Many of the approved chemotherapy drugs are either collected from natural sources or inspired by molecules existing in nature. Natural products with anticancer activity can be readily identified in cytotoxicity assays and other phenotypic screens, but before they can be developed into drugs, it is essential to characterize the molecular mechanism of action that underlies their cytotoxicity. However, the mechanism of action of many natural products that display promising anticancer activities in phenotypic screens remains uncharacterized because this represents a challenging and time consuming endeavour.
An example of such a natural product is ophiobolin A (OPA), a compound naturally made by pathogenic fungi of the Bipolaris genus in order to attack plant cells. It has been shown that OPA displays cytotoxicity at nanomolar concentrations against a range of cancer cell lines and induces paraptosis, a form of non-apoptotic cell death, in glioblastoma cells. Interestingly, OPA also displays antitumor activity in a mouse model for glioblastoma, which is a type of brain cancer notoriously difficult to treat with existing medications. Unfortunately, the development of OPA into a potential drug for glioblastoma treatment has been impeded by the lack of any identified cellular targets.
In our recent study published in eLife, we employed a genome-wide screen in human cells to identify genes required for OPA to exert cytotoxicity. We used insertional mutagenesis in the near-haploid human cell line KBM7, a technique pioneered by Jan Carette and Thijn Brummelkamp at the Whitehead institute, to generate a genome-wide collection of loss-of-function mutants and then selected for growth of cell lines resistant to OPA treatment. We discovered that inactivation of the pathway for de novo synthesis of phosphatidylethanolamine (PE), also named the Kennedy pathway, confers resistance to OPA in human cells. We also observed that resistance to OPA through Kennedy pathway inactivation was always associated with a reduction in cellular PE levels.
Our efforts to characterize the cellular target responsible for this genetic interaction led to the surprising discovery that OPA doesn’t target a protein constituent of the Kennedy pathway, but rather targets the lipid molecule PE directly. PE is the second most abundant phospholipid in human lipid bilayers and plays an important role structuring the membrane due to the biophysical properties of its ethanolamine headgroup. We discovered that OPA reacts with PE to form stable covalent adducts. We first characterized the reaction in vitro using liquid chromatography-mass spectrometry (LC-MS) and were next able to detect the formation of PE-OPA adducts in human cells treated with OPA. Modification of PE to PE-OPA adducts substantially changes the biophysical properties of this membrane lipid by modifying its head group from small and polar to bulky and hydrophobic. We thus hypothesized that PE-OPA adduct formation could lead to destabilization and permeabilization of cellular lipid bilayers and be the main cause of OPA cytotoxicity. We tested this hypothesis using lipid vesicles composed of variable amounts of PE and loaded with a fluorescent dye. We observed that OPA treatment indeed caused extensive vesicle leakage, and that the extent of leakage was directly dependent on both the PE content and on the OPA concentration. Importantly, we observed that vesicles with no PE or with very low PE content remain intact even at high OPA concentration, indicating that the leakage is likely specifically due to the formation of PE-OPA adducts.
In summary, we used an unbiased genetic screen in human cells to decipher the mechanism of action of OPA, an interesting candidate anti-cancer drug. Based on our results, we believe that when OPA is applied to cells it accumulates in the phospholipid bilayer of the plasma membrane, as it is a lipophilic compound. Due to high local concentrations or the hydrophobic environment (or both), OPA efficiently reacts with the primary amine head group of PE. The formation of PE-OPA adducts in the membrane changes the biophysical properties of PE, leading to membrane permeabilization and ultimately cell death. Differences in PE composition and distribution between cancerous and normal cells have been observed, notably by the Thorpe lab, and could be the basis for the increased sensitivity of cancer cells to OPA. Our discovery of this unusual cytotoxicity mechanism opens the exciting possibility to use OPA as an effective chemotherapy tool.