The Goldilocks metabolite: why tissues need just enough L-2-HG

L-2-hydroxyglutarate, or L-2-HG, has long been viewed as a “toxic” or harmful metabolite. Children born with nonfunctional forms of the L-2-HG-clearing mitochondrial enzyme L-2-hydroxyglutarate dehydrogenase (L2HGDH) develop a neurological disorder called L-2-hydroxyglutaric aciduria¹. L-2-HG also accumulates in some kidney cancers².

More than a decade ago, we observed that mitochondrial dysfunction in a cancer cell line increased L-2-HG levels³. Other groups demonstrated that low oxygen, which can also impair mitochondrial function, and acidic pH can elevate L-2-HG levels in cells^4–7. We later found that mitochondrial dysfunction in mouse hematopoietic stem cells and regulatory T cells was associated with increased levels of this metabolite^8,9. At the time, these observations seemed to fit the prevailing view of L-2-HG as harmful.

But this view felt incomplete. The enzymes that produce L-2-HG, as well as the enzyme that clears it, are conserved from bacteria to humans. We wondered why evolution would preserve a metabolite that is only toxic. L-2-HG is present at low micromolar levels in normal cells and tissues. We hypothesized that L-2-HG may have a physiological function when its levels rise acutely in certain contexts, whereas chronic elevation to high levels would be pathogenic.

This idea fits into our lab’s broader interest in understanding how mitochondrial signaling controls physiology and disease. We have been particularly interested in how reactive oxygen species (ROS) and metabolites control cell function and fate in physiological contexts. One challenge in studying metabolite signaling is that metabolites often have multiple functions, making it difficult to disentangle one role from another. L-2-HG offered a unique opportunity. Unlike many central metabolites, it does not make a substantial direct contribution to energy production or macromolecule synthesis. That made us wonder whether its main physiological role might be signaling.

To address this, we held L-2-HG to three criteria that a physiological signaling metabolite should satisfy. First, its levels should be actively regulated. Second, it should act on defined targets at physiological concentrations. Third, it should have a physiological effect.

The first part of the story took us into mitochondrial redox biology. Mitochondrial respiration helps maintain the balance between NADH and NAD⁺. When the mitochondrial electron transport chain (ETC) is functional, NADH is continuously converted back to NAD⁺. If mitochondrial respiration is impaired, NADH builds up, and the NADH/NAD⁺ ratio increases. We found that this shift in the NADH/NAD⁺ ratio drives the mitochondrial enzyme malate dehydrogenase 2 (MDH2), which normally interconverts malate and oxaloacetate, to reduce 2-oxoglutarate (2-OG) to L-2-HG in cells with impaired mitochondrial respiration. In that sense, L-2-HG behaves like a metabolic readout of mitochondrial state.

Then came one of the surprises of the project. The common assumption was that L-2-HG clearance by L2HGDH depends on a functional ETC. We tested this by overexpressing L2HGDH in cell lines and blocking the ETC at multiple points, genetically or pharmacologically. We also placed cells at 0% oxygen for 24 hours. L2HGDH kept working. This told us that the current model of L2HGDH function is incomplete. The enzyme must be able to hand off electrons in another way. There could be an alternative electron acceptor, or multiple electron acceptors. We do not yet know the exact identity of the electron acceptor in the absence of a functional ETC.

The next question was what L-2-HG does inside cells. We used an unbiased proteomic approach, which nominated the KDM4 family of histone demethylases as potential targets of L-2-HG. These enzymes remove H3K9me3, a repressive chromatin mark that helps keep certain regions of the genome silent. Reported IC₅₀ values (the concentrations needed to inhibit half of enzyme activity) for inhibition of KDM4 family members by L-2-HG are in a physiologically plausible range, whereas other proposed targets require much higher concentrations.

This finding provided a plausible model: physiological L-2-HG may act as a brake on KDM4 activity, thereby helping preserve H3K9me3 at specific genomic regions. Consistent with this, we observed that elevated L-2-HG levels in mouse embryonic stem cells primarily repressed nascent transcription, with 854 genes downregulated and only 20 upregulated. These repressed loci also gained H3K9me3.

The most striking result came when we tested whether physiological L-2-HG levels matter during mouse development. We engineered mice to overexpress the L-2-HG-clearing enzyme L2HGDH throughout the body from early development. If basal L-2-HG levels were physiologically irrelevant, these mice should have been fine. However, all but one L2HGDH-overexpressing mouse died by six weeks of age. They were smaller than their littermates and failed to gain weight. Across tissues, the kidneys appeared most abnormal histologically, with signs of early fibrosis. We also observed biochemical evidence of impaired renal function.

We found that in normal mice, the kidney L-2-HG level does not stay flat after birth. It rises during early postnatal life, peaks around postnatal day 11, and then falls. Such regulation of L-2-HG levels suggested that L-2-HG is part of a normal developmental program. In kidneys with lowered L-2-HG, we also found a selective loss of H3K9me3 at a class of retrotransposons called L1MdTf elements.

Retrotransposons are ancient viral-like sequences buried in the genome and usually remain silent. When H3K9me3-mediated silencing was weakened following L2HGDH overexpression and L-2-HG depletion, L1MdTf elements became activated. Activation of such elements can trigger inflammation and an integrated stress response, both of which we found upregulated in L-2-HG-depleted kidneys.

Overall, our study reveals that L-2-HG is not simply good or bad. Instead, it follows a “Goldilocks” principle: too little or too much can be harmful. We started with a low-abundance metabolite widely considered toxic and found that it is also necessary for normal physiology. Our study also raises the possibility that other metabolites known mainly for their toxicity may also have hidden physiological roles.

This study benefited from important collaborations, including contributions from Yuki Aoi in Ali Shilatifard’s lab at Northwestern and Jonathan Van Vranken in Steve Gygi’s lab at Harvard.

References:

1. Barth, P. G. et al. L-2-hydroxyglutaric acidemia: A novel inherited neurometabolic disease. Ann. Neurol. 32, 66–71 (1992).
2. Shim, E.-H. et al. L-2-Hydroxyglutarate: An Epigenetic Modifier and Putative Oncometabolite in Renal Cancer. Cancer Discov. 4, 1290–1298 (2014).
3. Mullen, A. R. et al. Oxidation of Alpha-Ketoglutarate Is Required for Reductive Carboxylation in Cancer Cells with Mitochondrial Defects. Cell Rep. 7, 1679–1690 (2014).
4. Intlekofer, A. M. et al. Hypoxia Induces Production of L-2-Hydroxyglutarate. Cell Metab. 22, 304–311 (2015).
5. Oldham, W. M., Clish, C. B., Yang, Y. & Loscalzo, J. Hypoxia-Mediated Increases in L-2-hydroxyglutarate Coordinate the Metabolic Response to Reductive Stress. Cell Metab. 22, 291–303 (2015).
6. Nadtochiy, S. M. et al. Acidic pH Is a Metabolic Switch for 2-Hydroxyglutarate Generation and Signaling. J. Biol. Chem. 291, 20188–20197 (2016).
7. Intlekofer, A. M. et al. L-2-Hydroxyglutarate production arises from noncanonical enzyme function at acidic pH. Nat. Chem. Biol. 13, 494–500 (2017).
8. Ansó, E. et al. The mitochondrial respiratory chain is essential for haematopoietic stem cell function. Nat. Cell Biol. 19, 614–625 (2017).
9. Weinberg, S. E. et al. Mitochondrial complex III is essential for suppressive function of regulatory T cells. Nature 565, 495–499 (2019).