Introduction

It is widely known that keratinocytes play a key role in the generation and activation of vitamin D31. The process is initiated when the epidermis is exposed to UVB light (280–320 nm) where 7-dehydrocholesterol (vitamin D3 precursor) is converted into pre-vitamin D3, which then isomerizes into vitamin D3 (cholecalciferol), or photoisomerizes to tachysterol3 (T3) or lumisterol3 (L3)2,3,4,5. Vitamin D3 is then hydroxylated by the enzyme cytochrome P450 family 2 subfamily R member 1 (CYP2R1) or cytochrome P450 family 27 subfamily A member 1 (CYP27A1), producing the 25-hydroxylated form (25(OH)D3). This process can occur both in the liver and to a lesser extent in keratinocytes6,7. The second hydroxylation, catalyzed by 1-alpha-hydroxylase (CYP27B1), converts 25(OH)D3 to the biologically active form of vitamin D3, 1,25-dihydroxyvitamin D3 (1,25(OH)2D3), also known as calcitriol8,9. Although this process takes place primarily in the kidneys, skin keratinocytes also contain CYP27B1 permitting the local synthesis of the active form of vitamin D3 directly in the skin8,10,11. Interestingly, the skin also expresses CYP24A1which is responsible for the catabolic inactivation of both 25(OH)D3 and 1,25(OH)2D3. The major pathway catalysed by this enzyme involves initial hydroxylation at C24 (C24-oxidation pathway). This leads to the formation of 24R,25-dihydroxyvitamin D3 (24,25(OH)2D3) or 1,24R,25-trihydroxyvitamin D3 (1,24,25(OH)3D3), which marks the beginning of vitamin D inactivation and is discussed further below12,13,14. 1,25(OH)₂D₃ controls numerous cellular processes by activating the vitamin D receptor (VDR)15. By binding to the VDR, 1,25(OH)2D3 enables its interaction with the retinoid X receptor (RXR) co-receptor and translocation of the VDR-RXR complex to the cell nucleus, thereby influencing the expression of over 3,000 genes in the human genome16,17. Keratinocytes are not only involved in the production and activation of vitamin D3 but also express the VDR. Most importantly, the active form of vitamin D3, 1,25(OH)2D3, together with calcium, regulates the process of renewal of the epidermal barrier1. When the VDR is absent or when the production of 1,25(OH)₂D3 is impaired (due to CYP27B1 mutations or deletion), the process of differentiation of epidermal keratinocytes is disrupted18,19. Eventually, this leads to excessive proliferation of keratinocytes, affecting both the formation of the skin’s permeability barrier and the innate immune response18. In addition to its local effects on the skin, vitamin D3 also has several systemic effects such as on the proper functioning of the musculoskeletal, immune, nervous and cardiovascular systems20,21,22,23. Vitamin D3 and its natural or synthetic metabolites show anti-cancer effects24, and vitamin D sufficiency as measured by the 25(OH)D3 concentration in blood, is strongly associated with a reduction in the incidence and severity of many types of cancer.25,26,27 Vitamin D3 not only modulates the activity of cancer cells by slowing proliferation and migration of cells, but also enhances the efficiency of anticancer drugs28,29,30. Despite the many beneficial properties of vitamin D3, its potential use in multiple diseases is limited due to its hypercalcemic properties31,32,33. Therefore, its low calcemic analogs which retain similar anticancer effects on cells to 1,25(OH)2D3, are under intensive investigation. Some of these analogs are already used in various medical fields, such as endocrinology, nephrology, and dermatology34,35,36.

1,24,25(OH)3D3 is the first metabolite in the inactivation of 1,25(OH)2D3 by the C24 oxidation pathway catalyzed by CYP24A1, ultimately leading to the excretory product, calcitroic acid12,13,14,37,38. 1,24,25(OH)3D3 can also be produced by the CYP27B1-catalyzed 1α-hydroxylation of 24,25(OH)2D3, the initial product of CYP24A1 action on 25(OH)D3 12. In vitro kinetic evidence indicates that 1,24,25(OH)3D3 accumulates (i.e. is released from the active site of CYP24A1) before it undergoes further CYP24A1-dependent oxidations leading to calcitroic acid12. Its concentration in human serum, measured by radioimmunoassay, was reported to be 9.3–18.5 pM 39. More recent measurements in human serum by mass spectrometry have given a mean concentration of approximately 35 pM, being 35–40% of the 1,25(OH)2D3 concentration and decreasing with kidney disease40,41. Early studies measuring competitive binding to the chick intestinal receptor (chick VDR) revealed that the relative binding of 1,24,25(OH)3D3 was 0.39-times that for 1,25(OH)2D3, and while it was still effective in promoting intestinal calcium absorption and bone calcium mobilization in the chick, it displayed lower potency than 1,25(OH)2D342. In rats, 1,24,25(OH)3D3 was found to have 20-fold less antirachitic activity than 1,25(OH)2D343. Another study in rats showed that it was 93% as active as 1,25(OH)2D3 for stimulating calcium transport in the intestine but had little activity in stimulating calcium mobilization from bone44. However, 1,24,25(OH)3D3 displayed comparable efficacy and potency to 1,25(OH)2D3 for stimulating the differentiation of HL-60 cells into monocytes/macrophages, with CYP24A1-derived metabolite activity only being lost after cleavage of the side chain between C23 and C24 . Our recent studies showed that 1,24,25(OH)3D3 displays anti-cancer properties on melanoma cell lines very similar to 1,25(OH)2D3, consistent with hydroxylation of the later at C24 by CYP24A1 not removing its activity46. To further define the biological activity of 1,24,25(OH)3D3 we investigated its genomic activity on human keratinocytes derived from healthy donors, using proliferation and VDR receptor translocation assays, as well as transcriptomics.

Results

1,24,25(OH)3D3 causes a comparable reduction in the proliferation of HPEKp keratinocytes to 1,25(OH)2D3

The antiproliferative effects of the hormonally active form of vitamin D3, 1,25(OH)2D3 and its first catabolite by the C24-oxidation pathway, 1,24,25(OH)3 D3, on the HPEKp keratinocyte cell line, were examined. The cells were treated with the secosteroids at concentrations of 1 nM, 10 nM and 100 nM, and their growth was recorded for 72 h (Fig. 1). Machine learning was used to count the visible cell nuclei without staining, in order to calculate the proliferation rate. Both secosteroids at the highest concentration tested, 100 nM, significantly inhibited the proliferation of HPEKp keratinocytes and their efficacies were similar (Fig. 1A). Therefore, in our subsequent experiments the two vitamin D3 derivatives were used at 100 nM, corresponding to the typical serum level of 25(OH)D3 (75–125 nM)47 which is used in the clinic as a biomarker for vitamin D3 status48. However, it has to be emphasized that the serum concentration of 1,25(OH)2D3 is at least 500 times lower49. On the other hand, keratinocytes can activate 25(OH)D3 through its hydroxylation in position C1 50 so the intracellular concentration of 1,25(OH)2D3 in these cells could be substantially higher than in the serum.

Fig. 1
figure 1

The effect of 1,25(OH)2D3 and 1,24,25(OH)3D3 on the proliferation of primary human epidermal keratinocytes (HPEKp). The cells were treated with 1, 10 or 100 nM of 1,25(OH)2D3 or 1,24,25(OH)3D3 for 72 h. Cell nuclei were marked via artificial intelligence for the calculation of the proliferation potential which was calculated using Olympus cellSens Software. (A), histogram showing the results for all concentrations of secosteroids tested. Statistical significance was determined using One-Way Anova and presented as ** p < 0.005, *** p < 0.0005 vs. control (B), images showing the results for the control (untreated cells), 100 nM 1,25(OH)2D3 and 100 nM 1,24,25(OH)3D3 for times 0 and 72 h, respectively.

Full size image

Both 1,25(OH)2D3 and 1,24,25(OH)3D3 stimulate VDR translocation to the nucleus of HPEKp keratinocytes

Our previous studies demonstrated that 1,24,25(OH)3D3 displays anticancer activities on human melanoma cell lines46, presumably via activation of the VDR. Here, stimulation of the translocation of VDR into the nucleus of HPEKp keratinocytes treated with 1,24,25(OH)3D3 was observed, and the effect was comparable to that of 1,25(OH)2D3 (Fig. 2). Increased translocation of the VDR to the nucleus was observed after just 1 h of incubation with either secosteroid compared to untreated cells. After 24 h of incubation of HPEKp keratinocytes with 1,25(OH)2D3, a 2-fold increase in the nuclear to cytoplasm ratio of VDR was observed, while for 1,24,25(OH)3D3, it was about 1.6 fold. (Fig. 2A). The effect was statistically significant for both 1,25(OH)₂D₃ and 1,24,25(OH)₃D₃ for all incubation time-points tested. Representative images after 24 h of incubation are shown in (Fig. 2B). Interestingly, this effect was weaker than in the previously studied A375 and SK-MEL28 melanoma cells, but it should be noted that the relative colocalization of VDR with the nucleus was higher in untreated - control keratinocytes than in untreated melanoma cells.

Fig. 2
figure 2

The effect of 1,25(OH)2D3 and 1,24,25(OH)3D3 on vitamin D receptor (VDR) translocation from the cytoplasm to the nucleus. Primary human epidermal keratinocytes (HPEKp) were treated with 100 nM 1,25(OH)2D3 or 1,24,25(OH)3D3 for 1, 4 or 24 h. (A), Histogram of VDR protein translocation to the nucleus of cells after the specified incubation times, presented as nucleus/cytoplasm ratio (mean of fold change +/- SD) and calculated from the fluorescence intensities. The statistical significance was estimated using a t-test and presented as **** p < 0,00001 vs. control. (B), Fluorescently labelled VDR protein and DAPI-stained nuclei images after 24 h of incubation with 100 nM 1,25(OH)2D3 or 1,24,25(OH)3D3. Superimposed images are on the right.

Full size image

1,24,25(OH)3D3causes a partially different effect on the gene expression profile of HPEKp compared to 1,25(OH)₂D₃

Our preliminary studies showed that the VDR-RXR complex in HPEKp cells reached the highest level of nuclear co-localization after 4 h of incubation with 1,25(OH)2D3 51. To compare the effects of 1,25(OH)2D3 and 1,24,25(OH)3D3 on gene expression in detail, two time points were selected: a short (4 h) and prolonged incubation (24 h). The following identification criteria were used: adjusted p-value < 0.05 and logFoldChange > 0.58 (corresponding to fold change > 1.5, fold change < -1.5). The PCA graph shows a similar transcriptome profile for HPEKp keratinocytes treated with 1,25(OH)2D3 or 1,24,25(OH)3D3 for both time points (4 h and 24 h) (Figure S1A) which is also confirmed by the heatmap of the distance matrix (Figure S1B).

In HPEKp cells treated with 1,25(OH)2D3 for 4 h, 141 differentially expressed genes (DEGs) were identified including only 7 DEGs with reduced expression and 134 DEGs with increased expression compared to the untreated control (Fig. 3A). In addition, a significant increase in the number of DEGs was observed in keratinocytes treated with 1,25(OH)2D3 for 24 h, 509 DEGs were detected, including 232 downregulated and 277 upregulated compared to the control. The same analysis was also performed for 1,24,25(OH)3D3. After incubation of HPEKp keratinocytes with this metabolite for 4 h, 130 DEGs were identified, including 10 downregulated and 143 upregulated compared to the untreated control. Similarly, an increase in the total number of DEGs was observed after prolonged (24 h) incubation of HPEKp keratinocytes with 1,24,25(OH)3D3. A total of 603 DEGs were detected, of which 307 were downregulated and 296 were upregulated compared to untreated cells (Fig. 3B). Interestingly, the ratio between upregulated and downregulated DEGs was comparable for both secosteroids at both time points.

Fig. 3
figure 3

The effect of incubation time with 1,25(OH)2D3 or 1,24,25(OH)3D3 on gene expression patterns of HPEKp cells. Venn diagrams show the distribution of the DEGs from HPEKp cells treated for 4 and 24 h with: (A) 100 nM 1,25(OH)2D3 (4 h, 29-up and 4 downregulated; 24 h, 172-up and 229 downregulated; see supplementary Table S1 and S2 for more details) or (B) 1,24,25(OH)3D3 (4 h, 8-upregulated; 24 h, 182-up and 299 downregulated; see supplementary Table S3 and S4 for more details). (C). Histograms show gene expression measured by real-time PCR for CYP24A1, CD14, ADGRD2, IGFN1, and SHE after 4 and 24 h of incubation with 1,25(OH)2D3 or 1,24,25(OH)2D3, vs. untreated control. Data are shown as means +/- SD of the experiment carried out in duplicate. * p < 0.05, ** p < 0.01, *** p < 0.001 vs. control (C).

Full size image

There were 108 DEGs in common between HPEKp keratinocytes treated with 1,25(OH)2D3 for 4 h and 24 h (Fig. 3A). The extension of the incubation of HPEKp cells with 1,25(OH)2D3 from 4 h to 24 h resulted in a marked (difference in log2fold change > 2) increase in the expression of genes UCA1, CD14, IGFN1, MMP3, TMEM156, IL1RL1, RIPOR3, SERPINB1 and uncharacterized LOC124907878.

Treatment of HPEKp cells with 1,24,25(OH)3D3 allowed the identification of 130 DEGs after 4 h and 603 after 24 h of treatment, including 122 common to both incubation times. Moreover, the treatment with 1,24,25(OH)3D3 for 4 h affected a lower number of DEGs than treatment with 1,25(OH)2D3 (8 versus 33), but the ratio was reversed after 24 h (481 versus 401) (Fig. 3B). Comparison of DEGs at two time points (4 h versus 24 h) for of HPEKp keratinocytes treated with 1,24,25(OH)3D3 revealed that the genes UCA1, SPP1, SERPINB9, RIPOR3, PADI2, MMP3, IL1RL1, IGFN1, CYP24A1, CD14 and uncharacterized LOC124907878 displayed a log2fold change greater than 2. The long-noncoding RNA (IncRNA) LOC124907878, which according to www.genecards.org is expressed in male germline stem cells was overexpressed in response to both 1,24,25(OH)3D3 and 1,25(OH)2D3, which might contribute to their pro-differentiating effect.

The RNA-seq data were verified by qPCR analysis of the expression of selected genes (Fig. 3C). CYP24A1 is a classical target gene for the VDR and it showed the largest log2fold change for both 1,25(OH)2D3 and 1,24,25(OH)3D3 in the RNA sequencing data. The expression of this gene at the RNA level measured by qPCR similarly showed a large increase at 4 h and 24 h for both 1,25(OH)2D3 and 1,24,25(OH)3D3 compared to untreated control keratinocytes. CD14, ADGRD2, IGFN1 and SHE were other genes that showed a large change in expression in response to 1,25(OH)2D3 and 1,24,25(OH)3D3 from the RNA sequencing data. The increased expression of these genes in response to the two secosteroids was also confirmed by qPCR where the increase was statistically significant for both 1,25(OH)2D3 and 1,24,25(OH)3D3 when HPEKp keratinocytes were cultured for 4–24 h, as shown in (Fig. 3C). The CD14 protein known as Cluster of Differentiation 14 is responsible for the activation of the innate immune response and our results showed that it was strongly upregulated by both 1,25(OH)2D3 and 1,24,25(OH)3D3, indicating a similarity in the mechanism of the immune effects of these two secosteroids52. The SHE gene encodes an adaptor protein that has a role in heart development, and stimulation of its expression is consistent with a role of both 1,25(OH)2D3 and 1,24,25(OH)3D3 in mediating HPEKp cell differentiation53. There is currently no information on the regulation ADGRD2 expression by vitamin D3, but as a member of the G protein-coupled receptor (GPCR) family, it contributes to cell adhesion and intercellular signaling54, potentially affecting processes such as cell differentiation and migration. Similarly, IGFN1 has not previously been identified as a target of vitamin D, but its encoded protein is involved in muscle cell fusion and differentiation and it plays a key role in maintaining the structural integrity of muscle fibers55.

There is strong overlap between the genes regulated by 1,25(OH)2D3 and 1,24,25(OH)3D3 with 107 common genes being noted at 4 h (Fig. 4A) and 403 at 24 h (Fig. 4B) of incubation. Based on Gene Ontology (GO) enrichment analysis, we identified 16 statistically significant common processes affected by both 1,25(OH)2D3 and 1,24,25(OH)3D3 (Fig. 4C). This analysis confirms the broad effect of 1,24,25(OH)3D3 on cellular processes that has been demonstrated by years of research on 1,25(OH)2D3. In addition to the effect on signaling via GPCR receptors, which influences far-reaching physiological processes such as neurotransmission, cellular metabolism, secretion and cell growth, immunological processes are also affected. These include signaling by interleukins, especially IL3, IL14, IL10, as well as signaling by interferon alpha and beta. Interestingly, keratinocyte-related processes, which are one of the key elements of 1,25(OH)2D3 activation such as keratinization and keratinocyte differentiation, are more strongly activated by 1,24,25(OH)3D3 compared to 1,25(OH)2D3 (Fig. 4C). Other important processes affected by the by the two secosteroids in relation to the functioning of cells in both the physiological and pathological states, are the regulation of IGF, protein phosphorylation and the organization of the extracellular matrix.

Fig. 4
figure 4

Comparison of the effects of 1,25(OH)2D3 and 1,24,25(OH)3D3 on gene expression patterns in HPEKp cells. Venn diagrams show the distribution of the DEGs from HPEKp cells treated for: (A) 4 h (1,25(OH)2D3, 29-up and 5 downregulated; 1,24,25(OH)3D3 18-up and 5 downregulated; see supplementary Table S5 and S6 for more details) or (B) 24 h (1,25(OH)2D3, 40-up and 66 downregulated; 1,24,25(OH)3D3, 59-up and 141 downregulated; see supplementary Table S7 and S8 for more details) with 100 nM 1,24,25(OH)3D3 and 1,25(OH)2D3. (C) Comparison of the effects of 24 h incubation with 1,25(OH)2D3 or 1,24,25(OH)3D3 on the fold enrichment of selected biological processes based on transcriptomic data.

Full size image

Some genes are regulated solely by just one of the two secosteroids

The incubation of HPEKp keratinocytes with 1,25(OH)2D3 for 4 h affected the expression of 141 genes while treatment with 1,24,25(OH)3D3 led to the change in expression of 130 genes and, as noted above, 107 genes were in common (Fig. 4A). There were 34 (29 up-regulated and 5 down-regulated) genes where expression was only affected by 1,25(OH)2D3 and 23 (18 upregulated and 5 downregulated) solely regulated by 1,24,25(OH)3D3. When cells were incubated for 24 h, 509 DEGs were detected for 1,25(OH)2D3 and 603 for 1,24,25(OH)3D3 with 403 genes being in common (Fig. 4B). There were 106 (40 upregulated and 66 downregulated) DEGs observed for just 1,25(OH)2D3 and 200 (59 upregulated and 141 downregulated) only associated with 1,24,25(OH)3D3 treatment.

The classical 1,25(OH)2D3 target gene, CYP24A1, showed the highest total log2FC for both 1,25(OH)2D3 and 1,24,25(OH)3D3 regardless of incubation time. In addition, other genes were also characterized by high expression in response to both secosteroids including EPHA10, ADGRD2, CD14, LGI3 TRPV6, IGFN1, IL1RL1, PARM1, LGI3 and BMP6. The most downregulated genes were EID3, SPRR2A, SFTA1P, SERPINB4, LY6G6C, BPIFC, IGF2, MMP10 and FLG. For the common genes whose expression was influenced by both 1,25(OH)2D3 and 1,24,25(OH)2D3, the direction of change was always the same, either upwards or downwards. Opposing effects were not observed. The range of differences in fold change for 1,25(OH)2D3 versus 1,24,25(OH)3D3 was between − 0.46 and 0.82 for 4 h and − 3.56 and 3.33 for 24 h.

The GO analysis of the upregulated genes dependent only on 1,24,25(OH)3D3 using the Reactome revealed the pathways with the highest p-value were somitogenesis, paraxial mesoderm formation and gastrulation. However, analysis of downregulated genes after 1,24,25(OH)3D3 treatment identified the processes with the highest p-value as being “Formation of the cornified envelope” and “Differentiation of Keratinocytes in interfollicular epidermis of mammalian skin” (supplementary data). Interestingly, 1,24,25(OH)3D3 decreased loricrin gene expression by more than 7 log2FC, with the loricrin protein being involved in the later stages of keratinocyte differentiation56.

1,25(OH)2D3and 1,24,25(OH)₃D₃ induce the opposite expression pattern for a group of cancer-related genes in HPEKp cells compared to the resulting cancer cells

It is well documented that 1,25(OH)2D3 has anti-cancer properties, as demonstrated in various cell and tissue models. Recently, we reported that treatment of the A431 squamous cell carcinoma (SCC) cell line with 1,25(OH)2D3 partially reversed the expression of head and neck squamous cell carcinoma driver genes57. In the current research we compared the list of DEGs detected in HPEKp treated with 1,25(OH)2D3 or 1,24,25(OH)3D3 for 24 h with the DEGs database for head and neck squamous cell carcinoma (HNSCC). Of the 158 genes common to HNSCC and HPEKp treated with 1,25(OH)2D3, the direction of the change in expression of 82 of them was opposite. Interestingly, there were almost 2 times more DEGs (327) for 1,24,25(OH)3D3 (compared to 1,25(OH)2D3) shared by HPEKp and HNSCC. This included 152 DEGs whose direction of change in expression were opposite (Fig. 5). Genes whose expression was upregulated with variability in the range of absolute Δlog2FC >4 in the HNSCC but downregulated after treatment of HPEKp with 1,24,25(OH)3D3 were: CXCL11, FBN2, FN1, GDPD2, IL24, ISG15, KANK4, LEMD1, MMP10, MMP9, NETO1. POSTN, PTHLH, RSAD2, RTP4, SCG5, SFTA1P, STC2 and SULT1E1. The highest change (absolute Δlog2FC) in the expression was observed for MMP10, which was 6.49. On the other hand, the expression of genes such as ATP8A1, CIDEA, CLDN11, CYP3A5, DPF3, DTNA, FOS, HRCT1, KLK13, KRT31, LNX1, PADI1, PADI2, SERPINB1, SLC1A1, TNNI2 and TRPV6 was downregulated in HNSCC but upregulated in HPEKp treated with 1,24,25(OH)3D3, with the highest difference being 10.16 in Δlog2F.

Fig. 5
figure 5

The distribution of the genes between HHSC from the TACCO database and HPEKp cells treated for 24 h with 1,25(OH)2D3 or 1,24,25(OH)3D3.

Full size image

Discussion

1,24,25(OH)3D3 is the first metabolite of the C24 oxidation pathway of deactivation of 1,25(OH)2D3 and displays weaker effects than 1,25(OH)2D3 on calcium metabolism in chickens and rats, but comparable effects to 1,25(OH)2D3 on HL-60 cell differentiation45. Our recent studies using a melanoma model have shown that it has comparable anti-cancer properties to 1,25(OH)2D3 46. The current research aimed to compare the biological activity and transcriptional changes in cells between treatment with 1,25(OH)2D3 and 1,24,25(OH)3D3 using the well-established model of human primary epidermal keratinocytes (HPEKp).

The present study confirms that 1,24,25(OH)3D3 exhibits antiproliferative activity comparable to that of 1,25(OH)2D3. Consistent with this, our study also shows that both 1,25(OH)2D3 and 1,24,25(OH)3D3 stimulate the translocation of the VDR from the cytoplasm to the nucleus, with approximately a 3-fold increase observed in a time-dependent manner, implicating the VDR in the mechanism of action of 1,24,25(OH)3D3. Chen and co-workers have shown that 1,25(OH)2D3 exerts a strong antiproliferative effect on keratinocytes by arresting the cell cycle at both G0/G1 and G2/M phase58. Similarly, another report showed that 20(OH)D3, effectively suppresses keratinocyte proliferation by inhibiting NF-KB activity, thereby reducing the expression of proinflammatory and proliferative genes59.In terms of our GO analysis, processes regulated by both 1,25(OH)2D3 and 1,24,25(OH)3D3 that may be involved in the inhibition of proliferation include Signalling by GPCR, GPCR ligand binding, NGF-stimulated transcription and Post-translational protein phosphorylation. Inhibition of the proliferation of keratinocytes by 1,25(OH)2D3 is commonly associated with promotion of differentiation58. It is well established that 1,25(OH)2D3 promotes keratinocyte differentiation and supports the renewal of the epidermal barrier18,60,61. In line with this, our transcriptomic data reveals that both 1,25(OH)2D3 and 1,24,25(OH)3D3 strongly enrich processes such as “keratinization” and “differentiation of keratinocytes in the interfollicular epidermis in mammalian’’ suggesting a potentially important role for 1,24,25(OH)3D3 in epidermal maturation and barrier formation. To our knowledge, this is the first report demonstrating that 1,24,25(OH)3D3 may actively contribute to these skin-specific differentiation processes, pointing to its previously underappreciated biological function and warranting further investigation.

A key finding of our study is that that in addition to modulating the expression of a substantial number of shared DEGs, both 1,25(OH)2D3 and 1,24,25(OH)3D3 exert some compound-specific regulatory effects on a distinct subset of DEGs. Among the 106 DEGs regulated exclusively by 1,25(OH)2D3, but not by 1,24,25(OH)3D3, Reactome pathway analysis revealed their involvement in several biological processes, including Response of EIF2AK1 (HRI) to heme deficiency, Cam-PDE 1 activation, and RUNX3 Regulates Immune Response and Cell Migration. Notably, the CAM-PDE 1 activation pathway is directly associated with calcium metabolism. One of its key genes, Phosphodiesterase 1 C (PDE1C), which hydrolyzes both cAMP and cGMP62, was significantly downregulated following 24 h of 1,25(OH)2D3 treatment (fold change: -1,29). However, other calcium-related genes, such as TRPV6 (a component of the Mineral absorption according to the KEGG database63,64,65 and annotated in UniProt with the Gene Ontology term66,67 as a calcium-selective epithelial channel), RYR1 (which mediates calcium release from the sarcoplasmic reticulum in skeletal muscle and is a central gene in the calcium channel pathway per Reactome) and SPP1 (which plays a key role in calcium binding)68 showed highly similar transcriptional effects. The expression patterns of these genes were consistent in both the direction and magnitude of changes in expression after treatment with each of these compounds. The GO analysis of the upregulated genes dependent only on 1,24,25(OH)3D3 using the Reactome revealed the major pathways affected were somitogenesis, paraxial mesoderm formation and gastrulation, but at this stage it is difficult to relate these effects to keratinocyte biology. However, analysis of genes that were downregulated by 1,24,25(OH)3D3 treatment identified the processes “Formation of the cornified envelope” and “Differentiation of Keratinocytes in interfollicular epidermis of mammalian skin”. This suggests that 1,24,25(OH)3D3 may further modulate the differentiation of keratinocytes to form the cornified envelope over the regulation exerted by 1,25(OH)2D3, with 1,24,25(OH)3D3 markedly decreasing loricrin gene expression, possibly slowing the later stages of keratinocyte differentiation56.

It is interesting to consider the mechanism by which 1,25(OH)2D3 and 1,24,25(OH)3D3 can each regulate an independent set of genes besides the ones in common. Some natural vitamin D metabolites have been reported to work, at least in part, via other nuclear receptors besides the VDR and hence they regulate a partially different set of genes to those regulated by 1,25(OH)2D3 69. For example, 20,23(OH)2D3 can exert biological effects by acting as an agonist on the aryl hydrocarbon receptor (AhR)16 while 20(OH)D3 can act as an inverse agonist on retinoid-related orphan receptors alpha and gamma16. However, pathway analysis of our transcriptomics data did not identify other nuclear receptor-mediated pathways for either 1,25(OH)2D3 or 1,24,25(OH)3D3. This suggests that the major effects of both 1,25(OH)2D3 and 1,24,25(OH)3D3 are mediated via the VDR with 1,24,25(OH)3D3 displaying biased agonism, meaning that it is functionally selective for certain response pathways16.

Recently, we showed that active forms of vitamin D influence mitochondrial structure and bioenergetics in normal and malignant keratinocytes70. We documented also, that it modulates the mitochondrial membrane potential in human melanoma cells71. Furthermore, the crucial role of 1,25(OH)2D3 and its low-calcemic analog 20,23(OH)2D3 72 in regulating mitochondrial functions has also been supported by Gene Set Enrichment Analysis. 16 However, our transcriptomic data presented here do not indicate activation of mitochondrial-related pathways. This absence of enrichment may be caused by the fact that many stem cells, including progenitor cells tested in our model, typically exhibit low mitochondrial content and a reliance on mitochondrial-independent glycolytic metabolism for energy73. In line with this hypothesis, it has been demonstrated that A431 squamous carcinoma cells and non-transformed HaCaT keratinocytes differ markedly in mitochondrial morphology, respiratory activity, membrane potential and reactive oxygen species generation70.

Carlberg and colleagues demonstrated a strong impact of 1,25(OH)2D3 on the immune system74. In our study among the 25 most significantly enriched relevant signaling pathways sorted by p-value, 8 can be assigned to the immune response, including interferon, interleukin, and cytokine signaling. Of 61 vitamin D3 target genes previously identified as part of key eight major pathways of innate immunity in peripheral blood mononuclear cells (PBMCs)75, our data showed that after 4 h of incubation 1,25(OH)2D3 modulated the expression of 7 genes, while 1,24,25(OH)3D3 influenced 6, with 5 genes overlapping (NFKBIA, FOSC2, BCL3, CD14 and THBS1). After 24 h, the number of DEGs increased to 13 for 1,25(OH)2D3 and 11 for 1,24,25(OH)3D3, with 10 shared targets identified (NFKBIA, FOSL2, TNFAIP3, S100A8, S100A9, CD14, THBS1, PADI2, H2AC6 and SLC25A4). The increasing number of DEGs over time supports the biological activity of 1,25(OH)2D3 and further demonstrates that 1,24,25(OH)₃D₃ exerts similar effects on key processes involved in immune system regulation. We cannot rule out that some of the changes in expression seen at 24 h compared to 4 h are due to metabolites of 1,25(OH)2D3 and 1,24,25(OH)3D3 rather than the compounds themselves, being produced for example by CYP24A1, CYP3A4 or 25-hydroxyvitamin D 3-epimerase13. However, we and others have proposed that secondary transcriptional changes may instead result from the activation of intracellular processes involving other transcription factors or long non-coding RNA. These possibilities have been investigated, with an elegant explanation provided by Carlberg74. Thus, while early transcriptional responses (after 4 h) are likely direct VDR–mediated effects, changes observed at later time points (24–72 h) may reflect secondary downstream processes influenced by vitamin D or by metabolites of 1,25(OH)2D3 and 1,24,25(OH)3D3.

Squamous cell carcinoma (SCC) and basal cell carcinoma, referred to as non-melanoma skin cancers, are the most common malignant tumors in humans and both originate from mutated epidermal keratinocytes76,77,78. Several studies in animals have demonstrated that vitamin D3 hydroxyderivatives can attenuate UVB or chemically induced epidermal cancerogenesis and inhibit the growth of SCC79. Our current data indicate that 1,25(OH)2D3 and 1,24,25(OH)3D3 regulate a set of genes whose expression is upregulated in HHSC, serving to oppose this and downregulate their expression. Furthermore, other analogues of vitamin D3 such as 20(OH)D3 and 20,23(OH)2D3 reduce the production of reactive oxygen species and NO induced by UVB in cells of epithelial origin, such as keratinocytes including normal human keratinocytes (HaCaT), as well as neural crest derived melanocytes. This could contribute to the reduction of the incidence of these types of cancers by protecting the DNA80. Results of another study investigating the role of vitamin D3 indicate that it increases the expression of Suppression Of Tumorigenicity 2 (ST2) at the mRNA level, encoded by the IL1RL1 gene, in keratinocytes and melanocytes. These findings suggest an important function of interleukin 33 (IL-33) as a ligand for ST2 whose mRNA level is reduced in psoriatic lesions81. Interestingly, our studies indicate a slightly greater upregulation of IL1RL1 expression with 1,24,25(OH)3D3 compared to 1,25(OH)2D3 (log2FC 10.11 vs. 9.97), which may make 1,24,25(OH)3D3 a potential candidate for further testing for psoriasis treatment. Such a study should include further investigation of the calcemic activity of 1,24,25(OH)3D3 due to the potential risk of hypercalcemia82,83,84,85. On current evidence, it shows reduced potency for regulating calcium metabolism compared to 1,25(OH)2D3 42,43,44.

This is the first comprehensive report demonstrating that 1,24,25(OH)3D3, traditionally considered a catabolic product in the vitamin D C24-deactivation pathway, exhibits biologically relevant activity in skin cells. Despite reports of its diminished calcemic activity, we found that 1,24,25(OH)3D3 regulates gene expression, modulates immune-related and differentiation pathways, and inhibits keratinocyte proliferation in a manner comparable to 1,25(OH)2D3. These studies are consistent with the equipotent activities of 1,24,25(OH)3D3 and 1,25(OH)2D3 reported for the stimulation of HL-60 cell differentiation45 and indicate that 1,24,25(OH)3D3 should be considered as a functionally relevant signalling molecule rather than as an inactivation product, with final 1,25(OH)2D3 inactivation only occurring following further CYP24A1-catalysed oxidations, especially side chain cleavage between C23 and C24 45. Further studies are warranted to explore the therapeutic potential and safety profile of 1,24,25(OH)3D3, particularly in the context of chronic inflammatory skin diseases and skin cancers. It should be noted that supraphysiological concentrations (100 nM) of both vitamin D metabolites were used in the present study as is widely used for in vitro studies to ensure detectable effects (see recent review86). Nevertheless, the in vivo relevance of these findings remains to be established. Interestingly, Milani at al. compared the transcriptional effects of 1,25(OH)2D3 at the concentrations of 0.5 nM and 100 nM in breast cancer organotypic culture. Both concentrations of 1,25(OH)2D3 had effects on genes expression, however considerably more DEGs were observed at the higher concentration (7 vs. 136 upregulated genes with 5 shared genes)87. Thus, it seems that 100 nM concentration of vitamin D analogs although supraphysiological, enables a comprehensive set of responsive genes to be identified for subsequent physiological studies. While the serum concentration of 1,24,25(OH)3D3 is low, lower than that of 1,25(OH)2D3 40,41, its intracellular concentration in keratinocytes could potentially be much higher given that it cannot only be produced by hydroxylation of 1,25(OH)2D3 at C24 by CYP24A1, but also by 1α-hydroxylation of 24,25(OH)2D3 by CYP27B188. 24,25(OH)2D3 is the major dihydroxyvitamin D3 species present in serum, being in the nM range, well above that for 1,25(OH)2D3 13.

In conclusion, this is the first report revealing the transcriptional activity of the primary catabolite of 1,25(OH)2D3, namely 1,24,25(OH)3D3 on keratinocytes. As well as sharing many DEGs with 1,25(OH)2D3, it also transcriptionally regulates a unique set of genes. These properties and its low calcemic activity compared to 1,25(OH)2D3 give this CYP24A1-derived secosteroid therapeutic potential which warrants further investigation. Finally, a functional validation of these transcriptomic findings, including studies on keratinization and immune signalling, remain to be carried out and will be the target of our future research.

Materials and methods

Cell culture

Human Primary Epidermal Keratinocytes (HPEKp) from pooled juvenile donors (a mixture of primary keratinocytes derived from at least three distinct donors, thus not representing a single clonal cell line) were purchased from CELLnTEC (Bern, Switzerland). The cells were cultured in Epidermal Keratinocyte Medium (CnT-07, CELLnTEC, Bern, Switzerland) containing a supplement mix (A, B, C), bisphenol A (BPE) and gentamycin, with low calcium (0.07 mM). TrypLE™ Express solution (Gibco, Life Technologies, Waltham, MA) was used for trypsinization when cells were at about 80–90% confluency, and cells from passages 3 and 4 were used for experiments.

Assay of cell proliferation

Cells were seeded at 10,000 per well on 8-well imaging chamber slides (MoBiTec Molecular Biology, Goettingen, Germany). Cells were cultured overnight and then treated with a serial dilution of 1,25(OH)2D3 or 1,24,25(OH)3D3 in DMEM medium supplemented with 2% charcoal-stripped FBS (Fetal Bovine Serum) and 1% penicillin/streptomycin. Both 1,25(OH)2D3 and 1,24,25(OH)3D3 treatments were performed under identical solvent conditions, with a final ethanol concentration below 0.02%. The slide was then placed in the incubation chamber of the microscope for live imaging at 37 °C under controlled humidity and an atmosphere of 5% CO2, for the whole experiment. An Olympus Cell Vivo IX 83 (Japan) was used for time-lapse experiments. Analysis of the proliferation rate was carried out via live imaging and images were collected every 30 min over 72 h. The results were calculated using Olympus cellSens software, normalized to 1.0 at the beginning of the experiments (n = 3).

Measurement of VDR translocation to the nucleus

The cells were seeded at a density of 10,000 per well on an 8-well chamber slide. After 24 h when the cells had formed a monolayer, they were treated with 100 nM 1,25(OH)2D3 or 1,24,25(OH)3D3), diluted in DMEM medium supplemented with 2% charcoal-stripped FBS and 1% penicillin/streptomycin. After incubation for 1–4 h, the slides were washed three times for 5 min and then fixed for 10 min with a 4% paraformaldehyde solution in PBS. After three further washes (5 min each), the cells were permeabilized with 0.2% Triton X100 in PBS for 5 min. Following additional washes, slides were treated with 1% BSA in PBS for 30 min at room temperature to block non-specific binding. The primary antibodies (anti-VDR, Santa Cruz Biotechnology, USA) were added, and the slides were incubated overnight at 4 °C. The next day, the primary antibody was removed and the slides washed three more times with PBS for 5 min each. They were then incubated with a secondary antibody (Alexa Fluor 488 anti-rabbit, Invitrogen, USA) in PBS for 1 h at room temperature in the dark. After three additional washes with PBS (5 min each), the slides were counterstained with 4′,6′-diamidinio-2-phenylindole (DAPI). Images were captured using an Olympus cellVivo IX83 microscope, and the translocation of VDR from the cytoplasm to the nucleus was analysed using Olympus cellSens software (Olympus, Japan).

RNA extraction

HPEKp cells were seeded on 6-well plates and then treated with 100 nM 1,25(OH)2D3 or 1,24,25(OH)3D3 for 4 and 24 h. RNA was extracted according to the manufacturer’s instruction with the use of an ExtractMe total RNA KIT (Blirt, Poland). The concentration of RNA was assessed with an EPOCH Microplate Spectrophotometer (BioTek, USA).

RNA sequencing

For RNAseq, the purity and quality of the RNA was assessed using the RNA ScreenTape assay using a 4200 TapeStation System (Agilent Technologies, Germany). The RNA samples with RIN (RNA Integrity Number) of 10 were sequenced using weSEQ.IT of the library preparation kit VAHTS Universal V6 RNA-seq Library Prep Kit for Illumina. The percentage of Q30 bases in each sample was above 92.81%.

Assay of gene expression by real-time RT-PCR

Real-time PCR was performed with a StepOnePlus™ Real-Time PCR System (Life Technologies Applied Biosystems, Grand Island, NY, USA) using the AMPLIFYME SG No-ROX Mix kit (Blirt, Poland). The primers used in the reactions were obtained from Merck, Germany. As our long-term studies have shown that RPL37 is the most stable reference gene in the context of vitamin D therapy61,81,89, it was used to normalize gene expression levels. The comparative ΔΔCT method was used for the analysis, and results were expressed as fold change ± SD. Primers sequences are summarised in Table 1.

Table 1 Primers sequences.
Full size table

Bioinformatic analyses of RNA sequences

The raw RNA sequence reads were subjected to an initial quality assessment and cleanup procedures using FastQC90 and Trimmomatic91. Subsequently, the reads were aligned to the human reference genome (GRCh38) using STAR92 followed by read counting using feature Counts 2.0.3 93. To compare the transcriptomic profiles across samples, Principal Component Analysis (PCA) was conducted on rlog-transformed count values93, revealing distinct major clusters.

Differential gene expression analysis was performed using DESeq2 94, with a predefined criterion of an absolute value of log2fold change (log2FC) ≥ 0.58 and a false discovery rate (FDR) adjusted p-value < 0.05 to identify significantly differentially expressed genes.

To identify the biological functions associated with these genes, a Gene Ontology (GO) enrichment analysis was carried out using the R package topGO94 with Fisher’s exact test. Gene names were mapped to GO terms utilizing the org.Hs.eg.db package and only GO terms with an adjusted p-value < 0.05 were considered significantly enriched.

To obtain a more complete picture of gene functions and interactions, gene enrichment was also performed against the Kyoto Encyclopedia of Genes and Genomes (KEGG) database using the clusterProfiler tool. KEGG categories exhibiting an adjusted p-value < 0.05 were considered significantly enriched. The results of the analysis were visualized using R (R Core Team, 2021). RNA-seq data generated as part of this study are openly available without restrictions. The data have been deposited in the Sequence Read Archive (SRA) under accession number PRJNA926032.