Keywords

Introduction

Pompe disease (OMIM #232300) is caused by a deficiency of lysosomal acid α-glucosidase (GAA; EC 3.2.1.20), resulting in glycogen accumulation that has a destructive effect on muscle (Hirschhorn and Reuser 2001). Patients with a severe GAA deficiency present in infancy (infantile-onset Pompe disease; IOPD) with cardio- and skeletal myopathy and succumb to cardiorespiratory disease within 2 years (Kishnani et al. 2006). Patients with an attenuated phenotype (termed late-onset or later-onset Pompe disease; LOPD) usually have measurable residual enzyme activity and no cardiac involvement. They present anywhere from early childhood to adulthood with a myopathy that progresses to respiratory insufficiency if untreated (Hagemans et al. 2005).

Enzyme replacement therapy (ERT) with recombinant human alglucosidase alfa (rhGAA) (Myozyme®, Lumizyme®, Genzyme, Cambridge, MA) is available for IOPD and LOPD. Evidence suggests ERT has the most favorable clinical outcome when started early in the disease process (Kishnani et al. 2009; Chien et al. 2009). Recognition of the importance of early diagnosis and treatment of Pompe disease has led to the development of newborn screening (NBS) assays for this condition and a recommendation by the US Discretionary Advisory Committee for Heritable Disorders in Newborns and Children for its inclusion in the recommended uniform newborn screening panel. A NBS pilot program in Taiwan identified 7 infants with IOPD and 13 with LOPD out of 344,056 infants screened between 2005 and 2009 (Chien et al. 2008, 2009, 2011). Survival, particularly ventilator-free survival, was improved for patients with IOPD diagnosed by NBS compared with those diagnosed after onset of symptoms. A high false-positive rate was the result of a pseudodeficiency allele of the GAA gene, c.[1726A;2065A] which is common in the Chinese population, but has no known clinical effects (Kroos et al. 2008; Kumamoto et al. 2009; Labrousse et al. 2010).

The glucose tetrasaccharide, Glcα1-6Glcα1-4Glcα1-4Glc (Glc4), is a limit dextrin of glycogen (Kumlien et al. 1988) and was shown to correlate with glycogen content in quadriceps biopsies in patients with IOPD (Young et al. 2009). Glc4 can be useful in the diagnosis of Pompe disease and for monitoring the response to ERT (Young et al. 2009, 2012). We assessed the usefulness of urinary Glc4, measured as the total hexose tetrasaccharide (Hex4) fraction in urine, in the follow-up of infants with low GAA activity identified by the Taiwanese pilot NBS program.

Materials and Methods

Materials

Whatman grade 903 filter paper (VWR, Batavia, IL); Sep-Pak® Vac 100 mg C18 cartridges (Waters Corporation, Milford, MA); d3-creatinine (Cambridge Isotopes, Andover, MA); creatinine standards, sodium cyanoborohydride, butyl 4-aminobenzoate, and glacial acetic acid (Sigma-Aldrich, St. Louis, MO); Glc4 standard (Glycorex AB, Lund, Sweden); HPLC grade solvents (VWR, West Chester, PA); and [13C6]-labeled glucose tetrasaccharide internal standard which was synthesized as previously described (Young et al. 2003).

Patients

This study included infants identified by the Taiwan NBS program between 2005 and 2009 with IOPD (NBS-IOPD, n = 7) and LOPD (NBS-LOPD, n = 13). The diagnostic confirmation and clinical status have been reported for all patients except NBS9 (Chien et al. 2008, 2009, 2011), and subject designations (see Table Supplemental Digital Content 1) are consistent with these previous publications. Four of the thirteen patients in the NBS-LOPD cohort were started on treatment at or before the age of 3 years because of the severity of their clinical condition (NBS-LOPD early treated), whereas the remaining nine patients did not require treatment within the first 3 years of life (NBS-LOPD-Group 2) (Chien et al. 2011). An additional patient with LOPD (L14), identified at birth because of a positive family history, was also included.

Four patients with IOPD, who were not part of the newborn screening pilot study, were diagnosed in early infancy (<5 months age) after the onset of clinical symptoms (NBS-CLIN) and were included as a prospective comparison group for the NBS-IOPD patients (Chien et al. 2009).

Additionally, 58 infants identified by NBS who had a pseudodeficiency of GAA in DBS (Labrousse et al. 2010) were evaluated. These infants had the following combinations of the pseudodeficiency allele c.[1726A; 2065A] and known or putative disease-causing GAA mutations:

  1. 1.

    One mutation plus one pseudodeficiency allele (pseudodeficiency group 1, n = 23)

  2. 2.

    One mutation plus two pseudodeficiency alleles (pseudodeficiency group 2, n = 19)

  3. 3.

    Two pseudodeficiency alleles only (pseudodeficiency group 3, n = 16)

Baseline urine samples were collected within the first 6 weeks of life, except for NBS-L4 and NBS-L11 in the NBS-LOPD cohort on whom samples were collected at 9 and 6 months, respectively (Table Supplemental Digital Content 1). Longitudinal urine samples were collected at regular clinic visits and stored at −20°C.

This study was approved by the Institutional Review Boards of National Taiwan University Hospital and Duke University Health System. Informed consent was obtained from parents of all patients.

Control Samples

Age-specific reference ranges were determined using anonymized clinical samples (n = 472, median age: 1.5 years, min–max: 0.0–68 years) from the Duke Biochemical Genetics laboratory and urine specimens collected from anonymous volunteers (n = 143, median age: 33 years (min–max: 3–78 years)) under a Duke University Health System IRB-approved protocol.

Glc4 Determination in Dried Urine Samples Using Stable Isotope Dilution-ESI-MS/MS

Urine specimens were soaked onto filter paper strips, dried, and mailed to the Duke Biochemical Genetics Laboratory at ambient temperature. None of the patients were on ERT at the time of the baseline sample collections. 2 × 2 cm diameter disks cut from the dried urine spot on filter paper were extracted with 1 mL DI-H2O by shaking at room temperature for 1 h. 500 μL of the filter paper extract was dried under nitrogen at 40°C, reconstituted with 50 μL DI-H2O and mixed with 25 μL 0.1 mmol/L [13C6]-labeled Glc4 internal standard. The mixture was derivatized with butyl-p-aminobenzoic acid, and Glc4 was determined as the total hexose tetrasaccharide fraction (Hex4) in urine by ultra performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) on an Acquity UPLC-Quattro Micro tandem mass spectrometer system (Waters Corp, Milford, MA), as previously described by Young et al. (2003, 2009). Glc4 comprises ≥90 % of the total hexose tetrasaccharide fraction in most patient and control samples (unpublished observation). In keeping with previous publications on this biomarker, Hex4 measurements will be referred to as Glc4. Glc4 concentrations were normalized to creatinine determined in an aliquot of the same filter paper extract using stable isotope dilution-tandem mass spectrometry (Young et al. 2009). Comparison studies have shown the equivalency of this assay when applied to liquid urine specimens and extracts of urine specimens dried on filter paper (Young et al. 2003).

Statistical Analyses

Descriptive statistics and Mann–Whitney comparison of median Glc4 values for the study cohorts were calculated using GraphPad Prism 5.04 software (La Jolla, CA). The relationship of age to Glc4 in the control cohort was examined using univariable linear regression with STATA 11.0 (College Station, TX) software. p-values ≤0.05 were considered to be significant.

Results

Urinary Glc4 in Control Samples

A significant negative correlation with age was observed for urinary Glc4 values in controls under aged 3 years (−2.956 [95 % confidence interval; −3.680, −2.232], p < 0.001, n = 287). No correlation with age was observed for controls over 3 years of age (−0.006 [−0.012, 0.001], p = 0.09, n = 328). Controls younger than 3 years were divided into different age bins based on a visual inspection of the data and statistical analysis. Upper limits of the reference ranges were defined as the 95th centile and were stratified according to the following age groups: 0–6 months age (95th centile: 20 mmol/mol CN, n = 132), 6–12 months (95th centile: 14 mmol/mol CN, n = 78), 1–3 years (95th centile: 8.3 mmol/mol CN, n = 77), and > 3 years (95th centile: 3.0 mmol/mol CN, n = 328). No significant correlation with age was observed for controls younger than 6 months of age (−3.217 [−9.770, 3.335], p = 0.34, n = 132), and there were no significant differences between median Glc4 values for the following subgroups: 0–1 month, 1–3 months, and 3–6 months (Table 1).

Table 1 Median and ranges for baseline urinary Glc4 values and age at the time of sample collection for the patient cohorts (NBS-IOPD, NBS-LOPD-A, NBS-LOPD-B) and pseudodeficiency groups as described in the methods section
Full size table

Baseline Urinary Glc4 Concentrations

A comparison of the median values and ranges of urinary Glc4 concentrations and ages at baseline in the patient, pseudodeficiency, and control cohorts is shown in Table 1, and individual patient values are compared in Fig. 1 and Table Supplemental Digital Content 1. Urinary Glc4 concentrations were at or above the 90th centile (18 mmol/mol CN) of the age-matched reference range (Table Supplemental Digital Content 1) for the six full-term infants in the NBS-IOPD group. The remaining patient in this group, born at 29 weeks gestation and started on ERT at 40 days of age because of cardiomegaly, had pretreatment Glc4 concentrations of 32 and 16 mmol/mol CN at 0 and 1.2 months of age (not adjusted for prematurity), respectively. An appropriate age-matched control range has not been evaluated for pre-term infants. The NBS-IOPD group was significantly younger than the CLIN-IOPD group at the time of pretreatment assessment (p < 0.05) and had a lower median Glc4 concentration, although the difference was not statistically significant (p = 0.07) (Table 1).

Fig. 1
figure 1

Urinary Glc4 concentrations in infants with Pompe disease identified by newborn screening. Urinary Glc4 values at first evaluation (baseline) for individual patients with Pompe disease identified through newborn screening. Circles: Patients with infantile-onset Pompe disease (NBS-IOPD); Triangles: Patients with late-onset Pompe disease who were treated before the age of 3 years (NBS-LOPD early treated); Squares: Patients with late-onset Pompe disease who did not require treatment before 3 years of age (NBS-LOPD-Group 2). Dashed line represents the upper limit of the reference range for 0 to 6 months age (Glc4 < 20 mmol/mol CN)

Full size image

Baseline Glc4 concentrations were within reference limits for all patients in the NBS-LOPD group, and median values were significantly lower than those of the NBS-IOPD cohort (p < 0.05). There was no significant difference in the median baseline Glc4 values for the NBS-LOPD early treated and the NBS-LOPD-Group 2 cohorts. Patients NBS-L3 and NBS-L9 in the early treated group had the highest values close to the upper limit of the reference range. Urinary Glc4 levels were within reference limits for infants with a GAA pseudodeficiency (Table 1).

Comparison of Urinary Glc4 and Serum Creatine Kinase at Baseline

Urinary Glc4 was significantly correlated with serum creatine kinase (CK) at the initial follow-up evaluation for patients with a confirmed diagnosis of Pompe disease (Pearson correlation coefficient = 0.624, p < 0.05; see Figure Supplemental Digital Content 2). As expected, the NBS-IOPD group had the highest CK values (excluding the premature infant, NBS9). Two of four patients in the NBS-LOPD early treated group (NBS-L3 and NBS-L9) had Glc4 and CK values that were comparable with those for three patients in the NBS-IOPD group. Infants in the pseudodeficiency groups had CK values within the reference intervals (data not shown).

Pretreatment Monitoring of Glc4 in Patients with a Late-Onset Phenotype

Longitudinal Glc4 measurements for 11 of the 13 NBS-LOPD patients on whom data were available are shown in Figure Supplemental Digital Content 3. Within the NBS-LOPD early treated cohort, Glc4 was elevated in two of the four patients prior to treatment. NBS-L3 had a persistent elevation of Glc4 prior to treatment at age 36 months and NBS-L9 had elevated Glc4 immediately prior to treatment at 1.5 months. In contrast, NBS-L1 and NBS-L6 did not have elevated Glc4 prior to initiation of ERT. These trends in Glc4 are consistent with the CK trends observed for these patients as previously reported (Chien et al. 2011), in that NBS-L3 and NBS-L9 had elevated CK prior to treatment and NBS-L1 and L6 did not. Glc4 elevations were not observed for the NBS-LOPD group 2 during the observation period of up to 4 years of age.

Discussion

The newborn screening program for Pompe disease in Taiwan has presented a unique opportunity to evaluate Glc4 in infants with infantile- and late-onset Pompe disease, prior to the appearance of clinical symptoms. Our results indicate that Glc4 concentrations correlated with phenotype early in the disease process; patients with IOPD had higher Glc4 concentrations than those with LOPD. Furthermore, median baseline Glc4 in the NBS-IOPD group was clearly lower than that of the slightly older (by approximately 4–8 weeks) clinical comparator CLIN-IOPD group. These observations are consistent with (1) baseline clinical manifestations in the NBS-IOPD group including cardiomyopathy and elevated CK, despite a normal physical exam and tone (Chien et al. 2009), (2) the notable increase in baseline Glc4 values with age in untreated patients with IOPD ascertained clinically before 12 months of age (Young et al. 2012), and (3) the rapidly progressive nature of the infantile form of the disease.

The variability of late-onset Pompe disease is demonstrated by differences in the age of onset of clinical signs and symptoms of the disease within the NBS-LOPD group, of which one third of patients required treatment with ERT before 3 years of age (Chien et al. 2011). Although these early treated patients with LOPD had baseline urinary Glc4 concentrations within the reference range, two had the highest values (86th and 90th centile of the reference range) observed within the entire NBS-LOPD cohort. This observation was concomitant with an elevation of the serum CK values at baseline in these two patients, interpreted as a sign of cell damage as previously reported (Chien et al. 2011).

In conclusion, our findings suggest that urinary Glc4 determination may be a useful component in the follow-up of a positive newborn screening result for Pompe disease, especially when the results of confirmatory enzyme and molecular testing are equivocal. An elevated Glc4 suggests an infantile-onset phenotype, whereas a value within the reference range is consistent with a late-onset phenotype or a pseudodeficiency of GAA. However, more studies are needed to determine whether inclusion of Glc4 in the follow-up algorithm, either as a first or later tier test, would provide additional benefit to other testing such as CK or echocardiogram. Glc4 is probably most useful for evaluating the disease status in newly diagnosed patients, especially when combined with clinical and other laboratory assessments and for periodic monitoring of asymptomatic patients. With the possible expansion of newborn screening for Pompe disease in the United States and elsewhere, more data from larger cohorts should be accessible and will allow further assessment of the sensitivity of Glc4 in the newborn period and its prognostic value for monitoring asymptomatic patients.