1,25-Dihydroxyvitamin D3 [1, 1,25D, Figure 1], the active form of vitamin D3, is already used clinically to treat renal osteodystrophy and various types of rickets and has been investigated, together with some analogues, for treatment of a wide variety of diseases, including breast and prostate cancers, autoimmune diseases, psoriasis, and osteoporosis. Nonetheless, its therapeutic application has been limited by the parallel induction of hypercalcemic effects. More than 4.000 vitamin D analogues have been synthesized in recent years in an attempt to find therapeutic agents with low calcemic activity, but only a few have reached the pharmaceutical market or advanced clinical trials.
Until recently, the available information on the structure-activity relationships of 1,25D analogues was rather limited and the design of new compounds was essentially based on systematic structural variations. Now there is continued and increasing interest aimed at the rational design of new 1,25D analogues with selective biological functions as possible drugs for medical use.
Recently, Dino Moras and coworkers published the crystal structure of an engineered ligand binding domain of the vitamin D receptor (VDR LBD) that lacks a flexible insertion domain between helices H1 and H3 (Mol. Cell, 2000, 5, 173-1795). The mutant VDR bound to 1,25D [VDR(Moras)-1,25D complex] exhibits similar conformation, transactivation ability, and biophysical properties than the wild-type counterpart. The crystal structure of VDR(Moras)-1,25 complex shows the hydrogen bonding nature of the interactions between each of the three hydroxyl groups of the ligand with the mutant vitamin D receptor (1-OH with both Ser-237 and Arg-274, 3-OH and both Tyr-143 and Ser-278, and 25-OH and both His-305 and His-397) (Fig 1).
Fig 1. Crystal structure of the hormone 1,25D in the VDR binding pocket.
We are now using the structural details of the binding pocket and docking calculations for the rational design of new active vitamin D analogues. For example, Fig 2 shows the superimposition of the three crystal structures corresponding to the complexes formed by VDR(Moras) and several vitamin D analogues including the natural hormone (superagonists MC1288 and KH1060, and 1,25D). The side chains of the three compounds in the binding pocket follow different pathways to reach carbons-25 at a common point. The observation of the side-chain conformations suggested that a new analogue with a tetrahydrofuran unit at the side chain (AMCR277A) might behave as a new superagonist.
Fig 2. Superimposition of the crystal structures of the superagonists MC 1288 and KH100, and the natural hormone 1,25D in complex with VDR(Moras).
Superimposition of the crystal structures of VDR(Moras) complexed to AMCR277A and AMCR277B indicates that the side chain of the first analogue is closer to 418-valin residue (Fig 3). Considering that this aminoacid residue is important to induce transcription, the analogue AMCR277A should be more active than AMCR277B. Remarkably, isomer AMCR277A induces transactivation of human VDR 12 times more efficiently than the natural hormone at concentrations of 10-10 M, indicating its superagonistic nature while isomer AMCR277B shows the same activity as the natural hormone (Chem. Biol. 2008, 15, 383-392).
Fig 3. Superimposition of the crystal structures of AMCR277A and AMCR277B in complex with VDR(Moras).