|
|
||||||||
Molecular Diagnostics and Genetics |
1 Department of Pediatrics, Center for Metabolic Disease, Katholieke Universiteit Leuven, B-3000 Leuven, Belgium.
2 Analis SA, Rue Dewez 14, B-5000 Namur, Belgium.
aAddress correspondence to this author at: Center for Metabolic Disease, University Hospital Gasthuisberg, Herestraat 49, B-3000 Leuven, Belgium. Fax 32-16-34-72-84; e-mail hubert.carchon{at}med.kuleuven.ac.be.
| Abstract |
|---|
|
|
|---|
Methods: We performed 792 CZE analyses of Tf, using the CEofixTM-CDT (carbohydrate-deficient transferrin) assay. Peak identification was based on relative migration times (RMTs) to reduce migration variability.
Results: Tf profiles comprised three main groups (AC). Groups A and B were characterized by one or two dominant tetrasialo-Tf peaks, whereas group C showed a widely variable Tf isoform composition. Group A was composed of four subgroups: a major group with a typical Tf profile (considered as reference group), two minor groups with decreased or moderately increased trisialo-Tf isoform, and a group showing the presence of unknown compounds with RMTs similar to mono- and disialo-Tf. However, these compounds were absent on IEF. Group C contained all profiles from patients with confirmed as well as putative CDG. From the reference group, 99% confidence intervals were calculated for the RMTs of the Tf isoforms, and percentiles representing the Tf isoform distributions were defined.
Conclusions: All patients with abnormal IEF results and confirmed CDG were identified by CZE; thus, this method can be used as a diagnostic alternative to IEF in a manner suitable for automation. Because whole serum is analyzed, it should be kept in mind that CZE profiles can show substances other than Tf.
| Introduction |
|---|
|
|
|---|
1-antitrypsin (6), has been used. IEF, however, is a labor-intensive and time-consuming technique not suitable for automation. In 1986, Stibler et al. (7) reported a relationship between carbohydrate-deficient transferrin (CDT) in serum and chronic alcohol consumption. In the search for a suitable clinical diagnostic test for alcohol abuse, several analytical procedures, such as IEF (8), IEF in combination with immunofixation(9), zone immunoelectrophoresis (10), Western blotting (11), anion-exchange chromatography(7), and chromatofocusing (12), have been used. The problems encountered in CDT analyses using capillary zone electrophoresis (CZE) (8)(13)(14) were later resolved by an improved CZE method requiring only iron saturation as sample pretreatment (15)(16)(17) and the use of dynamically coated capillaries (18). The CEofixTM-CDT assay (Analis) is the first commercial buffer system combining the advantages of dynamic coating with simplicity of sample pretreatment and analysis (19)(20)(21)(22).
The present study aimed to evaluate CZE as an analytical tool for CDG diagnosis. The results obtained from 792 profiles indicate that CZE can be used as a diagnostic technique. However, if profiles contain unexpected peaks, CZE analysis with immunosubtraction and/or Tf IEF with immunoprecipitation should be performed.
| Materials and Methods |
|---|
|
|
|---|
For the present study, 792 samples were analyzed by CZE. A total of 743 samples (426 from males, 317 from females) were obtained from the patients and 49 (22 from males, 27 from females) were from the parents of some of the patients (Table 1
).
|
reagents
Phenylalanine (SA) and N-benzoyl-glycyl-histidyl-leucine (SB) were obtained from Sigma. Anti-human Tf was purchased from Dako. Agarose-IEF, Ampholine pH 57, GelBond film for agarose, and Coomassie Brilliant Blue R250 were obtained from Amersham Biosciences AB. All reagents used were of analytical grade.
ief analysis of TF
In the present method, based on the method of van Eijk et al. (23), 22 mg of agarose was suspended in 2.2 mL of cold water; the mixture was then heated until all of the agarose was dissolved and cooled to 73 °C. After the addition of 110 µL of Ampholine, the mixture was poured onto GelBond for agarose, left for 15 min at room temperature, and stored in the refrigerator overnight. The Tf sample consisted of 15 µL of serum, 70 µL of a solution containing 9 g/L NaCl, and 15 µL of 10 mmol/L Fe(III) citrate. Serum Tfs were separated on an Amersham Biosciences PhastSystem. After electrophoresis, gels were covered with undiluted antiserum (25 µL/cm2) for 30 min and kept overnight in saline. Gels were then washed with water, placed between paper layers under gentle pressure for 30 min, air dried, stained with Coomassie Brilliant Blue R250, and destained.
cze analysis of TF
CZE analyses were performed in uncoated fused-silica capillaries (50-µm i.d.; effective length, 50 cm) on a Beckman Coulter P/ACE 5000 system equipped with a single-wavelength ultraviolet absorbance detector and an interference filter at 214 nm. The buffer system and separation procedure of the CEofix-CDT assay (Analis) were used. Before each analysis, equal volumes of serum sample and iron solution (supplied with the assay reagents) were mixed and centrifuged for 1 min at 13 000g. The capillary was rinsed sequentially under pressure with a polycation-containing buffer (initiator) and a separation buffer containing a polyanion for 1.5 min and 2 min, respectively. An external standard solution (5 mmol/L SA and 2.5 mmol/L SB in 0.2 mol/L Tris-HCl buffer, pH 8.5), water, the pretreated serum sample, and again water were then injected under pressure for 3, 1, 7, and 1 s, respectively. Separation proceeded for 9 min at a constant voltage of 28 kV. Finally, the capillary was rinsed with 0.2 mol/L NaOH for 1.5 min, followed by a rinse under pressure at constant current (100 µA) for 1 min. Corrected peak areas were calculated by the Beckman integration software.
reference group
Samples obtained from patients (n = 534) and parents (n = 41) formed a reference group because of their normal IEF and CZE profiles (20). This group should not be considered as a "normal" group.
data management and evaluation
Microsoft® Visual C++, Ver. 6.0, was used to build appropriate software for data management. Data representing CZE electropherograms, migration times, and corrected peak areas were obtained from the Beckman software. The threshold and minimum peak width for integration were set at 0.01. A minimum area of 125 was used for peak acceptance. This value corresponded to 0.05% of the mean total Tf area.
Because migration times in CZE are often subject to variability, two external standards (SA and SB) were selected so that their electrophoretic mobilities sandwiched the electrophoretic mobilities of the Tf isoforms. The corresponding migration times (MTs) were used to calculate relative migration times (RMTs). Two types, indicated as RMTA and RMTB, are proposed dependent on whether a contribution of the protein has to be taken into account. The RMTA of compound X (RMTA.X) is defined as:
![]() | (1) |
![]() | (2) |
Eq. 1
eliminates only experimental variation, whereas Eq. 2
also eliminates variation resulting from the protein core.
statistics
The level of significance was set at 0.01. A single-factor ANOVA was used for comparing mean RMTB values of groups differing in age or sex.
We used n x 2 contingency tables to evaluate whether the frequency of 0 area values for monosialo-Tf were biased by the total Tf concentration or whether the distribution of the asialo-Tf isoform with age in group A was uniform, respectively.
| Results |
|---|
|
|
|---|
|
All CZE profiles could be assigned to three main groups with several subgroups. Figs. 2
and 3
show several types of CZE profiles and their corresponding IEF patterns. In group A (Fig. 2
), characterized by a single tetrasialo-Tf peak, A1 represents the reference group with a "typical" Tf profile and shows the references SA and SB. In group A, trisialo-Tf had the highest variability. Therefore, two more subgroups were defined: A2, with trisialo-Tf less than disialo-Tf; and A3, with trisialo-Tf greater than pentasialo-Tf but <15% of the total Tf amount. In A4, small compounds were located at the site of mono- or disialo-Tf but are absent on IEF. In group B, two tetrasialo-Tf peaks were present, corresponding to genetic variants of Tf (24). In group C (Fig. 3
), subgroups C1C3 represent the profiles obtained from CDG patients. C1 comprises profiles from CDG-Ia and other CDG-I patients, whereas C2 combines all profiles but one from CDG-II patients. C3 contains a single patient with CDG-IIa (25). Subgroups C4 and C5 combine patients with unexplained CDG. In subgroup C4, trisialo-Tf is high (1847% of the total Tf area), and C5 combines heterogeneous Tf profiles, one of which is shown.
|
|
TF isoform identification
As shown in Fig. 1C
, Tf isoforms can be identified by their RMTB because of the small dispersion. Although the group was heterogeneous in terms of age and sex (Table 1
), there was no statistical difference between sexes, not only within each age group but also between the age groups. The entire group was therefore used to calculate the mean RMTB value and corresponding SD (Table 2
). From these data, 99% confidence intervals were calculated and used for Tf isoform identification (Fig. 4
). The algorithm describing the identification procedure was verified by comparing the result with the corresponding IEF pattern.
|
|
TF isoform distribution
The relative concentration of a particular Tf isoform in the reference group was expressed as a percentage of the total Tf area. Asialo-Tf was not taken into account. The monosialo-Tf isoform was not biased by the total amount of Tf present in the sample. Table 3
summarizes the most common percentiles.
|
According to Arndt (26) and under physiologic conditions, asialo-Tf is present at concentrations <0.5% of the total Tf. This conclusion, applied to the data for group A, gave the values shown in Table 4
. Statistically, the frequency of asialo-Tf was not evenly distributed as a function of age.
|
| Discussion |
|---|
|
|
|---|
Although the variability of migration times in CZE may be largely reduced by the use of dynamically coated capillaries (30), variability remains significant over a longer period of time and after a large number of analyses (Fig. 1A
). Consequently, the use of real times in CZE is not appropriate. Fig. 5
shows the IEF and CZE profiles of two CDG patients, A and B, with unknown types of CDG. In the IEF profiles, lanes 1, 2, 3, and 4 represent the normal IEF pattern of four different reference individuals, whereas lanes A and B represent the patterns of the corresponding patients. Both CZE Tf profiles look rather similar, the three main Tf isoforms differing only in absorbance. In CZE profile A, the migration times are 5.35, 6.01, 6.14, 6.27, and 7.86 min for peaks SA, A1, A2, A3, and SB, respectively. In CZE profile B, the migration times are 5.20, 5.99, 6.12, 6.23, and 7.67 min for peaks SA, B1, B2, B3, and SB. The RMTA values for A1, A2, A3, B1, B2, and B3 are 0.26, 0.31, 0.37, 0.32, 0.37, and 0.42, respectively. Peaks A2 and A3 match most likely with peaks B1 and B2, which is in agreement with the IEF patterns.
|
RMTA values make the Tf profile independent of experimental variation such as length of the capillary or applied voltage, but the impact of the protein core remains unaffected. Fig. 6
shows the IEF and CZE profiles of two patients, A and B, and the mother (M) of patient A. The IEF and CZE profiles indicate that patient A and her mother have the same Tf protein variants.
|
In the present study, group A comprised all individuals with Tf profiles showing a single tetrasialo-Tf isoform. Most of these profiles would be classified as normal on IEF. However, the variability of trisialo-Tf justifies the individualization of subgroups A2 and A3. At present the reason for this variability is not clear.
The frequency of asialo-Tf in the subgroups of group A as a function of age is shown in Table 4
. According to Arndt (26), asialo-Tf is present at concentrations <0.5% of total Tf under physiologic conditions. The frequency observed in children under 1 year of age was significantly higher than in the older patients and in the parents. This is obviously not caused by increased alcohol intake. Both parents (one mother and one father) were healthy persons, and for the remaining five patients, the referring physicians reported no indications regarding alcohol intake.
An apparent increase in a-, mono-, and disialo-Tf observed by IEF may be less pronounced in CZE. In our opinion this may be a combined effect of sample amount and the detection system. In our IEF procedure, 0.2 µL of serum is applied on the gel, whereas in CZE this is
1 nL. Densitometry of the higher Tf isoforms may therefore be out of the linear range of the method, enhancing the importance of the smaller Tf isoform fractions.
As shown in Fig. 1C
, the CZE method separates the Tf isoforms. Using the mean and SD (Table 2
), we could calculate a confidence interval for each Tf isoform by which that isoform could be identified in the electropherogram. As shown in Fig. 4
, the 99% confidence intervals calculated for the Tf isoform in the reference group and the frequency distributions of the Tf isoform for different subgroups were superposed. It is remarkable that the RMTB values of penta- and hexasialo-Tf remain constant regardless of the changes of the profile in the lower Tf isoforms. For several patients in subgroup C1, trisialo-Tf was outside the reference interval, whereas for the others, as well as for the patients in subgroups C2 and C4, trisialo-Tf was located near the upper limit of the interval. Likewise, RMTB values for a- and monosialo-Tf in subgroup C1 seemed to be different from the corresponding values for subgroup C2. In our opinion, this is the limit of information that CZE can provide, and further investigation will require other techniques, such as glycan analysis.
A combination of "peak location" (Table 2
) and "peak distribution" (Table 3
) can be used for data reporting. Fig. 7
shows a combined presentation of Tf profile, Tf isoform identification, and distribution. Lines located within (horizontal arrows) the confidence intervals (upright bars) are identified Tf isoforms. Lines located outside the confidence intervals (vertical arrows) are unidentified peaks in the electropherogram. The line height may visualize either the percentile value obtained from the appropriate reference group or the peak area (percentage) with respect to the total Tf area, regardless of whether the Tf isoform had been identified. The numeric values in the upper horizontal bar represent the percentiles of the Tf isoforms under consideration. The values in the lower horizontal bar indicate the peak areas (percentage) of the unknown compounds with respect to the total Tf area.
|
It must be kept in mind that full serum is used and that other compounds (endogenous or exogenous) can be detected in the Tf region. As seen in subgroup A4, we regularly observed small compounds located at the positions of mono- and disialo-Tf in CZE that were absent on IEF. Fig. 8
shows the IEF and CZE profiles of two patients with CDG-Ia (A and B) and a normal adult (M; mother of A). Profiles A and B were very similar, although the profile for patient A showed an additional peak (arrow) that was also present in the serum of the mother but absent on IEF. In a separate analysis, anti-human Tf was post-injected with the serum Tfs (profiles D and E from individuals A and M, respectively), and immunosubtraction was performed (20). The additional peak remained in both profiles, indicating that the corresponding compound was not Tf. Therefore, and we consider this as an important rule, if the Tf isoform profile of a patient shows abnormal compounds in the Tf region, serum from the parents should also be investigated and an additional separation in the presence of anti-human Tf should be performed. Alternatively, IEF can be used.
|
As shown in Figs. 2
and 3
, in CDG, a huge variability exists in Tf profiles. The Tf isoforms of interest for CDG are a-, mono-, di-, and trisialo-Tf, and they may all increase to a variable extent at the expense of tetrasialo-Tf. The relative concentrations of penta- and hexasialo-Tf remain remarkably constant. The values in Table 5
also reflect the variability of the distribution of these Tf isoforms. The maximum relative concentrations observed for a-, mono-, di-, and trisialo-Tf were 46.6%, 8.2%, 58.5%, and 50.6%, whereas the minimum relative concentration for tetrasialo-Tf was 10.2%.
|
In summary, because all patients with confirmed CDG could be identified, CZE of serum Tf can be used as a diagnostic technique for the detection of CDG. In our experience,
95% of the samples analyzed for serum Tf have a normal profile. CZE analysis with immunosubtraction, Tf IEF with immunoprecipitation, or glycan analysis should be performed only for the remaining 5%, if necessary. In addition, the possibility of automation may enable much more intensive screening for the known and the many unknown glycosylation disorders that await identification.
| Acknowledgments |
|---|
| Footnotes |
|---|
1 Nonstandard abbreviations: CDG, congenital disorders of glycosylation; Tf, transferrin; IEF, isoelectric focusing; CDT, carbohydrate-deficient transferrin; CZE, capillary zone electrophoresis; SA, phenylalanine; SB, N-benzoyl-glycyl-histidyl-leucine; and RMT, relative migration time. ![]()
| References |
|---|
|
|
|---|
The following articles in journals at HighWire Press have cited this article:
![]() |
R. Zeevaert, F. Foulquier, B. Dimitrov, E. Reynders, R. Van Damme-Lombaerts, E. Simeonov, W. Annaert, G. Matthijs, and J. Jaeken Cerebrocostomandibular-like syndrome and a mutation in the conserved oligomeric Golgi complex, subunit 1 Hum. Mol. Genet., February 1, 2009; 18(3): 517 - 524. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. Van Maldergem, M. Yuksel-Apak, H. Kayserili, E. Seemanova, S. Giurgea, L. Basel-Vanagaite, E. Leao-Teles, J. Vigneron, M. Foulon, M. Greally, et al. Cobblestone-like brain dysgenesis and altered glycosylation in congenital cutis laxa, Debre type Neurology, November 11, 2008; 71(20): 1602 - 1608. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. R. Delanghe, A. Helander, J. P.M. Wielders, J. M. Pekelharing, H. J. Roth, F. Schellenberg, C. Born, E. Yagmur, W. Gentzer, and H. Althaus Development and Multicenter Evaluation of the N Latex CDT Direct Immunonephelometric Assay for Serum Carbohydrate-Deficient Transferrin Clin. Chem., June 1, 2007; 53(6): 1115 - 1121. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. A. Carchon, C. Nsibu Ndosimao, S. Van Aerschot, and J. Jaeken Use of serum on guthrie cards in screening for congenital disorders of glycosylation. Clin. Chem., April 1, 2006; 52(4): 774 - 775. [Full Text] [PDF] |
||||
![]() |
A. Helander, J. Bergstrom, and H. H. Freeze Testing for Congenital Disorders of Glycosylation by HPLC Measurement of Serum Transferrin Glycoforms Clin. Chem., May 1, 2004; 50(5): 954 - 958. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |