Reply to Struys: Role of biomarker of 2-hydroxyglutarate in acute myeloid leukemia

We agree with E. A. Struys’ point (1) that 2-hydroxyglutarate (2-HG) consists of two distinct forms, l-2-hydroxyglutarate (L2HG) and d-2-hydroxyglutarate (D2HG), and that these two endogenous metabolites should be differentiated in characterizing their association with isocitrate dehydrogenase1/2 (IDH1/2) mutations and with clinical consequences in acute myeloid leukemia (AML). In fact, in our report (2) we discuss that there are two different 2-HG enantiomers, D2HG and L2HG, previous experiments showed only D2HG was increased in IDH1/2 mutations, and that the determination of the D2HG/L2HG ratio as a more sensitive and specific clinical test may be developed in the future.

Recent studies have shown that either D2HG or L2HG, when increased, is implicated in tumorigenesis (3, 4). Although elevated serum D2HG can be oncogenic in IDH mutations associated with AML and gliomas (4), a high L2HG expression level resulting from l-2-hydroxyglutarate dehydrogenase (L2HGDH) mutations is highly associated with brain tumors (3). Although the relationship between increased L2HG and AML remains unclear, we regard both of them as potential oncometabolites or cancer markers, instead of D2HG alone.

The current 2-HG method presented in our report appears to be robust and simple in stratifying patients with good or poor prognosis with the increased serum level of 2-HG, independent of cytogenetic and molecular status. We established a cutoff of the mean plus 5 SD of 2-HG values from 405 healthy controls. Considering the mixture nature of two metabolites in 2-HG measurement, we further separated high 2-HG levels in two subsets of very high and moderately high groups, and showed that only the very high 2-HG group was associated with IDH mutations (27 of 31, 87%), and only 9 of 31 patients (29%) with moderately high serum 2-HG levels had these mutations. Other genetic events may have contributed to the increase of 2-HG, especially in patients with moderately high serum 2-HG levels. On the other hand, the measurements of serum 2-HG could not exclude the contribution of an increase of L2HG, which might yield false-positive results, as indicated by Struys (1). However, we believe that in pathological conditions, oncogenic 2-HG increases substantially regardless of its origin, as evidenced by a report from Sellner et al. that 2-HG was increased by 100-fold in patients with IDH-mutated AML compared with the negative ones (5).

Although the measurement of 2-HG may be more practically feasible, an analytical method that quantifies D2HG and L2HG, respectively, is needed—as Struys suggests—to fine-tune such a measurement. We are in the process of evaluating a more accurate analytical method to quantify the two molecules and obtain the ratio of D2HG/L2HG, and compare the two methods (combined or separated D2HG and L2HG) for AML patient stratification and prognosis.