Drug approval marks culmination of a marathon trek from sea sponges to clinic.
The latest breast-cancer chemotherapy to hit the market is more than just a triumph for patients in desperate need of treatment. Approved by the US Food and Drug Administration on 15 November, the highly complex molecule Halaven (eribulin mesylate) is the product of nearly 25 years of struggle in the lab. It represents a hard-won victory for the total synthesis of natural products, a field of chemistry that, although still popular in academia, had gone out of fashion for many in the pharmaceutical industry.
Eribulin is a synthetic compound that mimics part of the structure of halichondrin B, a molecule found in the sea sponge Halichondria okadai. Researchers learned that halichondrin B has potent tumour-fighting activity shortly after its discovery in 1986. But it is present in very low concentrations, making it difficult to isolate. The compound also has a fiendishly complicated structure — at the time of its discovery, producing it from scratch was well beyond the abilities of chemists.
A few years later, however, organic chemist Yoshito Kishi of Harvard University in Cambridge, Massachusetts, eyed the halichondrin B structure and decided to take a crack at it. His team had little interest in its anticancer properties, he says. They were simply looking for a project to test a chemical reaction — the Nozaki–Hiyama–Kishi reaction — that could be used to build bonds between carbon atoms.
Kishi's team had set themselves an enormous challenge with halichondrin B. Natural products often contain carbon stereocentres, in which surrounding atoms can be arranged in two mirror-image configurations. "If you don't get the stereocentres set up perfectly, it generates a mixture" of different molecules that can be extremely troublesome to separate, says Ian Paterson, a chemist at the University of Cambridge, UK, who works on natural-product synthesis. Although two mirror-image forms of a molecule are indistinguishable for most chemical reactions, they can produce completely different biological effects.
Halichondrin B has a staggering 32 stereocentres, meaning that there are 232 — more than 4 billion — possible forms, or isomers, of the molecule. "It's just ridiculous," says Robert Salomon, an organic chemist at Case Western Reserve University in Cleveland, Ohio, whose lab spent four years unsuccessfully trying to synthesize the compound in the early 1990s.
Nevertheless, Kishi's team succeeded. By the time he published a method for synthesizing the compound in 1992 (T. D. Aicher et al. J. Am. Chem. Soc. 114, 3162–3164; 1992), researchers at the Natural Products Branch of the US National Cancer Institute (NCI) in Frederick, Maryland, had discovered that halichondrin B fights cancer cells by inhibiting a protein component of the cytoskeleton — the internal latticework of rods and filaments that gives a cell its shape. That protein, called tubulin, is needed to support the rapid growth of cancer cells and is the target of several other cancer chemotherapies, including Taxol (paclitaxel).
But Kishi's synthesis was practical for generating only small quantities of halichondrin B, unlikely to be enough to usher the compound through preclinical and then clinical testing, says David Newman, now chief of the NCI's Natural Products Branch. Newman decided that he would simply isolate the compound from natural samples. So he headed for the sea to hunt for the prized compound.
Newman and his team collected more than one tonne of Lissodendoryx, another type of sponge containing halichondrin B, from the deep waters off New Zealand. He also teamed up with researchers to grow more of the sponges, flying seaplanes out to remote aquatic farms where the sponges grew attached to lines dangling 40 metres beneath buoys. The reward for his efforts: just 300 milligrams of halichondrin B, the equivalent of a few grains of rice. "My hair turned white as a result of halichondrin B," he jokes.
Meanwhile, Tokyo's Eisai Pharmaceuticals had licensed the patent on Kishi's method and began synthesizing hundreds of analogues of the compound. Newman's haul from New Zealand was just enough to conduct comparative studies with some of these analogues. One of them, eribulin, is more potent than halichondrin B yet also substantially smaller and easier to make. But it still has 19 stereocentres (see structure), and production of eribulin on a commercial scale seemed unfathomable.
Eisai says that eribulin takes 62 steps to synthesize — a remarkably long process for a marketable drug. The company was initially apprehensive about the project, says Kishi. But once the phase I study results had shown that the drug was safe — and revealed hints of clinical efficacy — "all the reservations disappeared", he says.
Further clinical trials showed that eribulin extends the lifespan of patients with late-stage breast cancer by an average of 2.5 months in those who are not benefiting from other chemotherapies such as Taxol, also a natural-product derivative. Analysts suggest that eribulin could command a US$1-billion market if it is approved for treatment of other cancers.
Few other pharmaceutical companies have been willing to bet on complex natural products. During the 1990s, many largely abandoned natural-product chemistry, focusing more on screening large libraries of synthetic chemicals for drug candidates, says Michael Jirousek, who once worked on halichondrin B synthesis and is now chief scientific officer and co-founder of Catabasis, a biotechnology company in Cambridge, Massachusetts. "Screening natural products and isolating the active ingredients is becoming a lost art," he says.
Proponents of total synthesis point to eribulin as proof that their approach, albeit arduous, can be highly successful. Phil Baran, a synthetic chemist at the Scripps Research Institute in La Jolla, California, says that more young investigators are entering the field and that improvements in chemical techniques are making it possible to synthesize additional complex molecules by commercially viable routes. "As advances in organic chemistry become greater and greater," he says, "I think we're going to see a lot more complex compounds being pursued by companies."
Related links in Nature Research
Related external links
About this article
Making better decisions during synthetic route design: leveraging prediction to achieve greenness-by-design
Reaction Chemistry & Engineering (2019)
Construction of Sulfonyl Phthalides via Copper-Catalyzed Oxysulfonylation of 2-Vinylbenzoic Acids with Sodium Sulfinates
The Journal of Organic Chemistry (2019)
A Cascade of Prins Reaction and Pinacol-Type Rearrangement: Access to 2,3-Dideoxy-3C-Formyl β-C -Aryl/Alkyl Furanosides and 2-Deoxy-2C-Branched β-C -Aryl Furanoside
European Journal of Organic Chemistry (2018)
Nature Reviews Chemistry (2018)
Journal of Natural Products (2018)