| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1
Central Laboratory of Biochemistry, Robert Debré Hospital, CHU of Reims, 51092 Reims cedex, France;
2
Laboratory of Cytogenics and Biology of Reproduction, Maison Blanche Hospital, CHU of Reims, 51092 Reims cedex, France;
3
Endocrinology Department, Robert Debré Hospital, CHU of Reims, 51092 Reims cedex, France;
a author for correspondence: fax 33 3 26 78 85 39, e-mail gillery{at}infobiogen.fr
Human seminal transferrin, a 79-kDa ß-glycoprotein synthesized mainly by Sertoli cells (1)(2), is considered an index of gonadal function because its concentration in seminal fluid correlates with sperm count (3)(4); it is therefore used as a reliable marker of testicular dysfunction (5)(6). This analysis is useful for exploring and monitoring the treatment of idiopathic infertility (7)(8). We considered the possibility of measuring seminal fluid transferrin by a simple peak-rate nephelometric method. Protocols used for measuring transferrin in serum or urine cannot be used, because the matrix of this fluid produces high-intensity nonspecific reactions, mainly as a result of interactions between samples and buffer components such as polyethylene glycol (PEG). We describe here a new, interference-free method, suitable for the routine measurement of seminal fluid transferrin.
The Array 360 nephelometer computer enhanced with Remisol UPC software, microtubes (cat. no. 448163), polyclonal antiserum against human transferrin (cat. no. 449420), calibrator 1 (cat. no. 449560), PEG-containing buffer (cat. no. 663600), diluent (cat. no. 663630), and control Vigil PRX level 2 (cat. no. 450125) were from Beckman Instruments.
Seminal fluids were obtained from healthy donors or patients after informed consent. The study protocol was approved by the ethics committee of our institution. Seminal fluids were liquefied at 37 °C for 10 min, centrifuged at 1500g for 10 min, and stored at -80 °C if not processed immediately.
The first step of the development of this method was to determine the lowest dilution ratio of the sample displaying no nonspecific reactions. The occurrence of a matrix effect was checked by making various dilutions of seminal fluid samples in buffer (1/1 to 1/99, by volume) and replacing the antitransferrin antiserum with buffer. No nonspecific reactions were found for dilutions from 1/3 to 1/99. Lower dilutions (1/1 and 1/2) showed a significant matrix effect, increasing with the dilution ratio. It should be emphasized that these nonspecific reactions were also noticed in higher dilutions when samples were diluted in the commercially available diluent (Beckman cat. no. 663630; data not shown). The major difference is that the buffer contains 45 g/L PEG, whereas diluent contains no PEG.
Consequently, the optimized conditions of our method were as follows. Before the assay, seminal fluid samples were diluted in buffer (1/4, by volume), then vortex-mixed for 10 s, chilled on ice for 15 min, and centrifuged for 5 min at 14 000g at room temperature. The assay was performed using 100 µL of the supernatant. Samples were dispensed in microtubes, placed on the external ring of the analyzer, and assayed using the following settings: optic gain, "33"; primary tube, "microtube"; mode AgX, "0%"; and sample, "undiluted".
The calibration curve was obtained by preparing serial dilutions of calibrator 1 in buffer (final transferrin concentrations between 7.5 and 250 mg/L) and then treating the dilutions as described above. The values of the signal rate ranged from 34 to 3020 and were input into REMISOL-UPC software. The linearity of the assay was verified by assaying 11 serial dilutions of seminal fluid samples containing high transferrin concentrations. A very good correlation between measured and expected values was found in the calibration range: y (measured value) = 0.99x (expected value) + 2.93 (r = 0.999).
The within-run imprecision (CV) was evaluated by 25 assays of three seminal fluid pools with transferrin concentrations of 12.3, 97.4, and 220 mg/L in single series, with intraassay CVs of 0.9%, 1.5%, and 3.0%, respectively. The between-run imprecision (CV) was determined by 25 assays of buffer-diluted Vigil PRX control (1/99, by volume) and of three seminal fluid pools in different series. The interassay CVs were 4.1%, 3.3%, 4.1%, and 5.6% for transferrin concentrations of 29.0, 32.8, 93.6, and 218 mg/L, respectively.
The analytical recovery was checked by adding exogenous transferrin (dilutions of the control serum Vigil PRX, with transferrin concentrations of 0–700 mg/L) to seminal fluid samples, using 20 µL of buffer-diluted Vigil PRX per 100 µL of seminal fluid. The average analytical recovery ranged between 95.0% and 103.3% for expected values ranging from 38 to 210 mg/L. The detection limit, determined as the mean ± 3 SD of 20 measurements of the signal rate generated by the buffer, corresponded to 3.81 rate units, which remained below the lowest point of the calibration curve (34 rate units).
The influence of preanalytical interferences was studied by measuring
several samples when they were received and after they were stored
under different conditions: 20 °C for 1 day; 4 °C for 1, 4, and 7
days; and -20 °C for 2 and 4 weeks. Measurements were also
performed after 1, 5, and 10 freeze-thaw cycles before analysis. No
significant difference was noticed in any of the conditions studied
(Table 1
). Samples (n = 45) were assayed by the new nephelometric assay
and a previously described RIA method (9). The method
comparison showed a significant correlation of the results
(r = 0.982; P <0.01; Fig. 1
).
|
|
This new method avoids pitfalls caused by the peculiar matrix of seminal fluid, which impedes the use of the nephelometric methodologies described for urine transferrin, although concentration ranges are comparable (6)(9). Nonspecific reactions attributable to the peculiar characteristics of seminal fluid interfere with the measurement of the rate signal generated by the antigen-antibody reaction. Indeed, PEG in the buffer, which is basically used to enhance the precipitation of antigen-antibodies complexes, is presumably reacting with various components of biological samples and generating light-scattering particles independently from any specific immunological reaction. This interference constantly occurs with seminal fluid. There are two ways to reduce nonspecific reactions in nephelometric assays. The first one is to use high dilutions of the sample in the diluent provided by the manufacturer. This procedure could not be used in our case, however, because of persistent nonspecific reactions in high dilutions (up to 1/19) and because of the insufficient sensitivity of the method in this dilution range. The occurrence of nonspecific reactions of a magnitude similar to that of antigen-antibody reaction may be particularly misleading and requires the use of a reliable dilution protocol. The second way to reduce nonspecific reactions is to prevent the occurrence of nonspecific reactions inside the flow cell, before the addition of the antibody (10). We have noticed that nonspecific precipitation could be accelerated at low temperatures by predilution of the samples in buffer instead of diluent; the precipitate could then be removed by centrifugation. We used such conditions in our optimized protocol.
Seminal transferrin proved stable in seminal fluid samples during storage. In addition, no significant change was noticed after as many as 10 freeze-thaw cycles. Finally, the results obtained with our technique were highly correlated with those of an RIA (9). In the absence of a reference method, we considered the fair agreement between the results provided by our new method and the RIA method used in many clinical studies (6)(9) as good evidence of the accuracy of our technique. This new method is easy to implement in any laboratory and may be used routinely for clinical purpose.
Acknowledgments
This work was made possible by grants from the "Délégation à la Recherche Clinique du CHU de Reims". We thank J.M. Guillaumin (Tours, France) for performing the RIAs; R. Circaud (Beckman, Gagny, France) for helpful discussions; and D. Deligne, E. Foulon-Guinchard, P. Ruiz, S. Etienne, and E. Deschamps for technical assistance.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
