| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
Department of Clinical Chemistry, Helsinki University Central Hospital, Haartmaninkatu 8, FIN-00290 Helsinki, Finland;
aauthor for correspondence: fax 358-9-47174804, e-mail ulf-hakan.stenman{at}hus.fi
Prolactin (PRL) is a 23-kDa protein hormone secreted mainly by the lactotrophs in the anterior pituitary gland and also at low concentrations by several other tissues (1). In serum, PRL occurs in various molecular forms, including the physiologically active monomeric form, also called little PRL (L-PRL; molecular mass, 23 kDa), big PRL (B-PRL; molecular mass, 40–62 kDa), and big-big PRL (BB-PRL; molecular mass, 150–170 kDa), which also is called macroprolactin (2)(3)(4). L-PRL is usually the dominant form in the serum of healthy and hyperprolactinemic individuals. The exact structures of B-PRL and BB-PRL are not fully understood: B-PRL is thought to consist of oligomers of PRL, whereas BB-PRL has been suggested to consist of an IgG complex with one or two PRL molecules (5)(6).
Hyperprolactinemia is mostly caused by a pituitary adenoma, but it may also be caused by macroprolactin (7)(8). Because the big forms of PRL have decreased bioactivity, they do not cause clinical symptoms of hyperprolactinemia (7)(9). Of hyperprolactinemic individuals, 15–26% are macroprolactinemic (10), which often leads to a false diagnosis of hyperprolactinemia. The results obtained by various PRL assays vary highly in samples containing macroprolactin (11)(12)(13). This has been thought to be caused by differences in the ability of the antibodies used to recognize big forms of PRL. It is generally thought that assays with high specificity for monomeric PRL best reflect the physiologic state of the patient.
Our aim was to characterize the immunoreactivity of the various molecular forms of PRL and to develop an assay that preferentially recognizes the biologically active monomeric form of PRL.
Serum was obtained from 41 patients with PRL concentrations >500 mIU/L (conversion factor, 1 ng/mL = 36 mIU/L) and 14 with concentrations within the reference interval. All samples were fractionated by gel filtration, precipitated with polyethylene glycol (PEG), and assayed with different incubation times. Of the samples, 24 contained mainly 23-kDa PRL, whereas 31 contained macroprolactin as the major form of PRL.
Serum samples were precipitated with PEG (10), and the supernatants were assayed for PRL by the DELFIA immunofluorometric assay (Perkin-Elmer-Wallac). The results were compared with those of the direct serum PRL assay, and recovery of PRL in the supernatant was calculated. Serum samples (0.2–0.5 mL) were fractionated by gel filtration at 4 °C on a 2 x 70 cm Sephacryl S-200 column (Amersham Pharmacia Biotech) at a flow rate of 15 mL/h. Fractions of 1 mL were collected and assayed for PRL. Serum samples were also fractionated on protein G-Sepharose (Amersham Pharmacia Biotech). Bound proteins were eluted with elution buffer (0.1 mol/L glycine-HCl, pH 2.7) in sixteen 1-mL fractions into tubes containing 50 µL of neutralizing solution (1.0 mol/L Tris-HCl, pH 9.0). The fractions were assayed for PRL with an incubation time of 90 min in one step. Incubation times of 5, 15, 30, and 60 min were also used. In some samples, we also measured PRL with the Bayer ACS180 PRL assay (Bayer Diagnostics). This assay uses a reaction time of 5 min with the labeled antibody and 2.5 min with both the labeled and solid-phase antibodies.
In gel filtration of serum, BB-PRL eluted as 150–200 kDa, whereas B-PRL eluted as 40–60 kDa, and L-PRL as 23 kDa. In two samples, the molecular mass of BB-PRL was 
200–250 kDa. Mean (SD) recovery of PRL in the fractions was 86% ± 11%. The samples were categorized into three groups; group A samples contained <4% BB-PRL and 15–30% B-PRL; group B contained 7–56% BB-PRL and 10–68% B-PRL; and group C contained mainly BB-PRL (60–90%).
In macroprolactinemic samples, the PRL results correlated inversely with the incubation time. With the 5-min assay, the apparent PRL concentration in group C was 44% ± 2% of that with 90-min incubation, in group B it was 61% ± 3%, and in group A it was 89% ± 3% (Fig. 1A
). Even in group A, which contained mainly L-PRL, the results were 10% lower with the 5-min incubation time than with the 90-min incubation time (P = 0.0042). When BB-PRL and L-PRL isolated by gel filtration were analyzed with the 5-min assay, the result for BB-PRL was 42% ± 3% of that obtained with 90-min incubation, whereas that for L-PRL was 97% ± 5% of that obtained with 90-min incubation. Sera containing various amounts of macroprolactin were precipitated with PEG and also fractionated by gel filtration. The PRL concentration in the PEG supernatant correlated strongly (r = 0.96) with that in the L-PRL fraction separated by gel filtration. We could identify macroprolactinemic samples by assaying them with both 5- and 90-min incubation times. If the result with the 5-min assay was <60% of that with the 90-min assay, the sample contained >50% BB-PRL (Fig. 1B
). Samples containing BB-PRL with a molecular mass of 200–250 kDa behaved in the same way as samples containing 160- to 200-kDa BB-PRL. The ratio of PRL measured before and after PEG precipitation correlated with ratio of PRL detected by the 5- and 90-min assays, respectively (r = 0.92; Fig. 1C
). These results show that it is possible to replace the PEG method with the short incubation assay for detection of macroprolactinemia.
|
Eighteen samples were analyzed with the ACS180 PRL assay, which has been reported to measure lower PRL concentrations than the DELFIA assay in macroprolactinemic samples (12). The differences between the ACS180 and the DELFIA assays increased as the proportions of BB-PRL and B-PRL in the sample increased. In samples containing >70% macroprolactin, the ACS results were 20–40% of those obtained by the DELFIA assay. However, even in samples containing mainly L-PRL, the results obtained by the ACS method were 20% lower, suggesting a difference in calibration. On the basis of this correlation, the ACS assay appears to underestimate BB-PRL by 50%.
Differences in PRL results for macroprolactinemic samples have been assumed to be caused by different recognition of macroprolactin by various antibodies (10)(12)(13). Although this most probably is a contributing factor, our results show that B-PRL and BB-PRL react more slowly than normal PRL in immunoassays and that incubation time therefore strongly affects the result for macroprolactinemic samples. When using a short incubation time with the DELFIA assay, we preferentially measured L-PRL. In samples containing 85% BB-PRL and B-PRL, the results with the 5-min incubation time were 47% of those obtained with the 90-min incubation. In samples containing mainly L-PRL, the corresponding value was 90%. The effect of incubation time was confirmed with BB-PRL and L-PRL separated by gel filtration. Our results suggest that differences in incubation time explain some discrepancies in PRL results obtained by various PRL assays (12). This is likely to be an important factor in the ACS assay, which uses a two-step incubation: 5 min with a labeled antibody and then 2.5 min after the addition of the solid-phase antibody. This assay appears to recognize BB-PRL approximately to the same extent as the 5-min DELFIA assay. We could not study the effect of incubation time in any other assay because the time settings cannot be modified in other analyzers. Because of this, we could not evaluate the effects of other differences in assay design. Interestingly, the Roche Elecsys assay, which uses a total incubation time of 18 min, has been reported to measure macroprolactin even more efficiently than the DELFIA assay (14). This could be attributed to the selection of antibodies used, but also to assay design. In the Elecsys assay, the labeled and biotinylated antibodies first react with prolactin in solution for 9 min. After the addition of streptavidin-coated microparticles, which bind the biotinylated antibody, incubation is continued for another 9 min before separation and measurement of the signal. Reaction in solution and efficient mixing of the microparticles may be expected to provide more rapid antibody binding than the microtitration plate format used in the DELFIA assay. The incubation time of 18 min is also sevenfold that of the second step in the ACS assay (2.5-min incubation with microparticle-bound antibody).
There is strong evidence that BB-PRL is a PRL-IgG complex, and our results confirmed that most of BB-PRL is retained on protein G-Sepharose. Recently, a PRL-binding protein was found in human serum and milk, and on average, 36% of the PRL in the circulation was found to consist of PRL bound to PRL-binding protein (15). In agreement with this, we found by gel filtration that 15–40% of PRL consisted of B-PRL.
Removal of macroprolactin by PEG precipitation is used as a screening method for macroprolactinemia (13)(16)(17), and our results confirmed that PRL measured after PEG precipitation consists mainly of L-PRL. The precipitation step is a manual procedure, which only some laboratories routinely perform. An alternative approach is to retest samples with increased PRL concentrations with a shorter incubation time. A much lower result with the short incubation time indicates the presence of BB-PRL. If our results can be reproduced with other assays, it should be possible to adapt this approach to certain immunoanalyzers to automatically recognize samples with increased PRL attributable to macroprolactinemia.
There are several possible explanations for the effect of incubation time on the PRL results. In addition to slower reaction kinetics of the big forms, BB-PRL and B-PRL may gradually dissociate from their complexes, exposing epitopes hidden in the complexes. In both cases, lower results may be expected with shorter incubation times.
In conclusion, the concentration of PRL measured by immunoassay is dependent on the incubation time if the sample contains macroprolactin. Measurement of PRL with two different incubation times is a potentially useful method for identification of samples containing macroprolactin.
Acknowledgments
This study was supported by a grant from Wilhelm and Else Stockmann’s Foundation.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) includes samples containing 5–59% BB-PRL, and group C (
) includes samples containing 60–90% BB-PRL. The error bars represent the SE. The significance of the differences between various time points is as follows: *, P <0.05; **, P <0.01; ***, P <0.001. (B), effect of the proportions of L-PRL and BB-PRL on the PRL concentration measured by the 5-min compared with the 90-min assay. (C), comparison of the proportion of PRL measured after precipitation with PEG and PRL and that measured with the 5- and 90-min DELFIA assays, respectively.