Variation in DNA enrichment among deuterium labeling studies is largely explained by different background corrections

Heavy water (D2O) labeling can be used to track the dynamics of circulating cells in vivo. D2O is administered via the drinking water, dividing cells become labeled by incorporating deuterium in their newly formed DNA, and the accrual of labeled deoxyribose molecules in the DNA is measured by gas-chromatography mass-spectrometry (GC-MS). Deuterium labeling studies need to correct for a background enrichment that is due to several naturally occurring isotopes, but the procedure for performing this correction differs between studies. We show that a background correction that is based upon subtracting the background enriched fraction (i.e., the atom percent excess, APE) underestimates the true enrichment, whereas one that is based upon subtracting the background enriched ratio (i.e., the tracer-to-tracee ratio, TTR) provides an enrichment that correctly describes the additional effect of the deuterium labeling. This difference is reflected in the so-called 'amplification factor', which quantifies the relative enrichment of deoxyribose in a fast reference population of cells to the enrichment in the body water. Using mechanistic binomials we reanalyze deuterium labeling data and show that previous studies based upon an APE-based background correction have underestimated this amplification factor by about 25%, which explains most of the variation among studies in their average amplification factor. We propose a novel model in which the amplification factor is defined by a binomial expression allowing for metabolic differences between volunteers in the contribution of body water to the hydrogens incorporated into deoxyribose during the synthesis of DNA.