Halaven earned an FDA approval in 2016 as the first drug to improve survival with advanced liposarcoma—by seven months.1 Just making this drug was, and is, a major molecular engineering effort.
There would be no man-made Halaven without its natural parent, Halichondrin B (HB, below).2 In the 1980’s, Japanese scientists isolated a rice grain-sized amount of HB from a 600 kilogram batch of Halichondria okadai sea sponge (above).3 As well as determining its structure they learned that at nanomolar concentrations it inhibits cancer cell growth, and that it improves the survival of cancer-stricken mice.4 Further investigation revealed its unique effects on microtubules, the structures that split apart chromosomes in mitosis—an essential act in quickly dividing cancer cells.4 Because of its potency and confidence in its mechanism, the U.S. National Cancer Institute (NCI) wanted to study HB in further detail and maybe test it on humans. Clearly they would need an efficient source.4
Before the NCI got so involved, Yoshito Kishi’s laboratory at Harvard University was working on the chemical synthesis of HB.5 Their “purely academic” goals at the outset were to test their skills and show off their bond-building reaction in a challenging context.6 The team succeeded and used their Nozaki–Hiyama–Kishi reaction to set five bonds in the final product. However, their route had over 100 steps, which is an order of magnitude longer than customary drug syntheses.7
Because chemistry seemed impractical, a New Zealand-based team began to farm sponges to meet the demand. The NCI calculated that 10 grams would be required for human trials, and if approved, a further 1–5 kilograms every year would be needed.4 So the New Zealanders worked for six years, testing different depths, locations, and temperatures. Their farms were so remote that they needed to take a seaplane to reach their sponges, which they grew on lines at least 40 meters below the sea (below).8 In the end, only 310 milligrams could be isolated from one metric ton of optimally grown sponge. Even though the aquaculture was “economically feasible” at 0.000031% mass yield, a better option was sought.9
Drug company Eisai sensed an opportunity in Kishi’s total synthesis. One of Kishi’s synthetic intermediates might contain the minimum structure needed for activity, and they could find it by screening these HB fragments in cell-based assays. None of the fragments in the western portion were active. But a fragment from the east (below) was found with similar potency to HB, and it shared the same cell-killing mechanism. The problem was that it showed no effect on tumors in mice.4 This situation is typical in medicinal chemistry and it can be overcome by putting a molecule through repeated cycles of refinement. Over the next five years, chemists created analogs with slight, usually one-at-a-time changes to keep and build in drug-like properties.4
Eisai chemists made nearly 200 analogs of the active intermediate and checked their progress along the way in cell-based tests. Early series A and B did not turn up any improvements. Series A members had small alterations to the pyranopyran system (green) like deletion, addition, inversion, or ring closure. Series B members had changes to the isolated pyran (red) and furan (purple), like exocyclic methylene and methyl deletions. When they simplified pyranopyran to pyran in series C or furan in series D, the molecules had the right combination of toxicity to cancer cells and ability to stop mitosis. The shown series C and D representatives were already more potent and simpler than the active intermediate, and they would be the starting points for more analogs.4
That series D member was morphed into E7389 by replacing an oxygen with methylene (red; below on left), which improved its stability in blood plasma, and perfecting its water-solubilizing chain (blue). Now the compound was superior to paclitaxel in animal models, at well below its maximum tolerated dose, and it had lower neurotoxicity. Although its manufacture took 62 steps,* optimized purifications—usually 60–80% of process chemistry expenses10—and modular construction (below on right) kept expenses low. This mix of features prompted a first-in-man trial.4
Phase I testing for E7389 began in 2002, and eight years later it was granted FDA approval for treatment of metastatic breast cancer.11 Each year in the U.S. 14,000 to 22,000 breast cancers are initially diagnosed at stage IV. Meanwhile 20 to 30% of existing cases are metastatic recurrences.12 Halaven (generic name eribulin) extends life by 2.5 months (P = 0.041) versus other monotherapy, with an objective response rate of 11%.13 Halaven won its second indication for inoperable and metastatic liposarcoma in January 2016.1 Annually about 1,000 new liposarcoma cases are diagnosed in the U.S.,14 sometimes with unclear margins or after spreading widely.15 Halaven showed an overall survival benefit [15.6 versus 8.4 months; HR (95% CI) = 0.51 (0.35, 0.75)] when compared against dacarbazine in patients with this particular soft tissue sarcoma.13
Halaven not only improves survival, it can also offer some quality of life. Intravenous infusions last 2–5 minutes and no premedication or solubilizing agent is needed.13 With paclitaxel, infusion can last up to 24 hours and a sometimes allergenic solubilizing agent is necessary.16
Halaven is the most complex wholly synthetic drug6,18 and the efforts to discover and develop this drug now seem worthwhile.
* Later on Eisai added 7 steps to their synthetic route, which actually reduced costs because it eliminated column chromatography. See this article from Chemistry World.
2. Pure Appl. Chem. 1986, 58, 701–710.
4. Yu, M. J. Kishi, Y., Littlefield, B. A., in Anticancer Agents from Natural Products, page 241; Editors Cragg, G. M., Kingston, D. G. I., and Newmann, D. J. Published by CRC press, Taylor and Francis group, Boca Raton, 2005.
5. J. Am. Chem. Soc. 1992, 114, 3162–3164.
7. (a) Beilstein J. Org. Chem. 2013, 9, 2265–2319; (b) Beilstein J. Org. Chem. 2011, 7, 442–495.
10. Chem. Eng. News. 2000, 78, 26.
13. FDA Label for Halaven (eribulin). http://www.accessdata.fda.gov/drugsatfda_docs/label/2016/201532s015lbl.pdf
16. FDA Label for Taxol (paclitaxel). http://www.accessdata.fda.gov/drugsatfda_docs/label/2011/020262s049lbl.pdf
17. J. Clin. Oncol. 2006, 24, 1633–42.
18. Synthesis of Heterocycles in Contemporary Medicinal Chemistry. Springer International Publishing. Editor Zdenko Časar. Series volume 44.
Halichondria okadai (Photo from Yasunori Saito) http://www.nature.com/news/2010/101130/full/468608a.html
Sponge farming picture http://www.marinecultures.org/en/gallery/
Prefabricated home construction http://rbahomes.com/photo-gallery-roofs-and-dormers.html
Please note that this is not medical advice.