Tandem Mass Spectrometry Enzyme Assays for Multiplex Detection of 10-Mucopolysaccharidoses in Dried Blood Spots and Fibroblast

Abstract

The mucopolysaccharidoses (MPSs) are a class of inborn errors of metabolism caused by deficiency of each of the enzymes involved in the lysosomal degradation of mucopolysaccharides. Newborn screening panels worldwide have been recently expanded to include one or more MPS disorders as treatments are available and are most efficacious if initiated early in life. Here we report the first multiplex assay of 10 enzymatic activities in dried blood spots and fibroblast lysates that allow newborn screening and diagnosis of all MPS disorders except the ultra-rare MPS-IX. The assay consists of incubation of enzyme-specific substrates with dried blood spot punches or fibroblast lysate followed by quantification of enzymatic products using liquid chromatography-tandem mass spectrometry (LC-MS/MS) together with internal standards. Assay of all MPS enzymes using fluorimetric or other methods has not been possible. The steps of the LC-MS/MS assay are sufficiently simple and rapid to be used in newborn screening and diagnostic laboratories. Assays showed acceptable precision, and enzymatic activities measured in confirmed MPS samples are well below the reference range.

Graphical Abstract

INTRODUCTION

The mucopolysaccharidoses (MPSs) are a group of recessive genetic disorders with a total incidence of 1:25000 live births. They are caused by deficiency of one of 11 enzymes involved in the stepwise degradation of glycosaminoglycans (also called mucopolysaccharides) in lysosomes.¹ For example, the degradation of heparan sulfate requires 7 distinct enzymes (3 glycosidases, 3 sulfatases, and one acetyltransferase), and MPS-I, -II, -IIIA-D, and VII diseases result due to the deficiency of a single one of these enzymes. MPS-IVA leads to accumulation mainly of keratan sulfate, whereas MPS-VI leads to accumulation of chondroitin sulfates. The enzyme deficiency and the consequent buildup of glycosaminoglycans causes progressive cellular damage and dysfunction of multiple tissues including those of the central nervous system.

As is the case with all other lysosomal storage diseases, it has not been possible in many cases to easily go from genotype to phenotype given the large number of partially penetrant DNA variations and the common occurrence of variations of unknown pathogenic significance. Thus, biochemical analysis still remains the best first-tier method for diagnosis. Since patients with different MPS types share common phenotypes, it is important to have a convenient biochemical assay that can measure all of the relevant enzymatic activities in a single multiplex panel.

Chamoles and co-workers were the first to show that at least some lysosomal enzymes remain active after storage in dried blood spots (DBS) and can be assayed by hydration of DBS punches in buffer containing specific substrates for the relevant enzymes.^2,3 We have been developing assays of lysosomal enzymes in DBS based on tandem mass spectrometry (MS/MS) as the readout of the enzymatic product.⁴ Fluorescence assay of lysosomal enzymes is also possible.^{2, 5–7} MS/MS is the only method so far that can be applied to all MPS disorders as will be shown in the present study. For example, an MS/MS assay, but not a fluorimetric assay, of MPS-IIIA using DBS has been reported.⁸ This is important for newborn screening since DBS is the only available sample. Also, DBS provides a much more convenient sample for diagnosis of already symptomatic patients.

Another option for detecting MPS disorders is to measure the accumulation of glycosaminoglycans either in blood or urine. This is not the method of choice for first-tier newborn screening of MPS disorders for two reasons. Analysis of glycosaminoglycans requires that these polymers be converted by chemical or enzymatic means to small fragments that can be analyzed by liquid chromatography-MS/MS (LC-MS/MS)^9–11 This analysis requires at least 6-7 minutes per sample, which is almost an order of magnitude too slow for newborn screening. Also, the false positive rate for newborn screening of MPSs by glycosaminoglycan analysis is estimated to be at least 40-fold higher than for enzymatic-based newborn screening.^12,13 It is becoming clear that glycosaminoglycan analysis in DBS is useful for reducing the false positives found in newborn screening based on measurement of enzymatic activities.¹⁴

In this article we report a multiplex LC-MS/MS assay of all types of MPS disorders except MPS-IX (only a handful of patients worldwide have been found for this disease) that is appropriate for high throughput newborn screening and diagnosis using a minimal amount of sample. Furthermore, all reagents for the multiplex assay are commercially available so that the new assay can be readily incorporated into newborn screening and diagnostic laboratories.

EXPERIMENTAL SECTION

Materials.

All studies with human samples were approved by the University of Washington IRB. DBS from affected patients were obtained with IRB approval from previously diagnosed patients. All DBS were stored at −20 °C in sealed plastic bags containing desiccant. Patient fibroblast cells were obtained from the Coriell Institute for Medical Research and were grown using their standard protocol. The specific cell lines used were GM03440 (healthy adult) and GM05093 (MPS-IIID). All substrates and internal standards are commercially available from PerkinElmer, Inc. (IDUA, IDS, NAGLU, GALNS, GLB1, ARSB, GUSB) or from GelbChem, LLC. (SGSH, HGSNAT, GNS, GLB1). Synthesis of the substrates and internal standards are described in the Supporting Information (for HGSNAT and GNS) or in our other articles.^12,15–17 NAG-thiazoline is from Toronto Research Chemicals (T293625).

Preparation of HGSNAT-NAGLU duplex assay cocktail.

The cocktail was prepared by mixing methanolic solutions of individual reagents. Solvent was removed in a vacuum centrifuge, and the residue was dissolved in sodium citrate buffer (50 mM, pH 5.0) containing acetyl-CoA (1.6 mM). Final concentrations are: NAGLU-S (0.5 mM), NAGLU-IS (10 μM), acetyl-CoA (1.6 mM), HGSNAT-S (0.5 mM), HGSNAT-IS (2 μM), NAG-thiazoline (0.1 mM). NAG-thiazoline is used as an inhibitor of hexosaminidase A/B which would otherwise convert trace amounts of β-anomer impurity in NAGLU-S to NAGLU-P, leading to anomalous NAGLU-P.¹² The cocktail can be stored frozen at −20 °C for at least 1 month.

Preparation of 7-plex and 8-plex cocktails.

The cocktails were prepared as above using sodium acetate buffer (50 mM, pH 5.0) containing Ce(OAc)₃ (5 mM). Final concentrations are: IDUA-S (0.5 mM), IDUA-IS (5 μM), IDS-S (0.5 mM), IDS-IS (5 μM), SGSH-S (1 mM), SGSH-IS (0.5 μM), GALNS-S (1 mM), GALNS-IS (2 μM), GLB1-S (0.5 mM), GLB1-IS (5 μM), ARSB-S (1 mM), ARSB-IS (2 μM), GUSB-S (0.5 mM), GUSB-IS (10 μM), NAG-thiazoline (0.1 mM). NAG-thiazoline is used as an inhibitor of hexosaminase A/B which would otherwise hydrolyze the GALNS-P and ARSB-P.¹² The 8-plex cocktail also contains GNS-S (1 mM), GNS-IS (1 μM) and 4-methylumbelliferyl-N-acetyl-α-glucosaminide (1 mM, Carbosyth EM06426). The cocktail should be prepared fresh before use; freeze-thawing the assay cocktail results in reduced sulfatases enzymatic activities.

9-plex assay on DBS.

DBS samples were punched in duplicate into identical wells in two 96-deep well assay plates. To each well of one plate was added 30 μl of the duplex assay cocktail and to each well of the other plate was added 30 μl of the 7-plex assay cocktail. The two plates were sealed with a silicone sealing matt and shaken for 16 h at 37 °C in an orbital shaker incubator (250 rpm). HEPES buffer (0.2 M, pH 8.0, 30 μl) was added to the 7-plex assays, followed by the addition of 60 μl of the acylating reagent (Figure 1) solution in acetonitrile (3 mM). The plate was covered with a silicone sealing matt and shaken at 37 °C (250 rpm) for 1 hr. NH₄OH (5%, 20 μl) was added, and the plate was shaken at 37 °C (250 rpm) for 30 min. Phosphate citrate buffer (0.4 M, pH 5.0, 200 μl) was added and after mixing up and down a few times with the pipette, and all of the mixture was transferred to the duplex assay wells. Ethyl acetate (400 μl, J. T. Baker, Cat No. 9828–03) was added, and the well contents were mixed by pipetting up and down ~10-times. The plate was centrifuged at 3000 g for 5 min, 150 μl of the upper layer was transferred to a well in a shallow 96-well plate, solvent was removed with a jet of nitrogen, and the residue was reconstituted in 150 μl of water:acetonitrile (7:3) containing 0.1% formic acid. The plate was covered with aluminum foil and placed in the autosampler chamber at 8 °C in preparation for LC-MS/MS analysis.

Figure 1.

Structures of the substrates, products and the enzymatic reactions in the designed MPS multiplex assay.

LC was carried out using a Waters Acquity binary solvent pump system with a CSH, Cl8, 1.7 μm, 2.1 × 50 mm column (Waters, 186005296) and a CSH, C18 VanGuard Pre-column (Waters, 186005296). The column temperature was set at 22°C. The solvent program was 70% A (water, 0.1% formic acid)/30% B (acetonitrile, 0.1% formic acid) to 56% B (T = 1.40 min) to 99% B (T = 1.41 min - 1.70 min) to 30% B (T = 1.75 min – 1.90 min) with flow rate of 0.4 mL/min. All LC solvents were Optima grade from Fisher Scientific. To minimize contamination of the ESI source, the LC eluent was diverted to the mass spectrometer only in the time region that analytes elute (0.80-1.40 min). MS/MS was carried out with a Waters Xevo TQ tandem-quadrupole instrument. MS/MS instrument settings are given in Supporting Information.

10-plex Assay on fibroblast lysates.

Fibroblast lysate (10 μl, 0.6 μg protein/μl in 0.9% saline) was added in duplicate into identical wells in two assay plates. To each well of one 96-deep well plate was added 30 μl of the duplex assay cocktail, and to each well of the other plate was added 30 μl of the 8-plex assay cocktail. The two plates were sealed with a silicone sealing matt and shaken for 16 h at 37 °C in an orbital shaker (250 rpm) incubator. The post-assay workup and LC-MS/MS method were the same as those for the 9-plex assay on DBS.

RESULTS AND DISCUSSION

Design of the MPS multiplex assay.

Figure 1 shows the substrates and enzymatic products for the 10 MPS enzymes reported in this study. Each substrate contains a portion that is structurally identical to the natural carbohydrate substrate for each enzyme, which is attached to an unnatural aglycone. The use of the natural substrate moiety helps ensure that each substrate is specific for the enzyme of interest (the test of specificity is ultimately the measurement of low enzymatic activity in DBS from patients deficient in the enzyme). This is not the case for previous substrates such as catechol and other aromatic sulfates that have been used for assay of sulfatases such as ARSB.¹⁸ The aglycones have been designed with the following properties: 1) Each product has a unique MS/MS signature to avoid isobaric interferences; 2) Each product ionizes efficiently in positive mode, electrospray mass spectrometry based on the ability of the products to readily bind a proton to form gas phase ions;¹⁹ 3) Each product extracts well from the aqueous buffer layer into ethyl acetate, which allows for simple pre-MS/MS removal of much of the substrate and thus less interference from substrate-to-product conversion in the electrospray source (see below); 4) Each enzymatic product ion undergoes a major fragmentation pathway leading to enhanced signal-to-noise (as opposed to distribution of the signal across multiple product ions); 5) Ease of reagent synthesis; 6) Ease of incorporation of deuterium for internal standard preparation.

Most of the enzyme substrates contain a structurally similar aglycone with a 4-amidophenoxy glucoside attached to each sugar (Figure 1). HGSNAT and GNS substrates have a coumarin-type aglycone since earlier studies showed this aglycone provides higher enzymatic activity.²⁰ We added a hydrophobic tail to the coumarin group to enhance binding to the LC column so that products and internal standards elute after the void volume and in the time window with the other analytes.

After incubation of a 3 mm punch of a DBS in suitable buffer containing a mixture of substrates and internal standards, reactions are subjected to liquid-liquid extraction with ethyl acetate. The upper organic solvent layer is taken, solvent is removed with a jet of air, and the residue is dissolved in solvent and injected onto the LC-MS/MS instrument. LC is used instead of flow-injection to introduce the sample into the mass spectrometer for the following reasons: 1) Some substrates undergo non-enzymatic conversion to products in the heated electrospray source of the mass spectrometer (for example desulfation of sulfatase substrates). LC leads to full separation of substrate and product, and only the MS/MS signal at the LC retention time of the product is used to calculate the enzymatic activity; 2) Impurities in the sample including DBS components and the breakdown product of the acylation reagent used to assay MPS-IIIA (see below) elutes near the void volume of the LC program and are diverted away from the electrospray source of the mass spectrometer with a switching valve. This in turn has the advantage that the electrospray ionization source requires less routine cleaning; 3) LC allows detection of the activity of enzymes that have the lowest level of activity (for example SGSH for analysis of MPS-IIIA). Flow-injection-MS/MS does not provide suitable signal-to-noise to detect the SGSH product.

Internal standards used in this study are structurally identical to the enzymatic products but contain deuterium in place of hydrogen so that they can be selectively detected by MS/MS. Use of this type of internal standard is important to ensure that the LC retention times of internal standards and products are identical; thus, internal standard and product undergo the exact same degree of ionization suppression in the electrospray source. Such suppression is caused by non-analyte components of the sample that co-elute with the analytes of interest, which is inevitable with complex biological samples including DBS. The internal standard also accounts for all loss of enzymatic product due to sample handing.

Two DBS punches (3 mm) are needed for a 9-plex MPS assay (all MPSs except MPS-IIID, see below). One punch is used in a single buffer to assay all enzymes except HGSNAT and NAGLU (7-plex). This buffer contains Ce(OAc)₃ which precipitates free phosphate and sulfate coming from the blood and would otherwise lead to significant product inhibition of the sulfatases (IDS, SGSH, GALNS, ARSB). Addition of acetyl-CoA, one of the substrates of HGSNAT, and HGSNAT-S to buffers containing Ce³⁺ leads to a precipitate; thus, a separate incubation is needed for MPS-IIIC. NAGLU is very active in DBS and consumes a significant fraction of the HGSNAT product and internal standard. We added NAGLU-S to the incubation for HGSNAT since this substrate blocks the action of NAGLU on the HGSNAT product and internal standard. After incubation, the two separate samples (7-plex and 2-plex) are mixed together and processed as a single sample, which is submitted to LC-MS/MS.

In order to include SGSH in the MPS multiplex assay, it was necessary to modify our previously reported assay,⁸ since SGSH-product elutes very early from the LC column, and including it would extend the length of time per sample to a level unsuitable for high throughput newborn screening. In the new procedure, an NHS-ester acylating reagent is added after the incubation, leading to acylation of the SGSH-liberated amino group on the glucosyl substrate (Figure 1). The acylation mixture is quenched with ammonium hydroxide solution to convert the remaining NHS ester to the amide. The amide, along with some free carboxylic acid elute near the void volume of the LC run and are shunted away from the mass spectrometer with a diversion valve (Figure 2).

Figure 2.

LC-MS/MS chromatogram for the 10 MPS products with the multiplex assay. Note that the side products of SGSH-P acylation elute early in the LC and can be diverted to the waste rather than to the ESI source. The S-to-P peak at ~0.82 min in the GLB1 channel is due to in-source breakdown of substrate. The small peak at ~1.25 min in the HGSNAT-P channel is due to a low abundance ion fragment from GLB1-P that is isobaric with HGSNAT-P (of no concern because of LC separation).

HGSNAT Assay Parameters.

For HGSNAT, a pH-rate profile showed pH ~ 5.0 as optimal (Figure S1, Supporting Information). The assay response is linear with incubation time of 0-24 hr (Figure S2, Supporting Information). Substrate concentration studies yielded a K_M of 48 μM (Figure S3, Supporting Information). Stability studies with DBS stored at room temperature, 4 °C and −20 °C showed no significant loss in HGNSAT activity over the first 30 days and ~ 25% loss of the activity over 105 days (Figure S4, Supporting Information). Thus, HGSNAT is adequately stable in DBS for newborn screening.

MPS 9-Plex Assay.

The final MPS 9-plex makes use of 2 DBS punches (3 mm each) and two incubations as noted above. Figure 2 shows the output of the single LC-MS/MS run (it also includes MPS-IIID, which is discussed below). It can be seen that the breakdown products of the acylation reagent used for SGSH elutes prior to the analytes of interest. A small product peak from breakdown of the GLB1-S in the electrospray source is seen at ~0.82 min. Similar S-to-P peaks occur for some of the other substrates but are not shown in Figure 2 because they elute outside of the 0.80-1.40 min time window where all products are observed and are diverted to waste with a switching valve. Thus, it is clear that use of LC eliminates any issues with in-source S-to-P breakdown. Using DBS from a single healthy adult, the 9-plex assay was repeated 8 times. The average enzymatic activities and the inter-assay coefficient of variations (CV) are provided in Table 1. The linearity of the response as a function of analytes quantity was determined by measuring the signal intensities for increasing amount of the internal standards in the 9-plex assay. The results suggest excellent response linearity for all analytes (R² > 0.993, Figure S5, Supporting Information). If substrates are omitted from the assay, the MRM responses for the products are undetectable showing that the sample does not contain enzymatic products or other compounds that have the same retention time, parent mass, and fragment mass.

Table 1.

Average enzymatic activity (μM/h) and inter-assay CV for the 9-plex assay with 8 different punches from a single adult DBS.

	IDUA	IDS	SGSH	NAGLU	HGSNAT	GALNS	GLB1	ARSB	GUSB
Activity	1.2	5.1	0.22	4.9	1.0	0.90	5.0	0.86	36.0
CV%	2.9	11.9	12.1	8.9	7.6	9.2	5.8	7.4	3.3

We applied the 9-plex assay method to DBS from 66 random newborns (presumed healthy), one DBS from an MPS-IIIC patient, one DBS from an MPS-IIIA patient, and 7 DBS from GM1-gangliosidosis patients (no MPS-IVB patient DBS were available). Results are summarized in Figure 3, and numerical values are given in Tables S1 and S2 (Supporting Information). The MPS-IIIC and MPS-IIIA patients showed ~3% residual HGSNAT activity and ~2% residual SGSH activity, respectively (percentage based on mean of 66 random newborn values). The GM1-gangliosidosis patients showed 7-23% residual GLB1 activity. For each patient, activities of the other 8 enzymes were in the normal range (Figure 3). We did not measure activities for the other types of MPS disorders since patient DBS are in short supply, and we established close to zero activities against these substrates in previous reports.^12,15,16

Figure 3.

MPS enzyme activities in DBS from 66 random newborns and patients, measured with 9-plex LC-MS/MS method.

Addition of GNS to give the 10-plex assay.

We have not been able to detect GNS in DBS despite multiple attempts and use of multiple substrates. GNS-S is shown in Figure 1. GNS is active on this substrate and was shown to be stable in DBS since addition of recombinant GNS to blood and preparation of DBS led to product formation that was readily detected by LC-MS/MS (not shown). We also studied the previous reported MS/MS substrate for GNS,²¹ but failed to detect activity in DBS that was not spiked with recombinant GNS. It appears that there is insufficient GNS activity in DBS to be detected with these substrates. We then turned to skin fibroblasts from a healthy patient and from an MPS-IIID patient. GNS activity was readily detected in the normal fibroblast lysate but was greatly reduced in the patient fibroblasts (10%). GNS-P increased linearly with incubation time (0-24 h) and also with increasing amount of fibroblast protein (Figures S5 and S6, Supporting Information). Variation of the concentration of GNS-S gave K_M = 1.9 mM (Figure S7, Supporting Information). Based on these results we constructed a 10-plex by adding GNS-S to the 7-plex described above. To block the action of NAGLU on GNS-P and internal standard we added a commercially available fluorimetric NAGLU substrate, 4-methylumbelliferyl-N-acetyl-α-glucosaminide. A second incubation consisted of HGSNAT-S and NAGLU-S, with the latter added to block the action of NAGLU on HGSNAT-P and HGSNAT internal standard. The two incubations were combined for a single LC-MS/MS run (10-plex), and the results are summarized in Table 2.

Table 2.

Enzyme activities (nmol/h/mg of protein) in fibroblast lysates with the 10-plex assay. The results are the average of duplicate measurements and corrected for the blanks.

	Healthy Adult	MPS-IIID patient
IDUA	9.37	16.74
I2S	7.95	8.83
SGSH	6.01	7.83
NAGLU	3.17	6.68
HGSNAT	0.94	1.44
GNS	0.150	0.015
GALNS	39.33	50.16
GLB1	17.80	37.39
ARSB	13.41	8.74
GUSB	32.29	34.07

CONCLUSION

The purpose of this study was to develop a multiplex panel assay for the full set of MPS disorders. We included all MPS types except MPS-IX, the rarest MPS type, with only 4 reported cases worldwide.²² The new assay makes use of 2 DBS punches (3 mm each) and two enzymatic assay cocktails. After incubation, samples are combined for a single and rapid LC-MS/MS readout of all enzymatic activities. This DBS assay does not include GNS for MPS-IIID as we were not able to detect enzymatic activity in DBS. However, GNS was readily detected in fibroblast lysates. MPS-IIID is a very rare subtype of MPS-III accounting for only about 6% of the MPS-III patients.²³ All reagents for the multiplex MPS assay are commercially available and thus appropriate for transfer to biochemical genetics laboratories.

Fluorimetric assays have been reported for only a subset of the MPS enzymes in DBS. There are no reports of DBS fluorimetric assays for MPS-IIIA, -IIIC, and -IIID. The fluorimetric substrate for MPS-IIIA, 4-methylumbelliferone α-glycoside, has been reported to work using leukocytes or fibroblasts,²⁴ but we found that it does not give a detectable assay response when DBS is used.⁸

The advantage of the LC-MS/MS assay reported here is that a single assay platform can be used for all the MPS subtypes (except MPS-IX). It has also been possible to multiplex the LC-MS/MS MPS assays with assays for other inborn errors including several additional lysosomal storage diseases, X-linked adrenoleukodystrophy, galactosemia, and biotinidase deficiency.²⁵ This has significance for newborn screening where consolidation of assays into a single platform is expected to save time and money. This is becoming increasingly important as newborn screening panels are expanded to include a constant increase in the availability of new treatments including gene therapy.

Abbreviations:

MPS

mucopolysaccharidosis

DBS

dried blood spots

IDUA

α-L-iduronidase

IDS

iduronato-2-sulfatase

NAGLU

α-N-acetylglucosaminidase

GALNS

N-acetylgalactosamine-6-sulfatase

GLB1

β-galactosidase

ARSB

arylsulfatase B

GUSB

β-glucuronidase

SGSH

N-sulfoglucosamine sulfohydrolase

HGSNAT

heparan-α-glucosaminide N-acetyltransferase

GNS

N-acetylglucosamine-6-sulfatase

NAG-thiazoline

N-acetylglucosamine thiazoline