Deslongchamps' Partial Synthesis of Hippuristanol

Hippuristanol is an antiproliferative agent that selectively targets an RNA helicase initiation factor [(eIF)4A] to block the translation stage of eukaryotic protein biosynthesis. This binding interaction triggers potent cytotoxic activity against cultured tumor cells and establishes RNA helicases as novel anticancer targets. Structurally, hippuristanol is a polyoxygenated steroid with a C22 spiroketal unit in a thermodynamically unfavorable stereochemical configuration (vida infra). The steroid framework is also hydroxylated on the beta-face at C11 and C20, that latter being a quaternary stereogenic position. Hippuristanol shares certain topological structural features with the ‘eastern’ substructure of cephalostatin 1, a dimeric steroidal pyrazine with uniquely selective and exquisitely potent anticancer properties. Pursuit of the currently unknown cellular target and mechanism of action of cephalostatin 1 is the subject of extensive research efforts in a number of academic laboratories and institutions.

A concise partial synthesis of hippuristanol, disclosed in 2010 from the laboratory of Pierre Deslongchamps, starts from hecogenin acetate, a plant-derived bulk chemical with side chain spiroketal functionality that can be rearranged and elaborated as needed. However, in order to make use of this inexpensive sapogenin as a precursor to hippuristanol, hecogenin’s C12 oxo moiety must be transposed to the C11 position. With this objective in mind, Deslongchamps and co-workers first employ osmium tetroxide-catalyzed dihydroxylation of the silyl enol ether derived from hecogenin to accomplish net alpha-hydroxylation at C11. The ketol thus obtained is then subjected to treatment with alkali under thermodynamically forcing conditions and rearrangement to the more stable ketol ensues, presumably via the intermediacy of the bracketed enediol shown above. This methodology was developed in the 1990’s by industrial chemists at Pfizer for the production of a steroidal cholesterol absorption inhibitor. The 11-keto-12beta-acetoxy intermediate then undergoes reduction with calcium metal, which effectively shifts the C12 keto group of hecogenin to C11. This reduction probably involves an initial epimerization of the C12 position to place the acetate in an axial position which would mechanistically facilitate the requisite 1,2-elimination.

In order to secure the C20 tertiary alcohol stereocenter, hecogenin’s C22 spiroketal must be truncated to a 20-oxo-pregnane. Historically, this has been achieved by implementation of Russell Marker’s degradation protocol, a three-step route that efficiently provides 3-hydroxypregna-5,16-dien-20-ones from various plant-derived sterols. The classical Marker degradation involves a base-mediated elimination (step 3) to install a delta(16,17) double bond. This results in the loss of stereochemical information (two stereocenters are abolished) and oxygenation at C16. In order to retain the beta-configuration at carbons 16-17 and retain the C16 hydroxyl, Deslongchamps has modifed step 3 of the Marker degradation by treating the intermediate derived from the chromic acid oxidation with vinylmagnesium bromide. This delivers an allylic alcohol that can then be oxidatively cleaved to provide the requisite beta-hydroxy ketone needed for elaboration to the natural product. Deslongchamps’ modification of the Marker degradation may be a generally useful process for the preparation of complex oxygenated steroids. In the next step, addition of a lithiated alkyne to the C17 methyl ketone stereoselectively affords the product of chelation control, the C20beta-tertiary alcohol.

Without question, the most interesting transformation in the synthesis of hippuristanol is the mercury(II)-catalyzed spiroketalization that generates semiprotected 22-epi-hippuristanol in one step from a 3-alkyn-1,7-diol precursor. A proposed mechanism was advanced by the authors involving intramolecular 5-exo-dig oxymercuration of the 16-hydroxyl onto the triple bond. A subsequent ketalization event establishes the C22 spirocarbon center. This latter mechanistic aspect is of great interest, as the reaction stereoselectively produces only one of two possible diastereomeric products (22S/22R = 99.9:0.1). Evidence is provided to suggest that this is a thermodynamically controlled process: treatment of hippuristanol under the aqueous mercury(II) reaction conditions leads to complete isomerization of the natural product to 22-epi-hippuristanol. Indeed, upon inspection of the possible diastereomeric products (stereo structures shown above), one can see that the observed product is stabilized by two anomeric effects and that hippuristanol’s bicyclic spiroketal is in the non thermodymanically-favored configuration (due to one anomeric effect). Therefore, the stereochemical outcome of the reaction is probably due to equilibration rather than selective kinetic ketalization of the nucleophilic hydroxyl onto the intermediate oxonium species from the alpha-face. Debenzylation and spiroketal isomerization completed the expedient synthesis of hippuristanol. The observation that partial epimerization of C22 to the natural configuration can only be accomplished in an aprotic acidic system implies the existence of an intramolecular hydrogen bond between the hippuristanol’s C20 tertiary hydroxyl group and a spiroketal oxygen atom.

Partial Synthesis of Withanolide A From Pregnenolone

            Gademann and co-workers have disclosed a stereocontrolled preparation of withanolide A from a readily available steroid precursor. Their partial synthesis proceeds in 13 steps with minimal use of protecting groups and accomplishes stereoselective elongation of the steroid (C17) side chain as well as oxidative elaboration of the western A/B decalin substructure. Among the synthetic challenges associated with this task is the establishment of stereocenters at C20, a quaternary center, and C22, which is incorporated into the alpa,beta-unsaturated valerolactone (dihydropyrone) in the side chain. Ikekawa’s partial synthesis of withanolide D (structure shown above) provides a strong precedent for pregnane side-chain homologation of this sort and Gademann clearly benefits from the knowledge gained in pioneering synthetic campaigns. What makes the synthesis of withanolide A significant is the stereocontrolled functionalization of rings A and B of pregnenolone, which entails installation of a very sensitive enone and an alpha-hydroxy epoxide. These structural motifs are unique as compared to other withanolides that have succumbed to synthesis and, thus, a novel synthetic strategy is required to access this important candidate for the therapeutic treatment of neurodegenerative disease (see previous post for details). Gademann and co-workers expertly apply the Schenck O₂ ene reaction and Wharton carbonyl transposition to complete the synthesis of the target molecule along with non-natural analogues in quantities sufficient for evaluation of neuritogenic properties and secretase inhibition activity.

The synthesis is initiated by a highly diastereoselective addition of lithiated 1,3-dithiane to the C20-keto group of 3beta-OTBS pregnenolone. Gademann and co-workers do not comment on the stereochemical outcome of this conversion except to say that the product is known. A 1978 review article by Jerzy Wicha provides a detailed rationale for the addition of sterically bulky nucleophiles to the 20-ketone of the steroid system, a reaction with a rich history dating back to Woodward’s total synthesis of cholesterol (JACS 1951). This stereochemical result is best explained by ‘steric approach’ control, which favors attack of the carbonyl from ‘outside of the molecule,’ or, in other words, from the C16 side. The observed diastereoselectivity, dictated by the conformational transition state model shown above, generally predominates for the addition of bulky nucleophiles to C17 non-hydroxylated (i.e. 17alpha-H) steroidal 20-ketones.

Two subsequent synthetic operations generate the MOM-protected alpha-hydroxy aldehyde which is elaborated to the lactone with excellent diastereoselectivity by a tandem vinylogous aldol/lactonization reaction. This method was developed by Ikekawa for the partial synthesis of withanolide D, recorded in 1984. The reaction seems to proceed through a transition state comparable to Cram’s cyclic chelate model.

Stereoselective oxyfunctionalization of the A/B ring junction position (C5) on the alpa-face is accomplished by implementation of the Schenck ene reaction with singlet oxygen, generated in situ from molecular oxygen in the presence of a porphyrin sensitizer with irradiation (Na lamp). Generally, diastereoselectivity in the photooxygenation of steroids is governed mainly by steric shielding by the angular methyl groups. Conformational rigidity also contributes to pi-facial selectivity, as hydrogen atoms perpendicular to the plane of the double bond are accessible only from one face of the molecule. In this case, the alpha-oriented allylic alcohol is generated in good isolated yield although the diastereomeric ratio of products is not provided by the authors.

The key step of the synthesis was the Wharton carbonyl transposition of the advanced bis-epoxy ketone intermediate. Exposure to hydrazine furnished the rearranged allylic alcohol and subsequent PDC oxidation completed the synthesis of withanolide A (50% yield over two steps). This partial synthesis effort is notable for its practicality (13 steps), high level of stereocontrol and pragmatic choice of a simple and abundant starting material derived from Nature’s chiral pool.

Withanolide A, a Lead Compound for the Treatment of Neurodegenerative Disease

The amyloid cascade hypothesis suggests an imbalance between the production and clearance of the peptide amyloid beta (A beta). The accumulation of A beta peptides in the brain leads to formation of the neuritic plaques that are a distinguishing characteristic of Alzheimer’s disease (AD). Indeed, the level A beta plaques in the brains of AD patients correlates with their degree of cognitive impairment. The A beta pathway implicates several potential AD therapeutic targets and opportunities for pharmacological intervention. Modulation of multiple A beta-regulating targets by a single chemical entity, as opposed to the present one-drug-one-target paradigm, is desirable.

The steroid lactone withanolide A (structure shown above), isolated from Winter Cherry (Withania somnifera), also known as Ashwaganda in ayurvedic medicine, has been shown to down-regulate the expression of beta-secretase 1 (BACE1) in a dose-dependent manner. The same molecule also dose-dependently enhances levels of mature ADAM10; the regulation of each of these targets is consistent with non-amyloidogenic processing of amyloid beta precursor protein (AbetaPP) which suggests that withanolide A may promote disease-modifying effects against AD, as opposed to the current symptomatic therapies. To determine the aforementioned protein expression levels, Chan and co-workers treated primary rat cortical neurons with various doses of withanolide A for 24 hours and then conducted western blot analyses of the cell lysates. It should be noted that the effects noted above were observed with relatively high concentrations of drug (5 – 100 uM). 

What is the precise role of BACE1 and ADAM10 in the amyloidogenic processing of AbetaPP? beta Amyloid is generated by sequential cleavage  events from a larger precursor protein (AbetaPP). First, the aspartic acid protease, beta-secretase (BACE1), cleaves AbetaPP in the extracellular region (see Figure below) to produce a secreted soluble fragment and a membrane-bound C-terminal fragment (C99). Subsequent cleavage of C99 by a separate protease, gamma-secretase, leads to the formation of A beta peptides. Hence, it is clear that downregulation of BACE1 would lead to decreased amyloidogenic processing of AbetaPP. A disintegrin and metalloprotease 10 (ADAM10), a type of alpha-secretase, cleaves AbetaPP within the A beta domain and thus precludes generation of intact A beta peptides. Therefore, upregulation of this enzyme promotes non-amyloidogenic AbetaPP processing by decreasing A beta production.

Protease cleavage of amyloid beta precursor protein.

The multilevel complementary activity of withanolide A may prove effective as a pharmacological therapeutic strategy for AD. Currently, the precise molecular mechanism by which withanolide A modulates levels of BACE1 and ADAM10 is not known. Withanolide A has also exhibited the rare ability to promote neurite outgrowth and reverse neuritic atrophy. In brief, cortical neurons were treated with A beta(25-35), an active partial fragment of amyloid beta, for four days to induce atrophy and then with withanolide A (1 uM). At four days after treatment, the cells were immunostained for an axonal marker or for a dendritic marker. Drug treatment induced regeneration of both axons and dendrites, and achieved reconstruction of pre- and postsynapses in the neurons. Moreover, lengths of the axons and dendrites treated with withanolide A were significantly extended relative to those treated with vehicle alone. Thus, it seems that the steroid lactone withanolide A, in addition to its A beta-related activities, also possesses the unique ability to reconstruct neuronal networks. This AD therapeutic lead compound also exhibited oral in vivo efficacy by recovering A beta(25-35)-induced memory deficits in mice. Memory-retention tests in male mice were quantified with an in vivo assay called the Morris water maze test.

In a forthcoming post, I will examine a practical partial synthesis of withanolide A starting from the readily available steroid precursor pregnenolone. The route was disclosed earlier this year in Angewandte Chemie Int. Ed. by the laboratory of Karl Gademann at the University of Basel.

Modern Steroid Science: An Introduction

Driven by the great medicinal potential of the steroidal sex hormones, anabolic substances and anti-inflammatory agents, the period of the 1930s through the 1950s has been called 'the golden age of steroid chemistry,' a time when researchers from industry (Syntex, Ciba and others) and academia (Johnson, Eschenmoser, Corey, Stork and others) alike fruitfully pursued and delineated the chemistry and biology associated with steroids and semisynthetic derivatives of great biomedical importance. Subsequently, the pharmaceutical industry shifted its focus to a library-based approach to drug discovery that favored small molecules whose physicochemical parameters fell within dimensions that were likely to confer oral bioavailability. In recent years, steroid research has seen a resurgence of activity, as a new generation of scientists has begun to recognize the pharmacologically privileged nature of the steroid molecular framework. This blog will highlight outstanding scientific achievements pertaining to steroid chemistry (organic synthesis and biosynthesis) and biochemistry (pharmacology / mechanism of action studies) published from 2010 to the present.