Ultra-Performance Liquid Chromatography–Tandem Mass Spectrometry Analysis of Urinary Oligosaccharides and Glycoamino Acids for the Diagnosis of Mucopolysaccharidosis and Glycoproteinosis

Abstract

Background

Mucopolysaccharidosis (MPS) and glycoproteinosis are 2 groups of heterogenous lysosomal storage disorders (LSDs) caused by defective degradation of glycosaminoglycans (GAGs) and glycoproteins, respectively. Oligosaccharides and glycoamino acids have been recognized as biomarkers for MPS and glycoproteinosis. Given that both groups of LSDs have overlapping clinical features, a multiplexed assay capable of unambiguous subtyping is desired for accurate diagnosis, and potentially for severity stratification and treatment monitoring.

Methods

Urinary oligosaccharides were derivatized with 3-methyl-1-phenyl-2-pyrazoline-5-one (PMP) and analyzed by ultra-performance liquid chromatography–tandem mass spectrometry (UPLC-MS/MS) together with the underivatized glycoamino acids. Novel biomarkers were identified with a semi-targeted approach with precursor mass scanning, the fragmentation pattern (if applicable), and the biochemical basis of the condition.

Results

A UPLC-MS/MS analysis with improved chromatographic separation was developed. Novel biomarkers for MPS-IIIA, IIIB, IIIC, and VII were identified and validated. A total of 28 oligosaccharides, 2 glycoamino acids, and 2 ratios were selected as key diagnostic biomarkers. Validation studies including linearity, lower limit of quantitation (LLOQ), and precision were carried out with the assay performance meeting the required criteria. Age-specific reference ranges were collected. In the 76 untreated patients, unambiguous diagnosis was achieved with 100% sensitivity and specificity. Additionally, the levels of disease-specific biomarkers were substantially reduced in the treated patients.

Conclusions

A multiplexed UPLC-MS/MS assay for urinary oligosaccharides and glycoamino acids measurement was developed and validated. The assay is suitable for the accurate diagnosis and subtyping of MPS and glycoproteinosis, and potentially for severity stratification and monitoring response to treatment.

Introduction

Lysosomal storage disorders (LSDs) are heterogenous conditions caused by deficient acid hydrolases, lysosomal membrane transporters, or proteins involved in posttranslational modification. More than 50 LSDs have been identified, including mucopolysaccharidosis (MPS), glycoproteinosis, sphingolipidosis, and neuronal ceroid lipofuscinosis (1).

MPS is a group of disorders caused by impaired degradation of glycosaminoglycans (GAGs), negatively charged linear polysaccharides with repetitive sulfated disaccharide units. Degradation of GAGs requires the concerted action of various highly specific exo- and endo-hydrolases (2). MPS is caused by defects in these exo-hydrolases, leading to the accumulation of the corresponding GAGs. Glycoproteinosis, also known as oligosaccharidosis, is a group of disorders caused by impaired degradation of oligosaccharides derived from N- and O-linked glycoproteins. Clinically, MPS and glycoproteinosis have overlapping features (1). Currently, many MPS and glycoproteinosis subtypes have disease-modifying treatments available or under development, including hematopoietic stem cell transplantation (HSCT), enzyme replacement therapy (ERT), and gene therapy (3, 4). With these new developments, early diagnosis is crucial for prompt treatment initiation to halt or prevent the devastating irreversible long-term complications (5, 6).

Currently, MPS and glycoproteinosis can be diagnosed biochemically and/or genetically. Biochemical testing involves measuring enzymatic activity and biomarker abundance. Elevated urinary GAGs are the hallmark of MPS. The traditional colorimetric assay is neither sensitive nor specific (7). New methods for GAG measurement have been developed, mostly using liquid chromatography–tandem mass spectrometry (LC-MS/MS) to detect the internal repeating disaccharide units generated by GAG digestion/depolymerization (8, 9). However, this approach cannot differentiate MPS subtypes accumulating the same GAG species. Alternatively, endogenous sulfated oligosaccharides and monosaccharides (also known as “endognoues GAG NRE” or “GAG fragments”), presumably derived from incompletely degraded GAGs, have been recognized as disease-specific markers, since their nonreducing ends (NREs) reflect the underlying enzyme defects and allow subtyping (10–12). Most recently, a Sensi-Pro assay involving lyase digestion and NREs measurement was described and was also suitable for MPS subtyping (13, 14).

Elevated urinary oligosaccharides and glycoamino acids are the hallmark of glycoproteinosis. The traditional thin-layer chromatography approach is not sensitive, specific, or structurally informative. Over the past decade, LC-MS/MS or matrix-assisted laser desorption/ionization time of flight (MALDI-TOF) mass spectrometry-based assays have been developed, where analytes typically undergo permethylation or reducing end derivatization (15–17).

Given that MPS and glycoproteinosis have overlapping clinical features, a multiplexed assay capable of simultaneous detection of both groups of diseases is desirable. Since oligosaccharides have been identified as biomarkers for both MPS and glycoproteinosis, we consolidated these 2 tests to improve throughput along with diagnostic sensitivity and specificity (11, 15). Novel biomarkers were also identified and validated to improve diagnostic sensitivity. In conclusion, we developed a multiplexed UPLC-MS/MS assay suitable for accurately diagnosing and subtyping both MPS and glycoproteinosis.

Materials and Methods

Materials

A total of 101 de-identified urine samples from MPS- or glycoproteinosis-affected patients were collected at the Children’s Hospital of Philadelphia (CHOP) and the Greenwood Genetic Center. Residual proficiency testing samples from the College of American Pathologists (CAP) and European Research Network for evaluation and improvement of screening, Diagnosis, and treatment of Inherited disorders of Metabolism (ERNDIM) were also included. Urines from 202 unaffected controls were residual samples from organic acid analysis. Using archived specimens for test development and validation at CHOP does not require consent (IRB 23-021290).

The internal standard (IS) working solution was 10 µM chondroitin disaccharide di-4S (dUA-GalNAc-4S, Biosynth), 6-a-D-glucopyranosyl maltotriose-¹³C₆ (¹³C₆-Hex4, TRC Canada), and β-D-GlcN[²H₃]Ac-(1→N)-Asn (d₃-GlcNAc-ASN, Omicron) in water.

The derivatization solution contained 0.25 M 3-methyl-1-phenyl-2-pyrazoline-5-one (PMP, Sigma) in 0.5 M ammonia in 24% methanol, pH 9.0 to 10, and was prepared fresh daily.

SampleDerivatization

Urine equivalent to 0.02 mg creatinine with 10 µL IS working solution was dried with a nitrogen evaporator at 20 to 35°C. Then, 100 µL derivatization solution was added before being derivatized at 70°C for 90 min. The sample was quenched with 100 µL 0.5 M formic acid in water and filtered through a 20 μm frit. Excess PMP was removed by liquid–liquid extraction using 400 µL water and 700 µL chloroform. The sample was mixed, then phase separation was achieved by centrifugation and 500 µL chloroform (bottom layer) containing unreacted PMP was removed. Then, 500 µL chloroform was added, followed by mixing, centrifugation, and removal of 500 µL chloroform, and this was repeated 2 more times. The aqueous phase was transferred to an autosampler plate and 5 µL was injected for UPLC-MS/MS analysis.

UPLC-MS/MS Analysis forOligosaccharides,GlycoaminoAcids, andHexNAc-S

The UPLC-MS/MS analysis was carried out on a Waters TQ-S triple quadrupole mass spectrometer coupled to an ACQUITY Classic UPLC system.

An ACQUITY UPLC HSS PFP column (100 Å, 1.8 µm, 2.1 mm × 100 mm, Waters) maintained at 25°C was used. Mobile phases A and B were water with 0.1% formic acid and acetonitrile with 0.1% formic acid, respectively. The weak needle wash was water with 0.1% formic acid. The strong needle wash was 1:1:1:1 (v:v:v:v) water:methanol:acetonitrile:isopropanol. All solvents were LC-MS grade. The flow rate was 0.5 mL/min. Two injections were performed per sample. The 7-minute gradient for oligosaccharide and glycoamino acids analysis was as follows: 0 to 2 min 5% B; 2 to 5 min, 5% to 70% B (linear); 5 to 5.1 min, 70% to 95% B (linear); 5.1 to 6 min, 95% B; 6 to 6.1 min, 95% to 5% B (linear); 6.1 to 7 min, 5% B. The 15-minute gradient for HexNAc-S analysis was as follows: 0 to 12 min, 5% to 20% B (linear); 12 to 12.5 min, 20% to 95% B (linear); 12.5 to 14 min, 95% B; 14 to 14.1 min, 95% to 5% B (linear); 14.1 to 15 min, 5% B.

The MS/MS analysis was performed in both electrospray ionization (ESI) positive and negative mode, with a source temperature of 120°C, capillary voltage of 3.5 kV (positive) and 2.0 kV (negative), desolvation temperature of 600°C, desolvation gas flow of 900 L/h, cone gas flow of 150 L/h, and nebulizer of 7.0 bar. For the 7-minute analysis, the first minute was in positive mode for glycoamino acids, then switched to negative mode for oligosaccharides. ESI negative mode was used for the 15-minute HexNAc-S analysis. The multiple reaction monitoring (MRM) transitions with individually optimized parameters are summarized in online Supplemental Table 1.

Calculation

Chromatography peaks were integrated using ASCENT (Indigo BioAutomation). The apparent concentration (µmol/mol creatinine) was calculated by multiplying the response ratio of the analyte to the IS by the amount of IS added (µmol), then dividing by the amount of creatinine (mol).

Validation

Linearity was assessed by mixing urine from affected patients and unaffected adults at varying ratios while maintaining a total creatinine of 0.02 mg. Mixed urine was diluted with normal urine to achieve low analyte concentration while maintaining a total creatinine of 0.02 mg. The sample was prepared 5 times per day for 5 days. Lower limit of quantitation (LLOQ) was the averaged concentration that yielded an averaged %CV ≤20% over 5 days.

Inter-day and intra-day precision were assessed with 3 quality control (QC) samples. QC1 was prepared with urine from unaffected controls. QC2 was prepared by combining urine from patients with various MPS subtypes. QC3 was prepared by combining urine from patients with various glycoproteinosis subtypes. A 20 × 2 × 2 study was performed, with each QC being prepared twice per day with 2 injections per preparation for 20 days.

Age-specific reference ranges were collected using 202 urine samples from unaffected controls. Four age groups were collected: <2 months, 2 months to 1 year, 1 to 3 years, >3 years. Reference ranges were either set at the LLOQ or established with the 95% confidence interval after log transformation.

Results

MethodDevelopment

The assay was semiquantitative due to the lack of reference materials and isotope-labeled IS. Three ISs were used for the whole panel. Each IS was assigned to analytes with similar properties, i.e., dUA-GalNAc-4S, ¹³C₆-Hex4, and d₃-GlcNAc-ASN to sulfated oligosaccharides, nonsulfated oligosaccharides, and glycoamino acids, respectively. IS was added to urine prior to being dried to account for subsequent sample loss and variation in derivatization efficiency. The oligosaccharides underwent PMP derivatization using a published protocol (11). Glycoamino acids remained underivatized but were compatible with the sample preparation and analysis and were therefore included in the panel.

PMP derivatization improved the chromatographic resolution and MS/MS sensitivity (10). The derivatized oligosaccharides were partially separated on the reverse-phase column based on their hydrophobicity, determined by the number, type, and linkage of the monosaccharide and sulfate groups. Two injections per sample were required. The 15-minute HexNAc-S analysis achieved close to baseline separation for the 3 isomers, namely GlcNAc-6S, GalNAc-4S, and GalNAc-6S (online Supplemental Fig. 1A). However, a 10-fold signal reduction was observed. Therefore, a 7-minute method was developed for general oligosaccharides and glycoamino acids analysis, which was sufficient for separating most isomeric oligosaccharides, including UA-HexNAc(1S) and UA(1S)-HexNAc for MPS-I and II, respectively (Supplemental Fig. 1B).

Although traditionally the glycoproteinosis-related neutral oligosaccharides were detected in ESI positive mode, negative mode was also suitable for their detection (15, 16). Conversely, negative mode was necessary for detecting MPS-related sulfated oligosaccharides. Therefore, all oligosaccharides were analyzed in ESI negative mode, whereas ESI positive mode was used for glycoamino acids since they eluted much earlier from the column.

Together, 84 oligosaccharides and glycoamino acids were monitored, 27 of which were selected as key diagnostic biomarkers based on analytical and clinical specificity and sensitivity. Novel biomarkers, described in the section below, were also validated and included. Key features for each condition are summarized in Table 1. Together, our approach largely improved the robustness, sensitivity, and specificity of the test.

Table 1.

Summary of the targeted conditions, including defective protein, number of patients included in the study, and the relevant biomarkers.

Condition	Defective protein	No. of patients included		Relevant biomarker(s)
		Untreated	Treated
MPS-I	alpha-L-iduronidase	6	5	UA-HexNAc-1S, #1; UA-HexN-UA-2S
MPS-II	iduronate 2-sulfatase	7	6	UA-HexNAc-1S, #2; UA-HexN-UA-2S
MPS-IIIA	N-sulfoglucosamine sulfohydrolase	5	0	HexN-UA-1S; HexN-UA-HexNAc-UA-2S
MPS-IIIB	N-alpha-acetylglucosaminidase	3	0	HexNAc-UA-HexNAc-1S; (HexNAc-UA)2
MPS-IIIC	heparan acetyl-CoA:alpha-glucosaminide N-acetyltransferase	3	0	HexN-UA-HexNAc-1S, #2; (HexN-UA)2-HexNAc-2S
MPS-IIID	N-acetylglucosamine-6-sulfatase	1	0	GlcNAc-6S
MPS-IVA	galactosamine-6-sulfate sulfatase	5	4	GalNAc-6S, GalNAc-6S/GalNAc-4S
MPS-IVB	beta-galactosidase	1	0	Hex3-HexNAc2; Hex5-HexNAc3
MPS-VI	arylsulfatase B	3	2	GalNAc-4S, GalNAc-6S/GalNAc-4S
MPS-VII	beta-glucuronidase	4	2	UA-HexNAc; HexNAc-UA-HexNAc; UA-HexN-UA-2S
Multiple sulfatase deficiency (MSD)	formylglycine generating enzyme	6	0	HexN-UA-1S; HexN-UA-HexNAc-UA-2S; GlcNAc-6S; UA-HexNAc-1S, #2; UA-HexN-UA-2S; GalNAc-6S; GalNAc-4S
GM1 gangliosidosis (GM1)	beta-galactosidase	7	0	Hex3-HexNAc2; Hex5-HexNAc3
Sandhoff disease	hexosaminidase A and B	3	0	Hex2-HexNAc3; Hex3-HexNAc4
Alpha-mannosidosis (aMan)	alpha-mannosidase	6	1	Hex3-HexNAc; Hex4-HexNAc
Beta-mannosidosis (bMan)	beta-mannosidase	3	3	Hex-HexNAc
Fucosidosis (aFuc)	alpha-L-fucosidase	2	1	Fuc-HexNAc
Sialidosis (Sia)	neuraminidase 1	2	0	Neu5Ac-Hex3-HexNAc2; Neu5Ac2-Hex5-HexNAc3
Galactosialidosis (GalSia)	protective protein/cathepsin A	1	0	Neu5Ac-Hex3-HexNAc2; Neu5Ac2-Hex5-HexNAc3; Hex3-HexNAc2; Hex5-HexNAc3
Mucolipidosis-II/III (ML-II/III)	N-acetylglucosamine-1-phosphotransferase	2	0	Neu5Ac-Hex3-HexNAc2; Neu5Ac2-Hex5-HexNAc3; Hex3-HexNAc2; Hex5-HexNAc3; Hex3-HexNAc; Hex4-HexNAc
Aspartylglucosaminuria (AGU)	aspartylglucosaminidase	2	0	GlcNAc-Asn
Schindler disease	alpha-N-acetylgalactosaminidase	3	0	GalNAc-Thr
Pompe disease	alpha-1,4-glucosidase	1	1	Hex4, Hex4/Hex3

Discovery ofNovelBiomarkers for MPS

(HexNAc-UA)2-1S and UA-HexN-UA-1S, the disease specific markers reported by Saville et al. (11), were below the LLOQ in one MPS-IIIB patient and one MPS-VII patient in our cohort, respectively (online Supplemental Table 2). This necessitated novel biomarker discovery, which was conducted using a semi-targeted approach using urine from affected patients. Given the accumulated GAG species and the repeating disaccharide units are known for each subtype, a list of potential biomarkers was compiled. Full-scan spectra in both positive and negative modes were collected for each subtype and cross-checked with the candidates. The scaffold for the oligosaccharides carrying zero or one sulfate group was deduced by the fragmentation pattern in the positive mode, given that mostly Y ions were generated in these PMP-derivatized oligosaccharides upon low-energy collision (18). Novel biomarkers were therefore tentatively annotated based on the precursor mass, the fragmentation pattern (if applicable), and the biochemical basis of the condition.

HexN-UA-HexNAc-1S was first identified as a marker for MPS-IIIC. However, an isomer was observed in MPS-IIIA urine at a different retention time (Supplemental Fig. 1C). The fragmentation patterns in ESI positive mode for both isomers are shown in online Supplemental Fig. 2. Both isomers were deduced to have a scaffold of HexN-UA-HexNAc with the HexN on the NRE, with uncertain sulfate position. Given N-sulfoglucosamine sulfohydrolase is deficient in MPS-IIIA, the marker was tentatively annotated as HexN(S)-UA-HexNAc in MPS-IIIA. The sulfate position remains uncertain for the MPS-IIIC–associated HexN-UA-HexNAc-1S but is not projected to be on the NRE.

HexNAc-UA-HexNAc-1S and (HexNAc-UA)2 were identified as markers for MPS-IIIB (online Supplemental Fig. 3). For HexNAc-UA-HexNAc-1S, the position of the sulfate remains uncertain but is not projected to be on the NRE, given the derangement of N-alpha-acetylglucosaminidase in MPS-IIIB. These 2 novel MPS-IIIB markers were found above the LLOQ in the 3 MPS-IIIB patients but absent in the non-MPS-IIIB subjects in our cohort (Fig. 1F and G), demonstrating a high degree of clinical sensitivity and specificity.

Fig. 1.

MPS-associated oligosaccharide abundance in urine from LSD patients. The “Other MPS” group represents MPS patients not displayed in other columns. The “Oligo” group represents all glycoproteinosis patients. In (C), the cross (X), open circle (O), and open square (▢) in the “Untreated MPSI/II/VII” and the “Treated MPSI/II/VII” groups represent MPS-I, II, and VII patients, respectively. In (D) and (E), the cross (X) and open circle (O) in the “MPSI/II/VII” group represent treated and untreated patients, respectively. The horizontal line indicates the median of each group. The grey area indicates the reference ranges for the 2 months to 1 year group. In plots where logarithmic scales are used, zero points cannot be displayed.

UA-HexNAc and HexNAc-UA-HexNAc were identified as markers for MPS-VII (online Supplemental Fig. 4). Given beta-glucuronidase is defective in MPS-VII, the reason for HexNAc-UA-HexNAc being an MPS-VII marker remains unexplained. However, it was consistently elevated in 4 untreated and 2 treated MPS-VII urine samples (Fig. 1K). These 2 novel MPS-VII markers were found markedly elevated in the MPS-VII subjects when compared to the non-MPS-VII subjects in our cohort, again demonstrating excellent clinical sensitivity and specificity.

MethodValidation

Given most analytes are only present in affected patients and the scarcity of positive samples, linearity, LLOQ, and precision were assessed with a subset of sulfated, unsulfated, and sialylated oligosaccharides, glycoamino acids, and HexNAc-S. The linearity study is summarized in online Supplemental Fig. 5, with R² > 0.95 for all interrogated analytes. The LLOQ for the sulfated, unsulfated, and sialylated oligosaccharides, glycoamino acids, and HexNAc-S were 8, 4, 5, 45, and 400 µmol/mol creatinine, respectively. Inter- and intra-day imprecision was <20% for most analytes. No obvious carryover or interference was observed. Age-specific reference ranges are summarized in Table 2 with most analytes being decreased with age.

Table 2.

Age-specific reference ranges for the key diagnostic biomarkers in µmol/mol creatinine.

	Apparent concentration, µmol/mol creatinine
	<2 months			2 months to 1 year			1 to 3 years			> 4 years
UA-HexNAc-1S, #1	0.0	-	8.0	0.0	-	8.0	0.0	-	8.0	0.0	-	8.0
UA-HexNAc-1S, #2	0.0	-	8.0	0.0	-	8.0	0.0	-	8.0	0.0	-	8.0
UA-HexN-UA-2S	0.0	-	9.7	0.0	-	13.3	0.0	-	8.2	0.0	-	8.0
HexN-UA-1S	0.0	-	49.7	0.0	-	49.7	0.0	-	41.0	0.0	-	24.6
HexN-UA-HexNAc-UA-2S	0.0	-	16.3	0.0	-	19.2	0.0	-	18.8	0.0	-	10.5
HexNAc-UA-HexNAc-1S	0.0	-	8.0	0.0	-	8.0	0.0	-	8.0	0.0	-	8.0
(HexNAc-UA)2	0.0	-	28.5	0.0	-	18.0	0.0	-	11.4	0.0	-	5.0
HexN-UA-HexNAc-1S, #2	0.0	-	8.0	0.0	-	8.0	0.0	-	8.0	0.0	-	8.0
(HexN-UA)2-HexNAc-2S	0.0	-	8.0	0.0	-	8.0	0.0	-	8.0	0.0	-	8.0
HexNAc-UA-1S, #2	0.0	-	52.8	0.0	-	42.9	0.0	-	34.2	0.0	-	10.0
HexNAc-UA-HexNAc-2S	32.8	-	483.5	43.6	-	251.5	19.8	-	159.6	0.0	-	87.8
UA-HexNAc	1470.0	-	9883.4	714.7	-	4600.7	355.1	-	2900.2	312.5	-	2144.3
HexNAc-UA-HexNAc	147.6	-	981.2	66.0	-	430.7	35.2	-	238.1	25.0	-	186.9
Hex3-HexNAc2	0.0	-	78.5	0.0	-	32.7	0.0	-	25.7	0.0	-	17.5
Hex5-HexNAc3	0.0	-	79.6	0.0	-	20.9	0.0	-	18.9	0.0	-	15.0
Hex2-HexNAc3	0.0	-	12.0	0.0	-	7.3	0.0	-	5.3	0.0	-	4.0
Hex3-HexNAc4	0.0	-	4.0	0.0	-	4.0	0.0	-	4.0	0.0	-	4.0
Hex-HexNAc	4674.7	-	24048.9	1612.4	-	18383.1	1474.9	-	10425.1	1124.5	-	5870.9
Hex3-HexNAc	99.2	-	462.1	94.4	-	449.7	53.4	-	325.6	35.0	-	204.6
Hex4-HexNAc	0.0	-	103.1	0.0	-	69.9	0.0	-	62.4	0.0	-	30.9
Fuc-HexNAc	244.5	-	1184.7	206.5	-	1217.4	124.9	-	768.5	55.9	-	467.4
Neu5Ac-Hex3-HexNAc2	0.0	-	34.3	0.0	-	30.1	0.0	-	22.6	0.0	-	20.2
Neu5Ac2-Hex5-HexNAc3	0.0	-	217.6	0.0	-	60.1	0.0	-	48.6	0.0	-	38.5
Hex4	0.0	-	19058.9	0.0	-	15977.0	0.0	-	9748.6	0.0	-	3132.1
Hex3	0.0	-	58454.4	0.0	-	59415.3	0.0	-	15621.0	0.0	-	4936.0
GlcNAc-Asn	0.0	-	10156.1	0.0	-	11192.9	0.0	-	11624.3	0.0	-	3649.5
GalNAc-Thr	0.0	-	2882.4	0.0	-	1903.3	0.0	-	771.6	0.0	-	297.1
GlcNAc-6S	0.0	-	6430.5	0.0	-	10318.7	0.0	-	3850.6	0.0	-	1773.0
GalNAc-6S	0.0	-	12025.3	0.0	-	4825.5	0.0	-	3666.9	0.0	-	2115.9
GalNAc-4S	0.0	-	64997.7	0.0	-	18435.7	0.0	-	15188.3	0.0	-	9952.0
GalNAc-6S/GalNAc-4S	0.22	-	0.64	0.18	-	0.43	0.16	-	0.34	0.13	-	0.29
Hex4/Hex3	0.00	-	1.71	0.00	-	2.41	0.00	-	3.19	0.00	-	2.90

Diagnosis of MPS,Glycoproteinosis, andOtherConditions

Results from the 101 affected patients are summarized in Figs. 1–4 and Supplemental Table 2. Two ratios, GalNAc-6S/GalNAc-4S and Hex4/Hex3, were also calculated (Figs. 3 and 4). Given the age at the time of sample collection was unavailable for most affected patients, the reference ranges for the 2 months to 1 year group are displayed in Figs. 1–4. Affected patients exhibited significantly increased disease-relevant biomarkers, allowing unambiguous identification of all MPS and glycoproteinosis subtypes (Figs. 1–4). A substantial reduction of disease-relevant biomarkers was observed in treated patients, suggesting the assay’s utility to monitor treatment efficacy (Figs. 1–3). The MPS-I, II, IIIA, IVA, and VI samples exhibiting the lowest levels of their relevant markers were the residual ERNDIM proficiency samples (Figs. 1–3). The markers may have been partially degraded given these lyophilized sample were stored at 2 to 5 °C for an extended period. The MPS-II ERNDIM sample displayed normal level of the specific marker, UA-HexNAc-1S, #2. However, an accurate diagnosis was achieved due to the elevation of UA-HexN-UA-2S, a marker associated with MPS-I, II, and VII, together with the normal levels of the MPS-I and VII specific markers (Fig. 1). MPS-I, II, and VII patients displayed moderate increases of MPS-IIIA–associated markers, consistent with the previous report (Supplemental Table 2) (11). Neu5Ac-Hex3-HexNAc2, Neu5Ac2-Hex5-HexNAc3, and Hex4 were also elevated in multiple conditions but typically to a lesser extent, suggesting these could be secondary changes reflecting general lysosomal dysfunction (Figs. 2 and 4) (12, 15, 19). However, given specific biomarkers were available for most conditions, secondary or unspecific findings did not complicate the diagnostic process, especially if a profiling approach was utilized. The relative abundance of the 3 HexNAc-S isomers was skewed in MPS-IIID, IVA, and VI. The GalNAc-6S/GalNAc-4S ratio was therefore calculated to better distinguish MPS-IVA and VI from the control (Fig. 3D). A similar approach was adopted with the Sensi-Pro assay, where endogenous and liberated HexNAc-S were measured altogether (14).

Fig. 2.

Glycoproteinosis-associated oligosaccharide abundance in urine from LSD patients. The “Other Oligo” group represents glycoproteinosis patients not displayed in other columns. The “MPS” group represents all MPS patients. The horizontal line indicates the median of each group. In (A), (B), (I), and (J), the cross (X) and the open circle (O) in the “GalSia/MLII/III” groups represent patients with galactosialidosis and ML-II/III, respectively. In (F) and (G), the cross (X) in the “Other Oligo” group represent patients with ML-II/III. The horizontal line indicates the median of each group. The grey area indicates the reference ranges for the 2 months to 1 year group. In plots where logarithmic scales are used, zeros cannot be displayed.

Fig. 3.

(A) GlcNAc-6S, (B) GalNAc-6S, (C) GalNAc-4S, and (D) GalNAc-6S/GalNAc-4S in urine from patients with MPS-IVA, VI, MSD, and other LSDs. The “Other MPS” group represents MPS patients not displayed in other columns. The “Oligo” group represents all glycoproteinosis patients. In (D), the cross (X) and the open circle (O) in the “MPSIVA” groups represent treated and untreated MPS-IVA patients, whereas the cross (X) and the open circle (O) in the “MPSVI” groups represent treated and untreated MPS-VI patients. The horizontal line indicates the median of each group. The horizontal line indicates the median of each group. The grey area indicates the reference ranges for the 2 months to 1 year group. In plots where logarithmic scales are used, zeros cannot be displayed.

Fig. 4.

(A) Hex4 and (B) Hex4/Hex3 ratio in urine from patients with Pompe disease. The horizontal dashed lines indicate the upper limits of the age-appropriate reference range for each patient.

The use of Hex4 for the diagnosis of Pompe disease is summarized in Fig 4. Among the 2 Pompe disease patients in our cohort, the one with a higher Hex4 level (18415.9 µmol/mol creatinine) was a 12-year-old patient with infantile-onset Pompe disease (IOPD) on ERT, whereas the one exhibiting a normal Hex4 level (1162.6 µmol/mol creatinine) was an asymptomatic patient with ambigous genotypes suspected of late-onset Pompe disease (LOPD) (Fig. 4). The Hex4/Hex3 ratio was further used to distinguish the 2 Pompe patients from the control, with the ratio in the treated IOPD patient being highly elevated and that in the LOPD patient at the higher end of the reference range (Fig. 4B).

Additionally, the overall oligosaccharide profile could pinpoint the underlying derangement for conditions involving multiple enzymes, including multiple sulfatase deficiency (MSD), galactosialidosis, and mucolipidosis II and III (ML-II/III). In MSD, the biomarkers associated with MPS-IIIA and IIID were significantly elevated to the level in those with isolated defects, while the markers associated with MPS-II, IVA and VI were mildly elevated and the GalNAc-6S/GalNAc-4S ratio were normal (Figs. 1 and 3). Galactosialidosis and ML-II/III patients displayed markedly increased Neu5Ac-Hex3-HexNAc2 and Neu5Ac2-Hex5-HexNAc3, 2 sialylated oligosaccharides, and Hex3-HexNAc2 and Hex5-HexNAc3, 2 presumably galactosylated species associated with GM1 gangliosidosis (Fig. 2). Hex3-HexNAc and Hex4-HexNAc, 2 presumably mannosylated species associated with alpha-mannosidosis, were mildly elevated in ML-II/III (n = 2) but within normal limits in galactosialidosis (n = 1) (Supplemental Table 2).

Discussion

The method described herein offered multiple improvements compared to the original method published by Saville et al. (11). First, adding IS in the earliest step helped to account for subsequent loss and variation in derivatization. Second, the chromatographic separation was greatly improved for the isomeric HexNAc-S, which was critical for accurate quantitation. It also allowed utilization of the GalNAc-6S/GalNAc-4S ratio to better differentiate MPS-IVA and VI from controls. Third, novel biomarkers were identified, some of which outperformed the ones reported by Saville et al. (11). The panel expansion further allowed monitoring 2 disease-specific markers for most targeted conditions to potentially improve the diagnostic sensitivity and specificity. We also demonstrated a generalizable strategy to identify novel oligosaccharide biomarkers. Last, the method was suitable for simultaneous and accurate diagnosis of a much larger battery of LSDs, with the potential for severity stratification and treatment monitoring.

MSD is caused by a deficient formylglycine-generating enzyme, encoded by the SUMF1 gene, that is essential for activating all sulfatases, including those involved in MPS-II, IIIA, IIID, IVA, and VI. A distinct pattern was observed in MSD urine, suggesting derangement of multiple sulfatases of varying degrees (Figs. 1 and 3). Similar findings have been recently reported (20).

GM1 gangliosidosis and MPS-IVB are caused by biallelic pathogenic variants in GLB1, which encodes β-galactosidase, a lysosomal hydrolase removing β-linked galactose from the NREs of glycans, keratan sulfate (KS), and GM1 gangliosides. Hex3-HexNAc2 and Hex5-HexNAc3, 2 previously reported markers originating from glycoproteins, were selected as the key biomarkers for GM1 gangliosidosis and MPS-IVB (15, 16, 21, 22). Hex3-HexNAc2 isomers were further reported as useful pharmacodynamic biomarkers to evaluate the efficacy of gene therapy for GM1 gangliosidosis (22). Saville et al. reported (Hex-HexNAc)2-(2S), presumably from KS, as a marker for these 2 conditions, but it was not detected using our platform (11). Conversely, we found Hex3-HexNAc2-S, presumably from KS, to be elevated in GM1 gangliosidosis and MPS-IVB but not controls (Supplemental Table 2). Given these markers derive from different macromolecules, i.e., glycoproteins and GAGs, it would be interesting to monitor them simultaneously to assess if they correlate with severity or certain clinical phenotypes.

Sialidosis is caused by isolated neuraminidase 1 deficiency. Galactosialidosis is caused by a combined deficiency of β-galactosidase and neuraminidase 1. ML-II/III is caused by failure in trafficking most lysosomal hydrolases to lysosomes, leading to the accumulation of a variety of undegraded substrates, including oligosaccharides and GAGs (23). As expected, sialidosis patients displayed an isolated increase of sialylated species (Fig. 2). The pattern observed in ML-II/III and galactosialidosis was consistent with impairment of neuraminidase 1 and β-galactosidase in both conditions, with additional alpha-mannosidase derangement in ML-II/III (Fig. 2). No apparent changes in GAG-derived oligosaccharides were observed in our ML-II/III cohort (Supplemental Table 2). Previous studies suggested all ML-II/III patients displayed a characteristic urinary oligosaccharide profile with elevation of species with terminal sialic acid, galactose, and mannose, whereas the increase of oligosaccharides with terminal N-acetylglucosamine and urinary GAGs was inconsistent and more common in severe cases (16, 23, 24). It is possible that the 2 ML-II/III patients in our cohort were not at the severe end of the spectrum. Therefore, Hex3-HexNAc and Hex4-HexNAc may differentiate ML-II/III from galactosialidosis, whereas other species, especially those associated with beta-mannosidosis, Sandhoff disease, and MPS may further stratify ML-II/III by severity, although a larger cohort is required to verify these findings (25).

Urinary Hex4, or Glc4, is a biomarker for Pompe disease (19, 26). The Hex4 level in IOPD <2 months or LOPD has been reported to overlap with age-matched controls, prompting the use of a reference range based on 0 to 90th confidence interval (19, 26). Using the Hex4/Hex3 ratio as an adjunct marker is promising based on our limited data but remains to be assessed with larger cohorts. Glcα1–3Glcα1-3Glcα1-2Man (Glc3Man), another Hex4 species, has been recognized as a specific marker for mannosyl-oligosaccharide glycosidase-related congenital disorder of glycosylation (MOGS-CDG) (12, 27). A retention time shift between Glc4 and Glc3Man is expected due to the difference in glycosidic linkages, but remains to be examined with positive samples (27).

N-glycanase (NGLY1) deficiency has been reported to display a similar glycoamino acid profile as aspartylglucosaminuria (AGU), since the defective N-glycanase is the cytosolic counterpart of the lysosomal aspartylglucosaminidase (12, 28, 29). Our assay should be capable of diagnosing NGLY1. Positive samples are required for validation and assessing if the abundance or pattern of glycoamino acids could further distinguish these 2 conditions.

Recently, measurement of sulfated oligosaccharides in dried blood spots (DBS) has been implemented as the second-tier test for the newborn screening of MPS and GM1 gangliosidosis, which outperformed the traditional internal disaccharide approach (30–32). A two-tier screening strategy can largely reduce false-positive rates and alleviate the unnecessary burden on families and healthcare systems. We observed all DBS markers reported by Herbst et al. in urine, most of which were included in our finalized panel (30). However, we note different markers were selected for MPS-IIIB, IVB/GM1 gangliosidosis, and MPS-VII due to differences in their relative abundance in blood vs urine (30). Translating the rest of the glycoproteinosis markers into DBS-based assays may be of interest, given novel treatments are being developed rapidly for LSDs, with newborn screening being the next logical step to ensure early diagnosis.

One major limitation is the lack of reference materials and isotope-labeled IS. Reference ranges should be carefully reassessed when transferring the method across instruments. The quantitation ability can be largely improved with a relative quantitation approach, where an unknown sample is derivatized by non-labeled PMP and compared to a reference sample derivatized by isotope-labeled PMP (33). However, we did not adopt this approach due to the substantial cost of isotope-labeled PMP and the lack of a large quantity of positive samples, since most oligosaccharides are only present in affected patients. We do note this as a potential option to improve the quantitative ability of the test. Alternatively, although not trivial, standards and isotope-labeled IS can be synthesized to assist structural confirmation and quantitation (22). Additionally, a larger cohort of patients with clinical information available would be ideal to assess if this test is suitable for severity stratification and monitoring treatment efficacy.

Conclusion

In conclusion, we developed and validated a multiplexed UPLC-MS/MS assay for measurement of urinary oligosaccharides and glycoamino acids. The assay is suitable for accurately diagnosing and subtyping MPS and glycoproteinosis and potentially for severity stratification and monitoring response to treatment.

Supplemental Material

Supplemental material is available at Clinical Chemistry online.

Nonstandard Abbreviations

UPLC-MS/MS, ultra-performance liquid chromatography–tandem mass spectrometry; MPS, mucopolysaccharidosis; LSD, lysosomal storage disorder; GAG, glycosaminoglycan; PMP, 3-methyl-1-phenyl-2-pyrazoline-5-one; LLOQ, lower limit of quantitation; NRE, nonreducing end; IS, internal standard; ESI, electrospray ionization; ML-II/III, mucolipidosis II and III.

Human Genes

SUMF1, sulfatase modifying factor 1; GLB1, galactosidase beta 1.

Author Contributions

The corresponding author takes full responsibility that all authors on this publication have met the following required criteria of eligibility for authorship: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved. Nobody who qualifies for authorship has been omitted from the list.

Parith Wongkittichote (Data curation-Equal, Formal analysis-Equal, Investigation-Equal, Writing—original draft-Equal, Writing—review & editing-Equal), Se Hyun Cho (Data curation-Equal, Formal analysis-Equal, Methodology-Equal, Resources-Equal, Validation-Equal), Miller Artis (Data curation-Equal, Formal analysis-Equal, Methodology-Equal, Resources-Equal, Validation-Equal), Kaitlyn King (Data curation-Equal, Formal analysis-Equal, Methodology-Equal, Resources-Equal, Validation-Equal), Zackary M. Herbst (Conceptualization-Equal, Investigation-Equal, Resources-Equal, Writing—review & editing-Equal), Zhimei Ren (Conceptualization-Equal, Formal analysis-Equal, Methodology-Equal), Michael Gelb (Conceptualization-Equal, Investigation-Equal, Methodology-Equal, Resources-Equal, Validation, Writing—review & editing-Equal), and Xinying Hong (Conceptualization-Equal, Data curation-Equal, Formal analysis-Equal, Funding acquisition-Equal, Investigation-Equal, Methodology-Equal, Project administration-Lead, Supervision-Lead, Writing—original draft-Equal, Writing—review & editing-Equal)

Authors’ Disclosures or Potential Conflicts of Interest

Upon manuscript submission, all authors completed the author disclosure form.

Research Funding

None declared.

Disclosures

M.H. Gelb received a grant from National Institutes of Health (NIH) (Grant R01 DK067859), which also supports Z.M. Herbst. M.H. Gelb is a co-founder of GelbChem, LLC. Z.M. Herbst is a member of GelbChem, LLC. Z. Ren was supported by the Office of Naval Research via grant N00014-20-1-2337 from July 2021 to June 2023 as a postdoctoral researcher, and was an external advisor at Bain & Company from July 2021 to June 2022.

Role of Sponsor

No sponsor was declared.

Acknowledgments

The authors gratefully acknowledge Dr. Laura Pollard (Greenwood Genetic Center) for providing positive samples. We are grateful to the Faculty of Medicine, Ramathibodi Hospital, Mahidol University for providing Research Career Development Awards to PW.