological role(s) of PDIA3 for different aspects of 1,25(OH) 2 D 3 signalling are not clear [21,43,48].Some findings have indicated crosstalk between PDIA3 and VDR in mediation of cellular effects [3,23].
The well-known anti-proliferative effect of 1,25(OH) 2 D 3 in many cell types has long been associated with processes involved in tumour development and progression of e.g., prostate, brain, and breast carcinomas [20,32,39].The levels of PDIA3 have been linked to altered cancer risk or prognosis in several studies on different types of carcinomas, although in most reports high levels of this protein is associated with a tumour-promoting effect [14,28,31,40].
In the present study we utilised siRNA-mediated silencing of PDIA3 in tumour cell lines of different origin to investigate the involvement of this protein in 1,25(OH) 2 D 3 -mediated cell responses.

siRNA-mediated silencing and treatment with 1,25-dihydroxyvitamin D 3
For silencing of PDIA3, Silencer™ Select siRNA (Thermo Scientific) was used to transfect PC3, DU145 or U2OS cells, seeded at a density appropriate for each assay.Silencer™ Select Negative Control No.1 siRNA ('siCtrl') or Silencer™ Select siPDIA3 (s6228, Thermo Scientific) ('siPDIA3') were used at a final concentration of 10 nM.In short, either siRNA was used to transfect PC3, DU145 or U2OS cells for 72 h, using siLentFect™ Lipid Reagent for RNAi (Bio-Rad Laboratories Inc.) in Opti-MEM™ I Reduced Serum Medium (Thermo Scientific), according to the manufacturer's instructions.Over 90% PDIA3 silencing was routinely achieved and was found to persist for at least 120 h following transfection.Treatment of cells with 1,25(OH) 2 D 3 (sc-202877A, Santa Cruz Biotechnology) for 24 h at a final concentration of 10 nM, or absolute ethanol (Thermo Scientific) ('EtOH') as vehicle control, was also performed.

Cell proliferation assay
Cell proliferation was examined using Click-iT™ EdU Cell Proliferation Kit for Imaging, Alexa Fluor™555 dye (Thermo Scientific).PC3 and DU145 cells were cultured on 8-well Nunc™ Lab-Tek™ II CC2™ Chamber Slides (Sigma-Aldrich) at a density of 200,000 cells/mL and transfected with either 'siCtrl' or 'siPDIA3', as described above.After 72 h, the cells were treated with 1,25(OH) 2 D 3 (or vehicle control) for 24 h, and simultaneously treated with 10 μM EdU for 16 h.The slides were fixed with 4% paraformaldehyde (PFA) (VWR) for 15 min on ice.The wells on the glass slides were encircled with ImmEdge Hydrophobic Barrier PAP Pen (Vector Laboratories), followed by 2 ×5 min wash with 3% w/v bovine serum albumin (BSA, Sigma-Aldrich) in 1x phosphatebuffered saline (PBS, Thermo Scientific) and permeabilised in 1x Trisbuffered saline (TBS, Thermo Scientific) 0.2% v/v Triton X-100 (Sigma-Aldrich) for 20 min, followed by 2 ×5 min wash with 3% w/v BSA in PBS.The cells were incubated with Click-iT™ Reaction Cocktail for 30 min in room temperature, according to the manufacturer's instructions, after which they were washed briefly with 3% w/v BSA in PBS.Blocking was performed with Intercept® (TBS) Blocking Buffer (LI-COR Biosciences) for 1 h at 37 • C in a humidified chamber, followed by incubation with primary antibody against PDIA3 (anti-PDIA3, 1:50, VMA00477, Bio-Rad Laboratories Inc.) diluted in Intercept® buffer, overnight at 4 • C in a humidified chamber.Afterwards, the cells were washed 3 ×10 min with 1x TBS 0.05% v/v Tween-20 (TBS-T, Thermo Scientific).Fluorophore-labelled secondary antibody (Donkey anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor Plus 647, A32795, Thermo Scientific) and Hoechst33342 (1:250, Thermo Scientific) were added for 1 h at 37 • C, followed by 2 ×10 min washing with 1x TBS and 1 ×5 min washing with 0.2x TBS and mounting with SlowFade Gold antifade reagent (Thermo Scientific).The slides were protected from light at all times.Images were acquired with Zeiss AxioImager M2 with a Zeiss Plan-Apochromat 63x NA 1.4 oil objective (Carl Zeiss AG) and then deconvolved with Huygens Essential (Scientific Volume Imaging, the Netherlands, http://svi.nl) using the Deconvolution Wizard option.Image analysis was performed with CellProfiler software version 4.1.3[11] on the deconvolved images.

Cell migration assay
Cell migration assays were performed using a scratch-wound protocol.PC3 and DU145 cells were seeded in a 96-well plate (Sarstedt) at a density of 720,000 cells/mL and transfected with either 'siCtrl' or 'siP-DIA3', as described above.After 72 h, the cells were subjected to wounding with a 10 μL tip and subsequently stimulated with 1,25 (OH) 2 D 3 (or vehicle control), as described above.Cell migration was observed over 24 h by acquiring images with 10% overlap around the wound area, that were later stitched together using the Grid/Collection Stitching Plugin in ImageJ version 1.53q [27].The distance the wounds spanned was measured in ImageJ version 1.53q and wounds between 250 and 350 μm for PC3 and between 350 and 450 μm for DU145 were included in downstream analysis.Images were acquired with ImageX-press® Micro (Molecular Devices).

Protein extraction and Western blot analysis
For the subcellular enrichment, protein was isolated from confluent cells, grown in 60 cm 2 dishes and lysed as described elsewhere [26].The cleared supernatant was collected as the cytoplasmic fraction and the nuclear pellet was further cleared.The resulting cleared supernatant was collected as the nuclear fraction.Total protein concentrations of the lysates were estimated using Pierce™ BCA Protein Assay Kit (Thermo Scientific), according to the manufacturer's instructions.The lysates were denatured at 95 • C for 5 min in 1x NuPAGE™ LDS Sample Buffer (Thermo Scientific) supplemented with 20 mM DTT to a final total protein content of 40 and 20 μg for the cytoplasmic and the nuclear protein pool lysates, respectively.For whole-cell lysates, cells were seeded at a density of 600,000 cells/mL in 6-well plates (Sarstedt), transfected with either 'siCtrl' or 'siPDIA3' and subsequently treated with 1,25(OH) 2 D 3 , as described above.The cells were washed 3x with ice cold 1x PBS and incubated in 1.2x NuPAGE™ LDS Sample Buffer supplemented with 20 mM DTT for 20 min before collection followed by denaturation at 95 • C for 5 min.Proteins were separated by SDS-polyacrylamide electrophoresis on a NuPAGE™ 4-12%, Bis-Tris, Mini Protein Gel (Thermo Scientific).Chameleon® Duo Pre-stained Protein ladder (LI-COR Biosciences) was used as a molecular weight marker.Proteins were transferred to Immobilon-FL polyvinylidene difluoride (PVDF) Membrane, 0.45 μm pore size (Sigma-Aldrich) after the gels were soaked in 1x Transfer Buffer (25 mM Tris-Base, 192 mM glycine, 20% v/v methanolall from Sigma-Aldrich) for 10 min.In the case where whole-cell lysates were used, Revert™ 700 Total Protein Stain (LI-COR Biosciences) was used for quantification analyses.The membranes were imaged on Odyssey® Fc with the Image Studio Lite version 5.2.5 software (LI-COR Biosciences).No total protein stain was performed for subcellular enrichment experiments.The membranes were subsequently blocked in Intercept® (TBS) Blocking Buffer for 2 h in room temperature under gentle agitation, before being incubated overnight at 4 • C with primary antibodies against CYP24A1 (anti--CYP24A1, 1:500, NBP1-85495, Novus Biologicals); PDIA3 (anti-PDIA3, 1:500, VMA00477, Bio-Rad Laboratories Inc.); VDR (anti-VDR/NR1I1/vitamin D receptor, 1:1000, PP-H4537-00, R&D Systems); anti-c-Jun (anti-Jun, 1:1000, 610326, BD Biosciences); anti-c-Myc (recombinant anti-c-Myc [Y69], 1:1000, ab32072, Abcam).The enrichment for the cytoplasmic and nuclear protein pools was evaluated by the detection of α-tubulin (anti-alpha-tubulin antibody, 1:5000, ab7291, Abcam) and lamin A/C (anti-lamin A/C antibody, 1:1000, #2032, Cell Signaling) as markers for the cytoplasm and the nucleus, respectively.The membranes were washed 3 ×10 min with 1x TBS-T before incubation with fluorophore-labelled secondary antibodies; Donkey anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 680 (A10038, Thermo Scientific); IRDye® 800CW Donkey anti-Rabbit IgG Secondary Antibody (926-32213, LI-COR Biosciences), followed by 3 ×10 min wash with 1x TBS-T, before imaging on Odys-sey® Fc with the Image Studio Lite version 5.2.5 software.

Real-time PCR
PC3 or U2OS cells were seeded onto 12-well plates (Sarstedt) at a density of 100,000 cells/mL, transfected with 'siCtrl' or 'siPDIA3' and treated with 1,25(OH) 2 D 3 (or vehicle control), as described above.Total mRNA was extracted using RNeasy® Mini Kit (QIAGEN) according to the manufacturer's instructions.mRNA concentration was measured using a ND-1000 spectrophotometer (Saveen & Werner AB). cDNA synthesis was performed with qPCRBIO cDNA Synthesis Kit (PCR Biosystems Ltd.), according to the manufacturer's instructions.Real-time PCR was used to quantify the mRNA levels for CYP24A1, CYP27B1, epidermal growth factor receptor (EGFR), FGF23, c-Fos, c-Jun, JunB, Klotho, PDIA3, c-Myc, transforming growth factor β receptor type I (TGFβR-I), TGFβ receptor type II (TGFβR-II), and VDR.β-Actin was used as housekeeping control.Primer pairs for CYP27B1, FGF23, PDIA3 (1,25D3-MARRS) and VDR were designed with Primer-BLAST, using National Center for Biotechnology Information (NCBI) sequence identification numbers.Primer pairs for c-Myc, TGFβR-I and TGFβR-II were kindly provided by Dr. Anders Sundqvist (Uppsala University) Additional primer pairs were those used in previously published studies: CYP24A1 [45], Klotho [9], EGFR [38], c-Jun [38], JunB [38], c-Fos [38].All primer pairs were purchased from Integrated DNA Technologies, IDT (for sequences, see Table 1).Real-time PCR was performed using the 2x qPCRBIO SyGreen Mix with Fluorescein (PCR Biosystems Ltd.) according to the manufacturer's instructions in the CFX Connect™ Real-Time System (Bio-Rad Laboratories Inc.).Quantification was performed based on the negative difference of the quantitative cycle (Ct) to the control ('siCtrl EtOH treated') normalised to the Ct for the housekeeping gene in the respective sample, raised to the power of 2 (2 -ΔΔCt ).

Cellular thermal shift assay (CETSA)
Potential binding or interaction between 1,25(OH) 2 D 3 and PDIA3 was evaluated using cellular thermal shift assay (CETSA).CETSA is based on the premise that a binding event or an interaction between a protein and a ligand has a stabilising effect on the protein, which remains folded and soluble during a heat challenge [17,34].PC3 cells were treated with 1,25(OH) 2 D 3 (or vehicle control) for 2 h, as described above.At the end of the treatment, 22,000,000 PBS-washed cells were harvested and pelleted by centrifugation at 300 g for 3 min at 4 • C, then resuspended in ice-cold 1x PBS followed by centrifugation at 300 g for 3 minutes at 4 • C. The cells were then resuspended in 750 μL 1x PBS, supplemented with 1x PhosSTOP™ and 1x cOmplete™ (Sigma-Aldrich) and further divided into 100 μL aliquots, followed by heating for 3 min at the following temperatures: 37, 40, 43, 46, 49 and 52 • C, using a Veriti thermal cycler (Thermo Scientific).After the heat treatment, the cells were incubated for 3 minutes at room temperature.To lyse the cells, they were then snap-frozen and subjected to three freeze-thaw cycles using liquid nitrogen and a heating block (Thermo Scientific) at a temperature of 25 • C. Soluble fractions were collected by centrifugation at 20,000 g for 20 min and denatured at 95 • C for 5 min in 1x NuPAGE™ LDS Sample Buffer, supplemented with 20 mM DTT and was further subjected to separation by SDS-polyacrylamide electrophoresis, electrotransfer, and immunoblotting, as described above.β-Actin (anti-β-actin antibody, 1:1000, ab8227, Abcam) was used as housekeeping control.

Statistical analysis
All data were tested for normality via Shapiro-Wilk test.Statistical variance was analysed using unpaired two-tailed t-test with Welch's correction for cell proliferation and western blotting; unpaired twotailed t-test with Welch's correction or unpaired two-tailed Mann-Whitney test for real-time PCR data, depending on normality; two-tailed Kruskal-Wallis test with Dunn's correction and Brown-Forsythe ANOVA test with Dunnett T3 correction for cell migration data and area under curve (AUC) comparisons, using GraphPad Prism version 9.3.1.

Table 1
Summary of primer pair sets.

Basal protein expression levels and subcellular localisation of PDIA3 and VDR in various cell lines
We tested an array of cell lines of different origin for the presence of PDIA3 and VDR and their localisation in the cytoplasm and/or the nucleus, using subcellular fractionation and Western blotting.All cell lines examined expressed both PDIA3 and VDR (Fig. 1).The data indicate that PDIA3 as well as VDR can be present in both the cytoplasm and the nucleus.In some cell lines, VDR appears to be predominantly nuclear and PDIA3 predominantly cytoplasmic.However, in others we detected bands corresponding to PDIA3 and VDR in both the cytoplasm and the nucleus.

Silencing of PDIA3 affects proliferation and migration in prostate carcinoma cells
In our initial studies we examined effects of PDIA3 silencing and 1,25 (OH) 2 D 3 treatment on proliferation and migration of prostate carcinoma PC3 and DU145 cells, as a means of investigating the potential role of PDIA3 for prostate cancer growth.As described in the Materials and methods section, the reduction of PDIA3 levels achieved was 90% or more and persisted for at least 120 h following siRNA transfection.
Silencing of PDIA3 resulted in significant reduction of proliferation in both PC3 and DU145 cells.The difference was about 20-25% compared to cells transfected with control siRNA ('siCtrl') for both cell lines (Fig. 2, Suppl.Effects of PDIA3 silencing on the migratory potential of PC3 and DU145 cells were studied by scratch-wound assay (Fig. 3).Calculation of area under curve (AUC), correlated to the wound closure rate, suggests an overall faster migration upon PDIA3 silencing in PC3 cells.PDIA3depleted PC3 cells showed significantly faster migration, particularly after 8-10 h, compared to cells treated with control siRNA ('siCtrl').The increase in the rate of migration upon PDIA3 silencing appeared to be reverted by treatment with 1,25(OH) 2 D 3 (Fig. 3a).In cells transfected with control siRNA ('siCtrl'), treatment with 1,25(OH) 2 D 3 did not affect migration.Silencing of PDIA3 in PC3 cells decreased the time for wound closure by approximately 2 h.In DU145 cells, which are much less motile than PC3 cells, no differences in migration could be detected by either silencing of PDIA3 or by treatment with 1,25(OH) 2 D 3 (Fig. 3b).

Effects of PDIA3 silencing on transcription factors involved in cell proliferation
Considering the observed effects by PDIA3 silencing on growth, we analysed the mRNA levels of transcription factors known to be of importance for cellular proliferation and tumour development, c-Jun, JunB, c-Fos and c-Myc, in PC3 cells depleted of PDIA3.Expression of c-Fos was not detectable in these cells.The mRNA levels of c-Jun and JunB were not altered by silencing of PDIA3 (data not shown).For c-Jun, protein levels were also examined, but showed no significant change in PDIA3-depleted cells (data not shown).However, silencing of PDIA3 significantly reduced c-Myc mRNA in PC3 cells, by about 20% (Fig. 4a).Additionally, downregulation was detected in the protein levels of c-Myc, by approximately 30% in PC3 cells depleted of PDIA3 (Fig. 4b).Thus, the decrease in c-Myc mRNA as well as c-Myc protein levels was in the same order of magnitude as the observed decrease in proliferation in PDIA3-depleted PC3 cells.

PDIA3 silencing affects 1,25-dihydroxyvitamin D 3 -mediated regulation of target genes involved in control of active vitamin D levels
As a means of investigating potential crosstalk of PDIA3 with VDRmediated pathways, we examined effects of PDIA3 silencing and 1,25 (OH) 2 D 3 treatment on the mRNA levels of VDR and VDR target genes.The levels of PDIA3 mRNA were examined in parallel to ensure efficient knock-down of PDIA3.Reduction of PDIA3 mRNA by silencing was more than 95% in all experiments (Fig. 5a).
The expression of VDR in prostate cancer PC3 cells was not altered by 1,25(OH) 2 D 3 treatment, however, the VDR levels decreased by approximately 50% upon PDIA3 silencing (Fig. 5b).Potential effects of PDIA3 silencing on VDR were also analysed by Western blotting.We did not detect any significant effects on VDR protein levels by depletion of PDIA3 under the conditions employed (data not shown).
In another set of experiments, we examined the expression levels of CYP24A1 and CYP27B1 in PDIA3-depleted PC3 cells.As described in the Introduction, CYP24A1 and CYP27B1 are important regulators of 1,25 (OH) 2 D 3 levels.Also, these genes are well known targets for VDR [6,42].As expected, treatment with 1,25(OH) 2 D 3 induced a robust upregulation of the expression of CYP24A1.However, upon PDIA3 silencing, although CYP24A1 induction by 1,25(OH) 2 D 3 still took place, it occurred at a much lower degree, only about 40% compared to control ('siCtrl') (Fig. 6a).CYP24A1 protein levels seemed to be affected similarly as the mRNA levels, although our data on CYP24A1 protein did not reach statistical significance, perhaps due to variation between experiments (Fig. 6b).As for CYP27B1 mRNA levels, we observed a trend towards downregulation by 1,25(OH) 2 D 3 treatment in cells transfected with control siRNA ('siCtrl'), albeit not statistically significant.In cells  with depleted PDIA3, the levels of CYP27B1 were significantly reduced compared to control ('siCtrl'), by about 30% (Fig. 6d).Treatment with 1, 25(OH) 2 D 3 appeared to slightly lower CYP27B1 mRNA levels also in PDIA3-depleted cells (cf.'siPDIA3 EtOH' and 'siPDIA3 1,25(OH) 2 -D 3 ').The lowest levels of CYP27B1 mRNA were found in cells depleted of PDIA3 and treated with 1,25(OH) 2 D 3 (Fig. 6d).As part of these experiments, we also attempted to analyse mRNA levels of FGF23 and Klotho, two proteins known to be involved in vitamin D-related regulation and important for calcium and phosphate homoeostasis [6].Expression of FGF23 was below the limit of detection in PC3 cells.Expression of Klotho appeared slightly enhanced by treatment with 1,25(OH) 2 D 3 , but was not influenced by silencing of PDIA3 (data not shown).
In order to assess if the observed effects on VDR, CYP24A1 and CYP27B1 mRNA levels might be specific to the 1,25(OH) 2 D 3 regulatory system or general transcriptional effects, we also measured the mRNA levels of epidermal growth factor receptor (EGFR), transforming growth factor β receptor type I (TGFβR-I) and transforming growth factor β receptor type II (TGFβR-II) in PDIA3-depleted PC3 cells in the absence or presence of 1,25(OH) 2 D 3 .The mRNA levels of these three receptors were, however, not significantly affected neither by PDIA3 silencing nor by treatment with 1,25(OH) 2 D 3 (data not shown).
To study whether the effects on CYP24A1 induction might be specific to prostate cancer PC3 cells or appear also in other cell types, we examined effects of PDIA3 depletion on CYP24A1 mRNA levels in human osteosarcoma U2OS cells in the presence and absence of 1,25 (OH) 2 D 3 .Significant effects by PDIA3 silencing on 1,25(OH) 2 D 3 -mediated CYP24A1 upregulation was found also in U2OS cells, similarly as in PC3 cells.In U2OS cells depleted of PDIA3, the upregulation of CYP24A1 by 1,25(OH) 2 D 3 was about half of that observed in control cells ('siCtrl') (Fig. 7).Basal levels of CYP24A1 mRNA (without 1,25(OH) 2 D 3 treatment) were not significantly altered by PDIA3 depletion in either PC3 or U2OS cells.

Discussion
In the current study we explored effects by silencing of PDIA3 in osteosarcoma and prostate carcinoma cell lines and the potential role(s) of this protein for 1,25(OH) 2 D 3 -dependent cell responses.
A role for PDIA3 in cellular growth has been previously reported in several cell types, but, to our knowledge not previously in prostate cancer cells [32,47].The results of the present study indicate decreased proliferation in PC3 and DU145 cells and increased migration in PC3 cells upon PDIA3 silencing, suggesting a potential role of this protein for growth, invasion and metastasis of at least some types of prostate cancer cells.Thus, PDIA3 or its downstream targets could constitute a potential target for the pharmaceutical treatment of certain kinds of prostate carcinomas.From the current findings, it may be hypothesised that during conditions where cellular levels of PDIA3 are high, proliferation might be favoured over migration, thus promoting signalling aimed  towards more growth and less spreading.Our results in PC3 cells also show significant downregulation of c-Myc, an oncogene associated with aggressive prostate cancer [30], both at the transcriptional and protein level, in cells depleted of PDIA3.The observed decrease of c-Myc protein was of a similar magnitude as the decrease in proliferation.Since the c-Myc transcription factor is important for proliferation, it may be speculated that the effect by PDIA3 on proliferation could be mediated via regulation of c-Myc levels, but other mechanisms cannot be excluded at the present stage.PDIA3 is known for its role as a chaperone and may affect the folding and localisation of many different proteins.Thus, more studies are needed to fully understand the molecular role of PDIA3 for proliferation in prostate cancer PC3 cells.
A potential connection between decreased PDIA3 levels and increased migration should not be unique for prostate cancer PC3 cells, since correlations between low levels of PDIA3 and increased migration are reported also in studies of gastric cancer and acute myeloid leukaemia [14,47].On the other hand, studies on PDIA3 depletion in squamous carcinoma A431 cells indicate that PDIA3 increases migration in this cell type, via a 1,25(OH) 2 D 3 -mediated mechanism [22].Thus, as illustrated by present and previous results, differential responses may occur in different carcinoma cell types, and also in cancer cells originating from the same organ.In squamous carcinoma, where PDIA3 has been proposed as a potential target for anticancer therapy, PDIA3 depletion influences sensitivity to 1,25(OH) 2 D 3 and affects the expression of many genes, including proliferation marker MKI67 and genes involved in cell cycle regulation [22].
In many studies, although not all, high PDIA3 expression in various carcinoma models has been associated with increased malignancy [14,28,31,40].Considering the proposed role for PDIA3 in 1,25(OH) 2 D 3 signalling, it should be kept in mind that effects of 1,25(OH) 2 D 3 in general are anti-proliferative [20,32,39].The effects of PDIA3 silencing on proliferation observed in our study was not influenced by treatment with 1,25(OH) 2 D 3 , indicating that these effects may be independent of 1,25(OH) 2 D 3 -mediated signalling.
In addition to the bioactivation of vitamin D into 1,25(OH) 2 D 3 , alternative metabolic pathways, initiated by the steroidogenic enzyme CYP11A1 and leading to several different 20-and 22-hydroxylated vitamin D metabolites, have been described [21,37].These metabolites are noncalcaemic but have antiproliferative and other properties and are reported to bind several nuclear receptors including VDR [29,[35][36][37].To our knowledge there are, however, so far, no studies reporting binding to PDIA3.Wasiewicz et al. [44] observed suppression of proliferation by the vitamin D analogue 21(OH)pD in melanoma SK-MEL-188b cells, which express PDIA3, but lack both VDR and CYP27B1, indicating that vitamin D-related signalling that control growth can involve different pathways.
To investigate the role(s) of PDIA3 in 1,25(OH) 2 D 3 -dependent signalling, we first analysed the effect of PDIA3 silencing on the classical vitamin D receptor VDR, a well-known mediator of vitamin D-related cellular effects and known as a target for 1,25(OH) 2 D 3 in several cell types [33].In the PC3 cells used in our study, mRNA levels for VDR were not affected by 1,25(OH) 2 D 3 treatment, at least not under the conditions employed.However, VDR mRNA levels in this cell type was markedly suppressed by silencing of PDIA3.Having observed this effect on VDR mRNA, we proceeded with looking into the expression levels of CYP24A1 and CYP27B1 in PDIA3-depleted PC3 cells, with and without 1,25(OH) 2 D 3 treatment.As outlined in the introduction, CYP24A1 and CYP27B1 are important for regulation of the physiological levels of the active vitamin D hormone (1,25(OH) 2 D 3 ).These genes are also well-known targets of VDR-mediated regulation.The effects on CYP24A1 and CYP27B1 by treatment with 1,25(OH) 2 D 3 were in line with previous concepts, showing decreased CYP27B1 levels and increased CYP24A1 mRNA levels in the presence of 1,25(OH) 2 D 3 .Interestingly, significantly altered responses to 1,25(OH) 2 D 3 were observed in these genes upon PDIA3 silencing.In PC3 cells depleted of PDIA3, the 1,25(OH) 2 D 3 -dependent induction of CYP24A1, important for removing excessive cellular levels of 1,25(OH) 2 D 3 , was substantially reduced.In addition, PC3 cells depleted of PDIA3 exhibited reduced expression levels of CYP27B1.These data may suggest that PDIA3 depletion decreases intracellular formation of 1,25(OH) 2 D 3 as well as CYP24A1-mediated elimination of 1,25(OH) 2 D 3 in PC3 cells.However, in prostate cells, uptake of active 1,25(OH) 2 D 3 , which is secreted from the kidneys into the circulation, may predominate over intracellular formation by CYP27B1 [39].CYP24A1, on the other hand, is considered of importance for catabolism and elimination of 1,25(OH) 2 D 3 in many cell types [10,21,39,6].Similar effects on 1,25(OH) 2 D 3 -dependent induction of CYP24A1 mRNA by PDIA3 depletion as in PC3 cells was also found in osteosarcoma U2OS cells, showing that the observed phenomenon is not unique for PC3 cells.To our knowledge, this study is the first to report a connection between PDIA3 and regulation of intracellular 1,25(OH) 2 D 3 levels.Taken together, the data indicate that elimination of 1,25(OH) 2 D 3 by CYP24A1, which is a response to rising cellular 1,25(OH) 2 D 3 levels, may be impaired in the absence of PDIA3.These findings indicate that PDIA3 may play a role for regulation and maintenance of adequate cellular 1,25(OH) 2 D 3 levels via effects on its metabolism.Also, our data suggest a potential link between signalling involving PDIA3 and signalling involving VDR.Considering the results obtained in PC3 and U2OS cells, it seems likely that the role of PDIA3 in vitamin D-related signalling is not limited to non-genomic effects but may also affect transcriptional events in these cell types.As mentioned above, effects of PDIA3 depletion on gene transcription have also been shown in squamous carcinoma cells [22].
As shown in the current study, as well as in previous studies, PDIA3 may be located in different subcellular compartments, where it assumes different functions [28,41,5,8].Localisation might vary between cell types, even ones with common tissue of origin.PDIA3 can be associated to biological membranes, such as the ER or the plasma membrane, and has also been shown to possess a nuclear localisation signal (NLS) in its structure [5].This indicates that PDIA3 may enter the nucleus where it might bind directly to DNA or function as a transcription factor co-activator [1,5].
CYP24A1 is able to metabolise not only 1,25(OH) 2 D 3 but also its precursor 25-hydroxyvitamin D 3 .Interestingly, Karlsson et al. [12] reported that inhibition of CYP24A1 and CYP27B1 in prostate cancer LNCaP cells potentiates the antiproliferative effect of 25-hydroxyvitamin D 3 .These authors proposed that reduced levels of CYP24A1 and CYP27B1 may be beneficial in treatment of prostate cancer.
Several published studies have supported the notion that PDIA3 might be important for 1,25(OH) 2 D 3 signalling [16,18,19,48].However, there are also contradictory studies, suggesting that PDIA3 might be of less importance and there is no clear evidence supporting high affinity binding of 1,25(OH) 2 D 3 to PDIA3 [16].In fact, it has been argued that  only VDR is able to bind 1,25(OH) 2 D 3 with high affinity [43,48].Whether VDR is the most important receptor for 1,25(OH) 2 D 3 or not, findings that indicate a role for PDIA3 in 1,25(OH) 2 D 3 -dependent events have been reported in cell types of different origin, including neurons, chondrocytes and skin cells [22,46,48].It has also been proposed that cooperation of VDR and PDIA3 may be essential for some 1,25 (OH) 2 D 3 -induced membrane responses [23].
In the present investigation, we used CETSA to study the potential connections between PDIA3 and 1,25(OH) 2 D 3 .CETSA is based on the premise that a binding event or an interaction between a protein and a ligand has a stabilising effect on the protein, keeping it folded and soluble during a heat challenge [17,34].Experiments using CETSA indicated increased thermostability of the PDIA3 protein in the presence of 1,25(OH) 2 D 3 .It is possible that this effect could be due to a direct interaction between PDIA3 and 1,25(OH) 2 D 3 .However, a perhaps more probable explanation for the results obtained is that PDIA3 might interact with proteins linked to one or more 1,25(OH) 2 D 3 -dependent signal transduction pathway(s).It is also possible that PDIA3 could be linked to 1,25(OH) 2 D 3 not directly but via proteins in the same regulatory complex.If so, such a complex may or may not include VDR.In the current study we found significantly decreased VDR mRNA levels in PDIA3-depleted cells although no measurable effect was detected on VDR protein levels.We also found that target proteins of VDR were affected by PDIA3 knock down.It is possible that a potentially small difference in VDR protein levels might not be detectable with Western blotting.However, the observed effect by PDIA3 depletion on CYP24A1 and CYP27B1 may also result from pathways independent of VDR.From the data available, PDIA3 and VDR may be perceived as acting in different cellular compartments since PDIA3 predominates in the cytoplasm, and VDR, at least upon ligand binding, is translocated to the nucleus.These findings may speak against closely related activities for these proteins.Still, it should be noted that for some of the cell lines examined in the present study, the levels of PDIA3 in the nuclear fraction are significant.Also, our data showing effects on gene transcription by PDIA3 knock down would seem to indicate that PDIA3 may play a role for actions in the nucleus.
At present, we cannot provide a detailed mechanistic explanation of the connections between PDIA3 and 1,25(OH) 2 D 3 signalling.Nevertheless, the results of our investigation strongly support a functional interaction between PDIA3 and some 1,25(OH) 2 D 3 -dependent cellular pathways.Thus, the action of PDIA3 clearly affects the vitamin D system.
In summary, the present results show that PDIA3 silencing interferes with the CYP-mediated regulation that is essential for control of the levels of 1,25(OH) 2 D 3. Furthermore, our study strongly supports the notion that actions of 1,25(OH) 2 D 3 and PDIA3 are linked to each other.Some of the effects of PDIA3 silencing observed in our study appeared dependent on and some independent of 1,25(OH) 2 D 3 .The results obtained support and extend the concept of an important role for PDIA3 in proliferation and migration.The previously unknown function of PDIA3 in control of adequate cellular 1,25(OH) 2 D 3 levels may play an important role for the many physiological actions of 1,25(OH) 2 D 3 , including those associated with proliferation and tumorigenicity.

Declaration of Competing Interest
The authors declare no conflict of interest.

Fig. 1 .
Fig. 1.Basal protein expression levels and subcellular localisation of PDIA3 and VDR in various cell lines.Cell lines PC3, DU145, and LNCaP (all prostate carcinoma), U2OS (osteosarcoma), HaCaT (immortalised keratinocytes), HepG2 (hepatocellular carcinoma), HEK293 (embryonal kidney) and SK-OV-3 (ovarian adenocarcinoma) were seeded on 60 cm 2 plates, grown to subconfluence and lysed as described in Materials and methods.A total protein content of 40 and 20 μg for the cytoplasmic ('cyt') and the nuclear ('nuc') protein pools, respectively, was subjected to SDS-polyacrylamide electrophoresis and analysed by Western blotting using antibodies against PDIA3 or VDR.Lamin A/C and α-tubulin were used as enrichment markers.Representative pictures are shown.Data were collected from at least 3 independent experiments.

Fig. 2 .
Fig. 2. Effect of PDIA3 silencing and 1,25-dihydroxyvitamin D 3 treatment on cell proliferation in PC3 (a) and DU145 (b) prostate cancer cell lines.Cells were depleted of PDIA3 and treated with 10 nM 1,25-dihydroxyvitamin D 3 for 24 h ('1,25(OH) 2 -D 3 ') or vehicle control ('EtOH') and 10 nM EdU for 16 h, as described in Materials and methods.Nuclei, proliferating cells and PDIA3 are shown in blue, green, and magenta respectively.Scale bars = 10 μm.Quantification of cell proliferation is shown as percentage of proliferating cells (EdU-positive) compared to the total number of cells per image.Representative images are shown.Data were collected from 4 independent experiments.Error bars represent SEM.p ⩽ 0.05 (*), p ⩽ 0.01 (**).

Fig. 3 .
Fig. 3. Effect of PDIA3 silencing and 1,25-dihydroxyvitamin D 3 treatment on cell migration in PC3 (a) and DU145 (b) prostate cancer cell lines Cells were depleted of PDIA3 prior to scratch-wound assay and then treated with 10 nM 1,25-dihydroxyvitamin D 3 ('1,25(OH) 2 -D 3 ') or vehicle control ('EtOH') for 24 h, as described in Materials and methods.Quantification of the cell migration is shown as percentage of the initial wound distance covered over time in PC3 cells (a) and DU145 cells (b).All distances were normalised to the initial length of the wound.Area under curve (AUC) measurements represent variation in cell migration between conditions, where lower AUC values correspond to increased rate of migration.A representative micrograph showing wound closure in PC3 cells is shown.Scale bars = 1 μm.Data were collected from 4 independent experiments.Error bars represent SEM.p ⩽ 0.05 (*), p ⩽ 0.001 (***), p ⩽ 0.0001 (****).

Fig. 4 .
Fig. 4. Effect of PDIA3 silencing and 1,25-dihydroxyvitamin D 3 treatment on c-Myc mRNA levels (a) and c-Myc protein levels (b) in PC3 cells.PC3 cells were treated with 10 nM siRNA against PDIA3 ('siPDIA3') or control siRNA ('siCtrl') for 72 h, prior to treatment with 10 nM 1,25-dihydroxyvitamin D 3 ('1,25(OH) 2 -D 3 ') or vehicle control ('EtOH') for 24 h, as described in Materials and methods.(a) c-Myc mRNA levels were assessed with real-time PCR and analysed based on the 2 -ΔΔCt method.Fold differences are shown relative to 'siCtrl EtOH'.(b) Whole-cell lysates were prepared as described in Materials and methods and analysed by Western blotting using antibodies against c-Myc.Band intensities were normalised against total protein stain and fold differences are shown relative to 'siCtrl EtOH'.Representative pictures are shown.Data were collected from at least 3 independent experiments.Error bars represent SEM.p ⩽ 0.05 (*).

Fig. 6 .
Fig. 6.Effect of PDIA3 silencing and 1,25-dihydroxyvitamin D 3 treatment on CYP24A1 mRNA levels (a), CYP24A1 protein levels (b), PDIA3 protein levels (c) and CYP27B1 mRNA levels (d) in PC3 cells.PC3 cells were treated with 10 nM siRNA against PDIA3 ('siPDIA3') or control siRNA ('siCtrl') for 72 h, prior to treatment with 10 nM 1,25-dihydroxyvitamin D 3 ('1,25(OH) 2 -D 3 ') or vehicle control ('EtOH') for 24 h, as described in Materials and methods.(a, d) All mRNA levels were assessed with real-time PCR and analysed based on the 2 -ΔΔCt method.Fold differences are shown relative to 'siCtrl EtOH'.(b) CYP24A1 protein levels in whole-cell lysates were examined by Western blotting as described in Materials and methods.(c) Protein levels of PDIA3 were examined in parallel, in order to ensure silencing efficiency.Band intensities were normalised against total protein stain and fold differences are shown relative to 'siCtrl EtOH'.Representative pictures are shown.Data were collected from 4 independent experiments.Error bars represent SEM.p ⩽ 0.05 (*), p ⩽ 0.01 (**).