Correlation and comparison of Nb 2 lymphoma cell bioassay with radioimmunoassay for human prolactin*

Transcription

1 FERTILITY AND STERILITY Copyright 1984 The American Fertility Society Printed in U.8A. Correlation and comparison of Nb 2 lymphoma cell bioassay with radioimmunoassay for human prolactin* Marappa G. Subramanian, Ph.D.t:!: Nicholas J. Spirtos, D.O.t Kamran S. Moghissi, M.D. t David M. Magyar, D.O.t Maria F. Hayes, M.D. t Richard R. Gala, Ph.D. C. S. Matt Center for Human Growth and Development, Wayne State University School of Medicine, Detroit, Michigan Serum samples from groups of men and women with normal and elevated prolactin (PRL) levels were assayed by radioimmunoassay (RIA) and by Nb 2 lymphoma cell bioassay (BA) for the presence of PRL. Because the Nb 2 lymphoma cells respond to both PRL and growth hormone, BA for PRL activity was carried out before and after neutralization of growth hormone in the serum samples. There were excellent correlations between RIA and BA both in euprolactinemic (r = ; P < 0.002) and hyperprolactinemic (r = ; P < 0.001) subjects. On an absolute basis, RIA and BA values were similar in the euprolactinemic group (6.6 ± 0.8 versus 6.2 ± 1.0), whereas in the hyperprolactinemic group, RIA values were significantly higher than the BA results (89.41 ± 22.4 versus 62.1 ± 212). The two assay systems also appeared to correlate better in women who were hyperprolactinemic, with obvious menstrual cycle disturbances, than in hyperprolactinemic women without menstrual cycle disturbances. Fertil Steril 42:870, 1984 Following the development of the rat Nb 2 node lymphoma cell bioassay (BA) for prolactin (PRL) and other lactogenic hormones by Tanaka et al. in 1980,1 it is now feasible to measure bioactive PRL in serum, milk, and pituitary homogenates. 2 A recent clinical report indicated that PRL values obtained by radioimmunoassay (RIA) and BA were in close agreement under a number of physiologic situations and after PRL stimulation test- Received July 6, 1984; revised and accepted August 13, *Supported in part by NIH grant HD (to R. R. G.). tdivision of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology. :t:reprint requests: Marappa G: Subramanian, Ph.D., C. S. Mott Center for Human Growth and Development, 275 East Hancock Avenue, Detroit, Michigan Department of Physiology. 870 Subramanian et al. Bioassay and RIA of hprl ing. However, in a woman who had hyperprolactinemia and also a larger proportion of "big, big" PRL, a high BA/RIA ratio was observed. 3 The purpose of this study was to investigate the relationship between the two assays comparing PRL levels in euprolactinemic and hyperprolactinemic individuals. While confirming that there is good correlation between RIA and BA, the present report provides evidence that on an absolute basis, the values obtained may be different between the two assays in some women evaluated for endocrinologic and/or infertility disorders. MATERIALS AND METHODS Serum samples from groups of men and women with normal PRL levels, hyperprolactinemic men, and hyperprolactinemic women with and Fertility and Sterility

2 without galactorrhea were assayed for PRL by RIA and BA before and after neutralizing growth hormone (GH) activity. RADIOIMMuNoASSAY Labeling of PRL was performed utilizing the glucose oxidase with lactoperoxidase method. 4 The double antibody technique was utilized for RIA to measure the PRL level in serum samples. 5 To minimize the variability in PRL values, all the samples were assayed in the same assay, and the intraassay variation was 5.7%. The reagents used in the PRL RIA were NIAMDD-hPRL-I-6 for labeling; human pituitary PRL preparation 75/504 from the National Institute for Biological Standards and Control, London, England, for the reference preparation; NIAMDD-anti-hPRL-3 at 1:100,000 dilution as the first antibody; and antirabbit gamma globulin, purchased from Antibodies, Inc., Davis, CA, as the second antibody. BIOASSAY The BA procedure has been previously described. 2 PRL activity was measured in the serum samples before and after neutralizing GH activity. Anti-GH serum (lot #AFP ) generated in rabbits was utilized to neutralize the GH activity in the serum. This antiserum had a crossreactivity of 0.4% against human PRL and < 0.1 % against human follicle-stimulating hormone, human luteinizing hormone, and thyroidstimulating hormone. Because the anti-gh serum contained Merthiolate, which was found to be toxic to Nb2 rat lymphoma cells, the Merthiolate was removed by dialyzing the anti-gh serum against 0.1% bovine serum albuminlphosphatebuffered saline for 3 days with buffer changes twice daily. Five milliliters of antiserum was dialyzed against 800 ml of 0.1 % bovine serum albumin/phosphate-buffered saline for each buffer change. Following dialysis, the anti-gh serum was diluted to 1:100, and 50 f.li of this diluted antiserum was added to the culture medium. This amount of anti-gh serum (50 f.li of 1:100) was capable of neutralizing 4 ng of GH in the BA well (equivalent to 80 ng of GH/ml of serum) and was more than adequate to neutralize the GH activity that was present in the serum used in the assay (normal GH in adults is < 5.0 ng/ml). For bioassaying serum samples which were not neutralized for GH activity, an equivalent amount of normal rabbit serum was added to the BA wells. At the dilutions used, normal rabbit serum did not stimulate cell growth. All serum samples were assayed at two dilutions each in duplicate; The dilution of serum samples ranged from 1:20 and 1:40 for euprolactinemic samples to 1:1000 and 1:2000 for hyperprolactinemic samples. The assay was performed by adding 50 f.li of various concentrations of human PRL standard (75/504) or serum samples diluted with Fischer's medium to duplicate wells of tissue culture plates. The reference preparation (75/504) used in the assay did not contain any significant amount of GH activity to influence the BA (650 miu ofprl activity and miu of GH activity per vial of material obtained). The plates were incubated in a humidity controlled incubator at 37 C in an atmosphere of 95% air, 5% CO2 for 72 hours. Then the cell suspension from each well (1 ml) was transferred with a Pasteur pipette into a clean polystyrene vial specifically designed for use with Coulter Counter (Fisher Scientific, Pittsburgh, PA) and rinsed with 1 ml of a particle-free electrolyte solution (Hematall, Fisher Scientific); an additional 8 ml of electrolyte solution was then added to the counting vial. The cells were then counted in a Model B Coulter Counter using a 70-f.Lm aperture and a 100-f.Ll mercury manometer. The cell counts from the Coulter Counter were manually entered in a MAAC minicomputer (Micromedic Systems Inc~, Horsham, PA) and analyzed using a modified point-to-point RIA program. The intraassay and interassay differences for the BA expressed as coefficient of variation were 8.8% and 9.6%, respectively. Statistical analysis of the data was performed by calculating the correlation coefficients and performing paired t-tests as indicated. RESULTS Pearson correlation coefficients and significance levels between RIA versus BA for PRL values before and after neutralization of GH activity in the euprolactinemic and hyperprolactinemic groups are summarized in Table 1. Highly significant correlations between RIA and BA after GH activity neutralization were observed in both groups, whereas the correlation between RIA and BA before GH activity neutralization was not significant in the euprolactinemic group. PRL values (mean ± standard error) obtained by RIA and by BA with and without the addition of anti-gh are summarized in Table 2. Paired Subramanian et al. Bioassay and RIA of hprl 871

3 Table 1. Pearson Correlation Coefficients and Significance Levels Between RIA Versus BA (PRL + GH) and RIA Versus BA (PRL) Group Euprolactin- 12 emic Hyperprolac- 14 tinemic n RIA versus BA (PRL + GH)" y p < ay, correlation coefficient; P, significance level. RIA versus BA (PRL)a y p < < t-tests were performed with these samples in two categories: euprolactinemic and hyperprolactinemic. In the euprolactinemic group the PRL levels estimated by RIA and BA following GH activity neutralization were similar (6.6 ± 0.8 versus 6.2 ± 1.0). The higher activity by BA obtained in the euprolactinemic group before the GH activity neutralization was attributed to the presence of GH in these samples. Similar comparison in the hyperprolactinemic group yielded somewhat paradoxical results: PRL activity by RIA was significantly higher than the BA activity measured in GH-neutralized serum samples (Table 2). Of the six women in the hyperprolactinemia without galactorrhea group, there was good agreement between the RIA and BA values for PRL in three women (K. M., R. G., and L. C.). These three women, who had elevated PRL levels by both RIA and BA, showed irregularity in their menstrual cycles or an infertility disorder. The other three women (S. G., T. M., and N. B.), who had high PRL levels by RIA but upper normal PRL levels by BA, showed normal menstrual cycles (Table 3). Close agreement between RIA and Table 2. Comparison of RIA and BA Group Description n RIA 1 Euprolactinemic males ± Euprolactinemic females ± Hyperprolactinemic males ± Hyperprolactinemic females ± 6.2 (no galactorrhea) 5 Hyperprolactinemic females ± 63.9 (galactorrhea) Euprolactinemic patients ± 0.8 (groups 1 + 2) Hyperprolactinemic patients ± 22.4 (groups ) BA was also observed in three of four women (D. A., C. Y., and B. T.) with hyperprolactinemia and galactorrhea. The one patient (W. A.) in this group with a normal menstrual cycle exhibited a BA value of < 50% of the value obtained by RIA (Table 3). DISCUSSION The BAiRIA ratios ranged between 0.30 and 1.67 for the serum samples tested in this study. This is in good agreement with those obtained by Rowe et al. 3 for human serum samples and also agrees well with those reported for plasma, serum, anterior pituitary homogenate, and milk samples in the rat. 2 The highly significant correlation between the RIA and BA adds further credence to the validity of the two assay systems for measuring PRL (Table 2). On an absolute basis, the results obtained with the two assay systems demonstrate some differences. The values obtained by the RIA and BA for the euprolactinemic group were similar (6.6 versus 6.2; Table 2). However, in the hyperprolactinemic group the value obtained by RIA was significantly higher than that obtained by the BA (89.41 versus 62.1; Table 2). This may possibly be a reflection of the two assay systems recognizing different regions of the secreted PRL molecule in hyperprolactinemic women. It was also noted that in the hyperprolactinemia without galactorrhea group, three of six women had elevated PRL levels by both RIA and BA, whereas the other three had elevated PRL levels by RIA but upper normal levels by the BA. These latter three women, ex- PRL levelsa BA (PRL + GH) BA (PRL) RIA versus RIA versus BA (PRL + GH) BA (PRL) 10.4 ± ± ± ± ± ± ± ± ± ± ± ± b ± ± O.OOl e P amean ± standard error. Values are in nanograms per milliliter. bp = 0.035, level of significance. cp = 0.001, level of significance. 872 Subramanian et ai. Bioassay and RIA of hprl Fertility and Sterility

4 Table 3. Serum PRL Levels Measured by RIA and by BA With and Without GH Activity Neutralized, BAIRIA Ratios, and Reproductive Cycle Stages in Women With Hyperprolactinemia Without Galactorrhea a~d Hyperprolactinemia With Galactorrhea Patient RIA PRL(ng/ml) BA(PRL + GH) Hyperprolactinemia + no galactorrhea K.M R. G L. C S.G T.M N.B Hyperprolactinemia + galactorrhea D.A C. Y B. T W.A BA(PRL) BA (PRL)/RIA Menstrual cycle and/or fertility status Amenorrhea, first-degree infertility Normal cycle, second-degree infertility Irregular and short cycles, firstdegree infertility Normal cycle, first-degree infertility Normal cycle, second-degree infertility Normal cycle, second-degree infertility Amenorrhea Oligomenorrhea/amenorrhea Normal cycle, occasional abnormal uterine bleeding Normal cycle hibiting hyperprolactinemia by RIA and euprolactinemia by BA, had normal menstrual cycles, whereas two of the three patients with hyperprolactinemia by both RIA and BA displayed irregularity of the menstrual cycle. It is of interest to note that a similar trend was also evident in the group of women with hyperprolactinemia and galactorrhea (Table 3). It is possible that treatment intervention may not be necessary in women hyperprolactinemic by RIA and euprolactinemic by BA, and further studies utilizing a larger patient population will be directed toward testing this hypothesis. The differences in the PRL levels by the RIA and BA may be related to the heterogeneity of circulating PRL.6 In a study which utilized RIA for measuring immunoreactivity and radioreceptor assay for measuring bioactivity ofprl, it was suggested that women with higher PRL activity by RIA and lower PRL levels by BA could have a larger proportion of "big, big" PRL, which may have lower biologic activity.7 Close correlation between RIA and BA was reported in normal, hyperprolactinemic, and under a variety of stimulated conditions; however, in one patient with a high proportion of "big, big" PRL, a BAIRIA ratio of 2.47 was observed 3 contradicting the study of Farkouh et al. 7 More recently, it was reported that both in normal and tumor patients, good correlation between RIA and BA was observed for the three molecular weight variants of PRL. 8 Subsequent investigatioll3 will examine the het- erogeneity profile of PRL in circulation, RIA/BA correlation, and its relationship to the menstrual cycle and reproductive disorders. Acknowledgments. We wish to thank Mr. Steven Waters for technical assistance and Ms. Jane Olson for statistical analysis. The gift ofhprl RIA materials by the National Pituitary Agency, National Institutes of Health, Bethesda, MD, and the National Institute for Biological Standards and Control, Holly Hill, Hampstead, London, England, is gratefully acknowledged. REFERENCES 1. Tanaka T, Shiu RPC, Gout PW, Beer CT, Noble RL, Friesen HG: A new sensitive and specific bioassay for lactogenic hormones: measurement of prolactin and growth hormone in human serum. J Clin Endocrinol Metab 51: 1058, Lawson DM, Sensui N, Haisenleder DH, Gala RR: Rat lymphoma cell bioassay for prolactin: observations in its use and comparisons with radioimmunoassay. Life Sci 31:3063, Rowe RC, Cowden EA, Faiman C, Friesen HG: Correlation of Nb 2 bioassay and radioimmunoassay values for human serum prolactin. J Clin Endocrinol Metab 57:942, Tower BB, Clark BR, Rubin RT: Preparation of polypeptide hormones for radioimmunoassay using glucose oxidase with lactoperoxidase. Life Sci 21:959, Midgley AR Jr: Radioimmunoassay: a method for human chorionic gonadotropin and human luteinizing hormone. Endocrinology 79:10, 1966 Subramanian et ai. Bioassay and RIA of hprl 873

5 6. Whittaker PG, Wilcox T, Lind T: Maintained fertility in a patient with hyperprolactinemia due to big, big prolactin. J Clin Endocrinol Metab 53:863, Farkouh NH, Packer MG, Frantz AG: Large molecular prolactin with reduced receptor activity in human serum: high proportion in basal state and reduction after thyrotropin-releasing hormone. J Clin Endocrinol Metab 48: 1026, Whitaker MD, Klee GG, Kao PC, Randal RV, Heser DW: Demonstration of biological activity of prolactin molecular weight variants in human sera. J Clin Endocrinol Metab 58:826, Subramanian et al. Bioassay and RIA of hprl Fertility and Sterility