Antiproliferative Activity of Non-Calcemic Vitamin D Analogs on Human Melanoma Lines in Relation to VDR and PDIA3 Receptors

Vitamin D is a precursor for secosteroidal hormones, which demonstrate pleiotropic biological activities, including the regulation of growth and the differentiation of normal and malignant cells. Our previous studies have indicated that the inhibition of melanoma proliferation by a short side-chain, low calcemic analog of vitamin D—21(OH)pD is not fully dependent on the expression of vitamin D receptor (VDR). We have examined the effects of classic vitamin D metabolites, 1,25(OH)2D3 and 25(OH)D3, and two low calcemic vitamin D analogs, (21(OH)pD and calcipotriol), on proliferation, mRNA expression and vitamin D receptor (VDR) translocation in three human melanoma cell lines: WM98, A375 and SK-MEL-188b (subline b of SK-MEL-188, which lost responsiveness to 1,25(OH)2D3 and became VDR−/−CYP27B1−/−). All tested compounds efficiently inhibited the proliferation of WM98 and A375 melanoma cells except SK-MEL-188b, in which only the short side-chain vitamin D analog—21(OH)pD was effective. Overall, 21(OH)pD was the most potent compound in all three melanoma cell lines in the study. The lack of responsiveness of SK-MEL-188b to 1,25(OH)2D3, 25(OH)D3 and calcipotriol is explained by a lack of characteristic transcripts for the VDR, its splicing variants as well as for vitamin D-activating enzyme CYP27B1. On the other hand, the expression of VDR and its splicing variants and other vitamin D related genes (RXR, PDIA3, CYP3A4, CYP2R1, CYP27B1, CYP24A1 and CYP11A1) was detected in WM98 and A375 melanomas with the transcript levels being modulated by vitamin D analogs. The expression of VDR isoforms in WM98 cells was stimulated strongly by calcipotriol. The antiproliferative activities of 21(OH)pD appear not to require VDR translocation to the nucleus, which explains the high efficacy of this noncalcemic pregnacalciferol analog in SK-MEL-188b melanoma, that is, VDR−/−. Therefore, we propose that 21(OH)pD is a good candidate for melanoma therapy, although the mechanism of its action remains to be defined.


Introduction
Vitamin D is a naturally occurring hormone precursor, which after activation serves as a pleiotropic regulator of homeostasis on cellular and whole organism levels [1,2]. It is produced in ( Figure 1) and found that 21(OH)pD inhibits melanoma growth through a mechanism independent from VDR or PDIA3.

New Vitamin D3 Analogs Effectively Inhibit A375 Cell Proliferation
Our previous studies on human melanomas SK-MEL-188 [10,39,57,59,62,[67][68][69], A375 [55], and WM98 [39,68] demonstrated the sensitivity of melanoma cells to vitamin D3 analogs with short and full side-chains. The efficacy of compounds was attenuated by melanogenesis, which was associated with the downregulation of VDR expression [56,68]. However, it was also shown that the antiproliferative activity of short side-chain analogs of vitamin D3, such as 21(OH)pD, were not fully dependent on VDR expression in rodent melanomas [56], consistent with the poor docking score on the ligand binding domain of the VDR and the poor translocation of VDR to the nucleus [69].
To better understand the effect of VDR expression on the differential action of vitamin D3 analogs, we used three human melanoma lines, A375 and WM98 and a subline b of SKMEL-188 which lost responsiveness to vitamin D during in vitro passaging [55], and detected the VDR expression and sensitivity to selected vitamin D analogs (1,25(OH)2D3, 25(OH)D3, 21(OH)pD and calcipotriol). As expected, these compounds [40,65] effectively inhibited the proliferation of A375 and WM98 melanomas expressing VDR receptor. IC50 values ranged from pM to µM, and the effects on cell proliferation were strongly dependent on the melanoma cell line and the nature of the compound. Interestingly, non-pigmented SK-MEL-188b was found to be resistant to vitamin D3 analogs with a full-length side-chain (1,25(OH)2D3, 25(OH)D3 and calcipotriol) and sensitive to short side-chained 21(OH)pD ( Figure 2). Overall, 21(OH)pD was found to be the most potent inhibitor of growth in tested melanoma cell lines.

New Vitamin D 3 Analogs Effectively Inhibit A375 Cell Proliferation
Our previous studies on human melanomas SK-MEL-188 [10,39,57,59,62,[67][68][69], A375 [55], and WM98 [39,68] demonstrated the sensitivity of melanoma cells to vitamin D 3 analogs with short and full side-chains. The efficacy of compounds was attenuated by melanogenesis, which was associated with the downregulation of VDR expression [56,68]. However, it was also shown that the antiproliferative activity of short side-chain analogs of vitamin D 3 , such as 21(OH)pD, were not fully dependent on VDR expression in rodent melanomas [56], consistent with the poor docking score on the ligand binding domain of the VDR and the poor translocation of VDR to the nucleus [69].
To better understand the effect of VDR expression on the differential action of vitamin D 3 analogs, we used three human melanoma lines, A375 and WM98 and a subline b of SKMEL-188 which lost responsiveness to vitamin D during in vitro passaging [55], and detected the VDR expression and sensitivity to selected vitamin D analogs (1,25(OH) 2 D 3 , 25(OH)D 3 , 21(OH)pD and calcipotriol). As expected, these compounds [40,65] effectively inhibited the proliferation of A375 and WM98 melanomas expressing VDR receptor. IC 50 values ranged from pM to µM, and the effects on cell proliferation were strongly dependent on the melanoma cell line and the nature of the compound. Interestingly, non-pigmented SK-MEL-188b was found to be resistant to vitamin D 3 analogs with a full-length side-chain (1,25(OH) 2 D 3 , 25(OH)D 3 and calcipotriol) and sensitive to short side-chained 21(OH)pD ( Figure 2). Overall, 21(OH)pD was found to be the most potent inhibitor of growth in tested melanoma cell lines.

Figure 2.
Effects of 1,25(OH)2D3, 25(OH)D3, calcipotriol and 21(OH)pD on growth of human SK-MEL-188b, A375 and WM98 melanoma cells. Cells were seeded into 96-well plates and incubated in a medium supplemented with serial dilution of vitamin D analogs from 0.01 nM to 1000 nM concentration (as described in Material and Methods). The statistical significance of results has been analyzed using one-way ANOVA (GraphPad Software, San Diego, CA, USA) and data are presented as means ± SEM for at least three independent measurements. The cutoff point of significance was defined as p < 0.05 (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001).

The Effect of Vitamin D3 Compounds on the Expression of Genes Related to Vitamin D3 Action or Metabolism
First, we investigated the expression of vitamin D3-related genes in untreated melanoma cells. As shown in Figure 3, PCR fragments characteristic for VDR, CYP27A1 and CYP27B1 mRNAs were below the level of detectability in SK-MEL-188b melanoma, which did not respond to vitamin D. Only the mRNA of PDIA3 was detected in this line. On the other hand, in A375 and WM98, cell lines corresponding to transcripts for all tested genes were detected.
Next, we tested the effects of vitamin D3 analogues on the relative mRNA levels of vitamin D3related genes in responsive melanoma cell lines (A375 and WM98). The treatment of WM98 melanoma with full length side-chain vitamin D3 analogs (1,25(OH)2D3, 25(OH)D3 and calcipotriol) is presented in Figure 4. Only 1,25(OH)2D3 stimulated the expression of VDR Co-receptor RXR and of CYP27A1 ( Figure 4B). All three analogs, however, stimulated the expression of CYP3A4, CYP2R1, CYP24A1 and its splicing variant CYP24A1sv. Interestingly, the strongest effect (approximately a 6fold induction) on CYP2R1 mRNA level was observed in cells treated with calcipotriol ( Figure 4C). Cells were seeded into 96-well plates and incubated in a medium supplemented with serial dilution of vitamin D analogs from 0.01 nM to 1000 nM concentration (as described in Material and Methods). The statistical significance of results has been analyzed using one-way ANOVA (GraphPad Software, San Diego, CA, USA) and data are presented as means ± SEM for at least three independent measurements. The cutoff point of significance was defined as p < 0.05 (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001).

The Effect of Vitamin D 3 Compounds on the Expression of Genes Related to Vitamin D 3 Action or Metabolism
First, we investigated the expression of vitamin D 3 -related genes in untreated melanoma cells. As shown in Figure 3, PCR fragments characteristic for VDR, CYP27A1 and CYP27B1 mRNAs were below the level of detectability in SK-MEL-188b melanoma, which did not respond to vitamin D. Only the mRNA of PDIA3 was detected in this line. On the other hand, in A375 and WM98, cell lines corresponding to transcripts for all tested genes were detected.
Next, we tested the effects of vitamin D 3 analogues on the relative mRNA levels of vitamin D 3 -related genes in responsive melanoma cell lines (A375 and WM98). The treatment of WM98 melanoma with full length side-chain vitamin D 3 analogs (1,25(OH) 2 D 3 , 25(OH)D 3 and calcipotriol) is presented in Figure 4. Only 1,25(OH) 2 D 3 stimulated the expression of VDR Co-receptor RXR and of CYP27A1 ( Figure 4B). All three analogs, however, stimulated the expression of CYP3A4, CYP2R1, CYP24A1 and its splicing variant CYP24A1sv. Interestingly, the strongest effect (approximately a 6-fold induction) on CYP2R1 mRNA level was observed in cells treated with calcipotriol ( Figure 4C). Furthermore, the level of CYP27B1 mRNA was not affected by tested secosteroids ( Figure 4G), while the expression of PDIA3 was stimulated only by 25(OH)D 3 ( Figure 4E).
The treatment of A375 melanoma with four vitamin D 3 analogs (1,25(OH) 2 D 3 , calcipotriol and 21(OH)pD) decreased the expression of VDR, RXR, PDIA3, CYP2R1 genes ( Figure 5). Interestingly, the strongest decrease in the mRNA levels of CYP3A4 and CYP2R1 was observed in A375 melanoma cells treated with 21(OH)pD. The induction of CYP24A1 gene was hundreds of times folds higher in cells treated with 1,25(OH) 2 D 3 and calcipotriol when compared to 21(OH)pD, showing minimal effect. Finally, the mRNA level of CYP27B1 was not affected by vitamin D analogues and CYP11A1 was elevated in A375 melanoma cells only after treatment with calcipotriol. Furthermore, the level of CYP27B1 mRNA was not affected by tested secosteroids ( Figure 4G), while the expression of PDIA3 was stimulated only by 25(OH)D3 ( Figure 4E) The treatment of A375 melanoma with four vitamin D3 analogs (1,25(OH)2D3, calcipotriol and 21(OH)pD) decreased the expression of VDR, RXR, PDIA3, CYP2R1 genes ( Figure 5). Interestingly, the strongest decrease in the mRNA levels of CYP3A4 and CYP2R1 was observed in A375 melanoma cells treated with 21(OH)pD. The induction of CYP24A1 gene was hundreds of times folds higher in cells treated with 1,25(OH)2D3 and calcipotriol when compared to 21(OH)pD, showing minimal effect. Finally, the mRNA level of CYP27B1 was not affected by vitamin D analogues and CYP11A1 was elevated in A375 melanoma cells only after treatment with calcipotriol.  Table S1 for primer sequences and the predicted length of PCR fragments.  Table S1 for primer sequences and the predicted length of PCR fragments.   It is well established that the antiproliferative effects of vitamin D3 analogs strongly depend on the expression of vitamin D receptor VDR [40,47,65]. As shown in Figure 3, the melanoma cell lines used in the study differ in their basal level of VDR expression, and the transcript is absent in SK-MEL-188b, which is not responsive to vitamin D. Theoretically, an alternative splicing of VDR pre-mRNA may results in the production of at least five alternative transcripts ( Figure 6A); thus, the unique set of primers was designed in order to detect specific variants of VDR mRNA (Table S1). Detection of specific variants of VDR was confirmed by sequencing of PCR fragments previously separated on agarose gel ( Figure 6B). WM98 and A375 melanoma cell lines express all four PCR fragments (Pr 1, Pr 2, Pr 4 and Pr 5) corresponding to VDR splicing variant c ( Figure 6A). In the WM98 melanoma, the cell line presence of an mRNA characteristic for isoforms a (Pr 3) and b (Pr 4, upper band) was also detected. Interestingly, in SK-MEL-188b melanoma cells, only one fragment of mRNA corresponding to the alternative promoter region of VDR (isoform c) was detected. Finally, the effects of two vitamin D3 analogs 25(OH)D3 and calcipotriol on the alternative splicing of VDR pre-mRNA was tested in WM98 cell line ( Figure 6C). It was show that treatment with calcipotriol, but not 25(OH)D3, resulted in at least 7-fold elevation of mRNA levels corresponding to all fragments of the VDR tested (fragments detected by a set of primers: Pr 1 to Pr 5). It is well established that the antiproliferative effects of vitamin D 3 analogs strongly depend on the expression of vitamin D receptor VDR [40,47,65]. As shown in Figure 3, the melanoma cell lines used in the study differ in their basal level of VDR expression, and the transcript is absent in SK-MEL-188b, which is not responsive to vitamin D. Theoretically, an alternative splicing of VDR pre-mRNA may results in the production of at least five alternative transcripts ( Figure 6A); thus, the unique set of primers was designed in order to detect specific variants of VDR mRNA (Table S1). Detection of specific variants of VDR was confirmed by sequencing of PCR fragments previously separated on agarose gel ( Figure 6B). WM98 and A375 melanoma cell lines express all four PCR fragments (Pr 1, Pr 2, Pr 4 and Pr 5) corresponding to VDR splicing variant c ( Figure 6A). In the WM98 melanoma, the cell line presence of an mRNA characteristic for isoforms a (Pr 3) and b (Pr 4, upper band) was also detected. Interestingly, in SK-MEL-188b melanoma cells, only one fragment of mRNA corresponding to the alternative promoter region of VDR (isoform c) was detected. Finally, the effects of two vitamin D 3 analogs 25(OH)D 3 and calcipotriol on the alternative splicing of VDR pre-mRNA was tested in WM98 cell line ( Figure 6C). It was show that treatment with calcipotriol, but not 25(OH)D 3 , resulted in at least 7-fold elevation of mRNA levels corresponding to all fragments of the VDR tested (fragments detected by a set of primers: Pr 1 to Pr 5). Statistical significance was estimated using t-test and data are presented as means ± SD (n = 3). The cutoff point for significance is defined as p < 0.05 (* p < 0.05, ** p < 0.01, *** p < 0.001).

Translocation of Vitamin D Receptor by Secosteroids
The ligand-induced translocation of VDR receptor into the nucleus was studied using A375 cells Statistical significance was estimated using t-test and data are presented as means ± SD (n = 3). The cutoff point for significance is defined as p < 0.05 (* p < 0.05, ** p < 0.01, *** p < 0.001).

Translocation of Vitamin D Receptor by Secosteroids
The ligand-induced translocation of VDR receptor into the nucleus was studied using A375 cells stably transduced with pLenti-CMV-VDR-EGFP lentiviral construct [70]. 1,25(OH) 2 D 3 or calcipotriol induced a dose-dependent and efficient translocation of VDR-GFP fusion protein into the nucleus (Figure 7). A lack of VDR-GFP translocation to the nucleus was observed in cells treated with 21(OH)pD.
Here, we present results showing the different sensitivities of three melanoma cell lines (SK-

Discussion
There is growing evidence that vitamin D 3 analogs are excellent candidates for melanoma therapy [40,47,65,71,72]. This is supported by the strong correlation between an adequate level of 25(OH)D 3 (>30 ng/mL in the serum) and a decreased incidence or severity of several cancers, including melanoma [62,[73][74][75][76][77]. However, the use of an active form of vitamin D 3 , 1,25(OH) 2 D 3 is limited due to its potential hypercalcemic effects; thus, the development and biological evaluation of new low calcemic vitamin D 3 derivatives is highly desirable [40,55,65,72].
Here, we present results showing the different sensitivities of three melanoma cell lines (SK-MEL-188b, A375 and WM98) to vitamin D 3 analogs. Recently, we described vitamin D 3 -resistant clone of SK-MEL1-88 (SK-MEL-188b) [55]. Contrary to parental SK-MEL-188, this clone was insensitive to the treatment with 1,25(OH) 2 D 3 , or with several analogs of vitamin D 2 , except of PRI-1731 with the inverse orientation of A ring in comparison to parental 1,25(OH) 2 D 3 by the introduction of (5Z,7Z) modification [55,78,79]. This clone does not express the coding region of VDR and of CYP27B1. Therefore, we call it SK-MEL-188VDR −/− CYP27B1 −/− . Interestingly, a short side-chain analog of vitamin D 3 with a pregnacalciferol-type configuration, such as 21(OH)pD, was shown to inhibit the growth of the SK-MEL-188b line. It has to be underlined that this analog, also showed high potency against two other melanoma cell lines in the study (A375 and WM98) and minimal stimulation of CYP24A1. This indicates that the antiproliferative properties of 21-OHpD are independent on its action on the VDR, which is consistent with our previous findings on pregnacalciferol derivatives acting as poor activators of VDR per VDR translocation studies and molecular modeling [69]. One of the possible mechanisms of 21(OH)pD action could include an activation of non-genomic pathways [25,80], including those linked to the PDIA3 protein [81,82]. However, interaction with other nuclear steroid receptors cannot be excluded. In addition, other receptors such as retinoic orphan acid receptors (ROR)α and γ can be considered since vitamin D hydroxyderivatives can act as reverse agonists on these receptors [62,83] and related pregnalumisterol derivatives can also act on RORα and RORγ [15]. These considerations require future in-depth investigations to define specific receptors for pD compounds (secosteroids with shortened side-chain).
Two other melanoma lines were responsive to vitamin D (WM98 and A375) and showed relatively high expression levels of VDR transcript. Thus, the studies presented above on three melanoma lines are consistent with reports showing that the expression of VDR is the key factor responsible for the antitumor activities of 1,25(OH) 2 D 3 [18,47,84,85]. Of note is that its decreased expression correlates with advanced melanoma staging, progression and decreased overall patient survival and disease-free survival time [45,46]. Accordingly, VDR polymorphism [41,42,72] and decreased levels of 25(OH)D 3 in the serum [33,86] positively correlate with melanoma prevalence and poor prognosis.
Here, we show that sensitivity to tested vitamin D 3 hydroxyderivatives with full side-chains depends on VDR expression and its alternative splicing. Interestingly, sensitive melanoma lines (A375, WM98) expressed all three major VDR splicing variants, suggesting that they may play an important but not necessarily identical function in vitamin D 3 signaling. Further studies are required to elucidate the significance of the expression of VDR splicing variants. Furthermore, the expression of vitamin D 3 -related genes was altered by vitamin D 3 analogs, and the effects on the expression of VDR, RXR, PDIA3, CYP2R1 or CYP24A1 were stronger in WM98 melanoma cells in comparison to the A375 line. This phenomenon could explain higher sensitivity of WM98 to 1,25(OH) 2 D 3 in comparison to A375 line.
In summary, low calcemic vitamin D 3 analogs such as 25(OH)D 3 , 21(OH)pD or calcipotriol showed similar antiproliferative activity to 1,25(OH) 2 D 3 in melanoma cell lines expressing VDR spliced variants. Furthermore, the short side-chain analog 21(OH)pD was found to be superior among vitamin D analogs and was the only one, which inhibited the growth of the melanoma subline negative for VDR, indicating a mechanism of action that is VDR-independent.

Cell Lines and Vitamin D Analogs
In our study, we determined the inhibitory effects of vitamin D 3 analogs against immortalized human melanoma cell lines SK-MEL-188b, WM98 and A375. SK-MEL-188b cells were cultured in F10 medium supplemented with 10% fetal bovine serum (FBS, Sigma, Poznan, Poland) and 1× antibiotic-antimycotic solution (Anti-Anti, Sigma). A375 and WM98 cell lines were cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% FBS and 1% anti-anti. To eliminate the influence of sterols present in fetal bovine serum in the experiments, we used 5% charcoal-stripped FBS (GE Healthcare Life Sciences, Warsaw, Poland). Vitamin D 3 analogs 1,25(OH) 2 D 3 , 25(OH)D 3 and calcipotriol were acquired from the Pharmaceutical Research Institute (Warsaw, Poland), while the new short side analog 21(OH)pD was synthesized in collaboration with ProChimia Surfaces (Gdynia, Poland) as described previously [49,50]. The chemical structures of these secosteroids are presented in Figure 1.

SRB Assay
Melanoma cells were seeded on 96-well plates at a density of 8000 per well in an appropriate medium supplemented with 5% charcoal-stripped FBS and 1× anti-anti solution. After 24 h, the medium was replaced with the fresh one containing vehicle and serial dilutions of vitamin D 3 analogs at concentrations from 10 pM to 1 µM. Following incubation for 48 h, 100 µL of 20% TCA (trichloroacetic acid) was added and cells were incubated for 1 h in 4 • C. The medium was discarded and cells were washed 5 times with sterile water. Afterwards, cells were dried and 100 µL of SRB (0.4 g sulforhodamin B in 100 mL 1% acetic acid) was added into the plate wells for 15 min incubation at room temperature. Next, cells were washed 5 times with 1% acetic acid, dried and resolved with 150 µL of 10 mM Tris (pH = 10.5) for 10 min in room temperature. The absorbance was measured spectroscopically at 570 nm with 96-well plate reader (BioTek, Winooski, VT, USA).

VDR Translocation
The melanoma A375 cell line was used to study the VDR receptor translocation to nucleus after vitamin D 3 analog treatment. Cells were transduced by pLenti-CMV-VDR-EGFP-pgkpuro plasmid, where VDR and GFP were expressed as a fusion protein in cytoplasmic compartment. Cells were cultured on 12 wells plate (0.5 × 10 5 cells/well) in DMEM supplemented with 5% charcoal-stripped FBS and 24 h later medium was replaced with DMEM containing 1,25(OH) 2 D 3 , calcipotriol and 21(OH)pD at 0.01 µM, 0.1 µM or 1 µM concentrations. After 1 h of incubation VDR translocation from cytoplasm to nucleus was determined with fluorescent Eclipse TE300 microscope (Nikon, Tokyo, Japan).

Statistical Analysis
SRB viability data was presented using one-way Anova test to compare experimental groups: cells were treated with vitamin D 3 analogs with a concentration from 0.01 nM to 1000 nM with control samples (GraphPad Software). The antiproliferative potency of vitamin D 3 analogs were compared by the calculation of EC50 for every compound (half maximal effective concentration).
Real-time PCR data was analyzed with the comparative ∆∆-Ct method normalized to the reference gene ACTB. As a control, a probe with sterile water was used instead of cDNA. Data are presented as mean ± SD (n = 4-6). Student's t-test (for two groups) or one-way ANOVA with appropriate post-hoc test (for more than two groups) were used to analyze data using Excel (Microsoft) or Prism 7.00 (GraphPad Software), respectively. Statistically significant differences are denoted with asterisks: p < 0.05 (* p < 0.05; ** p < 0.01; *** p < 0.0001).