Pilot study of newborn screening for six lysosomal storage diseases using Tandem Mass Spectrometry

Background There is current expansion of newborn screening (NBS) programs to include lysosomal storage disorders because of the availability of treatments that produce an optimal clinical outcome when started early in life. Objective To evaluate the performance of a multiplex-tandem mass spectrometry (MS/MS) enzymatic activity assay of 6 lysosomal enzymes in a NBS laboratory for the identification of newborns at risk for developing Pompe, Mucopolysaccharidosis-I (MPS-I), Fabry, Gaucher, Niemann Pick-A/B, and Krabbe diseases. Methods and Results Enzyme activities (acid α-glucosidase (GAA), galactocerebrosidase (GALC), glucocerebrosidase (GBA), α-galactosidase A (GLA), α-iduronidase (IDUA) and sphingomyeline phosphodiesterase-1 (SMPD-1)) were measured on ~43,000 de-identified dried blood spot (DBS) punches, and screen positive samples were submitted for DNA sequencing to obtain genotype confirmation of disease risk. The 6-plex assay was efficiently performed in the Washington state NBS laboratory by a single laboratory technician at the bench using a single MS/MS instrument. The number of screen positive samples per 100,000 newborns were as follows: GAA (4.5), IDUA (13.6), GLA (18.2), SMPD1 (11.4), GBA (6.8), and GALC (25.0). Discussion A 6-plex MS/MS assay for 6 lysosomal enzymes can be successfully performed in a NBS laboratory. The analytical ranges (enzyme-dependent assay response for the quality control HIGH sample divided by that for all enzyme-independent processes) for the 6-enzymes with the MS/MS is 5- to 15-fold higher than comparable fluorimetric assays using 4-methylumbelliferyl substrates. The rate of screen positive detection is consistently lower for the MS/MS assay compared to the fluorimetric assay using a digital microfluidics platform.

Blank factor 1 above was measured by analyzing substrate and internal standard by liquid chromatography-MS/MS under conditions in which substrate and product/internal standard are baseline separated during chromatography. Blank factor 2 was measured by running the incubated assay with a filter paper punch (no blood), and then measuring the amount of product by liquid chromatography-MS/MS. Blank factor 3 was measured by running the incubated assay with a DBS punch but in the absence of substrate and analyzing for product by liquid chromatography-MS/MS. Blank factor 4 was obtained by flow injection-MS/MS on a substrate/internal standard mixture. Note that blank factors 1-3 were obtained from liquid-chromatography-MS/MS rather than flow injection-MS/MS. Since substrate and product are baseline separated during liquid chromatography, the product MRM signal is devoid of any signal coming from in-source breakdown of substrate. These studies show that by far the major contribution to the enzyme-independent blank is blank factor 4 (blank factors 1-3 account for < 5% of blank factor 4). This study shows that the analytical range is properly calculated from the enzyme activity measured for the incubated assay with DBS and substrate and the enzyme activity measured for the incubated assay with filter paper only (no blood) and substrate. This method of calculating the analytical range has been previously reported [1] .
Fluorimetry. To calculate the analytical range for fluorimetric assays with 4-methylumbelliferyl substrates, a different method is needed. Components of blood (hemoglobin, etc.) substantially quench the fluorescence and have to be taken in to account. Both the blood sample and the substrate may display intrinsic fluorescence. Finally, the substrate may contain a trace of product as an impurity and also suffer non-enzymatic breakdown to product in the absence of enzyme.
As described previously [1] , we prepared two samples to measure the blank. Tube A contains DBS punch and a volume of assay buffer equal to half the volume in the complete assay. Tube B contains substrate at twice the concentration as in the complete assay and in a volume of buffer equal to that in the complete assay. Both tubes are incubated as for the complete assay. After incubation, half the volume of tube B is transferred to tube A, and the mixture is immediately quenched and processed for fluorimetry as for the complete assay. In this way the blank will display fluorescence from the following processes: 1) from the substrate itself, 2) from the buffer, 3) from the blood; 4) from product present in the substrate as an impurity; 5) from product generated non-enzymatically from substrate. Also quenching of the fluorescence by blood will occur to the same extent as for the complete assay. Studies show that the predominant factor contributing to the blank fluorescence is the intrinsic fluorescence of the substrate itself. The deprotonated form of the 4-methylumbelliferone product is ~1000-fold more fluorescent per mole than the substrate, but since only ~1% of the substrate is converted to product by the small amount of enzyme in the DBS, the increase in fluorescence is on the order of 10-fold [1] .
To estimate the analytical range for digital microfluidics fluorimetry (the values have not been reported), we note that the intrinsic fluorescence of the 4-methylumbelliferyl substrate occurs for all fluorimetric methods where the fluorescence of the product, 4-methylumbelliferone (or a derivative) is measured. Quenching by blood components is shown to be less for digital microfluidics presumably because the optical pathlength in the microdroplets is much smaller than in a 96-well plate reader [2] . But since the analytical range is limited by the intrinsic fluorescence of the substrate, the blood quenching does not significantly contribute to the analytical range. The incubation times for the digital microfluidics assays for Pompe, Fabry, and MPS-I are in the 1-3 hr range [2] compared to the overnight incubations done for the MS/MS and the plate reader fluorimetric assays. Thus the analytical ranges for the digital microfluidics assays are expected to be less than the values in the main text.

Buffer Optimization.
Buffer conditions were optimized by a 5-factor, 3-level design experiment to optimize pH and buffer concentration, as well as the concentration of sodium taurocholate, sodium oleate, and zinc chloride. Optimization of the inhibitor concentrations for acarbose, N-acetylglucosamine and D-saccharic acid 1,4lactone was performed last and included confirmed inhibitor-specific disease-positive patients along with presumed healthy neonatal samples. Three different concentrations were screened for each inhibitor using a one-factor-at-a-time approach, and the optimal amount was determined by maximizing the resolution between the healthy neonatal enzyme activity and that of the confirmed disease-positive sample.