Overlapping ranges in levels indicate that hexosylsphingosine is not a clinically relevant biomarker for GBA1 associated Parkinson disease

We were pleased to see the ongoing discussion of whether glucoslysphingosine provides a useful biomarker for GBA1 associated Parkinson disease (PD), and read with great interest the recent letter regarding a more genotypically diverse PD cohort carrying GBA1 variants from Kopytova et al entitled “Could blood hexosylsphingosine be a marker for Parkinson’s disease linked with GBA1 mutations?” This investigation focuses on hexosylsphingosine (HexSph), which while easier to measure, is less precise, as it includes both glucosylsphingosine (GlcSph) and galactosylsphingosine (GalSph).

The communication broadens the discussion, but has several findings that are incongruent with the field and warrant further investigation. The first concerns results of their glucocerebrosidase enzymatic activity assay. While the data presented stratify the severity of the mutant alleles (‘severe’ L444P and ‘mild’ N370S), the enzyme activity results are markedly similar in both groups. In fact, for the GBA+PD− group, the “severe” mutation carriers actually had a higher residual enzyme activity than those with “mild” mutations, which is quite surprising. This is in contrast to previously published work where the severe and the mild mutant alleles are reported to have residual GCase activity of 13-24% and 32-38% respectively (1). Furthermore, when the authors divide GBA1 variants into “mild” GBA1 carriers (GBA+/PD− «mild») vs mild PD carriers (GBA+/PD+ «mild»), they report a significant increase in blood HexSph concentration that is associated with PD, but only in carriers with “mild” GBA1 mutations (p<0.006). While compelling, this does not appear logical, as lipids levels should be higher with more severe mutations. Surface et al also evaluated the “mild” N370S mutation, concluding that ‘glucosylsphingosine levels do not distinguish between GBA1 N370S mutation carriers with and without PD, and we never claimed that glucosylsphingosine levels can be used for this purpose (2).’ Perhaps the difference in the study by Kopytova et al is that the authors used a combination of mild carrier variants [11 N370S carriers and 4 additional ‘mild’ variants (N227S, L327P, M124T, G241R)] when comparing HexSph levels to GBA+/PD+ <<mild>> patients harboring only the N370S allele. Actually, the functional consequences of the GBA1 variants N227S, L327P, M124T, G241R are largely unknown. A more accurate representation would be to stratify this analysis by comparing the 11 GBA+/PD− «mild» carriers who solely harbor the N370S allele with the N370S GBA+/PD+ «mild» patients (N=8).

Lastly, it is important to note with the findings provided by Alena and colleagues, that even if the median values are statistically significant, the range in HexSph values overlap between GBA+/PD+ «mild» and GBA+/PD− «mild». Clinically, it is impractical to rely on a biomarker where the ranges closely overlap between subjects with and without the PD phenotype. A graphic representation of their data would make this even more apparent. Thus, once again, even if there is a statistical difference between mild mutation carriers with and without PD, the closely overlapping HexSph ranges do not make for a clinically relevant GBA1 associated PD biomarker.