Novel Vitamin D Analogues 17,20S(OH)₂pD and 17,20R(OH)₂pD are Noncalcemic and Exhibit Antifibrotic Activity

LETTER TO THE EDITOR

The active form of vitamin D, 1,25(OH)₂D₃, has a number of desirable properties in addition to its classical effect on bone and calcium metabolism. These include immunomodulatory, anti-inflammatory and anti-fibrotic properties, however, its use is limited because of the toxic (calcemic) effect at therapeutically relevant doses (Bikle, 2010; Szodoray et al., 2008). Our goal of this and related studies has been to identify natural noncalcemic vitamin D analogues for use as therapeutic agents to treat diseases like scleroderma and/or autoimmune diseases.

We embarked upon a strategy to produce such agents with a short or long side chain. Specifically, we defined a novel steroidogenic pathway initiated by cleavage of the side chain of 7-dehydrocholestrol by P450scc to produce 7-dehydropregnenolone (7DHP), which could be further converted to hydroxy-5,7-dienal intermediates by classical steroidogenic enzymes (Slominski et al., 2004; Slominski et al., 2009). An example of the latter is 17,20(OH)₂7DHP (Shackleton et al., 2002), for which we established a chemical route of synthesis with UVB induced transformation to vitamin D–like 5Z,7E-3β,17,20-trihydroxy-9,10-secopregna-5,7,10 triene (17,20(OH)₂pD) and lumisterol-like structures (Zmijewski et al., 2009).

Since 17,20(OH)₂7DHP is produced in the body, at least under pathologic conditions (Smith Lemli Optiz Syndrome) (Shackleton et al., 2002), and it can be transformed in the skin to corresponding secosteroids when exposed to the UVB (Slominski et al., 2004; Zmijewski et al., 2009), these vitamin D-like hydroxyproducts are good candidates to serve as natural products, which by definition would have low toxicity. Supporting this, shortening of the side chain of vitamin D eliminates its calcemic activity (Holick et al., 1975). Thus, these novel vitamin D analogues could serve as candidates for treatment of autoimmune or fibrosing diseases since vitamin D3 displays well documented effects on these conditions in animal models (Moro et al., 2008).

17,20(OH)₂D and lumisterol derivatives were synthesized as described previously (Zmijewski et al., 2009). For details see Figure S1.

Female C57BL/6 mice 6 wks old were purchased from Jackson Labs and maintained on a regular laboratory chow diet. Mice were divided into groups of three to receive one of the following treatments by daily intraperitoneal injection for 14 days: Sterile sesame oil (50 µl) (Sigma Chemical Co., St. Louis, MO) or 3 µg/kg 17,20R(OH)₂D₃ or 17,20S(OH)₂D₃ dissolved in 50 µl sterile sesame oil. Mice tolerated the injections well without signs of clinical distress. On day 15 mice were euthanized, and sera were obtained for measurement of total Ca⁺⁺ by atomic absorption spectrophotometry (Bhattacharya, 1977), inorganic phosphate (iP) by Malachite Green Assay Kit (Cayman Chemical) and fibroblast growth factor (FGF) 23 by ELISA (Immuntopics, San Clemente, CA). Bone mineral density (BMD) and bone mineral content (BMC) were determined for whole body (excluding head) and excised soft tissue-cleaned femurs by dual-energy X - ray absorptiometry (DEXA) using a GE Lunar PIXImus (GE Healthcare); quality control and calibration were carried out within 24 hours of each scanning period. An Instron Universal Test System (Instron Corp) was used for three-point bending (gap size, 7 mm) to determine the flexure stress of mouse femoral diaphyses. The load was applied at a constant displacement of 5 mm/min to failure. The femurs broke in the mid-diaphyseal region, and data were recorded as flexure stress in megapascals (MPa). Hearts, livers, spleens and kidneys were processed for histology.

Human dermal fibroblasts were grown in 24 well Costar tissue culture plates for experiments to assess effect of vitamin D analogues on total collagen protein (Rahow et al., 1987), hyaluronan (Postlethwaite et al, 1989), type I collagen by ELISA (Chondrex) or Col1A1 mRNA by quantitative real time qPCR.

We previously showed that 20(OH)D₃ at a dose as high as 3.0 µg/kg in rats, has no calcemic activity (calcium = 10.4 ± 1.5 mg/dL(mean +SD) vs 9.3 ± 1.3 mg/dL for control) whereas 1,25(OH)₂D₃ at the same dose had the expected strong calcemic effect raising calcium to 16.0 ± 1.2 mg/dL (Slominski et al., 2010). In this study, we have found that 17,20S(OH)₂pD and 17,20R(OH)2 pD at doses of 3 µg/kg had no significant effects on sera total Ca⁺⁺ levels (Table 1). Examination of histologic sections of the hearts, kidneys, liver and spleen revealed lack of calcifications or identifiable toxicity (not shown). Thus, 17,20R(OH)₂pD and 17,20R(OH)₂pD show the same noncalcemic properties as the structurally similar 20(OH)pD (Holick et al., 1975) and 20(OH)D₃ (Slominski et al., 2010).

Table 1.

Serum Calcium Levels, Serum, FGF-23 Levels and Bone Measurements in Mice

Group	Ca++* mg/dL ± SEM	iP mg/dL ± SEM	FGF23 pg/mL ± SEM	BMD* Right Femur g/cm2 ± SEM	BMC* Right Femur g	BMD* Total g/cm2 ± SEM	BMC* Total g	Flexure Stress* mPa
Sesame Oil	9.5 ± 0.3	7.35 ± 0.36	452 ± 93	0.0421 ± 0.0016	0.0100 ± 0.0006	0.0333 ± 0.0011	0.188 ± 0.0110	51.3167 ± 0.602
17,20R(OH)2D3	9.4 ± 0.3	6.77 ± 0.28	980 ± 135**	0.0464 ± 0.0027	0.0117 ± 0.0009	0.0352 ± 0.0004	0.209 ± 0.0120	49.8713 ± 0.711
17,20S(OH)2D3	8.9 ± 0.2	6.32 ± 0.03	634 ± 143	0.0450 ± 0.0021	0.0110 ± 0.0010	0.0342 ± 0.0009	0.196 ± 0.0153	48.0900 ± 0.836

^*Serum total Ca⁺⁺ and iP, levels, Bone Mineral Density (BMD) for right femur or total skeleton, Bone Mineral Content (BMC) for right femur or total skeleton, and Flexure Stress in 17,20R(OH)₂D₃ or 17,20S(OH)₂D₃ treated mice were not statistically different from sesame oil control treated mice by ANOVA.

^**P = 0.032 by ANOVA

Serum levels of the phosphatonin factor, FGF23 were not different in mice treated with sesame oil or 17,20S(OH)₂pD (Table 1). However, 17,20R(OH)₂pD did induce a significant elevation in serum levels of FGF23 (Table 1). When rats are administered 1,25(OH)₂D₃ serum levels of FGF23 increase dramatically [>5-fold increase by 0.3 µg/kg 1,25(OH)₂D₃], and iP in sera is markedly reduced (Saito et al., 2005). In vitro, addition of 1,25(OH)₂D₃ to cultures of osteoblasts increases FGF23 mRNA levels (Liu et al., 2006). In this study, levels of iP in sera were not significantly different between mice treated with sesame oil vs 17,20R(OH)₂pD or 17,20S(OH)₂pD (Table 1). The data suggest that 17,20R(OH)₂pD has retained a weak ability to modestly increase FGF23 levels in serum but not sufficiently high to significantly reduce serum iP, while 17,20S(OH)₂pD does not possess this FGF23 regulatory property of 1,25(OH)₂D₃. There were no differences in BMD, BMC or flexure stress between 17,20S(OH)₂D₃, 17,20R(OH)₂D₃ and sesame oil treated mice (Table 1).

Table 2 shows that R and S forms of 17,20(OH)₂pD and 17,20(OH)₂pL, like 1,25(OH)₂D₃, significantly inhibit TGF-β1-induced total collagen protein and hyaluronan production by human dermal fibroblasts. We found no significant differences in fibroblast numbers per well and no significant differences in trypan blue exclusion. In agreement, 10₋₇ M 17,20S(OH)₂pD or 17,20R(OH)₂pD inhibited the TGF-β1-stimulated expression of Col1A1 mRNA (Figure S2), and secreted type I collagen protein (Figure S2, insert).

Table 2.

Inhibition in Human Dermal Fibroblasts Stimulated by TGF-β1^* of Total Collagen and Hyaluronan Production by Novel Secosteroids

Condition hrTGF-β1 5ng/ml; secosteroids at 10−10M	Collagen CPM/Wel (mean ± SEM) × 10−5 FB	p Value	Hyaluronan CPM/Well (mean ± SEM) × 10−5 FB	p Value
TGF-β1 + vehicle	829 ± 17*	<0.001	1556 ± 51*	<0.001
TGF-β1 + 17,20 R(OH)2pD	330 ± 17†	<0.001	163 ± 8†	<0.001
TGF-β1 + 17,20 S(OH)2pL	266 ± 6†	<0.001	97 ± 46†	<0.001
TGF-β1 + 17,20 S(OH)2pD	95 ± 10†	<0.001	662 ± 189†	0.01
TGF-β1 + 17,20 R(OH)2pL	320 ± 3†	<0.001	222 ± 13†	<0.001
TGF-β1 + 1,25(OH)2,D3	310 ± 100†	0.007	297 ± 18†	<0.001

After 2 hour preincubation, hr TGF-β1 (R and D systems) was added at a final concentration of 5ng/ml. To measure total collagen and hyaluronan production, after 48 hours of culture, plate wells were paused with 1µ CI ³[H]-proline. Then after 24 hours, culture supernatants were harvested and collagenase sensitive protein was determined as we have previously described (Raghow et al., 1987).

^*Significantly different (p<0.001) from PBS + vehicle

^†Significantly different (p<0.001) from TGF-β1 + vehicle

n=3; statistical analysis was performed using Student’s t test

In conclusion, novel dihydroxysecosteroidal derivatives with a short side chain are noncalcemic and exhibit antifibrotic activity; therefore they are excellent candidates for further testing in in vivo models of fibrosis and scleroderma.

Abbreviations

1,25(OH)₂D₃

1,25-dihydroxyvitamin D3

20,23(OH)₂D₃

20,23-dihydroxyvitamin D3

17,20(OH)₂pD

17,20-dihydroxypregnacalciferol

20(OH)D₃

20-hydroxyvitamin D3

7DHP

7-dehydropregnenolone

17,20(OH)₂pL

17,20-dihydroxypregnalumisterol

17,20(OH)₂7DHP

17,20-dihydroxy7DHP

FGF

fibroblast growth factor

TGF-β1

Transforming growth factor –β1