Case-control study finds altered lipoprotein metabolomic signature in m.3243A>G carriers

A single change in mitochondrial DNA, one letter swapped for another in a gene carried by thousands of families across the world, has now been linked to a strikingly specific metabolic fingerprint in the blood. In a case-control study published in the journal Metabolomics, researchers in Denmark report that carriers of the m.3243A>G mitochondrial DNA variant show a pronounced dysregulation of lipoprotein metabolism, with elevated triglycerides across multiple lipoprotein subclasses and altered concentrations and composition of very-low-density lipoprotein particles. The findings, based on high-throughput nuclear magnetic resonance metabolomics of fasting serum and urine, offer the most detailed picture yet of the systemic metabolic disturbances that accompany this mutation, and they hint at new ways to identify carriers and understand their elevated risk of diabetes.

The m.3243A>G variant sits in the MT-TL1 gene, which encodes a mitochondrial transfer RNA essential for building the protein complexes of the oxidative phosphorylation system. When the mutation is present, mitochondria produce less adenosine triphosphate, the chemical fuel that powers nearly every energy-demanding process in the cell. Because each mitochondrion contains multiple copies of the mitochondrial genome, the severity of disease depends on heteroplasmy, the proportion of mutant copies relative to normal ones. Carriers can develop a bewildering spectrum of symptoms, including hearing loss, muscle weakness, stroke-like episodes, cardiomyopathy, and a form of diabetes known as maternally inherited diabetes and deafness, or MIDD. Despite decades of clinical recognition, the systemic metabolic alterations that drive this heterogeneity have remained incompletely mapped.

To fill that gap, a team led by Simone Rask Nielsen of Aalborg University Hospital recruited 28 adult carriers of the m.3243A>G variant from 13 different families, drawing on the hospital’s clinical genetics department and an established Danish carrier cohort. Each carrier was compared with a healthy control individually matched for age and sex, with controls additionally required to have a body mass index below 26 kilograms per square meter to exclude the confounding effects of obesity-related metabolic disturbance. Sixteen of the carriers had been diagnosed with diabetes according to American Diabetes Association criteria, while 12 had not. Fasting blood and urine samples were collected in the early morning under a standardized protocol, and participants abstained from their usual morning medications before sampling.

The analytical centerpiece of the study was an automated nuclear magnetic resonance metabolomics platform developed by Nightingale Health, which quantified 169 metabolites in serum and 51 in urine in absolute concentrations. Unlike mass spectrometry, NMR spectroscopy excels at robust, reproducible measurement of lipoprotein subclasses, apolipoproteins, fatty acids, amino acids, and low-molecular metabolites such as ketone bodies and glycolysis intermediates, all from a single measurement. Heteroplasmy levels in whole blood were quantified separately using droplet digital polymerase chain reaction, allowing the team to relate mutation burden to metabolic output.

The results were unambiguous. Univariate analysis identified 25 serum metabolites and 16 urine metabolites that separated carriers from healthy controls. In serum, the differences were dominated by the lipoprotein system: carriers displayed increased triglyceride content in multiple lipoprotein subclasses and elevated concentrations of very-low-density lipoprotein particles, particularly the small and very small subclasses. Circulating lactate and pyruvate, two classic markers of impaired mitochondrial function and stalled glycolytic flux, were also increased. In urine, 16 metabolites were reduced in carriers, spanning pathways that include glycolysis, the tricarboxylic acid cycle, glutathione metabolism, one-carbon metabolism, and nucleotide metabolism. Reduced urinary levels of citrate and glutamine pointed to a struggling citric acid cycle, while diminished pyroglutamate suggested compromised antioxidant defense through the glutathione system.

Perhaps the most eye-catching result came from an exploratory machine learning analysis of urine. Three metabolites in particular, uracil, hypoxanthine, and 1-methylnicotinamide, showed strong discriminatory power between carriers and controls. When the researchers fed these three markers into four different classification algorithms, including random forest, linear support vector machine, partial least squares discriminant analysis, and logistic regression, the models achieved areas under the receiver operating characteristic curve between 0.94 and 0.99, with cross-validated prediction accuracies of 0.81 to 0.93. In practical terms, a simple three-molecule urine test could distinguish carriers from healthy individuals with remarkable consistency. A related four-metabolite blood panel developed previously by other researchers had achieved an area under the curve of 0.94, suggesting that metabolic signatures may become genuine diagnostic adjuncts for mitochondrial disease. The authors caution, however, that these exploratory findings require validation in independent cohorts before any clinical application can be contemplated.

When the team compared carriers with diabetes to those without, a different signal emerged. Carriers with diabetes, who tended to be older and to carry a heavier burden of clinical manifestations including myopathy, cardiomyopathy, and hearing impairment, showed higher levels of branched-chain amino acids, namely isoleucine, leucine, and valine, than carriers without diabetes. Elevated branched-chain amino acids are a well-established correlate of insulin resistance and a predictor of type 2 diabetes in large population studies, and their prominence in diabetic carriers suggests that the insulin resistance long associated with m.3243A>G may leave a recognizable amino acid trail. Carriers with diabetes also had increased triglycerides within medium-sized high-density lipoprotein particles, higher omega-3 fatty acids, and higher glucose. Serum models could not reliably discriminate between the two carrier groups, whereas urine metabolites such as pseudouridine, creatinine, ethanolamine, and glucose achieved moderate discriminatory performance.

The lipoprotein findings carry clinical weight. Triglyceride-rich lipoproteins and altered particle concentrations are associated with an increased risk of type 2 diabetes and cardiovascular disease, including atherosclerosis, myocardial infarction, and ischemic stroke, and recent analyses of the UK Biobank linked elevated triglyceride content in specific lipoprotein subclasses to all-cause mortality in people with type 2 diabetes. Cardiovascular mortality is a major cause of premature death in m.3243A>G carriers, though it usually stems from left ventricular hypertrophy or conduction defects rather than atherosclerosis. Even so, prior work with induced pluripotent stem cell-derived endothelial cells from high-heteroplasmy carriers found elevated oxidized low-density lipoprotein and pro-atherogenic behavior, raising the possibility that the mutation confers an intrinsic susceptibility to vascular disease. The mechanisms behind the dysregulated lipoprotein metabolism remain unresolved, but plausible candidates include insulin resistance that fails to suppress hepatic VLDL production, impaired mitochondrial fatty acid beta-oxidation, and increased hepatic de novo lipogenesis.

The study has limitations that the authors acknowledge candidly. Twenty-eight carriers is a small sample, an unavoidable consequence of the rarity of m.3243A>G-associated disease, and some participants came from the same families, which may introduce shared metabolic characteristics. Carriers with and without diabetes were not matched for sex, and the diabetic group was older and more severely affected, so the branched-chain amino acid finding could reflect disease severity rather than diabetes itself. Medication use, including antidiabetic, antihypertensive, and lipid-lowering therapies, may have influenced metabolite concentrations even though participants withheld morning doses before sampling. Urinary metabolites were analyzed as absolute concentrations rather than creatinine-normalized values, a deliberate choice because creatinine metabolism itself is altered in mitochondrial disease, though residual variation in urine concentration cannot be excluded. Finally, the cross-sectional design means no conclusions about causality or long-term clinical consequences can be drawn.

Even with those caveats, the work represents, to the authors’ knowledge, the first investigation to combine serum and urine metabolomics with detailed lipoprotein subclass profiling in m.3243A>G carriers, and its standardized fasting protocol and careful matching of controls lend the findings credibility. The picture that emerges is of a mutation that does far more than drain cellular energy supplies: it reshapes the transport of fats between organs, perturbs the machinery of glycolysis and the citric acid cycle, undermines antioxidant defense, and disturbs nucleotide and one-carbon metabolism. The three-molecule urinary signature of uracil, hypoxanthine, and 1-methylnicotinamide offers a tantalizing glimpse of a future where a simple urine test could help identify carriers or monitor disease progression, while the link between branched-chain amino acids and diabetes in carriers may illuminate how this rare genetic form of diabetes overlaps with the far more common insulin-resistant type 2. The researchers call for longitudinal prospective studies in larger cohorts to validate these metabolic signatures and to determine whether the dysregulated lipoprotein profile ultimately contributes to the premature mortality that shadows carriers of this mutation.