Preliminary diagnoses of FD were based on the presence of a myeloid body in histological studies and/or decreased α-gal A enzyme activity; patients with FD exhibit α-gal A enzyme activity ≤37.0 nmol/ml/h/mg (30–32). FD diagnosis was confirmed by genetic sequencing, as previously described (32). Between January 2012 and December 2014, 38 patients with FD (17 females and 21 males, 1:1.24; age, 34.7±16.0 years) and 120 healthy volunteers (60 females and 60 males, 1:1; age, 40.8±10.9 years) were enrolled at Ruijin Hospital, The Medicine School of Shanghai Jiao Tong University (Shanghai, China). The present study was approved by the Ethics Committee of Ruijin Hospital and all participants provided their informed consent prior to participation in the current study. Written informed consent was obtained from the parents or guardians of children enrolled. All EDTA-anti-coagulated blood samples from patients with FD were collected prior to enzyme replacement therapy (ERT).

For all patients with FD, demographic data, including age, sex, blood pressure and body mass index, as well as their clinical symptoms were recorded. Estimated glomerular filtration rate (eGFR) was estimated using the abbreviated chronic kidney disease epidemiology collaboration equation (33) in adults and the Schwartz formula (34) in children. End-stage renal disease was defined as an eGFR <15 ml/min or the requirement of kidney transplant or dialysis. Abnormal brain magnetic resonance imaging (MRI) results were indicated by the presence of white matter lesions and lacunar infarctions. Left ventricular hypertrophy was identified by echocardiography. Cornea verticillata and high frequency sensorineural hearing loss were confirmed by the slit-lamp examination and pure tone audiogram, respectively. The presence of acroparesthesia was determining by assessing the medical history of patients and identifying the presence of neuropathic pain in the palms and soles. Physical examination and history taking were used to detect angiokeratoma corporis diffusum and abnormal sweating. Based on these data, the Mainz severity score index (MSSI) was scored to evaluate overall disease severity (35). Patients included in the current study either exhibited classical FD phenotypes or atypical FD phenotypes, which were defined by the diagnostic criteria described in a previous study by our group (32).

For LC-MS/MS, lyso-Gb3 was purchased as standard (Matreya LLC, State College, PA, USA) whereas dimethyl psychosine was selected as the IS (Avanti Polar Lipids, Alabastar, AL, USA). Pure water (AppliChem GmbH, Darmstadt, Germany), acetonitrile and methanol (Thermo Fisher Scientific, Inc., Waltham, MA, USA), formic acid and ammonium formate (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) and all other chemicals were HPLC grade. To detect the α-gal A enzyme activity, 4-methylum-belliferyl-α-D-galactopyranoside (4MU-gal), 4MU and N-acetyl-D-galactosamine (galNAc) were purchased from Sigma-Aldrich; Merck KGaA.

A total of 3 ml whole blood was drawn from the elbow vein of patients with FD and centrifuged at 1,500 × g for 10 min at 4°C. Plasma was separated and stored at −20°C until use. The neat lyso-Gb3 standard was dissolved as recommended by the supplier and diluted with methanol to a working solution concentration of 10 µg/ml. Aliquots of 100 µl standard were dried and stored at −20°C for later use. For each analysis, a blank plasma pools from healthy controls were used for preparing the calibration standards and quality control (QC) samples. After adding appropriate volumes of the working solution to blank plasma matrix, the final calibration standards concentrations of lysoGb3 were 200, 100, 50, 25, 12.5, 6.25, 3.13, 1.56, 0.78 and 0 ng/ml. Final QC samples concentrations of lysoGb3 were 4, 20 and 160 ng/ml. Glycolipids were extracted from 50 ml plasma samples according to simultaneous protein precipitation (36), which was performed by adding a mixture of acetone and methanol (1 ml, 1:1) to a plasma sample and centrifuging the mixture at 14,000 × g for 10 min at 4°C to remove the sediment. The supernatant was collected and evaporated for LC-MS/MS analysis without the requirement of further sample preparation, such as Solid-Phase Extraction. For analysis, samples were re-dissolved in 100 µl pure methanol with subsequent sonication at 40 kHz for 10 min, followed by vortexing at 1,200 rpm for 3 min and centrifugation at 14,000 × g for 10 min at 4°C. The supernatant was transferred into glass microvials and 5 µl was injected into the LC-MS/MS system.

Quantitative LC-MS/MS analysis was performed on an AB SCIEX API 4000 triple quadrupole mass spectrometer (AB Sciex LLC, Framingham, MA, USA) operating in positive mode. An XDB-C18 column (2.1×100 mm, 3.5 µm particles) was used in combination with a VanGuard pre-column of the same material (Agilent Technologies, Inc., Santa Clara, CA, USA). The mobile phases consisted of (A) water and (B) acetonitrile/methyl alcohol (15/85=v/v). Each mobile phase contained 0.1% formic acid and 2 mM ammonium formate. The two mobile phases formed the following gradient: 0–0.25 (50% B) min, 0.25–2.0 min (50–99%, B), 2.0–4.5 min (99%, B), 4.5–4.75 min (99–50% B) and 4.75–8.5 min (50%, B). The MS conditions were as follows: ESI mode was positive, ion spray voltage was 5.5 KV, nebulizer gas and auxiliary gas were 60 psi, curtain gas was 25 psi, the desolvation temperature was 600°C and the flow rate is 0.3 ml/min. The following multiple-reaction monitoring (MRM) transitions were monitored, with a dwell time of 100 ms: 786.5>282.4 (lyso-Gb3) and 490.5>292.3 (IS). Collision energy was 48.6 V and 38 V in the MRM traces of lyso-Gb3 and IS, respectively (Fig. 1).

α-gal A activity in leukocytes was detected using the fluorimetric method, as described by Desnick et al (31). Briefly, 4 ml whole blood was added to 50-ml conical tubes. Leukocyte pellets were isolated and resuspended in 100 µl sodium phosphate buffer. Following 10 cycles of freezing/thawing, leukocyte lysates were centrifuged at 2,000 × g for 10 min at 4°C. Supernatants were stored at −20°C to allow assaying at a later period. The mixture of 4MU-gal plus galNAc solution was added to the leukocyte lysate and incubated at 37°C for 30 min. Reactions were quenched using 0.1 M Na-Glycine buffer (pH 10.7). Fluorescence readings were performed using a Turner Fluorometer (Turner Designs, San Jose, CA, USA) using the following wavelengths: Excitation at 360 nm and emission at 450 nm. Protein values were determined using a standard BCA assay and were used to calculate the specific enzyme activity of the lysates. Normal α-gal A activity was defined as >37.0 nmol/ml/h/mg (32).

The distributions of quantitative variables were assessed for normality. Results are presented as mean ± standard deviation or median (range). All analyses were performed using IBM SPSS Statistics, ver. 22 (IBM Corp, Armonk, NY, USA). A Mann-Whitney U-test was used to assess differences between two non-parametric variables, and a t-test was used to compare two normally distributed variables. To analyze the relationship between two variables, the correlation coefficient was calculated using Spearman's correlation coefficient. A receiver operating characteristic (ROC) curve was used to define the pathological cutoff point of lyso-Gb3 and compare the diagnostic value between it and enzyme activity. P<0.05 was considered to indicate a statistically significant difference.

The demographic variables of the 38 patients with FD prior to ERT are presented in Table I. Compared with female patients, males exhibited an earlier age of onset (13 vs. 33 years). Cornea verticillata was the presenting symptom most frequently reported by males and females (70% of males and 50% of females reported experiencing this symptom). Males with FD were more likely to report a clinical event, including end-stage renal disease and life threatening cardiovascular or cerebrovascular complications, compared with females. Male patients exhibited significantly higher levels of lyso-Gb3 accumulation and lower enzyme activity (each, P<0.001) and also presented with significantly higher MSSI scores than females (P<0.05).

An individual from each group was selected as a typical representative; the chromatography of LC-MS/MS results revealed that a male patient displayed with a lyso-Gb3 level of 106.5, 6.04 ng/ml for a female patient and 0.61 ng/ml for a healthy male control (Fig. 2). Diagnostic assay validation was performed according to the criteria of the Clinical and Laboratory Standards Institute (37,38). Linearity was tested by a dilution series of blank plasma pools analyzed in three separate runs. In this way, the limit of detection representing the assay background (0.22±0.04 ng/ml) was calculated from all subsequent sample dilutions that did not lead to further signal reduction. Recoveries of 86.3–97.0% were measured in QC samples prior to and following extraction (Table II). Intra-assay variation, determined by analyzing a pooled sample in one batch, was 4.6–9.9%. Inter-assay variation, determined by analyzing individually prepared samples in different runs over different days, was 2.8–18.9%. Methanolic stock solutions of the standard and IS were stable for ≥2 months at −20°C. Samples and controls were also frozen at −20°C and did not exhibit a significant decrease following 2 months storage. There were no noticeable influences on quantitative results ≤3 freeze-thaw cycles. These results reveal that using the LC-MS/MS method to quantify plasma lyso-Gb3 levels is accurate, reproducible and highly sensitive.

No significant difference in α-gal A enzyme activity was observed between male and female controls (male, median 104.7 nmol/ml/h/mg; range, 43.2–361.8 nmol/ml/h/mg; female, median 90.8 nmol/ml/h/mg; range, 35.5–175.1 nmol/ml/h/mg; Fig. 3A). By contrast, enzyme activities in leukocytes of male patients with FD (median, 0.3 nmol/ml/h/mg; range, 0.0–23.2 nmol/ml/h/mg) were significantly lower than that of female patients with FD (median, 64.4 nmol/ml/h/mg; range 14.8–152.4 nmol/ml/h/mg) (P<0.001; Table I). The significant differences were also observed between male patients and male healthy controls, and between female patients and female controls.

To verify the results of this assay and determine the normal range of lyso-Gb3, 120 healthy controls were included in the present study. No significant differences in lyso-Gb3 levels were observed between healthy males and females. Only trace amounts of it were present in healthy females (median, 0.44 ng/ml; range, 0.25–0.74 ng/ml) and healthy males (median, 0.42 ng/ml; range, 0.24–0.81 ng/ml). By contrast, median concentrations in male patients were significantly higher than that in female patients (14.50 vs. 2.79 ng/ml; P<0.001; Table I), as expected. Although male and female patients had significantly higher plasma lyso-Gb3 levels than healthy groups (both P<0.001), there was a slight overlap between female patients with FD and healthy individuals (Fig. 3B). The pathological threshold for lyso-Gb3 was set to 0.81 ng/ml with 100% specificity and 94.74% sensitivity, since all samples from healthy controls were clearly below this cut-off value analyzed using a ROC curve (Fig. 4).

To determine the usefulness of plasma lyso-Gb3 measurement in confirming FD diagnosis, enzyme activity was measured in all plasma specimens. Linear correlation analysis of all patients revealed a weak-negative correlation between lyso-Gb3 levels and enzyme activity (r=−0.593; P<0.001; Fig. 5). In males, FD diagnosis could be confirmed by measuring plasma lyso-Gb3 concentration or enzyme activity, as males with FD exhibit markedly higher plasma lyso-Gb3 concentrations and markedly decreased enzyme activity (25). The results of the current study indicated that the sensitivities of two diagnostic indicators were 100% in males. However, the sensitivity of enzyme activity was markedly lower than plasma lyso-Gb3 (23.5 vs. 82.4%) in females with FD (Table III).

Due to incomplete results from examinations, 15 patients lacking MSSI scores were excluded. In male patients, plasma lyso-Gb3 and enzyme activity were correlated with MSSI (P<0.01; Fig. 6A and B). Lyso-Gb3 was observed to be more strongly correlated with MSSI than enzyme activity (r=0.711 vs. r=−0.687; Fig. 6). However, the analysis of correlation between lyso-Gb3 or enzyme activity and MSSI showed no significant difference in female patients (Fig. 6C and D).