Stereodynamic investigation of labile stereogenic centres in dihydroartemisinin

Abstract

Since its identification in the early 1970s, artemisinin, as well as semi-synthetic derivatives and synthetic trioxanes, have been used in malaria therapy. Reduction of artemisinin by NaBH4 produced dihydroartemisinin (DHA), and yielded a new stereochemically labile centre at C-10, which, in turn, provided two interconverting lactol hemiacetal epimers (namely alpha and beta), whose rate of interconversion depends on buffer, pH, and solvent polarity. Since interconversion of the two epimers occurred on a chromatographic time-scale, this prompted a thorough investigation of the phenomenon as a crucial requisite of any analytical method aimed at quantitating this family of drugs. In this critical review we discuss the current importance of the on-column epimerization of DHA in the development of analytical methods aimed at quantifying the drug, with the purpose of identifying the optimal conditions to minimize on-column epimerization while achieving the best selectivity and efficiency of the overall separation.

Figure 1

Chemical structures of artemisinin (1), dihydroartemisinin (DHA, 2), artemether (3), arteether (4), artelinic acid (5), artesunic acid (6), and a ubiquitous thermal decomposition product of 2 designated as diketoaldehyde (DKA, 7).

Figure 2

Chemical structures of the two interconverting epimers of DHA (2): the 2α-epimer bears the hydroxyl group in the equatorial position (absolute stereochemistry at C-10: R), whereas the 2β-epimer possesses an axial hydroxyl group.

Figure 3

Polytube models of the two interconverting epimers of DHA (2): the 2β-epimer (right) was obtained by computer editing of the X-Ray data of crystalline 2β [10]; the model for the 2α-epimer (left) was derived by molecular mechanics optimization (MMFF force field as implemented in SPARTAN 04) by inverting the configuration at C-10.

Scheme 1

Primary (vertical arrows) and secondary (horizontal arrows) equilibria taking place during chromatography of two interconvertible species A and B on a stationary phase. Left: actual interconversions occurring in the mobile and stationary phases; right: the apparent equilibria. k’_A and k’_B are the retention factors of the first (A) and second (B) eluting species, k₁^m and k₋₁^m are the rate constants for the forward and backward interconversion in mobile phase, respectively, and k₁^s and k₋₁^s are the rate constants for the forward and backward interconversion in stationary phase, respectively.

Figure 4

DHPLC traces of a tertiary amide in the form of variable temperature HPLC.

Figure 5

Computer-simulation of experimental deformed profiles of two interconverting species. Red line: experimental profile. Blue line: computer-simulated profile. Green lines: computer-simulated profiles of the individual interconverting species.

Figure 6

Typical room temperature chromatograms obtained for a standard mixture of 1 and 2 (containing 7 as impurity). Peak 1 corresponds to 7, peak 2 to the 2α-epimer, peak 3 to the 2β-epimer, and peak 4 to 1. Reproduced from Figure 3 in Reference [41] with permission from Elsevier. Copyright 2008.

Figure 7

(A) Schematic representation of the integration mode used for the calculation of 2α- and 2β-epimers areas. (B) Computer simulated profiles of 2α (dotted line), 2β (dashed line), and of the mixture 2α+ 2β (solid line). Reproduced from Figure 5 in Reference [41] with permission from Elsevier. Copyright 2008.

Figure 8

Variable temperature chromatographic profiles obtained on the Symmetry C₁₈ column. Solid line: experimental chromatograms. Dotted line: computer simulated profiles obtained with the measured free energy activation barriers for the on-column epimerization process. Reproduced from Figure 7 in Reference [41] with permission from Elsevier. Copyright 2008.