Thunder god vine may not be a useful herbal medicine but the compounds isolated from it are fascinating – if not as medicines, then most certainly as laboratory tools. Nature Chemical Biology recently published an article where a research team from Johns Hopkins, the University of Colorado at Boulder, and Drew University in New Jersey, has determined the molecular mechanism of action of triptolide, an unusual triepoxide compound from the plant.

Tripterygium wilfordii Hook F, or thunder god vine, is known as lei gong teng in Chinese traditional medicine and has a history of use as an anti-inflammatory herb. As with many traditional medicines, usage patterns do not necessarily indicate scientific validity. In fact, a Cochrane review published just last month on herbal therapies for rheumatoid arthritis indicated that the efficacy of thunder god vine was mixed. More concerning is that the herb had significant adverse effects in some trials, from hair loss to one case of aplastic anemia.

Nevertheless, the herb’s components have been studied since the 1970s for since they also appears to kill tumor cells in culture with nanomolar potency and have immunosuppresant activity in animal models. The group of the late natural products chemist at the University of Virginia, S. Morris Kupchan, first identified the unusual structures of triptolide and tripdiolide from Tripterygium wilfordii as described in this 1972 paper from the Journal of the American Chemical Society. Cytotoxic activity toward tumor cells in culture was used to guide the chemical fractionation of extracts. The unusual presence of three consecutive epoxides in the structures of both compounds led Kupchan to hypothesize later in Science that they target leukemia cells by covalent binding to cellular targets involved in cellular growth.

As an aside: Epoxides are chemically reactive groups composed of an oxygen atom bonded to two carbons; the constraints of this triangular structure and the electrons on the oxygen favor the opening of this ring and attack of other atoms such as sulfur, often present in regulatory regions of enzymes. The Wikipedia entry gives a pretty nice primer. The reactivity of epoxides also makes these compounds highly useful intermediates in industry, particularly in the manufacture of ethylene glycol antifreeze and industrial paints and adhesives (e.g. epoxy resins).

Conventional wisdom would drive most scientists to take one look at triptolide and say that this stuff is a royal mess – so chemically reactive that it couldn’t possible have a specific cellular target. It’s probably too “dirty” – so promiscuous in its binding that it probably attacks all manner of sulfhydryl-containing enzymes and blows the cells to smithereens.

However, several groups have shown over the last 10 or 15 years that some epoxide-containing natural products have very specific cellular targets. Epoxides are not so wildly reactive that they bind everything in their midst. Instead, the environment in which the epoxide exists seems to provide some binding specificity. For example, the group of Jun O. Liu, then at MIT, showed in 1998 that another epoxide-containing natural product, fumagillin, exerted its antiangiogenic activity by binding to a protein called methionine aminopeptidase 2 (MetAP2). Similarly, Brent Stockwell at Columbia University and the Howard Hughes Medical Institute recently published a tour de force in another Nature Chemical Biology paper showing that a reactive 2-chloromethylketone compound specifically targets protein disulfide isomerase, preventing neuronal cell death from misfolded proteins with potential use in Alzheimer’s or Huntington’s diseases.

Jun Liu was again at the helm in the current thunder god vine study in Nature Chemical Biology. The group started with a simple approach to narrow down the target of triptolide from thunder god vine: they treated the venerable HeLa cervical carcinoma cell line with the drug and examined the incorporation of the building blocks of DNA, RNA, or protein. Triptolide was several orders of magnitude more potent in rapidly inhibiting RNA synthesis.

In an elegant series of experiments, the researchers progressively dissected the modulation of RNA synthesis – first identifying the multiprotein complex of RNA polymerase II (RNAPII) as the target but acting via a mechanism different from known RNAPII inhibitors such as the mushroom toxin, ?-amanitin.

Further experiments revealed that triptolide bound to a transcription factor component of the RNAPII complex called TFIIH. Then, finally, the investigators demonstrated that triptolide specifically attacked a component of TFIIH called XPB. XPB is a type of DNA unwinding enzyme called a helicase and is involved in DNA repair. The group then made semi-synthetic chemical analogs of triptolide to determine how inhibition of the ATP hydrolyzing activity of XPB correlated with potency in killing HeLa cells. While the rank order of potency of the compounds correlated, the drugs were less potent in attacking the enzyme activity of the XPB protein than in killing HeLa cells. So, it’s unclear as to exactly how binding to XPB is leading to cell killing. The investigators do note that triptolide may have other cellular targets that are less abundant than XPB that contribute to its activity.

Of course, we can’t tell right now if triptolide selectively kills tumor cells relative to normal cells. Again, conventional wisdom would argue that a drug that hits such a crucial target as a transcription factor is unlikely to have selective activity. After all, the classic RNAPII inhibitor ?-amanitin is well-known as a lethal toxin responsible for legendary poisonings by the death cap mushroom, Amanita phalloides. However, low concentrations of such a compound might indeed have some selectivity when given together with a DNA-damaging anticancer drug. But that’s a very fine tightrope to walk.

In the end, triptolide may end up “just” being a useful laboratory tool for understanding the basics of gene transcription and DNA repair. But if normally disregarded epoxides do indeed have some specificity in their action on cellular targets, perhaps analogs can be made with selective action against tumor cells. Many triptolide analogs have been synthesized over the years and should certainly be revisited in the context of cancer treatment. But this finding should also serve to warn us that the indiscriminate use of the herb as an anti-inflammatory should be revisited, particularly if the dose of the herb gives variable concentrations of compounds with a very low margin of safety.

Titov, D., Gilman, B., He, Q., Bhat, S., Low, W., Dang, Y., Smeaton, M., Demain, A., Miller, P., Kugel, J., Goodrich, J., & Liu, J. (2011). XPB, a subunit of TFIIH, is a target of the natural product triptolide Nature Chemical Biology, 7 (3), 182-188 DOI: 10.1038/nchembio.522