Serum pattern of different molecular forms of prolactin during normal

Human Reproduction vol.8 no.10 pp.1617-1622, 1993

Serum pattern of different molecular forms of prolactin during normal human pregnancy

Fernando Larrea1, Isabel Mendez, Adalberto Parra2 and Antonio Espinosa de los Monteros2

'To whom correspondence should be addressed

In this study, the molecular heterogeneity of prolactin was analysed in serum from normal women throughout pregnancy. Lectin affinity chromatography and denaturing polyacrylamide gel electrophoresis under reducing and nonreducing conditions, followed by Western blotting and immunostaining were used to resolve and identify the molecular variants of prolactin. During the first trimester, large molecular forms (64 and 53 kDa) and those corresponding to glycosylated and non-glycosylated prolactin (25 and 23 kDa, respectively) were present either under reducing or non-reducing conditions. The 64 and 23 kDa were the predominant species at this stage of gestation. As pregnancy progressed, the 64 kDa variant, which did not bind to concanavalin A, decreased until disappeared at the third trimester of gestation. The unbound/bound ratio of serum prolactin to concanavalin A increased only at the third trimester; however, the relative proportions of concanavalin A-bound prolactin did not show statistically significant changes along the gestational period. The results demonstrated the occurrence of changes in the heterogeneity of prolactin during gestation and further confirmed previous observations that various forms of non-glycosylated prolactin are indeed the predominant species in serum from normal women throughout pregnancy. Key words: glycosylated prolactin/pregnancy/prolactin heterogeneity

Introduction It is now well established that human prolactin is encountered in a variety of molecular forms that include the 23 kDa monomer, the polymeric forms, the glycosylated 25 kDa (G-prolactin) and the cleaved variants (Mittra, 1980; Sinha etal., 1985; Sinha, 1992). All these forms have been detected in the pituitary gland and plasma but their proportions and biological significance are not yet fully determined. In women with macroprolactinaemia (Whittaker etal., 1981; Andino etal., 1985; Larrea etal., 1985), we have demonstrated that the 150 kDa polymeric form © Oxford University Press

Materials and methods Subjects and samples The study protocol was approved by the Human Investigation Committee of the Institute Nacional de la Nutricidn SZ, and all subjects were informed and asked if they were willing to participate in this study. Serum samples were obtained from 19 women with clinically normal pregnancies (first trimester, n = 6; second trimester, n = 7; third trimester, n = 6), ranging in age from 19 to 37 years. Blood samples were obtained from an antecubital vein as part of routine hospital procedure. The serum obtained after centrifugation at 2000 g at 4°C was stored at -20°C until assayed. Materials Acrylamide, A^-dMyltartardiamide (DATD), A^-methylenebisacrylamide(bis), Af^^^-tetrarnethylethylenediamine (TEMED), ammonium persulphate, SDS, affigel blue, nitrocellulose paper, 2-mercaptoethanol, and Coomassie Blue R-250 were obtained from Bio-Rad Laboratories (Richmond, CA, USA). Bovine 1617

Downloaded from http://humrep.oxfordjournals.org/ at Universidad Nacional AutÃ³noma de MÃ©xico on January 15, 2014

Department of Reproductive Biology, Institute Nacional de la Nutrition Salvador Zubiran, Vasco de Quiroga 15, Mexico 14000 D.F. and department of Endocrinology, Institute) Nacional de Perinatologia, Mexico D.F.

of serum prolactin remains increased during pregnancy with a shift towards monomeric prolactin near delivery (Larrea et al., 1989). It is believed that this increase in the small-molecularweight variety forms at the expense of high-molecular-weight forms during pregnancy involve thiol-disulphide interchange mechanisms among prolactin molecules in the pituitary (Beneviste et al., 1979; Lorenzon et al., 1984; Larrea et al., 1987; Mena etal, 1992). Changes in prolactin heterogeneity have also been described in normal women during pregnancy (Pansini etal., 1985; Markoff et al., 1988). Under this condition, there is a gradual increase in the prolactin species of low molecular weight and an apparent decrease in the relative proportions of G-prolactin compared to monomeric 23 kDa prolactin as pregnancy progresses. Since pregnancy represents a physiological condition that stimulates prolactin secretion and changes its heterogeneity, we thought it was of importance to analyse the molecular heterogeneity of serum prolactin with regard to lectin absorbance in normal women during different stages of pregnancy. To accomplish this objective, we have used sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) under reducing and non-reducing conditions followed by immunostaining with specific prolactin antiserum. This technical approach allowed us to distinguish between the major molecular components of serum prolactin.

F.Larrea et al.

serum albumin (BSA), Tris, glycine, protein A and Tween-20 were purchased from Sigma Chemical Co. (St Louis, MO, USA). Concanavalin A (Con A)—Sepharose, and molecular-weight markers for various proteins were obtained from Pharmacia Fine Chemicals (Uppsala, Sweden). Rabbit anti-human prolactin antiserum (NIDDK-anti-hPRL-3, AFP-C11580), human prolactin antigen (AFP-9900) and human prolactin reference preparation (AFP-2312C) for prolactin radioimmunoassay (RIA) were a gift from the US National Hormone and Pituitary Program. Antihuman prolactin antiserum (VLS-2) used in immunoprecipitations from serum and immunostaining was kindly supplied by Dr Y.N.Sinha, Whittier Institute (La Jolla, CA, USA).

Fractionation of serum samples by Con A affinity chromatography Serum samples were loaded on a 3-ml bed volume of Con A —Sepharose in a 1.0 X 20 cm column. The column was equilibrated and washed with 0.05 mol/1 phosphate-buffered saline (PBS), pH 7.5 at 4°C. Bound material was eluted with 0.2 mol/1 methyl-a-D-mannopyranoside (Sigma Chemical Co.) in PBS (Larrea et al., 1992). The proportion of prolactin in the free and bound areas was calculated by RIA after lyophilization of each immunoreactive peak. Prolactins in the free and bound fractions were immunoprecipitated, electrophoresed and immunoprobed, as described below. Western blotting and immunostaining Human prolactin from serum samples or column eluates was immunoprecipitated using the VLS-2 anti-human prolactin

Statistical analysis Student's unpaired and paired f-tests were performed as appropriate to determine the significance of differences between quantitative cohorts. A statistical significance (a) level of P < 0.05 was chosen a priori.

Results Figure 1 shows representative Con A elution profiles of serum samples from women at different trimesters of pregnancy. As depicted, serum from the first (A), second (B) and third (C) trimesters have two peaks of immunoreactive prolactin on Con A chromatography, which we have called the unbound and bound fractions, respectively. The immunological identity of prolactin in sera from subjects throughout different stages of pregnancy was assayed over a range of dilutions. Serial dilutions of whole serum from women during pregnancy [weeks 6 (slope -2.34), 10 (slope -2.34), 20 (slope -2.35) and 28 (slope -2.38)] resulted in displacement curves parallel to one another and to the standard curve (Rp-1) (slope —2.39), which implies a close

60

70

60

30

30 30

10

10

10

10

20

10

FRACTION

20

SO

10

20

30

NUMBER

Fig. 1. A representative elution profile of human serum (HPRL) after chromatography on Con A—Sepharose. Serum from women during the first (A), second (B), and third (C) trimesters of pregnancy were applied on 3-ml bed volume of Con A—Sepharose, and the free and bound fractions were pooled after their identification by specific radioimmunoassay. 1618


Hormone measurements Prolactin in serum and column eluates was measured by double antibody RIA using reagents and protocols from the US National Hormone and Pituitary Program, as described previously (Larrea et al., 1989). The results were expressed as /*g/l human prolactin Rp-1. The intra- and inter-assay coefficients of variation were 5.2 and 7.3%, respectively. Serum samples were assayed at two different dilutions per duplicate. In some instances, serum samples and column eluates were gradually diluted and assayed per triplicate to ascertain parallelism with the standard curve.

antiserum for 72 h at 4°C. The precipitates, obtained after the addition of goat anti-rabbit y-globulin as previously described (Sinha et al., 1984), were resuspended and heated for 5 min at 95°C in Laemmli's sample buffer (Laemmli, 1970) in the presence or absence of 10% 2-mecaptoethanol. Samples were electrophoresed on 10% polyacrylamide gels containing SDS using the discontinuous buffer system of Laemmli (1970). After electrophoresis, proteins were electroblotted onto nitrocellulose membranes in PAGE buffer containing 20% methanol without SDS (Towbin, 1970). Membranes were fixed, blocked and incubated overnight at room temperature in 1:1000 final dilution of anti-human prolactin serum VSL-2 in 0.01 mol/1 Trisbuffered saline, pH 7.4, containing 0.1% Tween-20 and 3% Carnation® powdered non-fat dry milk. Immunocomplexes were visualized by autoradiography after incubation in 125I-labelled protein A (200 000 c.p.m./ml). The excess radioactivity was removed by several washed in Tris-buffered saline, and the paper was dried and autoradiographed.

Prolactin heterogeneity in pregnancy

Figure 3 illustrates the pattern of serum prolactin during the first trimester using SDS-PAGE under reducing conditions. In total serum (lane l), four specific prolactin bands were identified 40

in all subjects with apparent molecular weights of 64, 53, 25 and 23 kDa. The 64 and 23 kDa forms represented the principal bands. Lanes 2 and 3 show the analysis of the free and bound fractions to Con A, respectively. As can be seen, the 64 and the 53 kDa forms Qanes 1 and 2) did not bind to the lectin (lane 3) and only the 25 kDa band was present when Con A-bound fractions were analysed. Lane 4 represents the non-specific band and contains a serum sample, as in lane 1, but immunoprecipitated with hyperimmune rabbit serum. Under non-reducing conditions (not shown) there were several non-specific high-molecularweight bands, which were also present when serum was substituted for buffer in the immunoprecipitation step. The 64, 53, 23 and a very light 25 kDa bands were observed in serum and Con A-free samples. The electrophoretical analysis of Con A-bound serum prolactins under non-reducing conditions showed the absence of the 64 kDa form in a similar manner as described in the presence of mercaptoethanol (Figure 3, lane 3). The analysis of prolactin molecular variants during the second trimester of gestation is shown in Figure 4. At this stage of pregnancy, prolactin heterogeneity analysed under reducing conditions was characterized by an apparent increase in the density of the 25 kDa form and by an important decrease in that corresponding to the 64 kDa prolactin band (Figure 4, lane 1). In some subjects during the second trimester, a 21 and a 16 kDa prolactin band appeared very faintly, which, as in the case of the 64, 53 and 23 kDa forms, did not apparently bind to Con

B

i A

30-

c

as

it

.-1

D 20 O CD

10 12.5 25 0.1

50

lOOjll

1.0 hPRL (ng/tube)

12.5

10.0

0.1

25

50

IOOJJI

1.0 hPRL (ng/tube)

Fig. 2. Dose-response curves of serum Con A-unbound (A) and bound (B) fractions obtained during the first (I), second (II), and third (HI) trimesters of pregnancy. Column eluates were dialysed, concentrated, diluted and assayed at four different dilutions per triplicate and compared to Rp-1 used as standard. HPRL = human prolactin.

64

53-

25 23-

Table I. Serum prolactin (PRL) concentrations and relative percentages of bound (G-PRL) and unbound/bound ratios of circulating PRL to concanavalin A (Con A) throughout pregnancy Stage

Serum PRL

G-PRL

Con A unbound/bound

Trimester 1 Trimester 2 Trimester 3

157 ± 56a 240 ± 90 308 ± 64 b

27.3 ± 7 28 ± 12 17 ± 7.3

2.7 ± 0.45 2.6 ± 0.98 4.8 ± 0.81 cd

"Mean ± SD of six individuals per group. b P < 0.05 versus trimester 1. C P < 0.05 versus trimester 1. A P < 0.005 versus trimester 2.

1 2

3

4

Fig. 3 . Autoradiography of serum samples corresponding to the first trimester of pregnancy. Total serum (Lane 1) and Con A-free (lane 2) and bound (lane 3) serum fractions were immunoprecipitated and analysed by SDS-PAGE under reducing conditions. Lane 4 shows the non-specific staining obtained when hyperimmune rabbit serum was used in the immunoprecipitation procedure. 1619


immunological similarity of the serum samples throughout gestation. Figure 2 shows the Con A-unbound (A) and bound (B) fractions of serum samples obtained at all trimesters of pregnancy (roman numbers) assayed over a range of dilutions and compared with a prolactin international standard (Rp-1). As in the case of whole serum the Con A-unbound and bound fractions were parallel to each other and to the standard curve at all the stages of pregnancy studied. The relative proportions of these fractions along pregnancy as expressed as the unbound/bound ratio of serum prolactin to Con A is shown in Table I. As can be seen, immunoassayable serum prolactin concentrations were augmented since the first trimester and continued to rise as pregnancy progressed. The relative unbound and bound proportions of serum prolactin were similar during the first and second trimester, as demonstrated by the similar Con A unbound/bound ratios. However, a significant increase in this ratio was observed at the third trimester of gestation. The relative proportions of Con A-bound prolactin [G-prolactin], expressed as the percentage of the total prolactin eluted from Con A (Table I), did not show statistical differences between the three trimesters studied. The percentage of G-prolactin at the third trimester was lower than that observed at early stages, but this difference did not reach statistical significance.

F.Larrea et at.

A. Under non-reducing conditions, an increase in the 25 kDa and a decrease in the density of the 64 kDa prolactin species was observed. In all women at this stage of gestation, the 23 kDa monomeric prolactin was the predominant prolactin species in blood.

*

-

•

-

92.5 66.2

64

Discussion

53

45 31 21 .5

Fig. 4. Autoradiography of total serum and Con A-free and bound serum fractions corresponding to the second trimester analysed by SDS—PAGE under reducing conditions. Legend specifications are as in Figure 3. The electrophoretical migration of molecular-weight standards are indicated on the right.

Human prolactin has been shown to be structurally heterogeneous in blood, pituitary extracts and other biological fluids. This heterogeneity has been attributed to genetic and post-translational events, as well as to structural rearrangements of prolactin molecules in the pituitary gland (Sinha, 1992) or in peripheral target tissues (Compton and Witorsch, 1983; Wong et al., 1986; Clapp, 1987; DeVito et al., 1992). The physiological significance of such a polymorphism is unknown but is probably related to the poly functional actions of the hormone. In the present study, we have further demonstrated the presence of several immunoreactive prolactin components in human sera during pregnancy. These components, which were mainly represented by the 23 and 25 kDa forms, have been previously observed under a variety of physiological conditions including pregnancy. The results presented herein demonstrated an increase in the unbound/bound ratio of prolactin to Con A only during the third trimester of pregnancy, which was probably due to

m 64

9 2 .5 6 6 .2

64 —

45 25 23 16

6—

3 1

w

21 . 5 14 . 4

1

2

3

4

5

6

7

Fig. 5. Autoradiography of serum samples from the third trimester of pregnancy analysed by SDS—PAGE under reducing conditions. Lanes 1 —4 as in Figure 3. Lanes 5 and 6 represent the same samples as in lanes 1 and 2 but overexposed. Lane 7 shows the non-specific prolactin.

1620


During the third trimester of gestation (Figure 5), total serum and Con A-free prolactin analysed under reducing conditions were mainly composed of the 23 kDa monomeric form (lanes 1 and 2, respectively). In some subjects at this stage of their pregnancy the 16 kDa species was evident only when total serum was immunoprecipitated, and a new 27 kDa species appeared in the Con A-bound fractions (lane 3). As depicted in Figure 5, the 53 and 64 kDa bands disappeared (lanes 1 and 2) and the 64 kDa was only evident, in total serum and Con A-free samples, when the X-ray film was overexposed (lanes 5 and 6, respectively). Under non-reducing conditions (data not shown), the main prolactin band was that corresponding to monomeric 23 kDa without the presence of the 64, 16 and 27 kDa forms.

Prolactin heterogeneity in pregnancy

Human prolactin components in total and Con A-free and bound serum samples from normal women during pregnancy were analysed by Western blotting and immunostaining. In all stages of pregnancy and irrespective of the experimental conditions used in this study, the major prolactin species in serum corresponded to monomeric 23 kDa prolactin. During the first trimester, the presence of an intense 64 kDa mercaptoethanolstable band could be demonstrated. This form was probably not the result of random molecular aggregation among prolactin species, since it was present under denaturing conditions and even when completely reduced by mercaptoethanol. This 64 kDa prolactin form, that did not bind to Con A and probably represents a mercaptoethanol-resistant dimer, showed an important reduction at the second trimester and almost disappeared at the third trimester of gestation. A similar observation was reported by Fukuoka et al. (1991) in maternal blood and amniotic fluid from full-term normal deliveries. It is not known whether the 50 and 60 kDa prolactins identified by gel permeation chromatography and reported by Pansini et al. (1985) and Shoupe et al. (1983), respectively, are the same as the 64 kDa form observed in this study. The observation that 50 and 64 kDa prolactin change as pregnancy progresses and the demonstration that 60 kDa but not 64 kDa prolactin bind to Con A, suggests that 50 and 64 kDa but not 60 kDa species in pregnancy are basically the same material. The 25 kDa prolactin form observed during pregnancy showed binding affinity for Con A; thus, it was considered as the glycosylated variant of prolactin (Markoff et al., 1988). G-prolactin species represented 20-30% of the total prolactin immunoreactivity in serum throughout pregnancy as determined by Con A affinity chromatography. This observation was confirmed in Western blots by the apparent stain intensity of the 25 kDa band when compared with that of 23 kDa band. Furthermore, in no instances did G-prolactin represent the main circulating species during pregnancy in any of the subjects studied. Recently, the development of an immunoradiometric assay for non-G-prolactin by Brue et al. (1992) allowed for the first time the glycosylated and non-glycosylated forms of prolactin

to be distinguished quantitatively. As reported by these authors, non-G-prolactin represented more than 70% of the total immunoreactive prolactin in sera from normal subjects. These observations agreed with our results that only a few prolactin species, besides 25 kDa prolactin, could bind to Con A. In some instances, a 25 kDa species was observed in Con A-free material (Figure 3, lane 2). This could be explained on the basis that there is apparently more than one form of G-prolactin and that these forms differed from Con A in their binding properties (Lewis et al., 1989). In this study we have also identified the presence of 27 kDa prolactin in serum from a pregnant woman during the third trimester of gestation. This species was found in Con A-bound fractions and may be similar to that described by Strickland and Pierce (1985) in ovine serum and by Markoff and Lee (1987) in human serum. Unfortunately, we were not able to identify prolactin species of high molecular weight as described by Fukuoka et al. (1991). The most reasonable explanation for this is that under nonreducing conditions a large amount of immunoreactive material was present at the top of the gels. Thus, it was not possible to ascertain whether this material represents prolactin variants or appeared as the result of the immunoprecipitation procedure used to concentrate prolactin from serum samples. In this regard, we have previously concluded that, in ovu.latory hyperprolactinaemic women (Larrea etal, 1989, 1992), prolactin species of large molecular weight (150 kDa) also show significant changes in relative proportions throughout gestation. The exact physiological significance of heterogeneity in prolactin is unknown. However, the consistent pattern of changes in the relative proportions of prolactin components in serum during normal pregnancy suggests that these modifications are physiologically important and may be correlated with the bioactivity of prolactin in order to fulfil specific biological needs.

Acknowledgements We thank Dr Y.N.Sinha (Whittier Institute for Diabetes and Endocrinology, La Jolla, CA, USA) for the anti-human prolactin antiserum VLS-2. The authors gratefully acknowledge the National Hormone and Pituitary Program, the National Institute of Diabetes and Digestive and Kidney Diseases, the National Institute of Child Health and Human Development, and the US Department of Agriculture for the human prolactin RIA reagents. This study was supported in part by grants from the WHO Special Programme of Research Development and Research Training in Human Reproduction (Geneva, Switzerland) and PLACIRH (Mexico).

References Andino.N.A., Bidot.C, Valdes.M. and Macado.A.J. (1985) Chromatographic pattern of circulating prolactin in ovulatory hyperprolactinemia. Fertil. Sterii, 44, 600—605. Benveniste.R., Helman.J.D., Orth.D.N., McKenna.T.J., Nicholson.W.E. and Rabinowitz.D. (1979) Circulating big human prolactin: conversion to small human prolactin by reduction of disulfide bonds. J. Clin. Endocrinol. Metab., 48, 883-886. Brue.T., Caruso,E., Morange.L, Hoffmann.T., Ervin.M., Gunz.G., Benkirane,M. and Jaquet,P. (1992) Immunoradiometric analysis of circulating human glycosylated and nonglycosylated prolactin forms: spontaneous stimulated secretions. J. Clin. Endocrinol. Metab., 75, 1338-1344.

1621


increasing concentrations of non-G-prolactin (Con A-free) without a significant concomitant change in G-prolactin (Con A-bound). In addition, non-G-prolactin represented the major circulating species throughout the different stages of pregnancy studied. These observations have been similar to those previously described in normo- and hyperprolactinaemic subjects studied individually throughout pregnancy (Larrea et al., 1992). In line with our results, Brue et al. (1992) found the non-G-prolactin as the major prolactin species in subjects under pregnant and nonpregnant conditions. Our results did not completely agree with those of Markoff et al. (1988) where 25 kDa G-prolactin represented the major circulating prolactin form in samples obtained during the first trimester of gestation. Possible explanations for these discrepancies could be the relative immunodetectability of G-prolactin (Lewis et al., 1985; Sinha et al., 1988; Pellegrini et al., 1988) as well as the relative amounts of G-prolactin in the pituitary human prolactin used as standard in the RIAs. These obervations raise the urgent necessity to develop more specific and reliable techniques to quantify the various molecular forms of human prolactin.

F.Larrea el al.

1622

Sinha.Y.N., Gilligan.T.A. and Lee.D.W. (1984) Detection of high molecular weight variant of prolactin in human plasma by a combination of electrophoretic and immunologic techniques. J. Clin. Endocrinol. Metab., 58, 752-754. Sinha.Y.N., Gilligan.T.A., Lee.D.W., Hollings worth, D.R. and Markoff.E. (1985) Cleaved prolactin: evidence for its occurrence in human pituitary gland and plasma. J. Clin. Endocrinol. Metab., 60, 239-243. Sinha.Y.N., Campion.D.R., Jacobsen.B.P. and Lewis.U.J. (1988) Glycosylated prolactin in porcine plasma: immunoblotic measurements from birth to one year of age. Endocrinology, 123, 1728 — 1734. Strickland.T.W. and Pierce.J.G. (1985) Glycosylation of ovine prolactin during cell-free biosynthesis. Endocrinology, 123, 1295 — 1298. Towbin.H., Stachelin.T. and Gordon.J. (1970) Electrophoretic traiwfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA, 76, 4350-4356. Whittaker.D.G., Wilox.T. and Lind.T. (1981) Maintained fertility in a patient with hyperprolactinemia due to big.big prolactin. J. Clin. Endocrinol. Metab., 53, 863-866. Wong.V.L.Y., Compton.M.M. and Witorsch.R.J. (1986) Proteolytic modification of rat prolactin by subcellular fractions of the lactating rat mammary gland. Biochim. Biophys. Ada, 881, 167 — 174. Received on March 1, 1993; accepted on June 10, 1993


Clapp.C. (1987) Analysis of the proteolytic cleavage of prolactin by the mammary gland and liver in the rat: characterization of the cleaved and 16-kD forms. Endocrinology, 121, 2055-2064. Compton,M.M. and Witorsch,R.J. (1983) Proteolytic fragmentation of rat prolactin by the rat ventral prostate. Prostate, 4, 231-246. DeVito.W.J., Avakian.C. and Stone,S. (1992) Proteolytic modification of prolactin by the female rat brain. Neuroendocrinology, 56, 597-603. Fukuoka,H., Hamamoto.R. and Higurashi,M. (1991) Heterogeneity of serum and amniotic fluid prolactin in humans. Horm. Res., 31, Suppl. 1, 5 8 - 6 3 . Laemmli.U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680-685. Larrea,F., Villanueva.C., Cravioto.M.C, Escorza.A. and del Real.O. (1985) Further evidence that big.big prolactin is preferentially secreted in women with hyperprolactinemia and normal ovarian function. Fertil. Steril, 44, 2 5 - 3 0 . Larrea.F., Escorza.A., Granados.J., Valencia,X., Valero,A., Cravioto.M.C. and P6rez-Palacios,G. (1987) Familial occurrence of big-big prolactin as the predominant immunoreactive human prolactin species in blood. Fertil. Steril, 47, 956-963. Larrea.F., Escorza.A., Valero.A., Hernandez,L., Cravioto.M.C. and Diaz-Sanchez, V. (1989) Heterogeneity of serum prolactin throughout the menstrual cycle and pregnancy in hyperprolactinemic women with normal ovarian function. J. Clin. Endocrinol. Metab., 68,982—987. Larrea.F., Mdndez.I., Escorza.A., Veayra.F., Carino.C. and Cravioto.M.C. (1992) Prolactin size variants during pregnancy in women with ovulatory hyperprolactinemia: characterization by isoelectric focusing and lectin affinity chromatography. Eur. J. Obstet. Gynecol. Reprod. Biol, 44, 91-100. Lewis.U.J., Singh.R.N.P., Sinha.Y.N. and Vanderlaan.W.P. (1985) Glycosylated human prolactin. Endocrinology, 116, 359—363. Lewis.U.J., Singh.R.N.P. and Lewis.L.J. (1989) Two forms of glycosylated human prolactin have different pigeon corp sacstimulating activities. Endocrinology, 124, 1558 — 1563. Lorenzon.M.Y., Miska.S.P. and Jacobs.L.S. (1984) Molecular mechanism of prolactin release from pituitary secretory granules. In Mena,F. and Valverde-R,C.M. (eds), Prolactin Secretion: A Multidisciplinary Approach. Academic Press, New York, p. 141. Markoff.E. and Lee.D.W. (1987) Glycosylated prolactin is a major circulating variant in human serum. J. Clin. Endocrinol. Metab., 65, 1102-1106. Markoff.E., Lee.D.W. and Hollingsworth,D.R. (1988) Glycosylated and nonglycosylated prolactin in serum during pregnancy. J. Clin. Endocrinol. Metab., 67, 519-523. Mena.F., Hummelt,G., Aguayo.D., Clapp.C, Martinez de la Escalera.G. and Morales,M.T. (1992) Changes in molecular variants during in vitro transformation and release of prolactin by the pituitary gland of the lactating rat. Endocrinology, 130, 3365-3377. Mittra.I. (1980) A novel 'cleaved prolactin' in the rat pituitary. I. Biosynthesis, characterization and regulatory control. Biochem. Biophys. Res. Commun., 95, 1750-1759. Pansini,F., Bergamini.C.M., Malfaccini.M., Cocilov.G., Linciano.M., Jacobs.M. and Bagni,B. (1985) Multiple molecular forms of prolactin during pregnancy in women. J. Endocrinol., 106, 81—85. Pellegrini,I., Gunz,G., Ronin.C, Fenouillet,E., Peyrat,J.-P., Delori.P. and Jaquet.P. (1988) Polymorphism of prolactin secreted by human prolactinoma cells: immunological, receptor binding, and biological properties of the glycosylated and nonglycosylated forms. Endocrinology, 122, 2667-2674. Shoupe.D., Muntz.F.J., Kletzky.O.A. and deZerega.G. (1983) Response to thyrotropin-releasing hormone stimulation of concanavalin A-bound and -unbound immunoassayable prolactin during human pregnancy. Am. J. Obstet. Gynecol., 147, 482-487. Sinha.Y.N. (1992) Prolactin variants. Trends Endocrinol. Metab., 3, 100-106.