Article Text
PDF
Abstract
Background Lysosomal protein profiling is being developed as a high throughput method to screen populations for lysosomal storage disorders (LSD).
Design 1415 blood spots from patients referred to a metabolic clinic for LSD were screened using a single multiplex assay for 14 proteins in a dried blood spot.
Results All patients with Pompe disease, metachromatic leukodystrophy, and mucopolysaccharidosis (MPS) type I, IIIA, IIIB and VI were identified by reduced lysosomal protein. Five samples were identified as possible pseudo-arylsulfatase A deficiency; four were confirmed. One multiple sulfatase deficiency patient was identified with multiple reduced sulfatase proteins. There were 10 MPS II patients identified with reduced iduronate 2-sulfatase, and one MPS II patient with iduronate 2-sulfatase in the unaffected range. For Fabry disease, 10 male patients were identified with reduced α-galactosidase and 2/6 female Fabry heterozygotes returned α-galactosidase concentrations in the male Fabry range. All 10 mucolipidosis II/III patients were identified with multiple raised proteins. For 79 blood spots with chitotriosidase >3.4 mg/l, a follow-up one-plex chitotriosidase assay enabled identification of all nine Gaucher patients.
Conclusion This study demonstrates the sensitivity and specificity of this technology to accurately identify 99% of LSD patients, with the exception of one MPS II false negative.
- Lysosomal storage disorder
- multiplex immune quantification, lysosomal protein
- screening
- diagnostics
- metabolic disorders
Statistics from Altmetric.com
- Lysosomal storage disorder
- multiplex immune quantification, lysosomal protein
- screening
- diagnostics
- metabolic disorders
Introduction
Lysosomal storage disorders (LSD) are progressive inborn errors of metabolism that manifest in severe physical and/or mental disabilities. Onset of pathology is highly variable and patients are most often not diagnosed until symptoms are apparent. Many patients present clinically with similar or overlapping clinical manifestations, including cardiac and respiratory insufficiency, neurological impairment, bone abnormalities, facial dysmorphism, and sight and hearing impairment. Despite these shared attributes, each disorder (up to 50) emanates from mutations in a different gene leading to a deficiency of a specific lysosomal enzyme, cofactor or transport protein.
The combined incidence of LSD is reported at 1 in 7700 births,1 but may be as high as 1 in 1000–2000 given the prevalence of individual LSD in some populations. For the LSD Fabry disease, Spada et al2 measured α-galactosidase activity in blood spots collected from 37 104 Italian male neonates and found that 12 neonates had deficient enzyme activity and Fabry mutations; one patient had a mutation consistent with classical early onset Fabry disease; 11 neonates had mutations in the α-galactosidase gene, some consistent with later onset disease and others novel. Investigation of extended family members of those neonates with novel mutations identified individuals who died or suffered from stroke but were not previously identified as having Fabry disease. This study suggests an incidence of 1 in 3100 for Fabry disease, with a majority (11/12) predicted as later onset. In the Taiwan Chinese population, 110 027 newborns were screened for Fabry disease by assaying α-galactosidase activity in blood spots. Low activity and eight different Fabry disease mutations were demonstrated in 45 newborns.3 The majority of patients (82%) had a known intronic mutation which upon maternal follow-up was identified in 11 grandmothers and nine grandfathers; three of the latter had hypertrophic cardiomyopathy. Furthermore, 25% of males suffering from hypertrophic cardiomyopathy were identified with the intronic mutation and low α-galactosidase activity. As well as finding a high prevalence of the cardiac variant of Fabry disease in males with idiopathic hypertrophic cardiomyopathy, the newborns screened suggest an incidence of 1 in 1600 males for Fabry disease.
These findings, of a higher incidence of LSD in the population than originally expected, support the likelihood of there being undiagnosed later onset LSD patients in the population. Coupled with clinical and biochemical reports demonstrating that the success of current and future therapies correlates with early commencement4 provides justification for screening programmes for LSD. Newborn screening measuring a specific lysosomal enzyme activity for Pompe disease is performed in Taiwan,5 and for Krabbe disease in the state of New York,6 and other newborn screening programmes for multiple LSD are being piloted.7 8 These use a series of substrates for determining enzyme activity rather than protein, with the products of the different enzyme assays measured by mass spectrometry in a multiplex assay.9 However, enzyme activities may reduce during shipment and storage whereas protein measurements will be far more robust as has been reported for arylsulfatase A.10 Therefore we have been developing newborn screening technology based on protein profiling11; using 14 lysosomal and related proteins in a multiplexed assay, we report here the identification of patients with LSD by screening patients referred from clinical settings to a diagnostic metabolic laboratory.
Methods
Materials and patient samples
Dried blood spots collected from 1415 patients referred to the National Referral Laboratory for the diagnosis of lysosomal, peroxisomal, and related genetic disorders were used in this study. Confirmation of LSD affected blood spots were confirmed by this laboratory using classic enzyme activity measurement and in some cases also mutation analysis. EDTA blood was spotted onto Guthrie cards (<48 h following collection), air dried and stored at −20°C sealed with dessicant. Blood spots were collected from 2003 to 2007 and analyses were performed between October 2006 and July 2007. The Human Research Ethics Committee of the Women's and Children's Hospital, North Adelaide, Australia, approved the use of the blood spots for this purpose. Chitotriosidase protein was a gift from JM Aerts (Amsterdam, The Netherlands), CD45 protein was sourced commercially from BIOMOL International LP (Plymouth Meeting, Pennsylvania, USA) and α-N-acetylglucosaminidase was expressed and purified as described previously.12 Polyclonal antibodies were produced in sheep following immunisation with purified chitotriosidase and α-N-acetylglucosaminidase, and two short peptides (residues 1229–1259 and 1280–1304) were used in place of recombinant CD45. Affinity purified antibodies were prepared from sheep sera as described previously.13 The antibodies were conjugated to carboxyl beads (Luminex Corporation, Austin, Texas, USA) and the antibodies were biotinylated with FluoReporter Biotin-XX labelling kit (Molecular Probes Inc, Eugene, Oregon, USA), according to the manufacturer's instructions.
Multiplex immune-quantification of 14 lysosomal and related proteins (14-plex)
Immune-quantification of 14 proteins was performed using microbead suspension array technology (Bio-Rad, Hercules, California, USA), essentially as described previously for 11 lysosomal proteins11 with the addition of α-N-acetylglucosaminidase, CD45, and chitotriosidase. The concentration of α-N-acetylglucosaminidase and CD45 antibodies were at 36 μg/2.5×106 carboxyl beads and chitotriosidase was at 9 μg/2.5×106 carboxyl beads. Reporter antibody concentrations for α-N-acetylglucosaminidase, chitotriosidase, and CD45 were 32 μg/l. The chitotriosidase reporter antibody was added in a separate 1 h incubation following an overnight incubation with the other 13 reporter antibodies. Concentrations of the individual proteins were determined from a five-parameter logistic regression calibration curve for each of the 14 proteins.
One-plex immune-quantification of chitotriosidase
Blood spots with chitotriosidase protein >3.4 mg/l were reanalysed in a single plex immune-quantification assay. Serially diluted blood spot eluates (100 μl) were incubated with 50 μl of anti-chitotriosidase coupled beads (5000/eluate) at 4°C overnight on a platform shaker. The beads were washed, resuspended, and incubated with reporter antibody (0.64 μg/l) for 1 h on a platform shaker at room temperature. The plates were washed, beads resuspended in assay buffer with 2 mg/l streptavidin-phycoerythrin, incubated, and fluorescence intensity of the beads was determined as described previously.11 The calibration curve consisted of 11 standards covering the concentration range 0.0061–6.25 μg/l.
Results and discussion
The calibration curves for CD45, chitotriosidase, and α-N-acetylglucosaminidase were linear over the biological range determined for the other 11 proteins.11 Intra-assay (n=12) and inter-assay (23 measurements over 11 days) coefficient of variation (CV) were <20% for all proteins. Table 1 shows the medians and ranges of 11 of the 14 proteins (LAMP-1, saposin C, and CD45 were omitted) in the 1415 blood spots which shows that, with the exception of Gaucher disease, the majority of LSD could be demarcated from the others by a reduction in the specific lysosomal protein. All 17 Pompe, 10 male Fabry, 11 metachromatic leukodystrophy (MLD), five mucopolysaccharidosis (MPS) I, three MPS IIIA, one MPS IIIB, and four MPS VI patients were identified by reduced α-glucosidase, α-galactosidase, arylsulfatase A, α-iduronidase, N-sulfamidase, α-N-acetylglucosaminidase and N-acetylgalactosamine 4-sulfatase, respectively. Of the 11 MPS II blood spots, 10 had iduronate 2-sulfatase values in the MPS II affected range (<3 μg/l), and in one patient iduronate 2-sulfatase was 17 μg/l, clearly in the unaffected range. The capacity to identify this MPS II patient was attempted with multivariate analysis of all 14 proteins but was unsuccessful (data not shown). This patient had deficient protein present within the normal range but in an inactive form and would therefore not be identified measuring protein alone. Potentially, this patient could have been identified by the direct measurement of iduronate 2-sulfatase activity.14
Lysosomal protein concentrations in blood spots
In addition to the 10 male Fabry blood spots there were six female Fabry blood spots that had been identified based on mutational analysis, and two of these returned α-galactosidase concentrations in the affected Fabry range of <9 μg/l. The other four were in the non-LSD range; two of these were below the mean with values of 14 and 20 μg/l, and the other two were greater than the mean at 65 and 95 μg/l. These individuals should be followed clinically for the development and extent of Fabry disease symptoms and whether the amount of α-galactosidase correlates with their clinical picture.
For arylsulfatase A measurements there were five blood spots with concentrations that fell between the non-LSD spots and the MLD patients at concentrations between 4–7 μg/l. Four of these five blood spots, all with values below the control range, were confirmed pseudo-arylsulfatase A deficiency by mutational analysis. Figure 1 shows the 10 mucolipidosis II/III blood spots could be identified by a number of raised proteins, including N-sulfamidase, acid sphingomyelinase, arylsulfatase A, and iduronate 2-sulfatase. Similarly, the single multiple sulfatase deficient blood spot was noted for the combination of reduced N-sulfamidase, arylsulfatase A, and N-acetylgalactosamine 4-sulfatase values.
High-low plots of lysosomal proteins in mucolipidosis II/III and multiple sulfatase deficiency compared with non-lysosomal storage disorders (LSD) blood spots. The concentration of N-sulfamidase (N-sulph), acid sphingomyelinase (ASM), α-iduronidase (IdUA), LAMP-1, α-glucosidase (GAA), β-glucosidase (β-gluc), saposin C (sap C), α-galactosidase (α-gal), arylsulfatase A (ASA), iduronate 2-sulfatase (I2S), N-acetylgalactosamine 4-sulfatase (4S), chitotriosidase (chito), CD45, and α-N-acetylglucosaminidase (NAG) in the 1323 non-LSD (shaded bars), 10 mucolipidosis II/III (open bars), and one multiple sulfatase deficiency (solid triangle) blood spots is shown. The median value of each protein is indicated by the circle and the highest and lowest values demarcated by the bars.
There were 79 blood spots with chitotriosidase concentrations >3.4 mg/l in the initial 14-plex assay. These concentrations were higher than the top standard and were therefore repeated upon dilution and reanalysed for chitotriosidase in a single plex assay. Of these 79 samples, nine could be identified as Gaucher disease based on chitotriosidase concentrations in the range 0.6–9.2 mg/l. There were also five patients with high chitotriosidase concentrations in the Gaucher range of 0.6–9.2 mg/l; one male Fabry, one Pompe, one MLD, and two pseudo-MLD which had been identified with reduced lysosomal protein in the 14-plex. The 65 remaining non-LSD blood spots were in the range 0–0.6 mg/l. It must be noted that there is a significant limitation to the use of chitotriosidase as a marker for Gaucher disease, as 40% of Caucasians are heterozygous or homozygous for a common null mutation.15 Chitotriosidase is a measure of macrophage inflammation and could be replaced in future studies by the recently reported chemokine CCL18.16
Previously, we have proposed that the concentration of the lysosomal proteins LAMP-1 and saposin C would increase as a result of lysosomal storage and that they may be useful biomarkers for the detection of many LSD.17 18 These data were generated from plasma, whereas in newborn dried blood spots there is no difference between LSD and the control group for either LAMP-1 or saposin C.19 This suggests that LAMP-1 and saposin C are actually derived from plasma rather than white blood cells. Neither proteins were elevated in any of the LSD groups compared to the non-LSD blood spots in the blood spots analysed here (data not shown) and of note were the broad control ranges for both LAMP-1 and saposin C—postulated to be due to differences in the amount of white blood cells in each blood spot. Therefore, we attempted to correct for the number of white blood cells in each of the blood spots by the inclusion of the marker CD45. However, normalisation with CD45 did not improve discrimination of the LSD affected from the non-LSD blood spots. Discriminate analysis with up to 14 of the protein markers provided no further separation of the non-LSD and affected groups for these disorders (data not shown).
This study has demonstrated the use of deficient proteins as markers for LSD to identify 99% of patients, with the exception of one MPS II patient. Our multiplex assay applying immune quantification using microbead suspension array technology has the potential to detect numerous LSD in a single analysis. This presents a considerable advantage to the diagnostic laboratory in not having to maintain multiple enzyme assays for each individual LSD. As with all screening assays, initial positive results would require follow-up conformational analysis to exclude false positives; this may include molecular genetic analysis like that to confirm Krabbe disease following newborn screening for this disorder in New York.6 As the identification of new mutations can make interpretation of such analysis difficult, confirmation of positive results could also be based on individual enzyme activity assays using either direct measure20–22 or immunoquantification.23 24
Although we have described this multiplex technology to improve the efficiency of LSD diagnosis, it also has potential application for newborn screening. The low prevalence of LSD makes it unlikely that screening programmes for individual LSD will be widely adopted, and performing multiple assays to cover a range of disorders will not be cost effective. The cost and turnaround time of our multiplex assay is comparable with current newborn screening tests. The ongoing development of enzyme replacement therapy and other treatments for an increasing number of LSD, combined with the growing evidence that early commencement of therapy improves outcomes, has increased the pressure for the introduction of newborn screening programmes and a number of pilot studies are ongoing.2 3 5 6 Criteria for the inclusion of metabolic diseases into newborn screening programmes are also being re-evaluated in the light of new clinical evidence changing perceptions of good versus harm.25 Benefits resulting from earlier diagnosis and the resultant genetic counselling to the family, even in the absence of available therapies, are being considered. However, we still have much to learn about phenotype prediction, particularly for the attenuated forms of LSD, and so the continued development of prognostic assays will be an important component towards the implementation of newborn screening. Ultimately, the decision of which LSD to include in a screening programme and when to introduce such programmes will be made by the relevant health authorities in each country, and the availability of flexible, adaptable screening technology that can select and incorporate multiple LSD in a single assay will be an important prerequisite.
Acknowledgments
The authors would like to than Janina Pacyna and Alison Whittle for technical assistance as well as Ed Wraith and Eugene Mengel for the provision of blood spots from affected patients. The authors also wish to acknowledge the NHMRC of Australia for funding.
References
Footnotes
Competing interests None.
Ethics approval This study was conducted with the approval of the Human Research Ethics Committee of the Women's and Children's Hospital, North Adelaide, Australia.
Provenance and peer review Not commissioned; externally peer reviewed.