Diagnostic and therapeutic applications of the glycan biomarker H3N2b in GM1 Gangliosidosis Authors: Pamela Kell, Sonali Mishra, Precilla D'Souza, Cynthia J Tifft, Dennis J Dietzen, Laura Duque Lasio, Jennifer Powers Carson, Xinying Hong, Raphael Schiffmann, María Dolores Ledesma, Nahid Tayebi, Patricia I Dickson, Forbes D Porter, Daniel S Ory, Xuntian Jiang Published: 2026 ABSTRACT AbstractGM1 gangliosidosis is a progressive lysosomal storage disorder caused by β-galactosidase deficiency, resulting in accumulation of GM1 ganglioside and related glycoconjugates. Current diagnostic approaches rely on enzymatic and genetic testing but do not provide dynamic measures of substrate accumulation or therapeutic response. The oligosaccharide H3N2b has previously been identified as a potential biomarker of impaired β-galactosidase activity. In this study, H3N2b was quantified in plasma, urine, and cerebrospinal fluid (CSF) from patients with GM1 gangliosidosis (n = 47 plasma/urine; n = 34 CSF) and age- and sex-matched controls using a validated liquid chromatography–tandem mass spectrometry (LC-MS/MS) assay. A limited set of comparator lysosomal storage disorder samples was also analyzed to provide preliminary assessment of biomarker specificity. Longitudinal H3N2b measurements were obtained from participants receiving intravenous adeno-associated virus serotype 9 (AAV9)-mediated GLB1 gene therapy. H3N2b concentrations were significantly elevated in GM1 gangliosidosis compared with controls in plasma (cutoff 6.2 ng/mL; 100% sensitivity and specificity), urine (0.65 ng/μg creatinine; 99.3% sensitivity, 100% specificity), and CSF (5.1 ng/mL; 98.9% sensitivity, 100% specificity). H3N2b levels correlated with disease severity and were not influenced by age or sex. Among the comparator disorders analyzed, H3N2b values were generally below GM1 diagnostic cutoffs, although the comparator analysis was limited by sample availability and did not include several relevant disorders and CSF. In gene therapy recipients, plasma, urine, and CSF H3N2b concentrations decreased following treatment and closely paralleled changes in β-galactosidase activity, indicating pharmacodynamic responsiveness. These findings provide a cross-matrix clinical evaluation of H3N2b as a diagnostic and therapeutic biomarker for GM1 gangliosidosis and support its utility as a candidate biochemical biomarker to aid diagnosis and monitor treatment-associated biochemical changes.Keywords: GM1 gangliosidosis, Pentasacharide H3N2b, B-Galactosidase, Diagnosis, Gene therapy, Treatment assessment Introduction GM1 gangliosidosis is an autosomal recessive lysosomal storage disorder caused by pathogenic variants in the GLB1 gene encoding β-galactosidase. Deficiency of this enzyme results in lysosomal accumulation of GM1 ganglioside and related glycoconjugates with terminal β-galactose, particularly in neurons, leading to progressive neurodegeneration. Clinical manifestations vary according to age of onset and residual enzyme activity and are generally classified as infantile (type I), late-infantile/juvenile (type II), and adult/chronic forms (type III). The infantile form is the most severe, characterized by rapid neurological deterioration as well as somatic features, including hepatosplenomegaly and skeletal abnormalities that may resemble mucopolysaccharidoses, whereas later-onset forms typically follow a more slowly progressive course [1–4]. Diagnosis of GM1 gangliosidosis currently relies on measurement of β-galactosidase activity and molecular genetic testing of GLB1. However, diagnosis may be delayed because of nonspecific early clinical features that are overlapped with other lysosomal and metabolic disorders [3,5–7], variability in enzyme assays [8–11], and challenges in interpreting genetic variants of uncertain significance [12,13]. In addition, quantitative biomarkers that reflect disease-related substrate accumulation remain limited, highlighting the need for reliable biochemical markers that can support diagnosis and facilitate monitoring of therapeutic interventions [2,14,15]. Recent glycomic studies have identified disease-associated oligosaccharides that accumulate in lysosomal storage disorders [16–23]. In clinical practice, GM1 gangliosidosis is often initially identified through metabolic screening approaches, such as thin-layer chromatography [24], which remains in use in many laboratories, or mass spectrometry-based methods [19–23] that detect abnormal oligosaccharides. We previously identified two pentasaccharides, H3N2a and H3N2b, as glycan biomarkers associated with impaired β-galactosidase activity. In that study, we synthesized H3N2b (Fig. 1), developed and analytically validated a liquid chromatography–tandem mass spectrometry (LC-MS/MS) assay, and demonstrated that circulating H3N2b concentrations decreased following adeno-associated virus serotype 9 (AAV9)-mediated gene therapy, supporting its potential as a pharmacodynamic biomarker of treatment response [25]. Subsequently, results from a Phase 1/2 clinical trial of intravenous AAV9-mediated GLB1 gene therapy in patients with GM1 gangliosidosis demonstrated biochemical and clinical evidence of therapeutic activity [26]. H3N2b structure. Building on these findings, the present study further evaluates the clinical utility of H3N2b as a diagnostic and therapeutic biomarker. We measured H3N2b concentrations in plasma, urine, and cerebrospinal fluid (CSF) from individuals with GM1 gangliosidosis and unaffected controls, examined a limited set of comparator lysosomal disorders to provide preliminary information on biomarker distribution outside GM1 gangliosidosis, and assessed its potential as a biomarker for therapeutic monitoring. Despite prior identification of H3N2b, its diagnostic performance, cross-matrix consistency, and distribution across available comparator samples have not been systematically characterized in a clinical cohort. Here, we report an expanded clinical evaluation of H3N2b in plasma, urine, and CSF, demonstrating marked elevation in GM1 gangliosidosis relative to controls and responsiveness to therapeutic intervention. Materials and methods Ethics Human studies were conducted in accordance with the principles of the Declaration of Helsinki and Title 45 of the U.S. Code of Federal Regulations, Part 46 (Protection of Human Subjects). Written informed consent was obtained from all participants or their legal guardians. Study protocols were approved by the Institutional Review Boards of Washington University School of Medicine, National Institutes of Health, Children's Hospital of Philadelphia, Texas Christian University, and Centro de Biología Molecular Severo Ochoa. All samples were de-identified prior to analysis, and use of de-identified specimens was approved by the Human Studies Committee at Washington University School of Medicine. Collection of clinical samples Plasma, urine, and CSF samples were collected across participating clinical centers using standardized protocols to minimize pre-analytical variability. For plasma, ≤10 mL of blood (representing <1% of total blood volume) was collected into K2EDTA tubes and centrifuged at 1500g for 10 min at 4 °C within 30 min of collection. Plasma was separated and aliquoted into polypropylene tubes (0.5 mL per aliquot). Urine samples were collected as midstream specimens into sterile collection cups and aliquoted into polypropylene tubes (1 mL per aliquot). CSF was obtained via lumbar puncture, with 2–12 mL collected depending on patient size under anesthesia, and aliquoted into polypropylene tubes (0.5 mL per aliquot). All aliquots were immediately stored at −80 °C until shipment. Samples were transported on dry ice to the central analytical laboratory (Washington University) and maintained at −80 °C prior to analysis. Stability studies performed during method validation confirmed analyte stability under these collection, processing, and storage conditions, supporting robustness against pre-analytical variation across sites [25]. Plasma, urine, and CSF samples from patients with GM1 gangliosidosis were obtained through a natural history study (NCT00029965) conducted at the National Human Genome Research Institute and from participants enrolled in a Phase 1/2 intravenous AAV9/GLB1 gene therapy trial (NCT03952637). Diagnosis of GM1 gangliosidosis was confirmed by molecular genetic testing and measurement of β-galactosidase activity. Control plasma, urine, and CSF samples were obtained from anonymized residual clinical specimens at Washington University School of Medicine. Controls were verified as unaffected based on medical records and were matched to cases by sex and age (±3 years). The characteristics of human subjects are given in Tables S1 – S4. Samples from a subset of non-GM1 lysosomal storage disorders were analyzed to assess biomarker specificity. Plasma samples were obtained from patients with Niemann–Pick disease type A (NPA, n = 5) and type B (NPB, n = 5) (Centro de Biología Molecular Severo Ochoa), Niemann–Pick disease type C1 (NPC1, n = 5) (Eunice Kennedy Shriver National Institute of Child Health and Human Development natural history study, NCT00344331), Tay–Sachs disease (TSD, n = 5) and Gaucher disease (GD, n = 21) (National Human Genome Research Institute), and Fabry disease (FD, n = 33) (Texas Christian University). Urine samples were obtained from patients with Pompe disease (PD, n = 1), Fabry disease (FD, n = 16), alpha-mannosidosis (AM, n = 4), and mucopolysaccharidoses, including MPS I (n = 3), MPS II (n = 2), MPS IIIB (n = 1), MPS IIIC (n = 2), MPS IVA (n = 2), MPS VI (n = 1), and MPS VII (n = 1) (Children's Hospital of Philadelphia). One additional urine sample from a patient with MPS II was obtained from Washington University School of Medicine. No CSF samples from other lysosomal disorders were available for analysis. In addition, urine samples from TSD and plasma or urine samples from Sandhoff disease were not available in this study. Quantification of H3N2b and β-galactosidase activity Quantification of H3N2b in plasma, urine, and CSF was performed using previously validated LC–MS/MS methods [25] in accordance with the FDA M10 Bioanalytical Method Validation and Study Sample Analysis guideline (2022) [27] and the FDA Bioanalytical Method Validation for Biomarkers guidance (2026) [28]. Method validation demonstrated acceptable linearity, selectivity, specificity, sensitivity, accuracy, precision, stability, carryover, and dilution integrity across all matrices, as previously reported in detail in the Supplementary Materials of reference 25. For clinical sample analysis, each analytical run included eight-point calibration standards (prepared in duplicate), a double blank, a blank, and low-, mid-, and high-quality control (LQC, MQC, HQC) samples, along with study samples. QC samples comprised at least 5% of the total study samples. Study samples were randomized using temporary identifiers for LC–MS/MS analysis, and QC samples were interspersed throughout the run sequence. The calibration range covered the expected concentration range of study samples; samples exceeding the upper limit of quantification (ULOQ) were diluted and reanalyzed. Dilution integrity was confirmed using dilution QC samples analyzed in replicate within each run. Run acceptance criteria followed FDA M10 guidance: back-calculated concentrations of calibration standards were required to be within ±15% of nominal values except for ±20% at the lower limit of quantification (LLOQ), with at least 75% of standards, including a minimum of six concentration levels, meeting these criteria. In addition, at least two-thirds of QC samples and at least 50% at each QC level were required to fall within ±15% of nominal concentrations. All analytical runs in this study met these acceptance criteria. Urinary H3N2b concentrations were normalized to creatinine that was measured with procedure reported previously [29]. β-Galactosidase activity in serum and CSF samples from participants in the gene therapy trial was measured using previously described enzymatic assays [26]. Statistical analysis Statistical analyses were performed using GraphPad Prism (version 11.0.0; GraphPad Software, San Diego, CA). Data normality was assessed using the Shapiro–Wilk test. For normally distributed data, comparisons between two groups were performed using a two-sample t-test. When data were not normally distributed, nonparametric tests were applied. The Mann–Whitney U test was used for comparisons between two groups, and the Kruskal–Wallis test was used for comparisons among multiple groups, followed by Dunn's multiple comparisons test as a post hoc test. A p value <0.05 was considered statistically significant. Reference intervals were calculated as the 2.5th to 97.5th percentiles of measured values in each group. Receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic performance of H3N2b in plasma, urine, and CSF, and the areas under the curve were calculated to determine optimal diagnostic cutoffs for discrimination of GM1 gangliosidosis from controls. Results Elevated H3N2b levels in plasma of GM1 gangliosidosis patients Plasma H3N2b concentrations were measured in 47 patients with GM1 gangliosidosis and 277 control subjects. H3N2b levels were significantly elevated in patients compared with controls (Fig. 2A). The plasma H3N2b concentrations ranged from 6.34 to 496 ng/mL for GM1 gangliosidosis and < 1–5.99 ng/mL for controls. Summary descriptive statistics, including the reference intervals, the mean (standard deviation, SD) and median, are provided in Table 1. Receiver operating characteristic (ROC) analysis demonstrated excellent diagnostic performance, with a cutoff value of 6.2 ng/mL providing 100% sensitivity and 100% specificity for distinguishing GM1 gangliosidosis from healthy controls (Fig. 2B). Plasma H3N2b concentrations in individuals with GM1 gangliosidosis. (A) Plasma H3N2b concentrations in GM1 gangliosidosis patients and controls. Values below the LLOQ (1 ng/mL) are shown for visualization; assay variability at these levels may exceed validation criteria. Data are shown as median ± interquartile range on a logarithmic scale. Group comparisons were performed using the Mann–Whitney U test. (B) ROC curve for diagnostic cutoff determination. (C) Plasma H3N2b stratified by clinical subtype (juvenile, late-infantile, infantile), shown as mean ± SD. Comparisons were performed using Kruskal–Wallis test with Dunn's post hoc test. (D) Effect of sex on plasma H3N2b (mean ± SD; Mann–Whitney U test). (E) Longitudinal relationship between plasma H3N2b and age in juvenile patients (n = 8); each colour represents an individual patient. ns, not significant; **p < 0.01; ***p < 0.001; ****p < 0.0001. Table 1Reference intervals from the 2.5th – 97.5th percentile, means (standard deviation, SD), and mediums for plasma, urine, and CSF H3N2b.MatrixPatient typeValue2.5th – 97.5th percentilesMean (SD)Medium PlasmaGM1 Gangliosidosisng/mL7.68–353.272.1 (91.0)36.6Controlng/mL< 1 – < 10.31* (0.59)< 1UrineGM1 Gangliosidosisng/μg creatinine0.93–176.325.0 (54.6)11.6Controlng/μg creatinine0.003–0.150.074 (0.55)0.017CSFGM1 Gangliosidosisng/mL6.5–62868.3 (175.1)21.6Controlng/mL<1–2.30.99* (1.0)< 1*The measured H3N2b concentrations that were below the LLOQ (1 ng/mL) were used in the calculation. Reference intervals from the 2.5th – 97.5th percentile, means (standard deviation, SD), and mediums for plasma, urine, and CSF H3N2b. The measured H3N2b concentrations that were below the LLOQ (1 ng/mL) were used in the calculation. Plasma H3N2b concentrations differed among clinical subtypes, following the pattern juvenile (n = 25) < late-infantile (n = 20) < infantile (n = 2) (Fig. 2C). Statistical analysis showed significantly higher levels in late-infantile patients compared with juvenile patients (p < 0.001), and infantile patients also had significantly higher levels than juvenile patients (p < 0.01). The difference between infantile and late-infantile groups was not statistically significant, likely due to the small number of infantile cases and variability in biomarker levels. Because both infantile patients were female, potential sex effects were evaluated only in the juvenile and late-infantile groups. No significant differences were observed between female (n = 17) and male (n = 8) patients in the juvenile group or between female (n = 11) and male (n = 9) patients in the late-infantile group (Fig. 2D). Plasma H3N2b levels were not significantly associated with age among GM1 gangliosidosis patients (Fig. 2E). Elevated urinary H3N2b levels in GM1 gangliosidosis Urinary H3N2b concentrations normalized to creatinine were measured in 47 GM1 gangliosidosis patients and 272 controls. H3N2b levels were significantly increased in patients compared with controls (Fig. 3A). The urinary H3N2b concentrations ranged from 0.84 to 326 ng/μg creatinine for GM1 gangliosidosis and 0.0005–8.04 ng/μg creatinine for controls. The reference intervals, corresponding mean (SD) and median values are summarized in Table 1. ROC analysis identified a cutoff value of 0.65 ng/μg creatinine, yielding 99.26% sensitivity and 100% specificity for diagnosis (Fig. 3B). Urinary H3N2b concentrations in individuals with GM1 gangliosidosis. (A) Urinary H3N2b normalized to creatinine in patients and controls. Data are shown as median ± interquartile range on a logarithmic scale (Mann–Whitney U test). (B) ROC curve for diagnostic cutoff determination. (C) Urinary H3N2b stratified by clinical subtype (mean ± SD; Kruskal–Wallis test with Dunn's post hoc test). (D) Effect of sex on urinary H3N2b (mean ± SD; Mann–Whitney U test). (E) Longitudinal urinary H3N2b trajectories in juvenile patients (n = 5); each dotted line represents an individual patient over time. ns, not significant; **p < 0.01; ***p < 0.001; ****p < 0.0001. As observed in plasma, urinary H3N2b concentrations varied by disease subtype and followed the order juvenile (n = 25) < late-infantile (n = 20) < infantile (n = 2) (Fig. 3C). Late-infantile patients had significantly higher levels than juvenile patients (p < 0.0001), and infantile patients had significantly higher levels than juvenile patients (p < 0.01). The difference between infantile and late-infantile groups was not statistically significant, likely due to the small number of infantile cases and variability. Sex did not significantly influence urinary H3N2b levels in juvenile or late-infantile patients (Fig. 3D). Additionally, urinary H3N2b concentrations were not significantly associated with patient age (Fig. 3E). Elevated H3N2b levels in cerebrospinal fluid H3N2b concentrations in CSF were measured in 34 GM1 gangliosidosis patients and 177 controls. CSF H3N2b levels were significantly elevated in patients compared with controls (Fig. 4A). The CSF H3N2b concentrations ranged from 5.4 to 888 ng/mL for GM1 gangliosidosis and < 1–9.0 ng/mL for controls. The reference intervals, mean (SD) and median values are shown in Table 1. ROC analysis demonstrated strong diagnostic performance, with a cutoff value of 5.1 ng/mL achieving 98.87% sensitivity and 100% specificity (Fig. 4B). CSF H3N2b concentrations in individuals with GM1 gangliosidosis. (A) CSF H3N2b concentrations in patients and controls. Values below the LLOQ (1 ng/mL) are shown for visualization; assay variability may exceed validation limits at these levels. Data are shown as median ± interquartile range on a logarithmic scale (Mann–Whitney U test). (B) ROC curve for diagnostic cutoff determination. (C) CSF H3N2b stratified by clinical subtype (mean ± SD; Kruskal–Wallis test with Dunn's post hoc test). (D) Effect of sex on CSF H3N2b (mean ± SD; Mann–Whitney U test). ns, not significant; *p < 0.05; ***p < 0.001; ****p < 0.0001. Consistent with plasma and urine findings, CSF H3N2b levels were associated with disease subtype severity, with concentrations increasing from juvenile (n = 22) to late-infantile (n = 10) to infantile (n = 2) patients (Fig. 4C). Late-infantile patients had significantly higher levels than juvenile patients (p < 0.001), and infantile patients had higher levels than juvenile patients (p < 0.01). The difference between infantile and late-infantile groups was not statistically significant, likely due to the small sample size and variability. Sex had no significant effect on CSF H3N2b levels in juvenile or late-infantile patients (Fig. 4D). The effect of age on H3N2b levels was not assessed because longitudinal samples were not available. Specificity of H3N2b for GM1 gangliosidosis To evaluate disease specificity, H3N2b levels were measured in a limited set of samples from individuals with other lysosomal storage disorders. Plasma samples from patients with NPA, NPB, NPC1, TSD, GD, and FD all showed H3N2b levels below the diagnostic cutoff established for GM1 gangliosidosis (Fig. 5A). H3N2b levels in patients with other lysosomal storage diseases. (A) H3N2b levels in plasma samples from patients with NPA (n = 5), NPB (n = 5), NPC1 (n = 5), TSD (n = 5), GD (n = 21), and FD (n = 33). Data are presented as mean ± SD. (B) H3N2b levels in urine samples from patients with PD (n = 1), FD (n = 16), AM (n = 4), MPS I (n = 3), MPS II (n = 2), MPS IIIB (n = 1), MPS IIIC (n = 2), MPS IVA (n = 2), MPS VI (n = 1), and MPS VII (n = 1). No CSF samples from other lysosomal disorders were available for analysis. Data are presented as mean ± SD. In urine, samples from patients with PD, FD, AM, and several mucopolysaccharidoses, including MPS I, MPS II, MPS IIIB, MPS IIIC, MPS IVA, MPS VI, and MPS VII, were also below the GM1 gangliosidosis diagnostic threshold (Fig. 5B), although elevations were observed in some cases, particularly among oligosaccharidosis. However, this analysis was limited by the availability of samples. CSF specimens from other lysosomal disorders were not available, and no urine samples from TSD or plasma/urine samples from Sandhoff disease were included. Accordingly, while H3N2b may serve as a useful biomarker to support the evaluation of GM1 gangliosidosis, these findings indicate that it is not necessarily disease-specific in isolation. Further studies including a broader range of lysosomal disorders and sample types are needed to more fully define its diagnostic specificity. H3N2b as a pharmacodynamic biomarker during gene therapy To evaluate the utility of H3N2b as a pharmacodynamic biomarker, plasma, urine, and CSF samples from participants enrolled in a Phase 1/2 intravenous AAV9/GLB1 gene therapy trial (NCT03952637) were analyzed longitudinally. This analysis includes previously unreported participants (IVGT013, IVGT015, and IVGT021), extended follow-up (up to 36 months for plasma, urine, and CSF H3N2b in IVGT001; up to 36 months for plasma H3N2b in IVGT004; and up to 48 months for CSF H3N2b in IVGT003, IVGT006, IVGT007, and IVGT008), and inclusion of infantile GM1 gangliosidosis cases (IVGT013 and IVGT015), enabling a more comprehensive assessment of biomarker dynamics across biological matrices and disease subtypes. Longitudinal trajectories for individual participants are shown to characterize within-patient biomarker dynamics, assess durability of response over extended follow-up, enable direct cross-matrix comparisons (plasma, CSF, and urine), and explore variability across disease severity, including infantile cases. Following treatment, β-galactosidase activity in serum increased, reaching a peak during the early post-treatment period (Fig. 6). Enzyme activity subsequently declined over time. Plasma H3N2b concentrations decreased after treatment and showed temporal changes that paralleled those observed for serum β-galactosidase activity across participants. Longitudinal changes in plasma H3N2b and serum β-galactosidase activity following gene therapy in AAV9/GLB1-treated participants. Each subfigure represents an individual patient over time. Blue lines indicate plasma H3N2b concentrations, and the green dotted line in each graph denotes the normal cutoff value. β-Galactosidase activity is shown as red lines and reported as fold of normal. A similar pattern was observed in CSF samples (Fig. 7). CSF β-galactosidase activity increased following treatment and gradually declined during follow-up, while CSF H3N2b concentrations decreased after therapy and tracked changes in CSF enzyme activity over time, including during extended follow-up of up to 48 months. Longitudinal changes in CSF H3N2b and CSF β-galactosidase activity following gene therapy in AAV9/GLB1-treated participants. Each subfigure represents an individual patient over time. Blue lines indicate CSF H3N2b concentrations, and the green dotted line in each graph denotes the normal cutoff value. β-Galactosidase activity is shown as red lines and reported as fold of normal. Urinary H3N2b concentrations showed a comparable trend (Fig. 8). Urinary H3N2b levels decreased following treatment and increased as serum β-galactosidase activity declined during longitudinal monitoring. Longitudinal changes in urinary H3N2b and serum β-galactosidase activity following gene therapy in AAV9/GLB1-treated participants. Each subfigure represents an individual patient over time. Blue lines indicate urinary H3N2b concentrations normalized to creatinine, and the green dotted line in each graph denotes the normal cutoff value. β-Galactosidase activity is shown as red lines and reported as fold of normal. Together, these observations demonstrate that H3N2b concentrations in plasma, CSF, and urine change dynamically in association with β-galactosidase activity following AAV9-mediated gene therapy, highlighting their utility for monitoring pharmacodynamic response, durability, and inter-individual variability across disease subtypes. Discussion In this study, we demonstrate that the glycan H3N2b is a promising biomarker for GM1 gangliosidosis with diagnostic and pharmacodynamic utility across multiple biological matrices. Using LC-MS/MS-based quantification, H3N2b concentrations are markedly elevated in plasma, urine, and CSF of individuals with GM1 gangliosidosis compared with controls, with clear separation from unaffected controls in this cohort. These findings substantially extend our previous work [25], in which H3N2b was identified as a glycan biomarker associated with impaired β-galactosidase activity. That initial study established the LC–MS/MS assay, provided analytical validation, and generated proof-of-concept pharmacodynamic evidence in a limited discovery cohort, including longitudinal data from a single treated participant [25]. The present study represents a substantially expanded independent validation cohort, enabling robust evaluation of diagnostic performance, reference intervals, and disease specificity across a larger patient population and multiple biofluids. However, this work constitutes a comprehensive clinical evaluation of a candidate biomarker rather than full regulatory-grade validation. Across plasma, urine, and CSF, H3N2b levels showed a consistent relationship with clinical severity, with infantile patients exhibiting the highest concentrations, followed by late-infantile and juvenile forms. This trend is consistent with the progressive metabolic burden associated with decreasing residual β-galactosidase activity [30], supporting H3N2b as a reflection of disease severity. No significant associations with age or sex were observed within patient groups; however, interpretation of age effects is limited by sparse longitudinal sampling, particularly in infantile patients, and incomplete age coverage. Given the collinearity between age and clinical subtype in early disease stages, the independent contribution of age to H3N2b levels cannot be fully resolved in this dataset. Importantly, the number of infantile GM1 gangliosidosis patients was limited (n = 2), reflecting the rarity of this phenotype and sample availability, which constrains the robustness of subtype-specific comparisons and interpretation of severity-associated differences in H3N2b levels. Within the broader biomarker landscape of GM1 gangliosidosis, multiple complementary modalities capture distinct layers of disease biology. At the most proximal level, β-galactosidase activity remains the primary diagnostic indicator [7–10], reflecting the underlying enzymatic defect. Closely aligned with this are substrate-derived metabolic biomarkers, including GM1 ganglioside [14,15,31] and its deacylated form lyso-GM1 [14,15,32], as well as disease-associated oligosaccharides such as H3N2b [25] and A2G2’ [33], which directly report impaired lysosomal degradation and substrate accumulation. Among these, H3N2b stands out for its robust elevation across plasma, CSF, and urine, large dynamic range, and marked responsiveness to therapeutic intervention, supporting its role as a sensitive pharmacodynamic marker of biochemical correction. In parallel, a second class of biomarkers captures downstream pathophysiological consequences, particularly neuroaxonal injury and neuroinflammation. These include neurofilament light chain (NfL) as a marker of axonal damage, along with aspartate transaminase (AST), lactate dehydrogenase (LDH), chitotriosidase, and inflammatory cytokines (e.g., MCP-1, MIP-1, TNFR2) [14,15], which reflect disease progression and CNS involvement. Considered together, these biomarker classes are highly complementary: metabolic markers provide direct insight into target engagement and substrate clearance, whereas neurodegenerative and inflammatory markers inform on clinical trajectory and tissue injury. Integration of these modalities may enable improved disease stratification, strengthen genotype–phenotype correlations, and support comprehensive longitudinal monitoring in both natural history studies and therapeutic trials. The assessment of disease specificity is limited by the restricted availability of comparator disorders and sample types. Although H3N2b levels in the evaluated non–GM1 lysosomal diseases were generally below diagnostic thresholds, CSF from other lysosomal disorders was unavailable, and urine from Tay–Sachs disease and plasma/urine from Sandhoff disease were not included. Elevated levels in selected oligosaccharidoses further indicate that H3N2b is not exclusively specific to GM1 gangliosidosis. Broader disease controls, including other metabolic, neurodegenerative, and dementing conditions affecting ganglioside metabolism, were not assessed due to limited samples. Therefore, H3N2b elevations may not be unique to GM1 gangliosidosis, and future studies incorporating a wider disease spectrum and additional matrices will be required to better define specificity. Accordingly, H3N2b should be interpreted in conjunction with other clinical and biochemical markers. From a mechanistic perspective, H3N2b accumulation reflects impaired lysosomal degradation of β-galactose–terminated glycoconjugates and is most directly associated with β-galactosidase deficiency, particularly GM1 gangliosidosis and Morquio B disease (mucopolysaccharidosis type IVB, MPS IVB) [34]. However, reduced β-galactosidase activity can also arise from defects in lysosomal trafficking (e.g., mucolipidosis II and III [35]) or impaired enzyme stabilization (e.g., galactosialidosis [36]), all of which converge on impaired glycan catabolism. Within this framework, H3N2b likely represents a downstream consequence of disrupted lysosomal function rather than a disease-specific signature. Related disorders affecting the same pathway may therefore also exhibit altered H3N2b levels, although their magnitude and clinical utility remain to be defined. Beyond its diagnostic utility, H3N2b also demonstrates promise as a pharmacodynamic biomarker for monitoring therapeutic response. In participants receiving intravenous AAV9-mediated gene therapy [26], plasma, urine, and CSF H3N2b levels decreased following treatment and closely paralleled changes in β-galactosidase activity. Initial increases in enzyme activity were consistent with effective target engagement, while subsequent declines over time may reflect mechanisms described in AAV gene therapy, including immune-mediated clearance of transduced cells and gradual loss or dilution of episomal vector genomes during cell turnover [26,37,38]. Regardless of mechanism, the temporal concordance between enzyme activity and H3N2b indicates that the biomarker sensitively reflects downstream biochemical changes in lysosomal function. These findings support its potential use in monitoring therapeutic efficacy and durability across disease-modifying strategies, including enzyme replacement, pharmacological chaperones, stem cell approaches, and gene editing therapies [1,2,39]. Importantly, urinary measurement of H3N2b provides a noninvasive approach for longitudinal monitoring of disease burden and treatment response. It is also important to note that circulating β-galactosidase activity may not fully reflect lysosomal enzyme activity within cells, where stabilization occurs within a multienzyme complex. In this study, leukocyte enzyme activity and pre-analytical stability of circulating enzyme measurements were not assessed; further work is needed to clarify the relationship between blood-based assays and intracellular lysosomal function. Importantly, longitudinal H3N2b measurements in treated participants showed a consistent decline following AAV9-GLB1 gene therapy, paralleling increases in β-galactosidase activity and reductions in GM1 ganglioside. These biochemical changes occurred in the context of largely stabilized clinical outcomes, as reflected by Vineland-3 and CGI-I assessments, which generally indicated maintenance of function or minimal change over time rather than overt improvement [26]. This is consistent with the natural history of GM1 gangliosidosis, in which progressive and partially irreversible neurodevelopmental impairment may limit the extent to which biochemical correction translates into measurable functional gains once disease is established. Thus, H3N2b appears to primarily capture upstream metabolic response to therapy, whereas clinical outcome measures reflect downstream, integrated, and time-lagged manifestations of disease biology. Although formal correlation analyses between H3N2b dynamics and clinical endpoints were not feasible due to limited sample size and heterogeneity of longitudinal follow-up, the qualitative concordance between biochemical correction and relative clinical stabilization supports the clinical relevance of H3N2b as a pharmacodynamic biomarker. Larger longitudinal studies integrating biochemical, imaging, and functional outcomes will be required to more precisely define the temporal and quantitative relationships between substrate reduction and clinical trajectories. Several limitations should be considered. The number of infantile GM1 gangliosidosis patients was limited due to disease rarity and sample availability. The absence of longitudinal infantile data and limited age coverage restrict interpretation of age-related effects on H3N2b levels. In addition, incomplete disease controls and missing sample types constrain assessment of diagnostic specificity. Larger, more diverse cohorts and longitudinal studies across all clinical subtypes will be essential to fully define biomarker behavior across disease stages and over time. In summary, H3N2b is a sensitive and applicable biomarker for GM1 gangliosidosis that can be measured in plasma, urine, and CSF. In this cohort, H3N2b distinguishes GM1 gangliosidosis from unaffected controls and shows responsiveness to gene therapy-associated biochemical changes. However, because comparator analyses were limited and several relevant disorders were not available for testing, its disease specificity remains incompletely defined. Further studies including Tay–Sachs disease, Sandhoff disease, and broader metabolic comparator cohorts will be required to establish the specificity and clinical role of H3N2b. Supplementary Material Acknowledgments This work was supported by grants U01NS114156, ZIAHG200409, and P30 DK020579 from the National Institutes of Health (NIH) and a grant from National Tay-Sachs & Allied Diseases Association Inc. Abbreviations: adeno-associated virus serotype 9 alpha-mannosidosis cerebrospinal fluid coefficient of variation Fabry disease Gaucher disease high-quality control Krabbe disease liquid chromatography–tandem mass spectrometry lower limit of quantification low-quality control mucopolysaccharidosis type I mucopolysaccharidosis type II mucopolysaccharidosis type IIIB mucopolysaccharidosis type IIIC mucopolysaccharidosis type IVA mucopolysaccharidosis type IVB mucopolysaccharidosis type VI mucopolysaccharidosis type VII middle-quality control neurofilament light chain Niemann-Pick disease type A Niemann-Pick disease type B Niemann-Pick disease type C1 Pompe disease quality control relative error standard deviation Tay-Sachs disease upper limit of quantification Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.ymgme.2026.110159. Footnotes Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: CRediT authorship contribution statement Pamela Kell: Methodology, Investigation, Data curation. Sonali Mishra: Methodology, Investigation, Data curation. Precilla D'Souza: Resources. Cynthia J. Tifft: Writing – review & editing, Formal analysis, Conceptualization. Dennis J. Dietzen: Writing – review & editing, Resources. Laura Duque Lasio: Writing – review & editing, Resources. Jennifer Powers Carson: Writing – review & editing, Resources. Xinying Hong: Writing – review & editing, Resources. Raphael Schiffmann: Writing – review & editing, Resources. María Dolores Ledesma: Writing – review & editing, Resources. Nahid Tayebi: Writing – review & editing, Resources. Patricia I. Dickson: Writing – review & editing, Resources. Forbes D. Porter: Writing – review & editing, Resources. Daniel S. Ory: Writing – review & editing. Xuntian Jiang: Writing – original draft, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Data availability Data will be made available on request. References Associated Data Supplementary Materials Data Availability Statement Data will be made available on request.