The Role of the Vitamin D Receptor and ERp57 in Photoprotection by 1α,25-Dihydroxyvitamin D₃

Abstract

UV radiation (UVR) is essential for formation of vitamin D₃, which can be hydroxylated locally in the skin to 1α,25-dihydroxyvitamin D₃ [1,25-(OH)₂D₃]. Recent studies implicate 1,25-(OH)₂D₃ in reduction of UVR-induced DNA damage, particularly thymine dimers. There is evidence that photoprotection occurs through the steroid nongenomic pathway for 1,25-(OH)₂D₃ action. In the current study, we tested the involvement of the classical vitamin D receptor (VDR) and the endoplasmic reticulum stress protein 57 (ERp57), in the mechanisms of photoprotection. The protective effects of 1,25-(OH)₂D₃ against thymine dimers were abolished in fibroblasts from patients with hereditary vitamin D-resistant rickets that expressed no VDR protein, indicating that the VDR is essential for photoprotection. Photoprotection remained in hereditary vitamin D-resistant rickets fibroblasts expressing a VDR with a defective DNA-binding domain or a mutation in helix H1 of the classical ligand-binding domain, both defects resulting in a failure to mediate genomic responses, implicating nongenomic responses for photoprotection. Ab099, a neutralizing antibody to ERp57, and ERp57 small interfering RNA completely blocked protection against thymine dimers in normal fibroblasts. Co-IP studies showed that the VDR and ERp57 interact in nonnuclear extracts of fibroblasts. 1,25-(OH)₂D₃ up-regulated expression of the tumor suppressor p53 in normal fibroblasts. This up-regulation of p53, however, was observed in all mutant fibroblasts, including those with no VDR, and with Ab099; therefore, VDR and ERp57 are not essential for p53 regulation. The data implicate the VDR and ERp57 as critical components for actions of 1,25-(OH)₂D₃ against DNA damage, but the VDR does not require normal DNA binding or classical ligand binding to mediate photoprotection.

Sunlight production of vitamin D is essential for survival but there are a number of damaging physiological effects of UV radiation (UVR). An important form of UVR-induced DNA damage is cyclobutane pyrimidine dimers (CPD). UVR breaks the carbon 5–6 double bond of adjacent pyrimidine residues that subsequently dimerize, resulting in a CPD (1). Inefficient repair of CPD can lead to mutations in the DNA sequence that may cause neoplasia (1). The most common form of CPD are thymine dimers because the majority of CPD occur at thymine-thymine pairs (2).

UVR also produces vitamin D₃ in the skin, which is converted through intermediates to the hormonal form 1α,25-dihydroxyvitamin D₃ [1,25-(OH)₂D₃] in kidney and a number of other tissues including skin (3). Previous findings indicate a photoprotective role for 1,25-(OH)₂D₃ in reducing UVR-initiated DNA damage in the form of thymine dimers in various cells including keratinocytes, fibroblasts, and melanocytes (4–6). There is also some evidence in keratinocytes to propose a role for increased expression of the tumor suppressor protein p53 in these photoprotective effects of 1,25-(OH)₂D₃ (7).

Like many steroid hormones, 1,25-(OH)₂D₃ is able to mediate its action via two signal transduction pathways: the classical genomic pathway and a nongenomic pathway. In the genomic pathway, 1,25-(OH)₂D₃ binds to a well-characterized (vitamin D receptor) VDR, which heterodimerizes with the retinoid X receptor. This complex subsequently binds to the vitamin D response elements in regulatory regions of target genes, regulating gene transcription. There have been numerous detailed reports on the nongenomic actions of 1,25-(OH)₂D₃, including the activation of the protein kinase C pathway, MAPK pathway, phosphatidylinositol 3-kinase pathway, and the opening of calcium and chloride channels (reviewed in Ref. 8). We have evidence, mainly from 1,25-(OH)₂D₃ analog studies, that the protective effect of 1,25-(OH)₂D₃ against UVR-induced DNA damage is mediated via the nongenomic signaling pathway because the analog 1α,25-dihydroxylumisterol₃ (JN), a nongenomic agonist, mimics the protective effects of 1,25-(OH)₂D₃ and 1β,25-(OH)₂-vitamin D₃ (HL), a nongenomic antagonist, abolishes these effects (5, 6). However, the role of the nongenomic pathway and the nature of the membrane-associated protein(s) through which 1,25-(OH)₂D₃ mediates these nongenomic protective biological responses remains controversial.

One group has reported that the nongenomic responses of 1,25-(OH)₂D₃ are mediated by the classic VDR located in caveolae, a subfraction of plasma membrane (9, 10). Caveolae are invaginations in the membrane, rich in proteins and lipids, that are involved in signal transduction and molecular transport. The caveolar-localized VDR has been found in chicks, rats, mice, and humans, in intestinal, chick kidney, and osteoblast cells (9). Also it has been shown that tritiated-1,25-(OH)₂D₃ localizes in the caveolae VDR in vivo (9). Knockout of the mouse VDR resulted in a loss of VDR in both the nucleus and the caveolae as well as the loss of VDR-mediated nongenomic responses (11). This is supported by a study in chick skeletal muscle cells that shows translocation of the VDR to the plasma membrane after activation by 1,25-(OH)₂D₃ (12). Results from molecular modeling studies have shown that the classic VDR can form interactions with vitamin D compounds in an alternate ligand-binding pocket, which is distinct from the genomic pocket (13–15). A 6-s-cis-locked analog of 1,25-(OH)₂D₃, JN, with no genomic trans-activating activity, was also observed to stably bind to the VDR for sufficient time to produce a nongenomic response. The naturally occurring 1,25-(OH)₂D₃ molecule was reported to form a favorably stable complex with this alternate pocket, similar to that of JN, but with different affinity to that formed when bound to the genomic pocket (14, 15). Unlike the genomic pocket that is selective for the 6-s-trans ligand, the alternate pocket accepts both the 6-s-trans and the 6-s-cis vitamin D sterols, facilitating nongenomic responses (14).

Evidence for an alternate 1,25-(OH)₂D₃ membrane receptor, endoplasmic reticulum stress protein 57 (ERp57), also known as membrane-associated rapid response steroid binding protein and protein disulfide isomerase family A, member 3, has also been reported (16, 17). Previous studies have shown that in the presence of Ab099, a neutralizing antiserum against the N terminus of ERp57, 1,25-(OH)₂D₃-mediated effects, such as augmented intracellular calcium, protein kinase C activity, and phosphate uptake, were abolished in chick intestinal epithelial cells (17, 18). Identification of a clone of this membrane complex provided strong evidence that this membrane-associated receptor has no sequence similarity to the VDR (19). Both ERp57 and the VDR have been associated with plasma membrane caveolae (9, 10, 20).

Skin fibroblasts from normal subjects and from selected patients with hereditary vitamin D-resistant rickets (HVDRR) were used to undertake studies of the effect of various VDR mutations on photoprotection by 1,25-(OH)₂D₃. HVDRR is a recessive genetic disorder due to mutations in the VDR gene. Hormonal resistance to 1,25-(OH)₂D₃ results in impaired calcium and phosphate absorption, secondary hyperparathyroidism, and early onset of rickets (21). The various mutations used in these studies result in differential loss of VDR function allowing us to pin point the requirements of the VDR to mediate photoprotection. To determine the contribution of ERp57, in parallel studies, we used Ab099 and small interfering RNA (siRNA) mediated against ERp57 in normal fibroblasts.

Results

UVR-induced thymine dimers

In vehicle-treated normal fibroblasts, nuclei staining positively for thymine dimers were 30 ± 7% of total nuclei after UVR (Fig. 1A). Both 1,25-(OH)₂D₃ at 1 nm (a concentration shown to produce a maximal effect in this system) and JN at 1 nm [shown to have equivalent activity to 1 nm 1,25-(OH)₂D₃ (6)] significantly reduced thymine dimers in normal fibroblasts (P < 0.001 and P < 0.01, respectively). There was no significant difference in thymine dimers after UVR, comparing 1,25-(OH)₂D₃ and JN-treated fibroblasts.

Fig. 1.

Reduction of UVR-induced thymine dimers by 1,25-(OH)₂D₃ is abolished in HVDRR fibroblasts expressing no VDR and in normal fibroblasts treated with Ab099 or ERp57 siRNA. Fibroblasts were exposed to solar-like UVR and (A–D) treated immediately after with vehicle [0.1% (vol/vol) EtOH], 1 nm 1,25-(OH)₂D₃, or 1 nm JN (A, C, and D) or (E) treated immediately after UVR with 1:500 Ab099 or 1:500 normal rabbit serum for 5 min at RT; medium was then replaced with medium containing vehicle [0.1% (vol/vol) EtOH] or 1 nm 1,25-(OH)₂D₃,or (F) transfected with 50 nm control siRNA or 50 nm ERp57 siRNA 24 h before irradiation and then treated immediately after UVR with vehicle [0.1% (vol/vol) EtOH], 1 nm 1,25-(OH)₂D₃. Thymine dimers were detected 3 h after UVR by immunohistochemistry using an antithymine dimer antibody. Image analysis quantified positive nuclei as a percent of total nuclei. A, E, and F, normal fibroblasts; B, VDR-null fibroblasts; C, DBD mutant fibroblasts; D, LBD mutant fibroblasts. Pooled data from a minimum of three independent experiments, mean ± sem. Significantly different from vehicle: ***, P < 0.001; **, P < 0.01; *, P < 0.05; from control serum + vehicle: ^###, P < 0.001; from Ab099 + vehicle: ^@, P < 0.05; from control serum + 1,25-(OH)₂D_3: ^%%%, P < 0.001; from control siRNA + vehicle: ^ΩΩΩ, P < 0.001; from ERp57 siRNA + vehicle: ^ΔΔΔ, P < 0.001; from control siRNA + 1,25-(OH)₂D₃: ^×××, P < 0.001. G, Fibroblasts were lysed 24 h after transfection with control siRNA (lanes 1 and 2) or ERp57 siRNA (lanes 3 and 4). Lysates were subject to SDS-PAGE electrophoresis and then probed by immunoblotting with an ERp57 antibody.

Dermal fibroblasts from a patient with HVDRR, expressing a nonsense mutation resulting in an early stop codon (R50X) in the VDR (VDR-null fibroblasts) (22), were then tested. This mutation effectively deletes the VDR (22). The findings in Fig. 1B clearly show that the photoprotective effects of 1,25-(OH)₂D₃ are abolished in VDR-null fibroblasts.

To examine the function of the VDR DNA-binding domain (DBD) in the reduction of thymine dimers by 1,25-(OH)₂D₃ and JN after UVR, we used HVDRR fibroblasts expressing a missense mutation in the DBD (DBD mutant fibroblasts). This mutation changed valine to methionine at amino acid 26 (V26M) and abolished genomic transactivation mediated by 1,25-(OH)₂D₃ (23). The proportion of thymine dimer-positive staining induced by UVR in vehicle-treated DBD mutant fibroblasts was 34 ± 4%, which was significantly reduced to 8 ± 3% by 1,25-(OH)₂D₃ (P < 0.001) and to 9 ± 4% by JN (P < 0.001) (Fig. 1C). The findings indicate that photoprotection could still be mediated by a VDR incapable of binding DNA and inducing genomic transactivation. These data, together with the loss of photoprotection in the VDR-null mutant cells, are highly suggestive that photoprotection is VDR dependent but not dependent on genomic transactivation by the VDR.

Dermal fibroblasts from a patient with HVDRR expressing a 5-bp deletion/8-bp insertion in helix H1 of the VDR [ligand-binding domain (LBD) mutant fibroblasts], which abolished classical ligand binding (24), were then tested to determine whether the classical (LBD) is implicated in the photoprotective effects of 1,25-(OH)₂D₃ and JN against thymine dimers after UVR. As shown in Fig. 1D, the proportion of thymine dimer-positive cells induced by UVR in LBD mutant fibroblasts treated with 1,25-(OH)₂D₃ or JN was reduced by more than 50% compared with vehicle (P < 0.05 and P < 0.01, respectively). Thus classical 1,25-(OH)₂D₃ binding in the LBD may not be required for (full) photoprotection.

Ab099, which neutralizes ERp57, had no effect on its own, but abolished the protective effect of 1,25-(OH)₂D₃ in reducing thymine dimers after UVR (Fig. 1E). When normal rabbit serum was used as a control instead of the polyclonal Ab099, 1,25-(OH)₂D₃ significantly attenuated UVR-induced thymine dimers from 19 ± 3% in vehicle-treated cells to 2 ± 1% in cells treated with 1,25-(OH)₂D₃ (P < 0.001). There was no significant difference in thymine dimers comparing control serum and vehicle with Ab099 and vehicle.

Figure 1G shows a marked reduction in expression of ERp57 protein in fibroblasts transfected with 50 nm ERp57 siRNA (lanes 3 and 4) compared with control siRNA (lanes 1 and 2). siRNA against ERp57 had no effect on its own but abolished the protective effect of 1,25-(OH)₂D₃ against thymine dimers after UVR (Fig. 1F). In cells transfected with control siRNA, 1,25-(OH)₂D₃ significantly reduced thymine dimers after UVR compared with vehicle (P < 0.001). There was no significant difference in thymine dimers comparing control siRNA and vehicle with ERp57 siRNA and vehicle.

Coimmunoprecipitation (Co-IP)

Having identified the VDR and ERp57 as mediators of the nongenomic photoprotective effects of 1,25-(OH)₂D₃, we wanted to determine whether these two proteins interact in the nonnuclear fraction of human fibroblasts. Nonnuclear cell extracts isolated from fibroblasts were subjected to Co-IP assay using an anti-VDR or anti-ERp57 antibody with agarose beads. Samples were electrophoresed, and associated proteins were detected by immunoblot. Co-IP of nonnuclear extracts with anti-VDR as the pull-down antibody followed by subsequent immunoblot with anti-ERp57 antibody confirmed that the VDR coimmunoprecipitated with ERp57 (band detected at 61 kDa) (Fig. 2, bottom panel, lane 1). Similarly, immunoprecipitation of nonnuclear extracts with anti-ERp57 followed by immunoblot with anti-VDR antibody showed that ERp57 coimmunoprecipitated with VDR in nonnuclear extracts (protein band detected at 50 kDa) (Fig. 2, top panel, lane 3). Co-IP with anti-VDR antibody followed by immunoblot with anti-VDR antibody showed the expected protein band at 50 kDa (Fig. 2, top panel, lane 1). Protein bands were evident at 61 kDa when the Co-IP and the immunoblots were performed with the anti-ERp57 antibody (Fig. 2, bottom panel, lane 3). Minimal to no protein was observed when the rabbit or mouse isotype control was used as the pull-down antibody, followed by immunoblot with anti-VDR or anti-ERp57 (Fig. 2, lanes 2 and 4).

Fig. 2.

VDR and ERp57 interact in nonnuclear extracts of fibroblasts. Fibroblast extracts were precleared with a rabbit (lane 2) or mouse (lane 4) isotype control and then incubated with an anti-VDR (lane 1) or anti-ERp57 (lane 3) antibody, respectively, together with agarose beads overnight at 4 C. The purified proteins were subjected to SDS-PAGE and then probed by immunoblotting with a specific VDR (top panel) or ERp57 (bottom panel) antibody.

p53

As shown in Fig. 3A, p53 staining after normal skin fibroblasts were treated with vehicle or 1 nm 1,25-(OH)₂D₃ showed nuclear expression of p53, and there was no staining in the cytoplasm of the cells. Expression of p53 was quantified by whole-cell ELISA. As shown in Fig. 3B, 1,25-(OH)₂D₃ up-regulated p53 expression in sham-irradiated normal fibroblasts by 2.5 ± 0.3-fold compared with vehicle (P < 0.001). p53 was not enhanced after UVR exposure compared with sham vehicle, and the increase in p53 with 1,25-(OH)₂D₃ treatment remained (P < 0.001), but was not further augmented in irradiated fibroblasts compared with sham 1,25-(OH)₂D₃-treated fibroblasts.

Fig. 3.

Up-regulation of p53 by 1,25-(OH)₂D₃ remained in HVDRR fibroblasts expressing no VDR and in normal fibroblasts treated with Ab099. Human fibroblasts were (A–E) exposed to UVR and treated immediately after with vehicle [0.1% (vol/vol) EtOH] or 1 nm 1,25-(OH)₂D₃ or (F) treated with 1:500 Ab099 or 1:500 normal rabbit serum for 5 min at RT and then medium was replaced with medium containing vehicle [0.1% (vol/vol) EtOH] or 1 nm 1,25-(OH)₂D₃. A, Immunohistochemistry was performed 6 h after UVR. Images are representative fields captured during image analysis. p53 expression is visualized as darkly stained nuclei. Black arrows point to positively stained nuclei: 1) sham vehicle; 2) sham 1 nm 1,25-(OH)₂D₃. B–F, Cells were fixed 6 h after treatment, and p53 expression was detected by whole-cell ELISA. Individual absorbance values were corrected for total cell protein by BCA assay before pooling. B and F, normal fibroblasts; C, VDR-null fibroblasts; D, DBD mutant fibroblasts; E, LBD mutant fibroblasts. Pooled data from a minimum of three independent experiments, mean ± sem. Significantly different from sham vehicle: ***, P < 0.001; **, P < 0.01; significantly different from UVR vehicle: ^%%%, P < 0.001, ^%, P < 0.05; significantly different from control serum + vehicle: ^###, P < 0.001; significantly different from Ab099 + vehicle: ^ΩΩΩ, P < 0.001.

The results in Fig. 3C show that in VDR-null fibroblasts, 1,25-(OH)₂D₃ still significantly augmented nuclear p53 in sham-irradiated (1.4 ± 0.1 fold) and UVR-irradiated VDR-null fibroblasts (1.3 ± 0.1 fold) compared with the respective vehicle-treated cells (P < 0.01 and P < 0.05, respectively). Although the fold increase was small, it was consistent and significantly different from vehicle in all the individual experiments. In DBD mutant fibroblasts, basal p53 expression was significantly increased 1.5 ± 0.1 fold when treated with 1,25-(OH)₂D₃ without UVR (P < 0.001) (Fig. 3D), and 1.7 ± 0.1 fold after UVR (P < 0.001), a similar change to that in sham-irradiated cells. In LBD mutant fibroblasts, 1,25-(OH)₂D₃ significantly up-regulated nuclear p53 compared with vehicle by 1.8 ± 0.1 fold in sham-irradiated cells (P < 0.001) and by 1.7 ± 0.1 fold in UV-irradiated cells (P < 0.001) (Fig. 3E). As shown in Fig. 3F, 1,25-(OH)₂D₃ significantly augmented nuclear p53 expression, compared with vehicle, in normal fibroblasts that were pretreated with Ab099 (P < 0.001) at concentrations of Ab099, which abolished the protective effect of 1,25-(OH)₂D₃ on thymine dimers. Cells that were pretreated with normal rabbit serum as a control also showed enhanced nuclear p53 expression when treated with 1,25-(OH)₂D₃ compared with vehicle (P < 0.001).

Discussion

The current study confirmed that 1,25-(OH)₂D₃ and JN decreased UVR-induced thymine dimers in normal human dermal fibroblasts. Human skin fibroblasts from patients with HVDRR, which expressed various loss of function mutations in the VDR, were used to study the effect of the different mutations on the reduction of thymine dimers by 1,25-(OH)₂D₃. 1,25-(OH)₂D₃-induced protection from thymine dimer formation was completely abolished in fibroblasts expressing a premature stop codon (R50X) in the VDR DBD (VDR-null fibroblasts). This mutation deleted most of the VDR protein, preventing specific 1,25-(OH)₂D₃-VDR binding and abolishing CYP24A1 expression, a marker of the 1,25-(OH)₂D₃ genomic response (22). Improved cell viability in the presence of 1,25-(OH)₂D₃ after UVR, seen in normal skin fibroblasts, was also abolished in VDR-null fibroblasts (Supplemental Fig. 1, a and b, published on The Endocrine Society's Journals Online web site at http://mend.endojournals.org), consistent with our proposal that 1,25-(OH)₂D₃-induced improved viability after UVR is a consequence of reduced DNA damage or at least requires similar pathways.

1,25-(OH)₂D₃-induced reduction of thymine dimers remained active in fibroblasts expressing a mutation in the DBD (V26M) of the VDR (DBD mutant fibroblasts). When these DBD mutant fibroblasts were treated with 1,25-(OH)₂D₃, they exhibited normal ligand binding but failed to induce CYP24A1 gene expression (23). Photoprotection against thymine dimers also remained in fibroblasts that expressed a mutation in the LBD of the VDR due to a 5-bp deletion/8-bp insertion in helix H1 (LBD mutant fibroblasts) (24). This mutation in the LBD impairs transactivation and induction of 24-hydroxylase gene expression by 1,25-(OH)₂D₃ (24). Molecular modeling studies have shown that the alternate ligand binding pocket is formed with helix H2 (14); thus the mutation expressed in these LBD mutant fibroblasts might still permit 1,25-(OH)₂D₃ binding to the alternate pocket of the VDR (13). Photoprotection in the LBD mutant fibroblasts was not as strong as that observed in normal skin fibroblasts. However, there is considerable variation in protection by 1,25-(OH)₂D₃ between donors (7) so it is not possible to attribute the lesser protection in the LBD mutant fibroblasts to the LBD mutation or to other unknown factors. The data implicate the VDR as a critical component of the pathway that is essential for the photoprotective actions of 1,25-(OH)₂D₃ in attenuating UVR-induced thymine dimers. These results complement a previous study showing enhanced rapid chloride fluxes by 1,25-(OH)₂D₃, at physiological nanomolar concentrations, in homozygous and heterozygous normal VDR mouse osteoblasts whereas no effect was seen in osteoblasts obtained from VDR knockout mice (25). The data in the current study also indicate that the VDR does not require classical ligand binding or DNA binding to mediate this photoprotective response against DNA damage. Although these studies were carried out in human skin fibroblasts, due to the availability of fibroblasts with various VDR mutations, caution must be applied before extrapolating these results to skin cancer; nevertheless, they provide further evidence that the photoprotective effects of 1,25-(OH)₂D₃ are mediated via the nongenomic pathway.

The current findings also complement a previous study that showed that control cells bearing wild-type VDR and fibroblasts bearing a spontaneous homozygous point mutation in exon 7 of the LBD (W286R) of the VDR responded to 1,25-(OH)₂D₃ with a transient increase in calcium concentration, within 5–10 sec. This effect was blocked after incubation of wild-type and W286R mutant cells with verapamil, a calcium channel blocker (26). The W286R mutation does not reside in helix H2 and therefore should not affect ligand binding to the alternate pocket (13). However, the study also showed abolition of 1,25-(OH)₂D₃-induced calcium concentration in cells bearing a homozygous missense mutation in exon 2 of the DBD (K45E) of the VDR. Those results suggest that the DBD, particularly lysine 45, may be important for the effects of 1,25-(OH)₂D₃ on intracellular calcium.

In the current study, when normal fibroblasts were pretreated with Ab099, a neutralizing antiserum against ERp57, or transfected with ERp57 siRNA, 1,25-(OH)₂D₃ failed to protect the cells from UVR-induced thymine dimers. These results implicate ERp57 as a necessary component of protection by 1,25-(OH)₂D₃. This is supported by a recent study showing no rapid effects of 1,25-(OH)₂D₃ on calcium uptake in enterocytes isolated from genetically engineered mice with a targeted knockout of ERp57 (16).

Although the cell extracts used in the Co-IP studies were not pure plasma membrane preparations (which are lengthy to perform and risk destruction of native protein structures), there is considerable published evidence that the classic VDR and ERp57 can be present in the plasma membrane (9, 10, 17, 20). Co-IP studies showed that when the VDR was pulled down from nonnuclear extracts of fibroblasts, the ERp57 protein coimmunoprecipitated with the VDR. The same was true when ERp57 was pulled down: the VDR was directly associated with the precipitated ERp57 protein. To our knowledge, this is the first study describing a physical interaction between the VDR and ERp57 proteins. Interaction between these proteins may explain why different studies have shown key roles for the VDR and ERp57 in mediating nongenomic actions of 1,25-(OH)₂D₃ (18, 25). How ERp57 is involved in mediating these actions of 1,25-(OH)₂D₃ is not clear, but the mechanism may depend on ERp57 binding to other proteins. ERp57 has also been shown to associate with nuclear factor κB (27), major histocompatibility class I molecules (28), and calnexin (29).

p53 may help facilitate DNA repair after UVR and other damage (30). It was therefore surprising that, unlike keratinocytes (7), p53 was not up-regulated in the current study after UVR alone in fibroblasts. A possible explanation for this is that skin fibroblasts reside in the dermis and because UVB does not penetrate the entire dermal layer (31), fibroblasts are not normally exposed to UVB. 1,25-(OH)₂D₃ enhanced nuclear p53 expression in sham-irradiated fibroblasts, unlike keratinocytes, and this remained but was not further enhanced in normal fibroblasts exposed to UVR.

Of particular interest, p53 controls XPC and DDB2, two proteins involved in nucleotide excision repair, the primary repair pathway for CPD (32); therefore it seemed likely that the 1,25-(OH)₂D₃-induced up-regulation of p53 described previously (7) might contribute to the mechanism of reduction of CPD in fibroblasts. If p53 up-regulation were a key mechanism for the attenuation of thymine dimers after treatment with 1,25-(OH)₂D_3, we would have expected to see no increase in p53 expression induced by 1,25-(OH)₂D₃ in VDR-null fibroblasts or in normal fibroblasts pretreated with Ab099. However, enhanced p53 expression by 1,25-(OH)₂D₃ was observed under both these conditions. Although the fold increase in p53 varied somewhat, the increase was seen reliably, and some variation could be expected as a result of differences between donors (7).

This study identifies the contribution of the VDR and ERp57 in the photoprotective effects of 1,25-(OH)₂D₃ against UVR-induced DNA damage in human skin fibroblasts. The results of this study indicate that the VDR and ERp57 are directly associated outside the nucleus in which they mediate at least some of the nongenomic actions of 1,25-(OH)₂D₃.

Materials and Methods

Human neonatal foreskin samples from patients without mutations in their VDR were cultured for the processing of fibroblasts similar to that described previously by Gupta et al. (7) for keratinocytes except that cholera toxin, hydrocortisone, and epidermal growth factor were omitted from the culture medium and fibroblasts were cultured in DMEM containing 2 mm glutamine, 44 mm NaHCO₃, and 5% (vol/vol) fetal bovine serum. These studies were approved by the Human Research Ethics Committee of the University of Sydney and South Sydney Area Health Services. Skin biopsies were taken from patients with HVDRR after informed consent was given under a Stanford University human subjects institutional review board approved protocol. The HVDRR fibroblasts used in this study were from patients with the following mutations: 1) a nonsense mutation resulting in a premature stop codon at R50 (R50X) in the VDR (VDR-null) (22); 2) a missense mutation converting valine to methionine at amino acid 26 (V26M) in the DBD of the VDR (DBD mutant) (23); or 3) a 5-bp deletion/8-bp insertion deleting two amino acids (H141 and T142) and inserting three amino acids (L141, W142, A143) in helix H1 of the LBD of the VDR (24). For all studies using HVDRR fibroblasts, normal fibroblasts were also used as a positive control and treated exactly the same throughout the experimental process.

For immunohistochemistry and whole-cell ELISA, fibroblasts, at passages 2–15, were seeded into 96-well plates in fibroblast culture medium at a density of 20,000 cells per well, except for immunohistochemical studies using ERp57 siRNA, which were seeded at 8000 cells per well. For immunohistochemical experiments, fibroblasts were plated onto 5-mm poly-l-lysine (0.1 mg/ml) coated glass coverslips. Cells were plated in quadruplicate wells for each treatment. For knockdown of ERp57, fibroblasts were transfected with 50 nm ERp57 siRNA or control siRNA diluted in siRNA transfection medium (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) using equal volumes of siRNA transfection reagent (Santa Cruz Biotechnology) according to the manufacturer's instructions. Normal fibroblast growth medium was added to the transfection mixture 6 h after transfection, and cells were incubated for a further 18 h. Medium was replaced with Martinez buffer (145 mm NaCl, 5.5 mm KCl, 1.2 mm MgCl₂·6 H₂O, 1.2 mm NaH₂PO₄·2 H₂0, 7.5 mm HEPES, 1 mm CaCl₂·2 H₂0, 10 mm d-glucose) immediately before irradiation. Cells were irradiated using a UVA and UVB fluorescent lamp (NEC, Tokyo, Japan; and Phillips, Eindhoven, Holland, respectively), emitting 203 mJ/cm² UVB and 1169 mJ/cm² UVA, filtered through a cellulose triacetate sheet (Eastman Chemical Products, Kingsport, TN) to eliminate wavelengths below 290 nm (7). All cells were treated identically throughout the procedure; however, the sham wells were shielded during irradiation. Martinez was replaced with medium containing treatments, 1,25-(OH)₂D₃ (Cayman Chemical Co., Ann Arbor, MI) or JN, immediately after UVR. To neutralize the ERp57 protein, cells were incubated with Ab099 (1:500), a rabbit polyclonal antibody recognizing the N terminus of ERp57 (17), diluted in medium for 5 min at room temperature (RT). Normal rabbit serum (1:500) (Santa Cruz Biotechnology) was also used as a negative control in parallel wells. These cells were then treated with vehicle or 1,25-(OH)₂D₃.

For cell viability studies shown in the supplemental data, a Cell Titer-Blue assay was performed according to the manufacturer's instructions (Promega Corp., Madison, WI) 6 h after UVR.

Immunohistochemistry

For detection of thymine dimers, fibroblasts were fixed 3 h after irradiation. Immunohistochemistry was performed as described previously by Gupta et al. (7) except that cells were blocked with 50% horse serum (vol/vol) in PBS and four images were captured for each UV-irradiated coverslip. The number of total nuclei and the number of positively stained nuclei from each image were identified by the Zeiss KS400 image analysis (Carl Zeiss, Thornwood, NY) program using greyscale thresholding and mechanical editing. Positively stained nuclei were evident as a result of a greater density (7).

Co-immunoprecipitation

Normal fibroblasts grown to 95% confluence in a 75-cm² cell culture flask were used for co-immunoprecipitation studies. Flasks were allowed to equilibrate to RT for 30 min before being washed with PBS. Cells were lysed with 50 mm Tris (pH 7.5), 150 mm NaCl, 1 mm EDTA, 1% (vol/vol) Nonidet P40, 0.2 mm phenylmethylsulfonylfluoride, 2 μm Bestatin · HCl and 2 μm pepstatin A. Lysed cells were spun at 113 × g for 10 min, and half the supernatant was incubated at 4 C for 30 min with 10 μg/ml normal rabbit IgG-agarose conjugated and the other half with 10 μg/ml normal mouse IgG-agarose conjugated to preclear the lysates (Santa Cruz Biotechnology). The suspension was spun at 402 × g for 1 min, and the supernatant was incubated for 1 h at 4 C with 2 μg/ml VDR rabbit polyclonal antibody or 2 μg/ml ERp57 mouse monoclonal antibody (both from Santa Cruz Biotechnology). Protein A/G PLUS-agarose (20 μl) (Santa Cruz Biotechnology) was added to the suspensions and incubated overnight on a gentle rotating platform at 4 C. The next day, the suspension was spun five times at 402 × g for 1 min, and the pellet was resuspended in 0.5 ml cell lysis buffer, as described above, after each spin. The washed beads were resuspended in SDS-PAGE sample buffer and analyzed by Western blot.

Western blot

To determine knockdown of ERp57, fibroblasts were seeded in 24-well plates at a density of 10,000 cells per well in fibroblast culture medium. Fibroblasts were lysed with 50 mm Tris (pH 7.5), 150 mm NaCl, 1 mm EDTA, 1% (vol/vol) Nonidet P40, 0.2 mm phenylmethylsulfonylfluoride, 2 μm Bestatin·HCl, and 2 μm pepstatin A, 24 h after transfection with ERp57 siRNA. Lysed cells were spun at 113 × g for 10 min.

A bicinchoninic acid (BCA) assay was performed according to the manufacturer's instructions to determine the concentration of protein in each sample. Equal amounts of protein were combined with SDS-PAGE sample loading buffer and heated to 95 C for 5 min. A benchmark molecular mass protein ladder (Invitrogen, Carlsbad, CA) and the samples were loaded onto a 10% (wt/vol) polyacrylamide gel (pH 8.8) and 4% (wt/vol) polyacrylamide stacking gel (pH 6.8) and separated at 100 V in SDS-PAGE running buffer under nonreducing conditions. The protein was electrophoretically transferred onto a nitrocellulose membrane (GE Healthcare, Buckinghamshire, UK) at 4 C for 90 min at 100 V. After blocking with 1% (wt/vol) heat-denatured casein in PBS containing 0.4% (wt/vol) thymol for 30 min at RT the membrane was washed three times with PBS, pH 7.2, containing 0.1% (vol/vol) Tween 20 (PBS-T) and then incubated overnight at 4 C with the primary antibody diluted in 5% (wt/vol) BSA in PBS-T. To detect the classic VDR, a rabbit polyclonal VDR antibody was used at a dilution of 1 μg/ml; for ERp57 detection, a mouse monoclonal antibody raised against ERp57 was used at a dilution of 1 μg/ml (Santa Cruz Biotechnology). The next day, the membrane was incubated with a goat antirabbit (Santa Cruz Biotechnology) or a goat antimouse (Cell Signaling Technology, Beverly, MA) secondary antibody diluted in blocking solution. After the primary and secondary antibody stages the membrane was washed with PBS-T. The membrane was incubated for 5 min with ECL detection reagents (Millipore Corp., Billerica, MA), and bands were detected with a FluorChem SP Digital Imaging System.

Whole-cell ELISA

For whole-cell ELISA, fibroblasts were plated and subjected to irradiation and then treated as indicated above. Cells were briefly rinsed with 100 μl of PBS. Fixation of cells with 3.7% (vol/vol) formaldehyde in PBS for 10 min at RT occurred at 6 h after UVR for the detection of p53 (7). The cells were washed three times with 0.1% (vol/vol) Triton X-100 in PBS on the plate shaker for 2 min at RT to permeabilize the cells. H₂O₂ (1% vol/vol) in PBS was applied to quench endogenous peroxidases before the cells were incubated with a blocker [10% (vol/vol) FBS in PBS] for 1 h at RT. For detection of p53, cells were incubated overnight at 4 C with 4 μg/ml rabbit polyclonal IgG p53 primary antibody (Santa Cruz Biotechnology) diluted in blocker. Rabbit IgG isotype (4 μg/ml) (R&D Systems, Minneapolis, MN) was added to the negative control wells for incubation overnight at 4 C. On the following day, cells were incubated with 0.16 μg/ml goat antirabbit IgG-horseradish peroxidase antibody (Santa Cruz Biotechnology) diluted in blocker for 1 h at RT. Cells were washed three times with PBS and then three times with citrate phosphate buffer (50 mm citric acid and 100 mm NaH₂PO₄, pH 5). Color development was initiated by the addition of citrate phosphate buffer containing 0.1 mg/ml 3,3′,5,5′-tetramethylbenzidine and 0.015% (vol/vol) H₂O₂. The substrate was allowed to develop for 30–40 min at RT before being stopped by the addition of 0.5 vol of 2 m H₂SO₄. Absorbance was read at 450 nm on a Polarstar microplate reader (BMG Labtech Pty Ltd, Mornington, Australia). Unless otherwise stated, the different stages of the whole-cell ELISA were separated by washing the cells three times with PBS-T. Raw absorbance values were corrected with the isotype control to account for nonspecific binding of the antibody.

A BCA assay (Pierce Chemical Co., Rockford, IL) to correct for total cell protein was performed according to the manufacturer's instructions, at the time of fixation. It was necessary to correct for total cell protein because there was a significant decrease in the number of cells after UVR exposure compared with sham-irradiated cells.

Statistical analysis

All experiments using normal human fibroblasts were repeated a minimum of three times with three different donors. Studies with HVDRR fibroblasts, which were each from a single patient, were tested in at least three independent experiments and showed similar results on each occasion. Data are presented as normalized results from pooled independent experiments. In each experiment, values were based on quadruplicate wells. Significant differences between treatment groups were determined by one-way ANOVA followed by Tukey-Kramer test using GraphPad Instat statistical program (GraphPad Software, Inc., San Diego, CA).