| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Department of Clinical Neuroscience, Karolinska Institutet & University Hospital, Stockholm, Sweden;
2 Department of Clinical Chemistry, Meander Medical Center, Amersfoort, The Netherlands;
aaddress correspondence to this author at: Alcohol Laboratory, L7:03, Karolinska University Hospital Solna, SE-171 76 Stockholm, Sweden; fax 46-8-51771532, e-mail Anders.Helander{at}cns.ki.se
The major form of the iron-transport glycoprotein transferrin in blood contains 2 N-linked disialylated biantennary oligosaccharide chains (glycans) and is named tetrasialotransferrin. Regular high alcohol consumption (mean of at least 50–80 g/day) generally alters the glycosylation profile of transferrin (1), increasing the relative amounts of glycoforms lacking one (disialotransferrin) or both (asialotransferrin) N-glycans (2)(3). The alcohol-related glycoforms are collectively referred to as carbohydrate-deficient transferrin (CDT). CDT measurements are widely used for identifying individuals with alcohol problems in various medical settings (e.g., addiction treatment) and for monitoring abstinence from alcohol in outpatient treatment programs (e.g., when drunk-driving offenders reapply for a driver’s license) (4). When drinking is discontinued, the CDT concentration normalizes with a half-life of 1.5–2 weeks (5)(6). The main advantage of CDT over the conventional alcohol biomarkers, such as the liver function test 
-glutamyltransferase, is the higher specificity for alcohol misuse with resulting lower risk for false-positive identifications (7)(8).
Since the discovery of CDT as an alcohol marker (1), a multitude of analytical techniques and methods have been applied for its measurement (1)(9). The most widely used assays worldwide today are the Axis-Shield %CDT immunoassay and various automated applications thereof, such as %CDT TIA from Bio-Rad and Tina-quant® %CDT from Roche (10)(11). These assays are based on ion-exchange minicolumn chromatographic isolation of the CDT fraction, separate measurement of CDT and total transferrin using the same transferrin antibody, and calculating CDT as a percentage of total transferrin (%CDT). Immunologic methods are convenient and time-efficient for routine use in central laboratories with high specimen throughput, but because these tests separate CDT from non-CDT moieties on the basis of differences in isoelectric point (pI), they will be disturbed by genetic transferrin polymorphisms (12) and by congenital disorders of glycosylation (13), which can cause falsely high or low results that may lead to false-positive or -negative identification of patients for alcohol misuse. Accordingly, when the %CDT ion-exchange immunoassay combination is used in medico-legal cases, such as traffic medicine (14)(15), there is a need to confirm the test result by an independent separation method to rule out analytical interferences as the cause of a high value (12)(16)(17).
This study compared the performances of 2 laboratory methods for confirmatory and routine testing of %CDT in serum, HPLC and capillary electrophoresis (CE). Both methods produce a fingerprint of the glycoform distribution and allow for quantitative and reproducible determination of single transferrin glycoforms. We compared the results with those obtained by a %CDT minicolumn immunoassay (Axis-Shield %CDT assay).
The 79 sera used in this study were selected from the routine samples pool to cover the range from low-normal to highly increased %CDT values (1.3%–24.2% by the Axis-Shield %CDT assay) as well as some genetic transferrin variants. At the Karolinska University Hospital in Stockholm, 42 anonymous surplus sera were collected and randomly analyzed in single determinations for transferrin glycoforms by HPLC. A portion of each of these specimens was sent frozen to the Meander Medical Center in Amersfoort for blinded CE analysis, also using single determinations. Another 37 serum samples were collected in Amersfoort, analyzed by CE, and sent to Stockholm for blinded HPLC analysis. Serum was separated by centrifugation, and samples were stored at –20 °C until analysis.
HPLC separation of transferrin glycoforms was performed on a SOURCE® 15Q anion-exchange column (Amersham Biosciences) by linear salt gradient elution, and quantification relied on measurement of the absorbance of the iron–transferrin complex at 470 nm (17). The relative amounts of single glycoforms to total transferrin were calculated as percentages of peak areas with use of baseline integration. The CE method was the CEofixTM CDT assay (version improved in November 2003; part no. 10-004260 Europe 844111046) from Analis (Belgium). The samples were iron-saturated, and electrophoresis was carried out with ultraviolet detection at 214 nm on a Beckman Coulter P/ACE 5000, according to the manufacturer’s instructions. The relative amounts of single transferrin glycoforms were calculated from peak areas by valley-to-valley integration, as suggested by the manufacturer. The %CDT was measured immunologically in single determinations by use of the Axis-Shield %CDT assay according to the manufacturer’s instructions. This assay measures the sum of CDT glycoforms with a pI >5.65 (i.e., mainly asialo-, monosialo-, and disialotransferrin) relative to the total amount of transferrin in the serum sample.
The HPLC and CE methods allowed for reproducible separation and quantification of single transferrin glycoforms with similar peak patterns. When analysis was performed on a transferrin-free serum sample obtained by immunosubtraction with rabbit anti-human transferrin (17), no interfering peaks were detected by either method (data not shown), thus indicating that the peaks observed in the HPLC chromatogram and CE electropherogram were indeed transferrin glycoforms. Shown in Fig. 1A
are examples of a typical HPLC chromatogram and the corresponding CE electropherogram for a serum sample with an alcohol-related increase in the relative amount of disialotransferrin and a visible asialotransferrin by HPLC. Rare genetic transferrin variants and glycoform types, including transferrin B homozygotes and BC and CD heterozygotes, another variant tentatively identified as "C2C3" (12), and serum samples containing high relative amounts of monosialo- and trisialotransferrin, all identified as actual or potential causes of falsely high or low %CDT results with the minicolumn immunoassays (12), were readily identified by both methods. Examples of an HPLC chromatogram and the corresponding CE electropherogram for a sample heterozygous for the genetic transferrin BC are shown in Fig. 1B
.
|
The relative amounts of disialotransferrin, the major glycoform in CDT, to total transferrin (i.e., total peak area for all transferrin glycoforms) obtained by HPLC and CE were highly correlated (r2 = 0.972; P <0.0001; Fig. 1C
). However, the HPLC method constantly yielded higher results, the values being 1.36% higher, on average (range, 0.15%–4.20%; P <0.0001), than the CE values (n = 68; genetic variants were excluded from these calculations; Fig. 1C
, inset). Use of the baseline integration mode for a subset (n = 32) of the CE results, instead of the valley-to-valley integration mode recommended with the CEofix CDT assay, only partly compensated for this difference; the HPLC disialotransferrin values was still 0.97% higher, on average (range, –0.01% to 2.75%; P <0.0001).
The %CDT results obtained with the minicolumn immunoassay were highly correlated (r2 = 0.904; P <0.0001) with the corresponding HPLC results (i.e., sum of asialo-, monosialo-, and disialotransferrin). However, several samples with genetic transferrin variants (e.g., transferrin CD and C2C3) produced falsely high %CDT results with the Axis-Shield %CDT assay (data not shown), which agrees with previous observations (12). It should be noted that the first direct immunoassay for %CDT (N Latex CDT; Dade Behring) will soon be commercially available. This new method is based on a monoclonal antibody that measures specifically the transferrin glycoforms lacking 1 or 2 N-glycans and on a simultaneous total transferrin assay (18). Accordingly, genetic transferrin variants B and D, albeit having primary structures different from that of the common transferrin C phenotype but a normal set of carbohydrate chains, are indicated not to interfere with the %CDT measurement by the N Latex CDT (19), which represents a major improvement over the minicolumn versions in terms of higher test specificity.
The present results show that the HPLC and CE methods for %CDT allow for visible and reproducible determination of transferrin glycoforms in small sample volumes. Most importantly, the genetic transferrin variants and glycoform types that often cause falsely high or low results with the current %CDT minicolumn immunoassays were readily identified by both methods. The relative amounts of disialotransferrin obtained by HPLC and CE were highly correlated, but the HPLC method constantly yielded higher values. One possible reason for this difference was related to the integration modes used for quantification of transferrin glycoform peak areas (baseline for HPLC and valley-to-valley for CE), but use of baseline instead of the recommended valley-to-valley integration with the CE method only partly compensated for this difference. It might be that the CE method used (CEofix CDT assay) is less sensitive for quantification of transferrin than the HPLC method, which agrees with previous results (17)(20). This could result from the CE method relying on nonspecific measurement of the peptide bond, whereas the HPLC method measures the selective absorbance of the transferrin–iron complex. However, improved CE methods with better resolution between disialo- and trisialotransferrin and thus producing results for the CDT glycoforms approaching those obtained by HPLC, have recently been published (21)(22).
In summary, these results support the use of HPLC and CE methods for confirmation of %CDT immunoassay results in medico-legal cases and for evaluation of borderline values. Several commercial HPLC and CE assays for %CDT testing with high capacity and reduced total analysis time are already, or will soon be, commercially available.
Acknowledgments
This work was supported by grants from the Karolinska Institutet (to A.H.).
References
-glutamyl transferase levels during disulfiram therapy. Alcohol Clin Exp Res 1996;20:1202-1205.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
