ABSTRACT

I use the Nernst equation, parameterised with experimental data, to predict that cancer cells will accumulate more of a lipophilic anion than normal cells. This effect is correlated to charge number. Model cancer cells accumulate *100 more of an anion, *10³ more di-anion, *10⁶ more tri-anion, *10⁸ more tetra-anion and *10¹⁰ more penta-anion (>>1 billion times more). The trend endures, conveying even greater specificity, for higher charge numbers. This effect could be leveraged for cancer therapy. Wherein the lipophilic anion is a toxin that targets some vital cellular process, which normal and cancer cells may even share. It delivers a high, lethal dose to cancer cells but a low, safe dose to normal cells. This mathematical finding conveys the prospect of a broad, powerful new front against cancer.

A new drug design rule: the more anionic a drug candidate, the greater its probable selectivity for cancer cells

I use the model to suggest that anionic character should be a new selection criterion for cancer drug design. The more anionic, the better. This stipulation can be applied alone – or combined with Lipinski's rule of five [58], which predicts “drug-likeness” (oral bioavailability) – to generate a new drug design rule for cancer specific drugs. This can guide novel drug screens e.g. systematically testing drugs that conform to the rule of five (or some variant of it) in turn: from the most to the least anionic. This should find some of the best drugs early. Indeed, I have conducted such a screen myself. Excitingly, there are some very anionic compounds that conform to the rule of five. For example, anions with −6 charge e.g. C₁₄H₂₂O₁₀P₂^-6 [59]. My model predicts that cancer cells will accumulate these >1 trillion times (!) more than normal cells. Some of the molecules I have found in my screens have had some bioactivity established, although not necessarily cancer-specific, by prior studies. But again, to repeat from earlier, the differential in accumulation is so great that targets which normal and cancer cells share are valid therapeutic targets for drugs of this class. For example, C₂₈H₁₉NO₁₅S₃^-4 [60] disrupts histone modification [61]. This is an important process in transcriptional control and DNA packaging. My model predicts that this chemical, and its disruption, will be targeted to cancer cells 100 million times more than normal cells. Or there is C₂₈H₁₇N₅O₁₄S₄^-4 [62], which inhibits human Flap Endonuclease 1 ([63]; FEN1; very important in genomic stability). Again, my model predicts that this poison will be targeted to cancer cells 100 million times more than normal cells.