Study of urinary proteomes in Anderson-Fabry disease

Abstract

Background: Anderson-Fabry disease (AFD) is an X-linked genetic disorder with deficient α-galactosidase A activity. The main aim of this work was to investigate possible differences in urine proteins between healthy controls and AFD patients and to identify abnormal proteins as potential biomarkers of disease. Material and methods: We studied 2D electrophoresis images of urine samples collected from AFD patients and healthy subjects. The proteins were separated using isoelectric focusing method followed by SDS-PAGE. The proteins were then visualized by silver staining and characterized by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). Results: We found out that the urinary spectra of all the Fabry disease samples included identical proteins with molecular weight around 20–40 kDa. The concentration of some proteins was more than three times higher in the AFD samples, compared to the controls. The abundant proteins were identified by MALDI-TOF MS and included the following: alpha-1-antitrypsin, alpha-1-microglobulin, prostaglandin H2 d-isomerase, complement-c1q tumor necrosis factor-related protein, and Ig kappa chain V–III. Possible glycosylation at Asn51 and Asn78 sites of the prostaglandin H2 d-isomerase was detected. Conclusions: AFD urinary proteomics revealed increased secretion of several proteins. We postulate that the observed difference in the amount of prostaglandin H2 d-isomerase and its position on two-dimensional gels might be related to different glycosylation in AFD subjects.

• Anderson-Fabry disease
• MALDI-TOF MS
• two-dimensional electrophoresis
• urinary proteomes

INTRODUCTION

Anderson-Fabry disease (AFD) is an X-linked glycolipid storage disorder that is caused by deficient activity of the lysosomal enzyme α-galactosidase A (α-gal A).^1,2 Globotriaosylceramide (Gb₃), the glycolipid substrate for this enzyme, progressively accumulates within vulnerable cells and tissues of affected patients, which leads to a wide clinical spectrum of clinical manifestations. Endothelial, perithelial, and smooth muscle cells of the vascular system, renal epithelial cells, myocardial cells, dorsal root ganglia, and cells of the autonomic nervous system are selectively damaged by Gb₃.^3–5 The diagnosis of AFD is usually established by detection of low α-gal A activity in plasma, leukocytes, or fibroblasts, or by detection of increased Gb₃ levels in urine kidney, skin, and heart, or by detection of gene mutations.⁶ The activity of α-gal A in hemizygous Fabry disease patients is hardly detectable, whereas in milder disease variants (cardiac), in atypical hemizygous patients, and in female carriers, the residual activity of α-gal A is detectable with variable intensity.⁴

The incidence of AFD is estimated to be 1:40,000,^7,8 in some studies up to 1:3500.⁸ As in other X-linked recessive disorders, symptoms in carriers are expected to be rare; however, in AFD, up to 80% of female carriers suffer from some typical disease symptoms and up to 20% of these women present with health-endangering symptoms. The classical AFD phenotypes suffer from a multiorgan disorder affecting blood vessels, kidneys, heart, and the central and peripheral nervous system. The clinical onset of the disease typically occurs during childhood or adolescence, with symptoms of recurrent episodes of severe pain in extremities. Vital organs are progressively affected with increasing age in hemizygous males with death, resulting from renal, cardiac, or cerebrovascular complications, usually occurring during the fourth or the fifth decade of life.⁹ Enzyme replacement therapy (ERT) with human recombinant α-gal A has become available since 2001.^10,11 Recent evidence suggests that the ERT is more efficacious if started early, before the development of advanced kidney involvement with overt proteinuria.^7,12 Until now, assessment of Gb₃ excretion and its concentration in plasma and urine has been used as a biomarker to monitor the disease response to the treatment. Recently, lyso-Gb₃ has been suggested as a more promising tool. Lyso-Gb₃ is a soluble compound, in contrast to Gb₃, that can easily move in and out of cells. It lacks a hydrophobic (acyl) fragment compared to Gb₃. Lyso-Gb₃ is virtually not detectable in plasma obtained from normal individuals, but relatively high concentrations occur in samples from Fabry males. Also in the case of symptomatic Fabry females, increased levels of lyso-Gb₃ were detected.¹³ However, there is a critical need to find new biomarkers for assessment of the disease severity and the ERT efficacy.¹⁴ Our study was designed to analyze potential differences in urinary proteomes between AFD patients and healthy controls, using modern analytical techniques, such as isoelectric focusing and two-dimensional (2D) electrophoresis in polyacrylamide gels. Employment of a similar technique of protein determination by 2D electrophoresis was shown in previous studies,^15–18 in which some protein differences specific to certain diagnoses (nephrotic syndrome, Fanconi syndrome, or various carcinomas) were detected.

The main aim of this work was to establish the differences between the healthy controls and the AFD subjects and to identify potential disease-specific proteins using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). We expect this finding to facilitate better understanding of the AFD.

MATERIAL AND METHODS

Study population

Urinary samples from 20 patients with AFD (11 males, 9 females, age range 18–69 years) were collected. Their proteinuria varied between 0.1 and 1 g/24 h; 13 patients were on ERT (nine of them were given agalsidase beta (Fabrazyme®) 1 mg/kg every other week, four of them were given agalsidase alfa (Replagal®) 0.2 mg/kg every other week) and 7 subjects received no therapy. Physiologic proteinuria is usually estimated to be between 30 and 50 mg/day and is mostly composed of tubular proteins, namely uromodulin (Tamm–Horsfall protein). Proteinuria of 150 mg/day may be accepted as an upper limit of “normal” proteinuria. So our patients had only small proteinuria, which would be tagged as normal and could be determined using the routine laboratory evaluation (Table 1). The control group consisted of 10 healthy individuals (five males and five females aged between 27 and 42 years of age). Samples (30 mL) from their 24-h urine collections were used for the analyses. Proteinuria is usually very low (up to 1 g/24 h) in patients with Fabry disease; therefore, most of the urine samples (proteinuria of less than 1 g/24 h) had to be concentrated (10 kDa, 5 min, 3000 ×g, 4°C)¹⁹ using Amicon Ultra (Millipore, Billerica, Massachusetts, USA) before their 2D analysis. The samples were either analyzed immediately, or the sample aliquots were stored at −80°C. Stability of the urine samples during their storage was confirmed by urine sample reanalysis 30 and 180 days after the first analysis. The protein spectrum analysis confirmed that freezing and subsequent storage did not affect the presence of proteins in the urine samples.¹⁵ The study was approved by a local ethical committee (Table 1).

Table 1.  Clinical data of Anderson-Fabry disease patients

Patient number	Patient age	Gender	Mutation GAL, genotype	Mainz score	Alphagal leukocyte (nmol/mg h)	ERT	Therapy from	Drug	DM	ACE inhib., AT block.	S_creatinin (μmol/L)	U_creatinin(mmol/L)	HD	Transplantation	Proteinuria (g/L)
1	57	F	N215S	19	30.3	N			N	Y	53	11.5	N	N	<0.07
2	60	F	N215S	20	27.8	N			N	N	76	8.40	N	N	<0.07
3	69	F	c.674del59	27	23.6	Y	12/2001	Fabrazyme	N	Y	77	11.10	N	N	0.14
4	39	F	R342Q	14	28.6	N			N	N	78	1.6	N	N	<0.07
5	60	F	R342Q	15	8.5	Y	11/2004	Replagal	N	Y	85	18.3	N	N	0.17
6	46	F	delEx2	19	25.4	N			N	N	78	13.7	N	N	<0.07
7	55	F	R342Q	7	17.9	Y	12/204	Replagal	N	Y	57	18.6	N	N	0.13
8	64	F	I317T	38	22.3	N			N	Y	73	6.8	N	N	0.08
9	61	F	R342Q	54	21	N			Y	Y	93	4.3	N	N	0.09
10	40	M	c.674del59	38	0.7	Y	12/2004	Replagal	N	N	75	5.6	N	N (nephrectomy2006)	0.18
11	32	M	c.674del59	14	0.7	Y	12/2003	Replagal	N	N	67	6.9	N	N	0.08
12	43	M	c.674del59	31	0.3	Y	12/2003	Fabrazyme	N	N	82	9.5	N	N	0.1
13	42	M	delEx2	22	2.5	Y	1/2004	Fabrazyme	N	N	73	10.5	N	N	0.14
14	40	M	R342Q	25	1.6	Y	2/2004	Fabrazyme	N	N	75	7.7	N	N	0.08
15	18	M	c.674del59	16	0,2	N	2007	Fabrazyme	N	N	69	3,3	N	N	<0.07
16	52	M	delEx2	34	0.22	Y	1/2004	Fabrazyme	N	Y	71	9.5	N	N	0.24
17	59	M	c.674del59	51	2.6	Y	7/2001	Fabrazyme	N	N	185	2.4	N	N	0.37
18	55	M	L294X	32	0.4	Y	11/2004	Fabrazyme	N	Y	645	6.3	Y	Y (1984, HD started from 2005)	0.61
19	51	M	R342Q	47	1.0	Y	11/2004	Fabrazyme	N	Y	218	6.1	Y	Y (HD started in 2001, next transpl. 2006)	0.38
20	48	M	c.674del59	34	1.82	Y	2/2004	Fabrazyme	N	Y	122	6.2	N	N	0.66

Note: DM, diabetes mellitus; ERT, enzyme replacement therapy; HD, hemodialysis; ACE, inhibitors of angiotensin-converting enzyme; AT, blocker of angiotensin receptor; Y, yes; N, no; F, female; M, male.

Determination of protein concentration

The Bradford spectrophotometric method²⁰ was used to determine protein concentration in urine in concentrated samples. Absorption was monitored at 595 nm against a standard. Then, a method described by Watanabe²¹ using pyrogallol red was used to determine a total protein concentration at the beginning of the analyses. Absorption was measured by the Modular spectrophotometric instrument (Roche, Basel, Switzerland).

2D electrophoresis

The proteins were separated by the isoelectric focusing method using 7 cm polyacrylamide strips with immobilized linear gradient, pH 3–10 (IPG strip pH 3–10 L, Amersham Biosciences, Upsala, Sweden; electrophoresis Multiphor II, Amersham Biosciences). The total amount of protein placed on each polyacrylamide strip was 20 μg. The isoelectric focusing was performed in three steps to reach 3500 V (Power Supply EPS 3501, Amersham Biosciences). The second dimensional separation was performed in the SE260 Mini-Vertical Electrophoresis Unit (Scie-Plas, Warwickshire, UK) in the 12% polyacrylamide gel. The proteins were then visualized by silver staining (Silver Bullit Kit, Amresco, St. Louis, Missouri, USA).

Gel comparison

The Phoretix 2D expression 2005 software was used for the protein analysis. A urine sample from a healthy control was used as a reference gel. This reference gel was then used to compare the proteins in the investigation gels from the AFD patients with those from the healthy controls. The protein content increase (in percentage) was calculated using the following formula:

% increase protein: [total protein in the sample (Fabry disease)/total protein in the sample (healthy control)] × 100

In-gel protein digestion

The silver-stained protein spots were excised from the 2D gels, cut into small pieces, and washed several times with fresh 1:1 solution of 30 mM potassium ferricyanide and 100 mM sodium thiosulfate. Upon complete destaining, the gel pieces were washed with water, dehydrated in acetonitrile (ACN), and again rehydrated in water. The supernatant was removed and the gel was partly dried in a vacuum concentrator. The gel pieces were then reconstituted in a cleavage buffer containing 0.01% 2-mercaptoethanol, 0.1 M 4-ethylmorpholine acetate, 1 mM CaCl₂, 10% ACN, and sequencing grade modified trypsin (50 ng/mL, Promega, Fitchburg, Wisconsin). After overnight digestion, the resulting peptides were extracted with 40% ACN/0.5% trifluoroacetic acid (TFA). For the low-abundance proteins, the peptides were purified and concentrated prior to mass spectrometric analysis according to manufacturer's user guide using ZipTip tips with 0.6 μL C18 resin (Millipore, Billerica, Massachusetts).

MALDI-TOF MS and protein identification

Ten milligrams per milliliter solution of α-cyano-4-hydroxycinnamic acid (Bruker Daltonik, Bremen, Germany) in 30% ACN/30% MeOH water solution containing 0.2% TFA was used as a MALDI matrix. A sample (1 mL) was placed on the polished steel MALDI target plate, allowed to air-dry at room temperature, and then 1 mL of the matrix solution was added. Positive ion mass spectra were obtained using a matrix-assisted laser desorption/ionization time-of-flight/time-of-flight (MALDI-TOF/TOF) mass spectrometer (Autoflex II, Bruker Daltonik) in reflectron mode, equipped with a 337 nm nitrogen laser. The generated spectra were within the mass range of 700–4500 Da. The data were calibrated against the Peptide calibration standard I (Bruker Daltonik) as an external calibration standard, using a seven-point calibration. Mass spectrometric sequencing of selected peptides was done by MALDI LIFT-TOF/TOF MS of the same prepared samples with the above instrument operated in MS/MS mode. Using the FlexAnalysis 2.4 program with the SNAP peak detection algorithm, peak lists in the XML data format were created. Statistical calibration was included in the program and no further adjustments were applied; the maximum number of assigned peaks was set to 50. After the peak labeling, all known contaminant signals were manually removed. The peak lists were examined against Swiss-Prot or NCBInr database subsets of human proteins, using the MASCOT search engine with the following search criteria settings: peptide tolerance of 100 ppm, the number of allowed missed cleavage sites was set to 1, fixed modifications of cysteine carbamidomethylation, and variable methionine oxidation. No restriction criteria were set with respect to protein molecular weight and pI values. Identity of a particular candidate protein was confirmed by MS/MS peptide sequencing, when either the probability-based Mowse score was only slightly higher than the threshold value calculated for the parameters used or the sequence coverage was too low.

Statistical analysis

Statistical analysis was performed using a Student's t-test, the p-values of <0.05 being considered significant.²²

RESULTS

Changes in proteins

Twenty urine samples collected from AFD patients were analyzed in this study. Two examples of electrophoretograms are shown in Figure 1. Urinary spectra of all the AFD samples included identical proteins, most of them with molecular weight around 20–40 kDa (Figure 1).

Figure 1. Comparison of 2D electrophoreograms of urine samples collected from patients with Fabry disease (A, B) and healthy controls (C, D). The urine samples were analyzed immediately after their collection and short-time centrifugation, using 7 cm polyacrylamide strips, pH 3–10 L, and SDS-PAGE. BenchMark^TM Protein Ladder (Invitrogen, Glasgow, UK) was used as a molecular mass marker.

There was no significant difference in the distribution of proteins on electrophoretograms in AFD patients undergoing ERT, compared to those without therapy or healthy controls (Figure 1). No α-gal A activity was detected in the urine samples examined. α-gal A activity was measured as published.²³

Quantitative analysis revealed significantly increased amount of some proteins with molecular weights about 30 kDa in AFD patients. This finding was then confirmed by software Phoretix 2D expression 2005. We picked out three proteins, the quantities of which were significantly higher in the AFD subjects, compared to the healthy controls; in the AFD samples, the levels of prostaglandin H2 d-isomerase and complement-c1q tumor necrosis factor-related protein were three times higher and the Ig kappa chain V–III levels were more than five times higher, compared to the controls (Table 2, Figure 2).

Table 2.  Changes of selected proteins observed using Phoretix 2D expression software version 2005

No.	Protein	Anderson-Fabry disease (average ± SD) %^a	Expected molecular weight (kDa)	Expected pI
1	Prostaglandin H2 d-isomerase	398 ± 25	28	6.1
2	Complement-c1q tumor necrosis factor-related protein	289 ± 17	23	5.8
3	Ig kappa chain V–III	591 ± 7	27	5.5

Note: *p < 0.05 versus healthy controls. Samples were collected from 20 different patients with Anderson-Fabry disease. ^aN = 20: [(amount of protein in the sample with Anderson-Fabry disease/amount of protein in healthy control) × 100].

Figure 2. Changes in the protein amount in the samples collected from 20 patients with Fabry disease compared with those in healthy controls. *p < 0.05 versus healthy controls.

Characterization of proteins

Some AFD proteins also included differentially expressed spots. These were then excised and subjected to in-gel tryptic digestion and identification by MALDI-TOF MS. In cases when the sequence coverage was too low, identity of the protein candidate was confirmed by MS/MS peptide sequencing. The abundant proteins in the gels were characterized as alpha-1-antitrypsin, alpha-1-microglobulin, prostaglandin H2 d-isomerase, complement-c1q tumor necrosis factor-related protein, Ig kappa chain V–III (Figure 3). Positions of all the proteins identified on 2D gels were compared with their theoretical isoelectric points and molecular sizes (Table 3). Most of the proteins were in the expected range, except prostaglandin H2 d-isomerase. This exception may be caused by glycosylation at Asn51 and Asn78 sites.

Figure 3. A representative 2D gel image of the Fabry disease proteome, depicting all identified proteins. All urine samples were analyzed immediately after their collection using 7 cm polyacrylamide strips, pH 3–10 L, and SDS-PAGE.

Table 3.  Comparison of experimental and theoretical molecular size and isoelectric point of identified proteins by MALDI-TOF MS

No.	Protein	Experimental pI	Theoretical pI	Experimental molecular weight	Theoretical molecular weight
1	Alpha-1-antitrypsin	5.1	5.4	48	47
2	Alpha-1-antitrypsin	4.9	5.4	47	47
3	Alpha-1-microglobulin/bikunin precursor	5.2	5.9	40	40
4	Alpha-1-microglobulin/bikunin precursor	5.1	5.9	40	40
5	Ig kappa chain V–III	5.5	5.8	27	23
6	Complement-c1q tumor necrosis factor-related protein	5.8	6.1	23	25
7	Prostaglandin H2 d-isomerase	6.1	7.7	28	21

Note: *p < 0.05 experimental versus theoretical parameter, N = 20.

DISCUSSION

Two-dimensional electrophoresis followed by MALDI-TOF MS is a powerful tool for studying protein changes and for their identification. The main aim of most clinical studies using this technique is to identify possible biomarkers typical for specific diseases. In our previous study,¹⁵ lucid spectra of proteins were investigated in patients with nephrotic proteinuria and other diagnoses, such as systemic lupus erythematosus with renal involvement (lupus nephritis) and anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis. The main aim of our study was to analyze the composition or structure of proteins in urine samples in AFD patients, compared to that in normal subjects. No significant differences in the number of spots were detected between the healthy controls and the AFD patients; therefore, we may conclude that the disease does not lead to any formation of new proteins or degradation products. This conclusion can also be confirmed by comparison of the spectra obtained from AFD patients with the protein spectra obtained from nephrotic patients, in whom the protein quantity on the gel is substantially higher than in healthy volunteers.

Although qualitative changes were not detected, quantitative differences in protein contents were noted. We hypothesize that the increased quantity of proteins in urine is caused by the disease, because AFD is usually associated with impairment of proximal tubular function. This proximal tubular function impairment results in a failure to reabsorb a variety of filtered substances, particularly low-molecular-weight proteins. Most of the identified proteins were of plasma origin and, under normal conditions, would be reabsorbed in the proximal tubule via receptor-mediated endocytosis. On the other hand, analysis of quantitative changes revealed substantially increased Ig kappa chain V–III, complement-c1q tumor necrosis factor-related protein, and prostaglandin H2 d-isomerase. The first two protein levels were more than three times higher and the prostaglandin H2 d-isomerase level was five times higher than the respective values in the control subjects. The latter protein plays an important role in the regulation of phospholipid metabolism⁶ and its excretion is significantly increased in patients with chronic renal failure.^24–26 The higher quantity of these three proteins could result from their increased synthesis or from increased glomerular basement membrane permeability.^27,28 Impairment of secretion–reabsorption tubular processes because of lysosomal vacuoles affecting the renal tubular system appears to be the most likely cause, as the composition of the proteins, identified in the AFD subjects, appears to be very similar to a proteome found in patients with tubular proteinuria.^6,29 Other changes in protein excretion were reported in a histochemical study³⁰ showing that urinary excretion of integrin β₃ originating from the vitronectin receptor was significantly increased in the AFD group, compared to a healthy control group.

Finally, the abundant proteins identified on 2D gels were subjected to MALDI-TOF MS analysis. All seven studied proteins (alpha-1-antitrypsin, alpha-1-microglobulin, prostaglandin H2 d-isomerase, complement-c1q tumor necrosis factor-related protein, Ig kappa chain V–III) were also found to be present in healthy individuals. Except for H2 d-isomerase, positions of all the other proteins on 2D gels corresponded to their theoretical isoelectric points and molecular sizes. We assumed that modifications such as glycosylation of H2 d-isomerase at Asn51 and Asn78 sites might have occurred, which could result in a change in molecular size of the enzyme and its isoelectric point. The glycosylation changes at the protein sites are of potential relevance and might serve as potential biomarkers. However, this presumption should be confirmed by future studies analyzing the relationship between the above findings and the disease severity and progression. Most of the articles published last year were focused on clinical studies of various types of therapy of the Fabry disease patients,³¹ studies in patients prior to their ERT initiation,³² or in patients treated with α-gal A (a glycoprotein of molecular size 101 kDa with a homodimeric structure) infusions, aimed at reducing the Gb₃ storage in tissues.¹¹

CONCLUSION

Our proteomic analysis revealed no significant qualitative differences between the urine samples collected from the treated and untreated AFD subjects and the samples collected from the control subjects. However, quantity of several proteins was substantially different in the AFD group and the control group. In addition, the molecular size of H2 d-isomerase was shown to be modified. These findings should be confirmed by future studies analyzing the relevance of the observed changes for prediction of the disease severity and outcomes.

Acknowledgments

The authors are grateful to K. Pelinková, MSc, and A. Břichnáčová, MSc, from the Institute of Clinical Biochemistry and Laboratory Diagnostics, First Faculty of Medicine of Charles University and General Teaching Hospital, for technical cooperation. This study was supported by research project of the Ministry of Health of the Czech Republic (No. 64165), the Grant Agency of Charles University (No. 203434/54), and by grants of the Ministry of Education, Youth and Sports of the Czech Republic (MSM 0021620806 and LC 06044).

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.