Abstract
The Δ16 structure as a vitamin D analog enhanced vitamin D receptor (VDR) binding affinity and induced significant cell differentiation, whereas its relative calcemic activity was reduced compared to 1α,25-dihydroxyvitamin D3 (1α,25(OH)2D3). Methodologies available to introduce a double bond at C16-C17 of the D-ring on the seco-steroidal skeleton were limited; therefore, a new synthetic strategy was developed to obtain not only the Δ16 structure, but also a new C15-functional group. Since C15-functionalization was unprecedented in vitamin D analog studies, the hybrid structure of Δ16 and the C15-OH group at the D-ring may provide important information on the structure-activity relationship with vitamin D analogs. The synthesized 16-ene-2α-methyl-1α,15α,25-trihydroxyvitamin D3 showed almost 3-times higher VDR binding affinity and an equipotent level of osteocalcin promoter transactivation activity in human osteosarcoma cells as compared to 1α,25(OH)2D3.
The 16-ene structure of the CD-ring part of vitamin D3, first appeared in the total synthesis of 1α,25-dihydroxyvitamin D3 (1α,25(OH)2D3, 1 in Figure 1) reported by the Hoffmann-La Roche group in 1982 (1). To elongate the side chain at C20 (steroid numbering) from the (17Z)-ethylidene CD-ring of compound 5 creating the 20R-natural stereochemistry, they utilized the ene reaction with ethyl propionate in the presence of Lewis acid EtAlCl2 (Figure 2). After the successful ene reaction (88% yield and the desired 20R configuration) (1-3), the 16-ene structure 6 was obtained, and the double bond was then reduced stereo-selectively to the saturated CD-ring system 7 that was included in the natural active vitamin D3 skeleton. A great variety of unique 16-ene vitamin D analogs have been synthesized using the ene reaction, and their biological activities were evaluated (4-11). It was shown that 16-ene-1α,25-dihydroxyvitamin D3 had higher binding affinity for the vitamin D receptor (VDR), lower affinity for the vitamin D binding protein in circulation and greater resistance to 24-oxo-mediated catabolism as compared with 1α,25(OH)2D3 (11-16).
Another strategy for 16-ene construction was developed by Mouriño's group in 1999. They utilized SN2′-syn-facial displacement toward (17Z)-olefin carbamate (8) by high-order cuprate(s) (Li2Cu3Me5 etc.) to generate the 20R-natural configuration and the 16-ene double bond as in 9 with high yield(s) (Figure 3) (17). The leaving group (a carbamate group) for the SN2′ reaction was located at the C16 position, and various types of substituents could be introduced at the C20 position, stereoselectively, through this route.
C15 substituted vitamin D analogs have not been synthesized yet, and a new synthetic route for 16-ene analogs also via an SN2′ reaction was developed. In order to create a C15-substituent, a new substrate for the SN2′ reaction was used, since a C15-C16 epoxide ring system on the D-ring was desirable, in which the leaving group was still at the C16 position, and importantly, a required functional group (a hydroxy group) could remain at the C15 position after the SN2′ reaction.
Materials and Methods
Chemistry. A new epoxide 11 was synthesized from known ketone 10 in four steps (Figure 4). The stereochemistry of 15α,16α-epoxide 11 was determined by X-ray crystallographic analysis of its acetate 12 (Figure 5). Oxidation of the C17-OH group of 11 to the C17-oxo group followed by Wittig reaction with ethylidenetriphenylphosphorane gave the key intermediate of (17Z)-ethylidene epoxide 14 with a good yield (18).
Structures of active vitamin D3 (1α,25(OH)2D3), 1; 16-ene-1α,25(OH)2D3, 2; 16-ene-1α,15α,25(OH)3D3, 3 and 2α-methyl-16-ene-1α,15α,25(OH)3D3, 4.
Part of the first total and chiral synthesis of active vitamin D3 (1α,25(OH)2D3), in which 16-ene structure 6 was available by the ene reaction (1).
The next step was the SN2′ reaction of magnesium cyanocuprate prepared from 5-bromo-2-methyl-2-pentanol methoxymethyl ether (19) toward compound 14. This key SN2′ reaction proceeded smoothly in anti-displacement to afford the 16-ene product with the 15α-hydroxy group and the 20R-natural configuration side chain as the major isomer 15a (20R:20S=11:1).
The CD-ring 15a was converted to the 8-oxo compound 16 and subsequent Wittig reaction with bromomethylenetriphenylphosphorane gave bromoolefin 17. Trost coupling reaction (20) of the C15-substituted 16-ene-bromoolefin 17 and enynes 18 and 19 (21-23) gave the protected target molecules 20 and 21, respectively. The deprotection step took place under mild acidic conditions using (+)-camphorsulfonic acid in MeOH to yield 3 and 4 (Figure 6).
Human VDR binding assay. Binding affinity to the VDR was evaluated using a 1α,25(OH)2D3 assay kit (Polarscreen Vitamin D Receptor Competitor Assay, Red, Cat. No. PV4569) purchased from Invitrogen (Grand Island, NY, USA). The test compound solution (1 mM in EtOH) was initially diluted 10-fold with DMSO and then further 50-fold with the assay buffer provided in the kit. The assay was performed according to the procedures described in the kit insert. All the compounds were evaluated with N=2 within the range from 10−6 M to 10−10 M. IC50 values were calculated using the average of the measured values. The activities of each compound are presented as the relative value in which the activity of the natural hormone 1 was normalized to 100%.
Transactivation assay of human osteocalcin promoter. VDR transactivation of the osteocalcin promoter in HOS (human osteosalcoma) cells (American Type Culture Collection (ATCC) Manassas, VA, USA) by the test compounds was compared to the natural hormone. The human osteocalcin gene promoter fragment −838/+10 was cloned into the reporter plasmid pGL3 (Promega; Madison, WI, USA) as reported previously (24). Human VDR and RXR (Retinoid X receptor) genes were cloned into expression vector pcDNA3 (Invitrogen). The HOS cells were maintained in phenol red free DMEM (Invitrogen) containing 10% fetal calf serum (FCS) (Invitrogen). Prior to the transfections, the cells were plated in a 96 well plate at a density of 400,000 cells per well in Opti-MEM (Invitrogen). The cells were transfected with human osteocalcin reporter vector (pGL3-hOc: 100 ng/well), human VDR and RXR expression vector (pcDNA-hVDR, pcDNA-hRXR: 10 ng/well) and internal control plasmid phRL-TK (Promega: 25 ng/well) using 0.45 μL of Lipofectamine 2000 reagent (Invitrogen). After incubation at 37°C for 3 h, the culture media were replaced with phenol red free DMEM containing 10% FCS. The cells were treated with ethanol vehicle or various concentrations of the compounds (from 0.1 pM to 100 nM). After incubation at 37°C for 24 h, the luciferase activity of the cells was quantified by luminometer (Berthold Japan; Tokyo, Japan) using a Dual-Glo luciferase assay system (Promega).
Results and Discussion
Chemistry. In the first total synthesis of active vitamin D3 (1), trans-(8-hydroxy)hydrindan-17-one (22) was alkenylated through the Wittig reaction with ethylidenetriphenylphosphorane to afford (17Z)-ethylidene 23 as the major product. The steroid skeletons also gave (17Z)-ethylidene 25 after the Wittig reaction (Figure 7) (2,3,25). Furthermore, 15β,16β-epoxy-17-oxosteroid (26) was converted to (17E)-olefin 27, in which the direction of the ethylidene group was the same as in compounds 23 and 25 (Figure 7) (26-28). (17E)-Ethylidene-15β,16β-epoxysterols (27) were previously studied on the synthesis of 15β-hydroxysteroids such as oogoniol (28) (steroidal sex hormone of the water mold Achlya) (29). On the other hand, pavoninin-5 (29) (shark repellent of the sole Pardachirus pavoninus) has the 15α-hydroxysteroid structure (30, 31).
Part of 16-ene-1α,25(OH)2D3 synthesis through SN2′-syn displacement at C20 with the allylic carbamate system (17). MOM: methoxymethyl; TBS: tert-butyldimethylsilyl.
(17Z)-Ethylidene-15α,16α-epoxyhydrindan (14) synthesis from ketone 10 (18).
ORTEP (Oak Ridge Thermal Ellipsoid Plot Program) stereo drawing of acetate 12.
Interestingly, however, in the present process the 15α,16α-epoxyhydrindan-17-one (13) was converted to (17Z)-olefin (14) by the same Wittig reaction (the nomenclature of Z stereochemistry is the same in both compounds 23 and 25 coincidentally, but the direction of the ethylidene section is different, Figure 4). In the present case, the epoxide ring at the α-side would select orientation of the phosphorane reagent by steric repulsion as shown in Figure 8, i.e., first, the phosphorane reagent would attack the 17-keto carbonyl group from the less hindered α-side due to the C18-methyl group, and then, repulsion would occur between the methyl group and the α-epoxide ring, B, between the PPh3 group (which was the biggest group among H, CH3, and Ph3P) and the α-epoxide ring, C, and between both methyl and PPh3 groups and the α-epoxide ring, D; therefore, A would be the most favored conformation to give (17Z)-ethylidene stereochemistry.
Introduction of the side chain to C20 with R-configuration and conversion to bromoolefin 17, and subsequent Trost coupling reaction with A-ring precursor enynes (18 and 19). The coupling product was deprotected under mild acidic conditions. CSA: (+)-camphorsulfonic acid.
The side chain elongation reaction at the C20 position of (17Z)-ethylidene epoxide 14 in the anti-stereochemistry of SN2′ manner, proceeded smoothly with the magnesium cyanocuprate reagent (Figure 6) (18). Fortunately, the natural configuration of 20R was available as the major isomer, which could be explained by Corey's theory of a stereoelectronic effect arising from bidentate binding involving a d-orbital of nucleophilic copper and σ* and π* orbitals of the ethylidene epoxide 14 (Figure 9) (32). The 20R stereochemistry of 15a was determined without ambiguity by chemical conversion of 15a to the known 25-OMOM Grundmann's ketone, which has the 20R configuration (18).
Human VDR binding affinity. The hVDR binding affinity of the new 16-ene-vitamin D3 analogs 3 and 4 having the C15-OH group was determined using the Invitrogen receptor assay kit. Compounds 3 and 4 showed 65% and 278% binding affinity relative to that of 1α,25(OH)2D3. It was reported that the original 16-ene-1α,25(OH)2D3 (2) exhibited 161% binding affinity for calf-thymus VDR (33) and 240% for rat intestine VDR (12). The introduction of a double bond to the ring D at the C16-17 position of 1α,25(OH)2D3 enhanced the binding affinity for the VDR, but further modification with the C15α-OH group decreased the VDR binding affinity of compound 3, however, additional modification at the C2α position with the methyl group recovered the affinity of compound 4.
Osteocalcin promoter transactivation activity. The EC50 values of transactivation activity of the osteocalcin promoter in the HOS cells were 0.08 nM for compound 3 and 0.12 nM for compound 4, and 0.09 nM for 1α,25(OH)2D3. Thus the 16-ene analogs with the C15α-OH group 3 and 4 showed comparable or even greater transactivation activity than 1α,25(OH)2D3.
Our previous X-ray co-crystallographic study on 15α-methoxy-1α,25-dihydroxyvitamin D3 revealed that the 15α-methoxy group made a new contact with Ser275Oγ at a distance of 3.1 Å (18); therefore, the 15α-OH group of the analogs 3 and 4 could form a hydrogen bond with Ser275Oγ to stabilize the VDR-ligand complex. Importantly, the C15-OH group could be converted to a wide variety of other functional groups, and would be useful for further structure-activity relationship studies on the D-ring moiety of the active vitamin D to develop potential therapeutic agents for cancer (34-36).
(17Z)-Ethylidene formation via the most favored conformation A at the Wittig reaction step.
Plausible reaction mechanism of anti-stereochemistry based on SN2′ reaction entering the full length side chain with 20R configuration.
Conclusion
The novel 16-ene-1α,15α,25-trihydroxyvitamin D3 analog 3 and its 2α-methylated analog 4 are synthesized efficiently utilizing the SN2′ reaction of organocuprate reagent toward (17Z)-ethylidene-15α,16α-epoxyhydrindan 14 with the anti-stereoselective displacement followed by Trost coupling with the A-ring precursor. These synthesized analogs show great VDR binding affinity and osteocalcin transactivation activity in HOS cells.
Acknowledgements
We thank Ms. Junko Shimode and Ms. Miki Takahashi (Teikyo University, Japan) for the spectroscopic measurements. This study was supported in part by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (No. 21590022 to AK and No. 23590015 to DS) and Grant-in-Aid for Young Scientists from the Ministry of Education, Culture, Sports, Science and Technology, Japan, (No. 23790021 to MT).
Footnotes
-
↵† Current Address: Faculty of Pharmaceutical Sciences, Hokkaido University, Sapporo 060-0812, Japan.
- Received September 9, 2011.
- Revision received November 17, 2011.
- Accepted November 17, 2011.
- Copyright© 2012 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved