Redox-cycling occurs because the ortho-quinone is easily reduced by cellular enzymes (e.g., NAD(P)H Quinone Oxidoreductase 1, NQO1) but the resulting ortho-hydroquinone, unlike a para-quinone, is not stable and auto-oxidises rapidly, reducing oxygen to superoxide. A large increase in reactive oxygen species and consumption of NAD(P)H reserves promotes cellular catastrophe.
The cell-killing mechanism of ortho-quinones requires an initial enyzmatic reduction, which show selective toxicity to cancer cells overexpressing reductases, such as NQO1. A notable example is β-lapachone, a compound derived from the bark of the South American ‘Lapacho’ tree, which was investigated in the early 2000s by the late David Boothman and co-workers as a ‘Kiss of Death’ therapy for treatment of NQO1+ solid tumours. However, NQO1+ selectivity is not sufficient, and this is conceivably the reason why, despite high interest, no ortho-quinone has yet reached further than Phase II clinical studies. Indeed, the quinone may be reduced by other reductive enzymes, such as those at mitochondrial membranes, and it is challenging to achieve sufficient doses for efficacy without dangerous haematological side-effects liked to the redox mechanism, notably methemoglobinemia.
We initially became interested in the idea of targeting β-lapachone as a cancer therapy. We realised that a chemical strategy to prevent redox-cycling in circulation would be key for applications of any ortho-quinone. We were struck by the limited protection strategies that are described for the compounds. For example, classical methodologies are described that protect the quinone as a hydroquinone ester, an imine or a hydrazone. We imagined that protection of ortho-quinones with a self-immolative benzyl linker might open-up the possibility of more stable protease-cleavable pro-moieties. Similar systems have been applied successfully for targeted therapies of amine- and alcohol- containing drugs via benzyl carbamate and benzyl ether linkers, respectively.
We tried to synthesise O-benzyl hydroquinone derivatives, but we quickly realised that O-benzyl ortho-hydroquinones are unstable and rearrange to their C-benzyl ketol isomers, as also described by Shurygina and co-workers. Instead of discarding the benzyl ketol products we, somewhat luckily (!), decided to investigate their properties. Surprisingly, we saw that deprotected para-aminobenzyl ketols also undergo self-immolative elimination, with their C–C bond breaking in a 1,6-elimination to reform the hydroquinone and release an aza-quinone methide side product. This was a new strategy for ortho-quinone protection.
In our article, we demonstrate this protection strategy with several ortho-quinones. The benzyl ketol derivatives are redox-inactive, which is key for ortho-quinone targeted therapies. We performed modelling to understand our results and work out the active fragmenting species responsible for the elimination. Unlike para-aminobenzyl ethers, para-aminobenzyl ketols have an acidic rate-pH dependence. The rate-pH dependence is modulated depending on the structure and electronic environment of quinone species protected, which could lead to future tuning of release rates and kinetics.
We also show how our strategy may be applied. Dipeptide prodrugs of β-lapachone are cleavable by protease cathepsin B, with drug elimination occurring via the para-aminobenzyl intermediate. Additionally, an antibody-drug conjugate with protected β‑lapachone demonstrated an anti-tumour effect in an in vivo model of acute myeloid leukemia, with β‑lapachone detected in tumour tissue by LC MS/MS, which shows the potential of the technology for quinone-tumour targeting.
With this new protection strategy, controlled masking of ortho-quinones is now more viable. We hope that this chemistry that will enable more applications of ortho-quinone containing compounds in medicine and other areas.