Читать книгу Successful Drug Discovery, Volume 5 - Группа авторов - Страница 23

While total synthesis was shown not to be a feasible production route for epothilone and Taxol derivatives, the approach still proved to be key for the development of another microtubule stabilizing agent. In 1986, Hirata and Uemura described the isolation of several family members of a novel class of natural products from the marine sponge Halichondria okadai [64]. This class, named halichondrins, consists of several family members that vary in their oxidation state. They show a remarkable structural complexity. Halichondrin B (Figure 1.10) possesses a staggering 32 stereocenters. In particular halichondrin B displayed outstanding cytotoxicity against a panel of 60 human cancer cell lines, which at that time was newly established at the NCI and became known as the NCI‐60. Even more importantly, it showed excellent activity in in vivo cancer models. However, while it could be also detected in a few sponges of the Axinella, Phakellia, and Lissodendoryx families, its availability was extremely limited, as it could only be obtained in minimal quantities from the harvested sponges. Owing to the high potency of the compound, calculations indicated that only 10 g should be sufficient to supply clinical development and future need for commercialization was estimated to be between 1 and 5 kg. However, the producer organisms are rare, and it was calculated that at the time the available world supply of halichondrin B derived through extraction of one ton of harvested Lissodendoryx n. sp. 1 would amount to only 300 mg. Lissodendoryx n. sp. 1 is only found in an area of about 5 km² at a depth of 80 to 100 m, south of the coast of New Zealand. Calculations performed in 1993 estimated the total available biomass of Lissodendoryx to be only (289 ± 90) tons [65]. Yoshoito Kishi from Harvard University became interested in the unique structure of halichondrin B and set out to develop a synthetic access route. His main motivation was actually not in the anticancer properties of the drug, but at demonstrating the utility of the Nozaki–Hiyama–Kishi reaction in complex real‐world examples. This was a grand challenge, but in 1992, Kishi and his coworkers succeeded in completing the first synthesis, which comprised a total of 128 steps [66]. Also in 1992, the NCI nominated halichondrin B for preclinical testing. Eisai decided to license the synthesis of halichondrin B patented by the Kishi laboratory and initiated a very unique and fruitful collaboration in which researchers at Eisai were supplied with advanced intermediates by the Kishi laboratory. This joint effort led to establishment of several analogues and the understanding of the scaffold's SAR. In the course of this exploration, the anticancer activity of halichondrin B could be associated with the right‐hand side of the molecule, allowing a significant simplification of the molecule and finally resulting in the identification of E7389, later termed eribulin (Figure 1.10). The SAR studies and associated synthetic challenges have been reviewed in detail [67]. Compound availability by total synthesis was essential to start clinical work. Preclinical data for eribulin were more than convincing, but for internal reasons Eisai could not pursue the compound at the time, so it was decided to explore the compound's effects through a NCI‐sponsored phase 1 clinical trial. The first results were positive, so Eisai decided to sponsor further trials [68]. The compound received FDA approval in November 2010, only eight months after submission of the application. Today it is available in 50 countries for treatment of advanced metastatic breast cancer. It is the first drug that has shown improvement of survival in women with heavily pretreated metastatic breast cancer.

Figure 1.10 Structures of halichondrin B, eribulin mesylate, and E7130.

Albeit structurally significantly simplified, eribulin still bears 19 stereogenic centers and represents a showcase for organic synthesis, enabling access to structural complexity. Thorough optimization by the Kishi group [69] and by the Eisai process development group [70] led to significant improvement of the synthesis. Eribulin is now accessible in a 62‐step synthesis and still represents the most complicated technical synthesis of a marketed drug to date. This record may be in danger, though, as the Kishi group recently reported the synthesis of an even more complex development candidate, termed E7130 (Figure 1.10). While the initial synthesis took a total of 109 steps, they managed to improve the syntheses to “only” 92 – significantly higher yielding – steps and obtained remarkable 11 g of material [71]. E7130 is now undergoing clinical trials and may eventually become a successor to eribulin.