Oxygen is the quintessential two-edged sword of molecular biology: essential for (animal) life, but at the same time a perennial source of damage to macromolecules. Reactive oxygen species (ROS), arising from both external sources and the intrinsic metabolic machinery of the cell itself, have been implicated in many aspects of cellular aging.

Of particular interest to human beings, especially those living in the rapidly aging post-industrial Western nations, is the relationship between oxidative damage and neurodegenerative illness. While most of the age-related neurodegenerative diseases are caused by accumulation of protein aggregates, it is becoming evident that ROS play an important role in exacerbating the underlying pathologies: e.g., DNA oxidation arises early in the pathogenesis of Alzheimer’s disease; and oxidative damage to a key anti-oxidant defense protein may generate a pernicious positive-feedback loop in the initiating events of Parkinson’s disease. It is straightforward to imagine how such damage could accumulate in nondividing cells like neurons, which neither synthesize new DNA nor dilute out damaged proteins by cell division.

I usually think of oxidative damage as occurring after macromolecules are initially made — moreover, made in a pristine, undamaged form. Reading between the lines of a recent paper on Parkinson’s disease, however, I found myself questioning this implicit model, and wondering whether sometimes the relevant damage might be present even before synthesis begins.

The paper, by Nakabeppu et al., describes a mouse knockout of MTH1, a gene involved in defense against oxidation of DNA. The MTH1 protein identifies a specific oxidatively damaged nucleotide triphosphate (8-oxo-dGTP) and takes it out of commission before it can be incorporated into DNA. The authors find that mice lacking MTH1 show a dramatic increase in oxidative DNA damage (specifically, of the type that MTH1 prevents) as well as increased cell death in response to oxidative stress. They argue that their model provides further evidence that oxidative damage to DNA plays a significant role in the onset of Parkinson’s disease.

Oxidative DNA lesions, such as 8-oxoguanine (8-oxoG), accumulate in nuclear and mitochondrial genomes during aging, and such accumulation can increase dramatically in patients with Parkinson’s disease (PD). To counteract oxidative damage to nucleic acids, human and rodents are equipped with three distinct enzymes. One of these, MTH1, hydrolyzes oxidized purine nucleoside triphosphates … to their monophosphate forms. … We have shown a significant increase in 8-oxoG in mitochondrial DNA as well as an elevated expression of MTH1 … in nigrostriatal dopaminergic neurons of PD patients, suggesting that the buildup of these lesions may cause dopamine neuron loss. We established MTH1-null mice and found that MTH1-null fibroblasts were highly susceptible to cell death caused by H2O2 … and that this was accompanied by an ongoing accumulation of 8-oxoG in nuclear and mitochondrial DNA. We also showed that MTH1-null mice exhibited an increased accumulation of 8-oxoG in striatal mitochondrial DNA, followed by more extreme neuronal dysfunction after 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine administration than that of wild-type mice. …

So far so good: a simple story in which a mutant with diminished antioxidant defenses is more sensitive to oxidative stress, and exhibits an acceleration in the sort of DNA damage that is present in (and possibly causative of) Parkinson’s. But there’s an intriguing complication, and it blew me away once I thought about it.

Remember, MTH1 is involved in eliminating oxidized dNTPs before they can be incorporated into DNA. So in order for MTH1 to play a direct role in limiting oxidative damage to DNA, it must be the case that oxidized dNTPs are being actively incorporated into DNA — in other words, DNA can be synthesized in a damaged form.

Another corollary to these findings is that neurons (which don’t divide) are constantly turning over their DNA, at least rapidly enough for 8-oxo-dGTP to be incorporated at a significant rate — n.b. the result of the hydrogen peroxide experiment, which resulted in higher levels of 8-oxoG in the nuclear DNA of mice that lacked MTH1. If all MTH1 is doing is clearing 8-oxo-dGTP before it’s incorporated into DNA, then it must be the case that the neurons are synthesizing nuclear DNA. Perhaps the synthesis occurs constantly, and perhaps it only happens after genotoxic damage (e.g., peroxide exposure), but either way it’s happening enough for pre-existing damage in the pool of nucleotide precursors to have an impact on the steady-state levels of damage in the nuclear genome.

I haven’t sorted through all the ramifications of this idea yet, but it seems major: We think of damage as occurring after synthesis, and of repairing damage as a return to a pristine original state. But if the reason that aged cells exhibit high steady-state levels of damage is not simply accumulation with time, but rather that macromolecules are increasingly being synthesized in a damaged state…well, it seems like a trip to the drawing board might be in order.