PDIA3 modulates genomic response to 1,25-dihydroxyvitamin D3 in squamous cell carcinoma of the skin

An active form of vitamin D3 (1,25-dihydroxyvitamin D3) acts through vitamin D receptor (VDR) initiating genomic response, but several studies described also non-genomic actions of 1,25-dihydroxyvitamin D3, implying the role of PDIA3 in the process. PDIA3 is a membrane-associated disulfide isomerase involved in disulfide bond formation, protein folding, and remodeling. Here, we used a transcriptome-based approach to identify changes in expression profiles in PDIA3-deficient squamous cell carcinoma line A431 after 1,25-dihydroxyvitamin D3 treatment. PDIA3 knockout led to changes in the expression of more than 2000 genes and modulated proliferation, cell cycle, and mobility of cells; suggesting an important regulatory role of PDIA3. PDIA3-deficient cells showed increased sensitivity to 1,25-dihydroxyvitamin D3, which led to decrease migration. 1,25-dihydroxyvitamin D3 treatment altered also genes expression profile of A431ΔPDIA3 in comparison to A431WT cells, indicating the existence of PDIA3-dependent genes. Interestingly, classic targets of VDR, including CAMP (Cathelicidin Antimicrobial Peptide), TRPV6 (Transient Receptor Potential Cation Channel Subfamily V Member 6), were regulated differently by 1,25-dihydroxyvitamin D3, in A431ΔPDIA3. Deletion of PDIA3 impaired 1,25-dihydroxyvitamin D3-response of genes, such as PTGS2, MMP12, and FOCAD, which were identified as PDIA3-dependent. Additionally, response to 1,25-dihydroxyvitamin D3 in cancerous A431 cells differed from immortalized HaCaT keratinocytes, used as non-cancerous control. Finally, silencing of PDIA3 and 1,25-dihydroxyvitamin D3, at least partially reverse the expression of cancer-related genes in A431 cells, thus targeting PDIA3 and use of 1,25-dihydroxyvitamin D3 could be considered in a prevention and therapy of the skin cancer. Taken together, PDIA3 has a strong impact on gene expression and physiology, including genomic response to 1,25-dihydroxyvitamin D3.


PDIA3 1,25-dihydroxyvitaminvitamin D3 Squamous cell carcinoma Transcriptome
A B S T R A C T An active form of vitamin D 3 (1,25-dihydroxyvitamin D 3 ) acts through vitamin D receptor (VDR) initiating genomic response, but several studies described also non-genomic actions of 1,25-dihydroxyvitamin D 3 , implying the role of PDIA3 in the process.PDIA3 is a membrane-associated disulfide isomerase involved in disulfide bond formation, protein folding, and remodeling.Here, we used a transcriptome-based approach to identify changes in expression profiles in PDIA3-deficient squamous cell carcinoma line A431 after 1,25-dihydroxyvitamin D 3 treatment.PDIA3 knockout led to changes in the expression of more than 2000 genes and modulated proliferation, cell cycle, and mobility of cells; suggesting an important regulatory role of PDIA3.PDIA3-deficient cells showed increased sensitivity to 1,25-dihydroxyvitamin D 3 , which led to decrease migration.1,25-dihydroxyvitamin D 3 treatment altered also genes expression profile of A431ΔPDIA3 in comparison to A431WT cells, indicating the existence of PDIA3-dependent genes.Interestingly, classic targets of VDR, including CAMP (Cathelicidin Antimicrobial Peptide), TRPV6 (Transient Receptor Potential Cation Channel Subfamily V Member 6), were regulated differently by 1,25-dihydroxyvitamin D 3 , in A431ΔPDIA3.Deletion of PDIA3 impaired 1,25dihydroxyvitamin D 3 -response of genes, such as PTGS2, MMP12, and FOCAD, which were identified as PDIA3dependent.Additionally, response to 1,25-dihydroxyvitamin D 3 in cancerous A431 cells differed from immortalized HaCaT keratinocytes, used as non-cancerous control.Finally, silencing of PDIA3 and 1,25-dihydroxyvitamin D 3 , at least partially reverse the expression of cancer-related genes in A431 cells, thus targeting PDIA3 and use of 1,25-dihydroxyvitamin D 3 could be considered in a prevention and therapy of the skin cancer.Taken together, PDIA3 has a strong impact on gene expression and physiology, including genomic response to 1,25dihydroxyvitamin D 3 .

Introduction
Protein Disulfide Isomerase Family A Member 3 (PDIA3), also known as ERp57 or GRP58, belongs to the oxidoreductase enzyme family (PDIs), which have thioredoxin-like (TRX-like) domains containing active sites [1].PDIA3, among other proteins from the PDIs family, is involved in multiple processes, including: protein folding, disulfide bond formation, and remodeling [2].PDIA3 is mainly localized in the endoplasmic reticulum (ER), but may also be found in different locations such as the nucleus or cell membrane [3].PDIA3 has binding sites for other ER chaperones like calreticulin and calnexin to form complexes that are necessary for its catalytic redox activity [4].Moreover, it plays a role in the activation of PERK in response to misfolded protein, leading to unfolded protein response [5].PDIA3 was also identified as an alternative membrane-associated receptor for 1,25-dihydroxyvitamin D 3 (and named membrane-associated, rapid response steroid-binding protein − 1,25D3-MARRS) [6].PDIA3 can form a complex with caveolin-1, activating downstream mediators like phospholipase A2-activating protein (PLAA).PDIA3/PLAA complex acts on phospholipase A2 (PLA2) together with Ca 2+ /calmodulin-dependent protein kinase II (CaMKII), leading to the activation of protein kinase C (PKC).CaMKIIα is required not only for mediating the rapid actions of 1α,25(OH) 2 D 3 from PLAA to PLA2 but also for mediating the Ca 2+ -dependent actions [7,8].
In recent years the involvement of PDIA3 protein in many pathological processes such as neurodegenerative diseases [9,10], cardiac diseases, and cancer [11,12] has been shown.PDIA3 is known to be dysregulated in various types of cancer and its expression levels can serve as a prognostic marker.Depending on the affected tissue either high or low expression levels can correlate with poor prognosis.The knockout of PDIA3 in mice is lethal, which emphasizes the importance of PDIA3 protein in cell functioning [13,14].Knockdown or overexpression of PDIA3 has a broad range of physiological effects such as apoptosis, proliferation, and motility [15][16][17].
Vitamin D directly or indirectly modulates the expression of around 3000 genes in the human genome including genes involved in cell differentiation, proliferation, and migration [18,19].A broad range of studies has shown the anticancer effects of an active form of vitamin D in various types of cancers e.g. head and neck skin carcinoma, melanoma, or breast cancer [20][21][22].Genomic activity of 1,25-dihydroxyvitamin D 3 is mediated through its nuclear receptor VDR (Vitamin D receptor), but numerous studies also describe the rapid, non-genomic activities, which involves PDIA3 protein as a membrane receptor for 1,25-dihydroxyvitamin D 3 [14,23,24].
In this study, the effects of 1,25-dihydroxyvitamin D 3 on biological features and gene expression in squamous cell carcinoma (SCC) cells line A431, in the presence or absence of PDIA3 protein was investigated.Bioinformatics was applied in order to compare transcriptomic profiles of A431 cell lines of immortalized human keratinocyte cell line (HaCaT) and head and neck squamous cell carcinoma (HNSCC) gene expression pattern available in a public database (TACCO) [25,26].

Deletion of PDIA3 in A431 cells affects response to 1,25-dihydroxyvitamin D 3 treatment
It is well established that deletion of PDIA3 is lethal in mice, however, we were able to establish human knockout of PDIA3 in human squamous cell carcinoma cell line A431 (A431ΔPDIA3) using gene edition technology (CRISP/Cas9).An analogous procedure was used to establish a stable cell line with impaired vitamin D receptor (VDR) expression (A431ΔVDR, [27]), which was used as an additional control.In order to investigate the impact of PDIA3 knockout on proliferation of A431 cells treated with an active form of vitamin D 3 , an SRB assay was performed.1,25-dihydroxyvitamin D 3 at 100 nM concentration inhibited proliferation of A431 WT cells maximally about 18% with half maximal effective concentration EC 50 = 39,355 nM after 24 h of incubation (Fig. 1A).For both A431 knockouts A431 ΔVDR, and A431 ΔPDIA3, only 10% of inhibition was observed.However, deletions of VDR or PDIA3 sensitized cells to 1,25-dihydroxyvitamin D 3 , as judged from the decrease in EC 50 values (1,346 and 0,019 nM, respectively) (Fig. 1B, C).Interestingly, under given conditions prolonged incubation (72 h) effect of 100 nM 1,25-dihydroxyvitamin D 3 on proliferation was diminished (not shown).In previous studies, it was shown that in HaCaT cells 1,25-dihydroxyvitamin D 3 treatment reduced proliferation rate around 20% with EC 50 = 0,089 nM [28].It was suggested that A431 cells are more resistant to 1,25-dihydroxyvitamin D 3 treatment.It was shown that 1,25-dihydroxyvitamin D 3 affected the distribution of A431 cells in different phases of the cell cycle, however, the effects were not statistically significant after incubation for only 24 h (Fig. 1D).Prolonged treatment of A431 cells for 72 h, significantly decreased the number of cells in the SubG1 and G1 phases of the cell cycle, regardless presence or absence of deletion of VDR or PDIA3 gene.Moreover, 1,25dihydroxyvitamin D 3 triggered cell cycle arrest in the S phase, and the result was the most prominent in A431 ΔVDR cells (Fig. 1E).In general, prolonged incubation with 1,25-dihydroxyvitamin D 3 resulted in limited accumulation of cells in subG1 representing dying cells and the effect was more pronounced in ΔPDIA3 cells (Fig. 1F).Next, to assess the impact of PDIA3 deletion on 1,25-dihydroxyvitamin D 3 biological activity, expression of genes associated with proliferation and cell cycle, was analyzed by qPCR.Expression of proliferation marker MKI67 was increased in VDR knockout cells, while in A431 PDIA3 cells, was greatly reduced after 1,25-dihydroxyvitamin D 3 treatment (Fig. 1G).The expression of cycle inhibitors, Cyclin Dependent Kinase Inhibitor 2A (CDKN2A) and Cyclin Dependent Kinase Inhibitor 1A (CDKN1A), was increased significantly in A431 WT treated with 1,25-dihydroxyvitamin D 3 .Deletion of VDR moderately increased expression of CDKN2A, but not CDKN1A¸ while the effect of 1,25-dihydroxyvitamin D 3 treatment was not observed (Fig. 1H).In PDIA3 deficient A431 cells no regulation of CDKN2A was observed after 1,25-dihydroxyvitamin D 3 treatment.Deletion of ΔPDIA3 effectively upregulated an expression of CDKN1A, but the effect was attenuated by 1,25-dihydroxyvitamin D 3 treatment (Fig. 1I).To study the physiological effects of PDIA3 deletion on motility and migration of cells, a wound closure assay was recorded live, with use of Olympus cellVivo IX83 for 72 h.In A431 WT cells, a full closure of the artificial wound was achieved after 50 h.In agreement with other studies [21] 1,25-dihydroxyvitamin D 3 inhibited cell motility, as full wound closure was not achieved at the end of the experiment (72 h).Interestingly, the effect of 1,25-dihydroxyvitamin D 3 was abolished in A431ΔVDR cells, but not in A431ΔPDIA3 cells.Surprisingly, the mobility of A431ΔPDIA3 cells was greatly reduced by PDIA3 deletion, with a further decrease observed after 1,25-dihydroxyvitamin D treatment (Fig. 1J, K).

PDIA3 knockout largely affects the transcriptome of A431 squamous cell carcinoma cell line
Previously, we have described the effects of the knockout of vitamin D receptor -VDR or its co-receptor RXRA on 1,25-dihydroxyvitamin D transcriptome in A431 SSC cells [27].
To establish the changes in expression profile after PDIA3 knockout, differentially expressed genes (DEGs) from A431WT cells and A431ΔPDIA3 were compared.Unexpectedly, the deletion of PDIA3 in A431 SCC cell line affected the expression of 2184 genes, of which were upregulated and 1157 were downregulated.Out of 1866 detected DEGs, 1766 belonged to well-characterized coding genes (890 upregulated and 976 downregulated), 66 represented long intergenic nonprotein coding RNA (LIN_ID) with 21 upregulated and 45 downregulated, and 252 were categorized as uncharacterized novel transcripts (116 upregulated and 136 downregulated) (Fig. 2A).None of the genes of uncertain functions were affected by PDIA3 deletion.All DEGs in A431ΔPDIA3 cells with significantly altered expression, in comparison to A431WT cells, were shown as a heatmap (Supplementary Fig. 2A).Volcano plot demonstrated DEGs with the most significant fold change and top ten (TOP10) upregulated and downregulated DEGs were listed in Fig. 2B.Gene ontology (GO) analysis showed significant enrichment of DEGs associated with cellular calcium-phosphorus signaling (bone mineralization, positive regulation of phosphorylation) in terms of biological processes and regarding to GO molecular functions: phospholipase C activity, MHC class II receptor activity, calcium-dependent phospholipid binding were considerably enriched (Fig. 2C).
Interestingly, detailed analyses of DEGs revealed that deletion of PDIA3 affects the expression of several classic targets for 1,25-dihydroxyvitamin D 3 transcriptional activity driven by VDR, including Cytochrome P450 Family 24 Subfamily A Member 1 (CYP24A1) [29], even in absence of the ligand.In addition, in A431ΔPDIA3 an increased expression of Cathelicidin Antimicrobial Peptide (CAMP) and a decrease of Transient Receptor Potential Cation Channel Subfamily V Member (TRPV6) were observed (Fig. 2D).
STRING online database [30] was used to perform network analysis of protein-protein interaction (PPI).Data revealed a dense interaction cluster between members of the Major Histocompatibility Complex Family and the second one between members of the Cytochrome P450 Family.Moreover, PPI analysis revealed interactions between proteins connected to cell adhesion (Fig. 2E).

Deletion of PDIA3 gene strongly influences response of A431 SCC cells to 1,25-dihydroxyvitamin D 3
Next, the effects of 1,25-dihydroxyvitamin D 3 at 100 nM concentration on the transcriptome of A431ΔPDIA3 were studied after 24 h incubation.The analysis revealed the presence of 704 DEGs of which 374 genes were upregulated and 331 downregulated.Among 704 DEGs, 532 were identified as genes of known function (295 upregulated and 275 downregulated); and 39 represented long intergenic non-protein coding RNA (LIN_ID), with 25 upregulated and 14 downregulated.Finally, among DEGs, we detected 135 uncharacterized novel transcripts (79 upregulated and 56 downregulated).Contrary to A431WT cells [27] none of the genes of uncertain function were identified (LOC_ID) (Fig. 3A).Lists of the TOP10 downregulated and upregulated DEGs found in A431ΔPDIA3 after incubation with 1,25-dihydroxyvitamin D for 24 h were shown on the Volcano plot (Fig. 3B).Gene Ontology analysis revealed significant enrichment of DEGs in A431ΔPDIA3 cells after 24 h treatment with 1,25-dihydroxyvitamin D 3 in terms of biological processes (phosphorus metabolic processes, MAPK cascade, and positive regulation of the apoptotic process) and in terms of molecular functions categories (identical protein binding, protein kinase binding, and calcium ion binding) (Fig. 3C, D).Venn analysis (Fig. 3E, Supplementary Table S2) showed that knockout of PDIA3 alone affected coding genes (lncRNA and uncharacterized genes were excluded), of which 1611 were not affected by 1,25-dihydroxyvitamin D 3 .Among 1,25-dihydroxyvitamin D 3 dependent DEGs detected in both A431WT and A431ΔPDIA3, 59 were solely found in A431 WT cells, which suggested that the regulatory effect of 1,25-dihydroxyvitamin D 3 depended on PDIA3 expression.On the other hand, 191 DEGs were unique for A431ΔPDIA3 cells, while 168 genes were commonly regulated by 1,25dihydroxyvitamin D 3 in A431WT and A431ΔPDIA3, thus their expression was not affected by PDIA3 presence.In A431ΔPDIA3 a group of DEGs were identified as 1,25-dihydroxyvitamin D 3 -dependent and PDIA3-regulated, while 109 DEGs were affected by 1,25-dihydroxyvitamin D 3 in both cell lines, but also deletion of PDIA3 influenced their expression.Finally, 58 DEGs, which were found in A431ΔPDIA3 cells were also modulated by 1,25-dihydroxyvitamin D 3 , but only in A431WT cells.Overall, amongst 117 PDIA3-dependent DEGs detected in A431WT cells, which were 1,25-dihydroxyvitamin D 3 -dependent, 98% were found to be also VDR-dependent, while around 80% RXR-dependent [27].PPI network analysis revealed three major clusters of A431ΔPDIA3 DEGs: the first one among positive regulators of the cellular biosynthetic process including Runt-related transcription factor 2 (RUNX2), JunB Proto-Oncogene (JUNB), MAF BZIP Transcription Factor (MAF).The second network was detected among the proteins involved in phosphorus metabolic processes including several cytochrome family The list was also supplemented with potential PDIA3-dependent and independent genes.Furthermore, human-immortalized HaCaT keratinocytes, which served as noncancerous control, were included in the study.Interestingly, in HaCaT keratinocytes, treated with 1,25-dihydroxyvitamin D 3 , CYP24A1 induction was significantly higher than in both A431 cell lines (Fig. 4A).The expression of CAMP was elevated after 1,25-dihydroxyvitamin D 3 treatment in all cell lines (Fig. 4B).The expression of another classical target for 1,25-dihydroxyvitamin D 3 -TRPV6, which was upregulated in HaCaT cells after treatment, but this effect was not observed in both A431 sublines.Furthermore, the expression of TRPV6 was almost completely attenuated in A431ΔPDIA3 cells (Fig. 4C).Deletion of PDIA3 increased expression of two genes (CAMP, PTGS2) in A431ΔPDIA3 subline and a decrease of baseline expression of five genes (CYP24A1, TRPV6, MMP12, FOCAD, and ZNF185) in comparison to A431WT.In general, the expression of PTGS2 (Fig. 4D), MMP12 (Fig. 4E), and FOCAD (Fig. 4F) genes was induced by 1,25-dihydroxyvitamin D 3 in HaCaT and A431WT, but not in A431ΔPDIA3.Of note, overexpression of PTGS2 was significantly higher in HaCaT cells treated with 1,25-dihydroxyvitamin D 3 , when compared to A431WT cells (6 vs. 2 folds, respectively).Further, those genes were assessed to be PDIA3dependent.Expression of NKAIN2 (Fig. 4G) and SERPINB7 (Fig. 4H) was upregulated in all cell lines.Interestingly, the expression of ZNF185 (Fig. 4I) was significantly reduced in cancerous cell lines compared to HaCaT cells.Moreover, expression of ZNF185 was affected by 1,25-dihydroxyvitamin D 3 treatment only in HaCaT and A431ΔPDIA3 cells, thus those genes were assessed to be PDIA3-independent.

Deletion of PDIA3 in A431 squamous cell carcinoma affects the expression of common DEGs found in head and neck squamous cell carcinoma
Previously, we reported that treatment of A431 SCC cell line with 1,25-dihydroxyvitamin D 3 , at least partially reverse the expression of signature genes for Head and Neck squamous cell carcinoma (HNSCC) from the TACCO database [27].Thus, here an impact of PDIA3 knockout and 1,25-dihydroxyvitamin D 3 on the expression of the DEGs characteristic for HNSCC were evaluated.As the previous experiment showed changes in the expression of DEGs in A431 cell lines with PDIA3 knockout (Fig. 5A) and after 24 h 1,25-dihydroxyvitamin D 3 treatment (Fig. 5B), the transcriptome of A431WT and A431ΔPDIA3 was compared with publicly available head and neck squamous cell carcinoma database (TACCO database).Analysis revealed that from 1027 DEGs upregulated in A431ΔPDIA3 cell line only 434 were commonly upregulated in HNSCC, while 377 were conversely expressed.Among DEGs that are downregulated after knockout of PDIA3, expression of 270 was increased and 435 decreased, when compared to the HNSCC expression pattern.1,25-dihydroxyvitamin D 3 treatment upregulated 217 DEGs, among which 102 had the same expression pattern, and 115 were conversely expressed after 1,25-dihydroxyvitamin D 3 treatment.In the group of downregulated DEGs, 220 were identified in comparison to the TACCO database; 148 were downregulated and 72 were upregulated.

Discussion
The present study aimed to identify the transcriptional changes in squamous cell carcinoma cells after the knockout of PDIA3 and its impact on response to 1,25-dihydroxyvitamin D 3 .To our knowledge, this is the first transcriptomic-based approach to study the contribution of PDIA3 in the regulation of the genomic activity of 1,25-dihydroxyvitamin D 3 .
PDIA3 is involved in a wide range of physiological processes, such as mediation of protein folding or modulation of store-operated Ca 2+ entry [31,32].Most importantly, it was linked with rapid response to 1,25dihydroxyvitamin D 3 [7,8].Furthermore, dysregulation of PDIA3 is associated with multiple types of cancers [11,33,34], thus our study was focused on squamous cell carcinoma cell line A431.
Here we are presenting data indicating, that deletion of either, the main receptor for vitamin D -VDR or potential modulator -PDIA3 changed the response of A431 cells to 1,25-dihydroxyvitamin D 3 treatment, in terms of proliferation, cell cycle, and migration.This is in agreement with our previous studies showing that silencing of VDR correlates with reduced responsiveness to 1,25-dihydroxyvitamin D 3 in melanoma cells [35].Moreover, others showed that inhibition of VDR expression induced hyperproliferation of cells [36].It is also supported by an impact of deletion of VDR or PDIA3 on the expression of cell cycle inhibitors: CDKN1A and CDKN2A observed here.
Recently, we have shown that deletion of VDR resulted in complete inhibition of the classic genomic pathway after incubation of A431ΔVDR cells with 1,25-dihydroxyvitamin D 3 for 24 h [27], and this results are in an agreement with other studies [37].However, after prolonged incubation (72 h), 20 most downregulated DEGs were detected [27], suggesting activation of the secondary (alternative) pathway(s) activated by 1,25-dihydroxyvitamin D 3 [38,39].Among those 20 genes, two were found to be VDR-independent and PDIA3dependent (Interferon Induced Protein 44 Like (IFI44L) and Epithelial Stromal Interaction 1 (EPSTI1)).It was postulated that PDIA3 is the major element of fast membrane signaling of 1,25-dihydroxyvitamin D 3 [8], activation of which may also affect the gene expression, as a secondary effect.Surprisingly, deletion of PDIA3 in human SCC cell line A431, regardless of the 1,25-dihydroxyvitamin D 3 , changed the expression of more than 2000 genes, and affects cell physiology, including proliferation and cell mobility.This observation underlines the importance of PDIA3 and explains why its deletion in mice models is lethal [40].Interestingly, PDIA3 was found in various intracellular locations, including the nucleus [4,38,41,42], and its activity as a transcription factor was postulated [43,44].Aureli and coworkers suggested that PDIA3 regulates the expression of MSH6, TMEM126A, LRBA, and ETS1 genes [43].However, neither deletion of PDIA3 nor 1,25-dihydroxyvitamin D 3 treatment affected the expression of those genes (and their direct targets) in our cellular model.On the other hand, deletion of PDIA3 was found to impair biological processes, such as bone mineralization, regulation of bone formation, or calcium homeostasis [40].For example, deletion of PDIA3 in primary osteoblast cultures decreased baseline expression of Odd-Skipped Related Transciption Factor 2 (OSR2) genecrucial for secondary palate growth and morphogenesis [45], which was shown to be induced by 1,25-dihydroxyvitamin D 3 [46].Nonetheless, 1,25-dihydroxyvitamin D 3 treatment did not affect OSR2 gene expression in A431 cells.Instead, we observed changes in the expression of genes connected with calcium homeostasis including TRPV6 [47], which levels were significantly reduced in PDIA3-deficient cells.Furthermore, after 1,25-dihydroxyvitamin D 3 treatment of A431ΔPDIA3 cells we observed changes in phosphorus metabolic process (ACKR3, CALM1, CAMK2A), MAPK cascade (MAPK8, MAP2K2, MAPK13) and positive regulation of apoptosis (BCL2L11, BAX, BAD).Taken together, data presented here indicate that PDIA3 is essential for the maintenance of calcium-phosphorus homeostasis, which links its function with vitamin D signaling, but the effect seems to be cell-type specific.It is worth mentioning, that deletion of PDIA3 in A431 cells, resulted in destabilization of cellular homeostasis triggering apoptosis, which additionally was enhanced by 1,25-dihydroxyvitamin D 3 (with up to 10% of apoptotic cells was observed).The latest study in melanoma cells has shown that 1,25-dihydroxyvitaminvitamin D 3 can induce apoptosisrelated proteins such as caspase-3, caspase 8, and caspase-9 [48] although apoptosis seems to be not the mayor mechanism of 1,25-dihydroxyvitamin D 3 activity.
Among 24 classic target genes associated with VDR transcriptional activity, an expression of CAMP and TRPV6 was significantly changed after PDIA3 deletion.Cathelicidin antimicrobial peptide (CAMP) plays a critical role in innate immune response and its expression is regulated by VDR transcription factor in response to 1,25-dihydroxyvitamin D 3 .Consequently, our data indicated strong upregulation of CAMP expression in both, normal keratinocytes (HaCaT) and cancer cells (A431).It is also in agreement with previously published data showing a concentration-dependent increase in CAMP mRNA after 1,25-dihydroxyvitamin D 3 treatment [49,50].TRPV6 is a highly Ca 2+ -selective channel, which plays a role in maintaining Ca 2+ ion homeostasis [51], but also is essential for Ca 2+ -induced differentiation of human keratinocytes [52].Furthermore, it was shown that 1,25-dihydroxyvitamin D 3 treatment in HaCaT cells increases TRPV6 expression, which lines up with the previous study [52].In A431 cells there was no increase in the expression of TRPV6 after 1,25-dihydroxyvitamin D 3 treatment and PDIA3 deletion resulted in almost complete attenuation of its expression.Those results suggest that squamous cell carcinoma cell line A431 is not sensitive to 1,25-dihydroxyvitamin D 3 induced expression of TRPV6 and deletion of PDIA3 deepens that effect.
Further, support of indirect involvement of PDIA3 in 1,25-dihydroxyvitamin D 3 signaling came from the detailed transcriptomic analysis.The deletion of PDIA3 affected the expression of 167 DEGs detected in A431 WT cells, treated with 1,25-dihydroxyvitamin D 3 .In addition, expression of the genes involved in classic 1,25-dihydroxyvitamin D 3 signaling, CYP24A1 was altered (increased) by PDIA3 knockout, but not VDR, while expression of RXRA was decreased.The PDIA3 knockout also significantly changed the top-ranked expressed gene in ΔPDIA3 cells after 1,25-dihydroxyvitamin D 3 treatment in comparison to A431WT [27].An increase in the expression of DEGs connected to cell-matrix adhesion Semaphorin 3E (SEMA3E) and Coronin 1B (CORO1B) after PDIA3 deletion may indicate a potential mechanism of inhibition of mobility of A431ΔPDIA3 cells.Indeed, expression of SEMA3E was inversely correlated with melanoma progression [53].However, we also noted enhanced expression of Vimentin (VIM), which is a marker of epithelial-mesenchymal transition after PDIA3 deletion [54].
Venn analysis of all 1,25-dihydroxyvitamin D 3 regulated DEGs allowed us to distinguished PDIA3-dependent (PTGS2, MMP12, FOCAD) and -independent (NKAIN2, SERPINB7, ZNF185) genes.Cyclooxygenase-2 (COX-2/PTGS2) is a membrane-bound prostaglandin (PG)-endoperoxide synthase 2 enzyme responsible for the generation of prostanoids [55].In our study, 1,25-dihydroxyvitamin D 3 , strongly stimulated the expression PTSG2 gene in HaCaT cells, which confirmed previous reports [56].Shirvani and coworkers showed that PTGS2 is downregulated after 6 months-long vitamin D 3 supplementation [57].Those differences might be explained by different models and experiment set up.In SSC cell line A431WT, the effect 1,25-dihydroxyvitamin D 3 treatment was partially attenuated, but still statistically significant, while in A431 cells with deletion of PDIA3, expression of PTGS2 was not affected by the treatment.Those results suggested that PTGS2 response to 1,25-dihydroxyvitamin D 3 is PDIA3-dependent, and our previous studies also indicated the involvement of RXRA in its expression [27].The MMP12 is a matrix metalloproteinase with a broad range of antitumor effects [58], such as a protective role in tumor progression [59] or generation of anti-angiogenic peptides [60].In our study, we noted elevated expression of MMP12 after 1,25-dihydroxyvitamin D 3 treatment in HaCaT cells and its upregulation was even greater in A431WT cells.However, the deletion of PDIA3 completely abrogated the effects of 1,25-dihydroxyvitamin D 3 on MMP12, indicating the involvement of PDIA3 in its expressional regulation.Interestingly, MMP12 expression was shown to be RXRA-independent [27].Focadhesin (FOCAD) functions as an anti-tumor protein, inhibiting cell proliferation and migration [61].It was shown, that the loss of FOCAD expression correlated with enhanced aggressiveness and worsen clinical outcomes in glioma patients [62].Our study indicates that in cancerous cells (A431WT and A431ΔPDIA3) effect of 1,25-dihydroxyvitamin D 3 treatment on expression of FOCAD was significantly reduced and PDIA3-dependent.Recently, we have shown that its expression is also VDR-and RXRAdependent [27].The Na+/K + transporting ATPase interacting 2 (NKAIN2) is a novel tumor suppressor identified in prostate cancer [63].Our study showed, that 1,25-dihydroxyvitamin D 3 -dependent NKAIN2 expression is reduced in cancerous cells (A431) compared to noncancerous (HaCaT).Moreover, the effect of 1,25-dihydroxyvitamin D 3 on the expression of this gene depends solely on VDR.The expression of SERPINB7, which belong to serine protease inhibitors, was elevated by PDIA3 deletion alone and furtherly stimulated by 1,25-dihydroxyvitamin D 3 .An overexpression of SERPINB7, along with other serpin family members, was associated with suppression of the invasiveness and motility of cancer cells [64].In our previous study, [27] it was shown that 1,25-dihydroxyvitamin D 3 impact on SERPINB7 expression is VDR/ RXRA-dependent and our current study suggested also the involvement of PDIA3.The ZNF185 is a zinc finger protein, highly expressed during keratinocytes differentiation.Interestingly, its downregulation was also described as a potential biomarker of squamous cell carcinoma [65].Our results showed that 1,25-dihydroxyvitamin D 3 upregulates the expression of ZNF185 in HaCaT keratinocytes, but not in SCC line A431.Interestingly, the deletion of PDIA3 reintroduced, to some extent, the effect of 1,25-dihydroxyvitamin D 3 , on ZNF185 expression.Thus, it seems, that PDIA3 in cancerous cells has an inhibitory effect on 1,25dihydroxyvitamin D 3 -induced expression of ZNF185.
In recent years, the PDIA3 protein has emerged as a potential prognostic marker for various types of cancer [34,66,67].Large-scale transcriptomic analysis of Head and Neck Squamous Cell Carcinomas (HNSCC) showed that overexpression of PDIA3 correlates with poor prognosis [68,69].In our study, around 39% of identified DEGs in A431ΔPDIA3 cells (A431ΔPDIA3 vs A431WT), showed the same expression trend as in HNSCC.Data presented here also demonstrated an impaired genomic response to 1,25-dihydroxyvitamin D 3 of cancer cells in comparison to HaCaT keratinocytes.Importantly, PDIA3 knockout radically changed the response of the cells to 1,25-dihydroxyvitamin D treatment, not only on the genomic level but also by sensitizing cells to 1,25-dihydroxyvitaminvitamin D 3 as shown by proliferation assay.Thus, PDIA3 is not only important for normal cellular physiology but also is necessary for the response and biological activities of 1,25-dihydroxyvitamin D 3 .For many years PDIA3 has been considered a membrane receptor for 1,25-dihydroxyvitaminvitamin D 3 [6], however, the nature of the interaction of PDIA3 with 1,25-dihydroxyvitamin D 3 still is under debate [70], because the crystal structure of the complex of PDIA3 with 1,25-dihydroxyvitamin D 3 was not determined, yet.Data presented here, however, open new possibilities.It could be postulated that PDIA3 affects directly or indirectly several intracellular pathways essential for genomic and/or nongenomic activity of vitamin D, modulating cellular response to this powerful secosteroid.Furthermore, it seems that PDIA3 could be also considered as a target for anticancer therapy as well as a modulator of response to 1,25-dihydroxyvitamin D 3 .

Cell cultures
Immortalized human basal cell carcinoma cell line (A431) was obtained from Synthego Corporation (Menlo Park, CA, USA), and immortalized human keratinocyte cell line (HaCaT) was obtained from CLS (Cell Line Service, Eppelheim, Germany).Cells were cultured in DMEM high glucose medium (4.5 g/L) with the addition of 10% FBS and penicillin (10000 units/ml) and streptomycin (10 mg/ml) (Sigma--Aldrich; Merck KGaA).Cell cultures were performed in the incubator with 5% CO 2 at 37 • C. Before treatment with 1,25-dihydroxyvitamin D medium was changed to DMEM with 2% charcoal-stripped FBS.

CRISPR/Cas9 knock-out cell lines
The knockout of PDIA3 gene was introduced to the squamous cell carcinoma cell line A431 using CRISPR/Cas7 technology [71].Specific guide RNAs (gRNAs) targeting the PDIA3 gene have been designed with the use of Synthego tools (https://www.synthego.com/guide/how-to-use-crispr/design-tool-tutorial).After selecting the appropriate gRNA, a pool of knockout cells (with an editing efficiency of at least 50 %) was purchased from Synthego (Menlo Park, CA, USA).For clonal selection, cloning discs were used (Bel-Art SP SCIENCEWARE).Briefly, cells in low density were seeded, on 6-well plates, and the growth of single-cell colonies was monitored every third day.When the colonies were large enough, trypsin-immersed discs were placed on the colonies and then transferred along with the cells to 12-well plates.The next day cloning discs were removed from the plate.Cells adhere to the discs after trypsinization were transferred to the individual wells and cultured for screening.The introduction of the knockouts was verified by Sanger sequencing (Oligo.pl,Institute of Biochemistry and Biophysics, Poland), qPCR, and Western blot analysis.The detailed analyses of the PDIA3 deletion in A431 SCC cell line were presented in Supplementary Fig. 1.The preparation and characterization of A431ΔVDR cells were described elsewhere [27].

Proliferation assay
The sulforhodamine B (SRB) assay was performed as described previously [28].Shortly, the A431 wild-type (WT) cells or A431ΔPDIA3 cells were seeded on 96-well plates in DMEM medium supplemented with 2% FBS charcoal and cultured overnight.Then, cells were treated with serial dilution of 1,25-dihydroxyvitamin D 3 (10 -6 -10 -12 M) for 72 h.Cells were fixated with 10% trichloroacetic acid and stained with the solution of 0.4% SRB (Sigma-Aldrich; Merck KGaA) in acetic acid.Tris Base buffer was used to solubilize SRB dye.The absorbance was measured at 570 nm on an Epoch™ microplate spectrophotometer (BioTek Instruments, Inc., Winooski, VT, USA).

Cell cycle analysis
The cell cycle analysis was based on the quantification of DNA content with flow cytometry as described previously [21].Shortly, cells were treated with 1,25-dihydroxyvitamin D 3 for 24 or 72 h, at 100 nM concentration.Cells were harvested and fixated with 70% ethanol, then treated with ribonuclease, and finally stained with propidium iodide (PI, Sigma-Aldrich; Merck KGaA) at 37⁰C.The fluorescence of the PI-stained cells was measured by flow cytometry (FACS Calibur™; Becton, Dickinson and Company, Franklin, Lakes, NJ, USA).The results were analyzed using the CellQuest™ Pro Software version 6.0 (Becton, Dickinson, and Company).

Wound healing assay
For motility assessment, A431-derived cell lines were seeded on 8 chamber slides.When cells reached 100% confluency mechanical wound was created by physical scraping with a pipette tip.Cells were treated with 1,25-dihydroxyvitamin D 3 at 100 nM concentration and migration of cells was observed for 72 h as live imaging with Olympus cell Vivo IX83.The cell-free area was calculated with Olympus cell Sens software with the use of TruAI technology.

Immunoblotting
A431-derived cell lines were treated with 100 nM 1,25-dihydroxyvitamin D 3 for 4, 8, and 24 h.The medium was removed from the plate and cells were washed twice with PBS and were scratched from the plate.The solution was moved to an Eppendorf tube and centrifuged at 16,000 × g for 10 min.The received cell sediment was dissolved in 100 µl of RIPA buffer (Thermofisher, Waltham, Massachusetts, USA).Concentration was determined by Bradford Assay.For SDS-PAGE electrophoresis 10% bottom gel and 5% upper gel were used.An equal amount of protein (20µg) was loaded into each well.Electrophoresis was run at 90-110 V in the Bio-Rad apparatus.Proteins were transferred to PVDF membranes with the use of the Trans-Blot Turbo system (Bio-Rad).After the transfer membranes were blocked in 5% milk dissolved in TBS-T.The membranes were incubated with primary antibodies anti-PDIA3/ ERp57 (Thermofisher, Waltham, Massachusetts, USA), anti-VDR (Santa Cruz Biotechnology, Dallas, Texas, USA), or anti-Cyp24A1 (Santa Cruz Biotechnology, Dallas, Texas, USA), overnight at 4 • C.Then, it was incubated with proper secondary peroxidase-conjugated antibodies (anti-rabbit antibody, Santa Cruz Biotechnology, Dallas, Texas, USA, or anti-Mouse antibody, Sigma-Aldrich, Saint Louis, Missouri, USA).For loading control membranes were stripped and reprobed using HRP conjugated anti-β-actin antibody (Santa Cruz Biotechnology, Dallas, Texas, USA).Bands were developed using enhanced chemiluminescence ECL Plus (Perkin Elmer, Waltham, Massachusetts, USA).Imaging was performed on ImageQuant LAS 500 (Cytivia, Mulhouse, France).

Sample preparation for qPCR and RNA sequencing
A431 wild-type cells and A431 knock-out (ΔPDIA3) cells were seed on 6 well plates and treated with 100 µM 1,25-dihydroxyvitamin D 3 or solvent (0.05% ethanol) (Sigma-Aldrich, Saint Louis, Missouri, USA) for 24 h.RNA was extracted according to the manufacturer's instruction with the use of ExtractMe total RNA kit (Blirt, Poland), and the concentrations were assessed with EPOCH Microplate Spectrophotometer (BioTek, USA).

RNA sequencing
For RNAseq, the RNA purity and quality were assessed using the R-NA ScreenTape assays on the 4200 TapeStation System (Agilent Technologies, Germany).The RNA samples with RIN (RNA Integrity Number) between 9,6-10 were sequenced in CeGaT (Tübingen, Germany).The amount of mRNA was 100 ng per sample.The library of RNA was prepared with the use of TruSeq stranded mRNA (Illumina), Q30 value greater than 91%, NovaSeq 6000, 2x 100 bp with usually 30-40 M sequenced raw fragments per sample.The initial quality control assessment confirmed that biological replicates and technical repeats were significantly clustered.

Real-Time PCR
Real-time PCR was performed using a StepOnePlus™ Real-Time PCR System (Life Technologies Applied Biosystems, Grand Island, NY, USA) with an AMPLIFYME SG No-ROX Mix kit (Blirt, Poland).All primers were purchased from Merck, Germany.All PCR reactions were conducted with the primers grouped in the table (Supplementary Table S1).The amount of amplified product for each gene was compared to that of the reference gene (RPL37) using a comparative ΔΔCT method and presented as a fold change ± SD [21].

Bioinformatics analyses
The raw reads were first quality-checked and cleaned up using FastQC [72] and Trimmomatic [73].Principal Component Analysis (PCA) for data quality control and cell line disparities checks, were performed.It detected major clusters distinct from each other, for each experimental condition and treatment.FastQC box plots of quality scores per reading position, Phred score greater than 30 were also performed.Samples clustering based on a distance matrix created from rlog-transformed gene expression values was done [27].RNA-seq data generated as part of this work are available without restriction.The data have been deposited in SRA under accession number PRJNA926032.
The reads were mapped to the human reference genome (GRCh38) using STAR [74].Next, the mapped reads were counted with feature Counts 2.0.3 [75].To compare transcriptomic profiles across samples PCA was conducted on rlog-transformed count values [76].Differential gene expression analysis was conducted using DESeq2 [76].The absolute value of log2fold change ≥ 1.0 and adjusted p-value < 0.05 were used as a criterion to identify differentially expressed genes.Gene ontology (GO) enrichment analysis was performed using the R package top GO [77] with Fisher's exact test to determine the functions of differentially expressed genes.Gene names were mapped to GO terms using the org.Hs.eg.db [78] package.Only GO terms with a p-value < 0.05 were considered significantly enriched.
The results of the analysis were visualized using the R package [79].The search Tool for the Retrieval Interacting Genes (STRING) database was used to construct the network of differentially expressed genes and proteins [80].

Statistical analysis
Statistical analysis was performed using GraphPad Prism version 7.05 (GraphPad Software, Inc., La Jolla, CA, USA).Data are presented as mean ± SD and were analyzed with a Student's t-test (for two groups) or one-way ANOVA analysis of variance with appropriate posthoc tests (for more than two groups) using Prism software.Statistically significant differences are illustrated with asterisks: *p < 0.05, **p < 0.01, ***p < 0.001, or ****p < 0.0001.

Fig. 1 .
Fig. 1.Changes in biological actions of 1,25-dihydroxyvitamin D 3 in A431ΔPDIA3 squamous cell carcinoma.The effect of 1,25-dihydroxyvitamin D 3 on the proliferation of (A) A431WT, (B) A431ΔVDR, and (C) A431ΔPDIA3 human squamous cell carcinoma.The cell lines were treated with serial dilutions (10 -6 -10 -12 ) of 1,25-dihydroxyvitamin D 3 for 24-72 h.Experiments were conducted in three independent experiments (n = 6 in each) ± SEM.Representative graphs for each cell line are shown.Statistical significance between plots was estimated using two-way ANOVA and presented as *p < 0.05; **p < 0.01, or ***p < 0.001.(D-F) The effect of 24-72 h incubation with 1,25-dihydroxyvitamin D 3 on the distribution of human squamous cell carcinoma sublines throughout the phases of the cell cycle.The data are presented as the mean ± standard deviation (n = 3).Statistical significance was estimated using two-way ANOVA followed by Tukey's multiple comparison tests and presented as *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.The results are representative of three experiments.Real-time PCR analysis of (G) MKI67, (H) CDKN2A, (I) CDKN1A mRNA expression levels in A431WT, A431ΔVDR and A431ΔPDIA3 cell lines.Results were normalized to none treated A431WT for all A431 sublines.(J) The effect of 72 h incubation with 1,25-dihydroxyvitamin D 3 on the migration rate in A431WT and A431ΔPDIA3 human squamous cell carcinoma cell lines.(K) Representative pictures of wound closure at 72 h.Statistical values were calculated with a one-way analysis of variance and Tukey's posthoc test and presented as *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.

Fig. 2 .
Fig. 2. Changes in gene expression in A431ΔPDIA3 human squamous carcinoma cells.(A) The amount of upregulated and downregulated DEGs after PDIA3 deletion.(B) Volcano plot of significantly upregulated and downregulated DEGs in A431ΔPDIA3 cells (|log2 FC| ≥ 1 and FDR < 0.01).(C) Gene ontology enrichment analysis of DEGs expressed in A431ΔPDIA3 cells according to biological processes and molecular functions.(D) Changes in the expression levels of genes connected with the Vitamin D receptor as a transcription factor.(E) Protein-protein interaction network of differentially expressed proteins in A431ΔPDIA3 cells in comparison to A431WT cells.The network nodes represent proteins while the edges represent predicted functional associations.There are 5 types of associations presented: neighborhood (green), experimental (purple), text mining (yellow), database (light blue), and co-expression (black) evidence.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3 .Fig. 4 .
Fig. 3. Changes in gene expression in human squamous carcinoma cells with the PDIA3 gene knockout treated with 1,25-dihydroxyvitamin D 3 for 24 h.(A) The number of regulated DEGs after 24 h 1,25-dihydroxyvitamin D 3 treatment in A431WT and A431ΔPDIA3 cell lines.(B) Volcano plot of significantly upregulated and downregulated DEGs in A431ΔPDIA3 cells treated with 1,25-dihydroxyvitamin D 3 (|log2 FC| ≥ 1 and FDR < 0.01).List of the top ten most regulated genes in A431ΔPDIA3 after 24 h 1,25-dihydroxyvitamin D 3 treatment.Gene ontology enrichment analysis of DEGs expressed in A431ΔPDIA3 cells treated with 1,25-dihydroxyvitamin D 3 in terms of (C) biological processes and (D) molecular functions.(E) Venn diagram showing the distribution of DEGs between A431WT and A431ΔPDIA3 cells treated with 1,25-dihydroxyvitamin D 3 and non-treated A431ΔPDIA3.(F) Protein-protein interaction network of differentially expressed proteins in A431ΔPDIA3 cells in comparison to non-treated or treated with 1,25-dihydroxyvitamin D 3 .The network nodes represent proteins while the edges represent predicted functional associations.There are 5 types of associations presented: neighborhood (green), experimental (purple), text mining (yellow), database (light blue), and co-expression (black) evidence.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5 .
Fig. 5. Deletion of PDIA3 in A431 squamous cell carcinoma affects the expression of common DEGs found in Head and Neck squamous cell carcinoma.Volcano plots of DEGs from A431 cell lines, the pink dots represent significantly (A) upregulated and (B) downregulated genes after PDIA3 deletion or (C) upregulated and (D) downregulated genes after 1,25-dihydroxyvitamin D 3 treatment, and black dots represent gene expression profiles from Head and neck Squamous cell carcinoma obtained from TACCO database.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)