ARTHRITIS & RHEUMATISMVol. 44, No. 6, June 2001, pp 1331–1335© 2001, American College of RheumatologyPublished by Wiley-Liss, Inc.

Molecular Heterogeneity of Prolactin in the Plasma of PatientsWith Systemic Lupus Erythematosus

Javier Cruz,1 Antonio Avina-Zubieta,2 Gonzalo Martınez de la Escalera,3

Carmen Clapp,3 and Carlos Lavalle4

Objective. Systemic lupus erythematosus (SLE)has been associated with high levels of prolactin in thecirculation of some patients. Although prolactin stimu-lates immune responses, the relationship between hy-perprolactinemia and the pathophysiology of SLE re-mains controversial. This study was undertaken toinvestigate whether circulating bioactive prolactin iso-forms are associated with the activity of SLE.

Methods. The molecular heterogeneity of prolac-tin was studied in the plasma of patients with active andinactive SLE and in healthy volunteers by radioimmu-noassay (RIA), enzyme-linked immunosorbent assay(ELISA), Nb2-cell bioassay, and immunoprecipitation–Western blots. The specificity of the bioassay determi-nations was assessed by neutralization of growth-promoting effects with antiserum to human prolactin.

Results. Significantly higher prolactin levels weredetected by bioassay and by ELISA than by RIA in bothsubsets of SLE patients and in normal individuals.Plasma prolactin levels in the SLE patients were signif-icantly greater than those in the normal controls whenmeasured by ELISA, but not by RIA or bioassay. Thebioassay:ELISA and bioassay:RIA ratios were similar

between SLE patients and controls, suggesting thatprolactin biopotency was not altered with the disease,and none of the 3 assays detected a difference inprolactin levels between patients with active SLE andthose with inactive SLE. However, the prolactin detectedin plasma was associated with immunoreactive proteinsof 130 kd and 23 kd, and the concentration of the 130-kdprolactin-like species was 10-fold higher in inactive SLEversus active SLE patients.

Conclusion. Discrepancies among assays sub-stantiate the molecular heterogeneity of circulatingprolactin. The prolactin isotype that is found inassociation with inactive SLE could be of potentialuse as a marker for the inactive form of the disease andas an index for the efficacy of treatment.

Many studies support the role of prolactin inmaintaining immune competence (1,2). Prolactinstimulates the proliferation or activation of cells ofthe immune system and restores immunocompetencein immune-compromised rodents and humans (1,2).The prolactin receptor, a known member of thehematopoietin/cytokine receptor superfamily, is ubiqui-tously expressed on hematopoietic cells, and lymphocytesubsets produce prolactin that can act as an autocrine orparacrine factor (2).

Hyperprolactinemia was first demonstrated inpatients with inactive systemic lupus erythematosus(SLE) (3). Since then, various reports have correlatedelevated prolactin levels with the activity of thisdisease, but others have failed to demonstrate thisassociation and a clear causal relationship betweenprolactin and SLE is lacking (4,5). These discrepanciesmay relate to the molecular heterogeneity of circulatingprolactin.

Prolactin represents a family of proteins in whichstructural differences result from posttranslational mod-ifications (such as glycosylation, phosphorylation, and

Dr. Clapp’s work was supported by grant 55000595 from theHoward Hughes Medical Institute, grant 27950-N from the NationalCouncil of Science and Technology of Mexico, and grant IN-226799from the National University of Mexico.

1Javier Cruz, MD: Hospital de Infectologıa, Centro MedicoNacional “La Raza,” Mexico City, Mexico; 2Antonio Avina-Zubieta,MD, MSc: Instituto Mexicano del Seguro Social, Centro MedicoNacional Siglo XXI, Mexico City, Mexico; 3Gonzalo Martınez de laEscalera, PhD, Carmen Clapp, PhD: Centro de Neurobiologıa, Uni-versidad Nacional Autonoma de Mexico, Queretaro, Mexico; 4CarlosLavalle, MD: Medica Sur, Mexico City, Mexico.

Address correspondence and reprint requests to Carlos La-valle, MD, Medica Sur, Puente de Piedra 150, Torre II-102, Col.Toriello Guerra, Tlalpan 14080, Mexico City, Mexico.

Submitted for publication July 26, 2000; accepted in revisedform February 6, 2001.

1331

proteolysis) of a protein with a molecular mass of 23 kd(6). Some prolactin isoforms have low affinity for anti-bodies directed against unmodified 23-kd prolactin (6).Bioactive prolactins with low immunoreactivity, but alsoimmunoreactive prolactins with low bioactivity, havebeen reported in the human circulation (6–8). Thus,prolactin isoforms may be a factor influencing the lack ofcorrelation between prolactin and the activity of SLE.Herein we investigated the heterogeneity of serum pro-lactin in patients with active and inactive SLE, asdetermined by Nb2-cell bioassay, radioimmunoassay(RIA), enzyme-linked immunosorbent assay (ELISA),and Western blot.

PATIENTS AND METHODS

Patients. Thirty-five patients (30 women, 5 men) whofulfilled the American College of Rheumatology criteria forthe diagnosis of SLE (9) were studied. Ten healthy volunteers(8 women, 2 men) who were matched for age were used ascontrols. Patients with renal or hepatic failure, as well aspatients who were pregnant, nursing, or taking any drug knownto elevate or inhibit circulating prolactin levels, were excluded.The mean (6SD) age of the SLE patients and of the controlswas 34 6 4 years. The activity of SLE was determined byevaluating the demographic, clinical, and serologic features ofthe patients according to the SLE Disease Activity Index(SLEDAI) (10).

Bioactivity of prolactin. The Nb2-cell bioassay wasused to analyze the bioactivity of prolactin, as previouslydescribed (11). Different dilutions of SLE or control plasma,or of prolactin standard (donated by A. F. Parlow, NationalHormone and Pituitary Program [NHPP], Harbor Universityof California Los Angeles Medical Center, Torrance, CA)were assayed with or without a 1:500 dilution of prolac-tin antiserum, which was generated in our laboratory andcharacterized as previously reported (11). Nb2-cell prolifera-tion was measured by the 3-(4,5-dimethylthiazol)yl-2,5-diphenyltetrazolium bromide colorimetric assay (11).

Immunoreactivity of prolactin. An RIA kit (RIA-gnostProlactin; CIS Bio International, Gif-sur-Yvette, France) or anELISA kit (Genzyme Diagnostics, San Carlos, CA) was usedto determine the immunoreactivity of prolactin. The ELISAwas standardized using the same prolactin reference prepara-tion (NHPP) as in the bioassay. Comparisons of the recoveryof RIA and ELISA standards indicated that the standard usedin the RIA was 1.5 times more potent than the standard usedin the ELISA.

Western blot analysis. Western blot was performed inplasma samples immunoprecipitated with antiprolactin anti-serum (donated by A. F. Parlow, NHPP), carried out underreducing conditions (with b-mercaptoethanol), and visualizedby autoradiography after incubation with 125I-labeled proteinA, as previously described (12).

Statistical analysis. The data were analyzed for statis-tical significance by Student’s t-test.

RESULTS

Bioactivity and immunoactivity of plasma pro-lactin. All plasma samples stimulated the proliferationof Nb2 cells in a dose-dependent manner. Stimulationparalleled that induced by the prolactin standard andwas prevented by prolactin antiserum (results notshown).

The level of prolactin was determined in the sameplasma samples by RIA, ELISA, and bioassay (Figure1). The levels of prolactin measured by ELISA and bybioassay were significantly higher (P , 0.05) than thosedetermined by RIA, both in patients with active or

Figure 1. Prolactin (PRL) levels in the plasma of control individuals(C) and systemic lupus erythematosus (SLE) patients or in SLEpatients with inactive (I) or active (A) disease. Determinations were inthe same samples by radioimmunoassay (RIA), enzyme-linked immu-nosorbent assay (ELISA), and bioassay (BA). The number of patientsanalyzed in each group is indicated in parentheses. Bars show themean and SEM. p 5 P , 0.05 versus RIA values; 1 5 P , 0.05 versusELISA values in control individuals.

1332 CRUZ ET AL

inactive SLE and in healthy individuals (Figure 1). TheRIA and bioassay revealed a prolactin concentrationthat was slightly higher in SLE patients than in controls,but this difference was not statistically significant. Con-versely, the ELISA displayed significantly higher levelsof prolactin in the plasma from SLE patients comparedwith controls (Figure 1 and Table 1).

The mean plasma bioassay:RIA and bioassay:ELISA ratios were evaluated as an index of relativeprolactin biopotency. Neither the bioassay:RIA nor thebioassay:ELISA values in SLE and control samplesshowed differences that were statistically significant(Table 1). The plasma prolactin levels in SLE patientswere evaluated in relation to the activity of the disease.A SLEDAI score .4 and #4 defined the subsets ofactive and inactive SLE, respectively. The RIA, ELISA,and bioassay detected no differences in the prolactinlevels between those with active SLE and those withinactive SLE (Figure 1) or between active SLE andhealthy control plasma samples (results not shown).

Prolactin isoforms in plasma. Prolactin in plasmawas associated with proteins that had apparent molecu-lar weights of 130 kd and 23 kd, which reacted withprolactin antiserum, but not with preimmune rabbitserum (Figure 2). The density of the 23-kd protein bandwas similar among the groups, whereas that of the130-kd protein was higher in the plasma of patients withinactive SLE than in that of patients with active SLE andcontrol subjects. Densitometric analysis revealed thatthe relative proportion of band intensity of the 130-kdprolactin protein with respect to the 23-kd prolactinprotein was 10-fold higher in the plasma of inactive SLEpatients compared with active SLE patients and normalsubjects (Figure 2).

DISCUSSION

Prolactin has been proposed to have a role instimulating autoimmune diseases, including SLE (1–4).Although high prolactin levels are prevalent in 15–30%of SLE patients (3,4), the correlation between hyper-

prolactinemia and the activity of SLE is controversial (5)and a clear causal relationship has not been proven. Themolecular heterogeneity of prolactin is a factor toconsider when trying to address this controversy. Pro-lactin isoforms with different chemical, immunologic,and biologic characteristics are present in the circulationof humans (6–8) and represent a burden to conventionalimmunoassay techniques that rely on the antigenic sim-ilarity between the molecular forms present in thebiologic sample and the isolated molecule used as atracer.

The biologic and immunologic properties of cir-culating prolactin in SLE patients were investigated byusing the sensitive Nb2-cell bioassay and 2 quantitativeimmunoassays, RIA and ELISA. The specificity of thebioassay determinations with regard to prolactin was

Figure 2. A, Western blot analysis of immunoreactive prolactin(PRL)–like proteins in the plasma of control individuals (C) andpatients with inactive (I) and active (A) systemic lupus erythematosus(SLE). Plasma samples were immunoprecipitated with anti-PRL anti-serum and subjected to Western blotting under reducing conditions.Immunoreactive PRL-like proteins of 130 kd and 23 kd are indicated(arrows). Negative results were obtained when normal rabbit serum(NRS) was used for immunoprecipitation. B, Ratio between thedensitometric values of the 130-kd PRL and 23-kd PRL bandsdetected in the plasma of control subjects (C) and active (A) andinactive (I) SLE patients. The number of patients in each group isindicated in parentheses. Bars show the mean and SEM. p 5 P , 0.05versus controls and active SLE patients.

Table 1. Levels of prolactin in the plasma of control subjects and systemic lupus erythematosus (SLE)patients as determined by radioimmunoassay (RIA), enzyme-linked immunoabsorbent assay (ELISA),and bioassay (BA)*

RIA ELISA BA BA:RIA BA:ELISA

Control (n 5 10) 10.2 6 1.7 21.4 6 7.5 37.3 6 4.5 4.8 6 1.1 6.1 6 3.1SLE (n 5 35) 12.2 6 0.9 43.6 6 4.8 44 6 4.3 4.1 6 0.5 2.3 6 0.6P NS ,0.05 NS NS NS

* Except where otherwise indicated, values are the mean 6 SEM ng/ml. NS 5 not significant.

MOLECULAR HETEROGENEITY OF CIRCULATING PROLACTIN IN SLE 1333

confirmed by the fact that prolactin-directed antibodies,able to block the stimulatory action of prolactin, abol-ished the stimulation of Nb2-cell proliferation by plasmasamples.

Prolactin levels measured by RIA in both theSLE and control samples were within the range of values(0–20 ng/ml) previously reported for this technique (4).Conversely, in the same samples, the ELISA and bioas-say measured equivalent prolactin levels (20–45 ng/ml)that were 2–4-fold higher than those determined byRIA. This disparity may be explained only in part by theuse of different standard preparations in the assays,since the ELISA and the bioassay, but not the RIA, werestandardized using a preparation with a potency 1.5-foldlower than that used in the RIA. Thus, the possibilitythat differences among the assays may be influenced bythe molecular heterogeneity of prolactin, i.e., that theRIA is not detecting all of the prolactin variants revealedby the ELISA or by the bioassay, is suggested.

The molecular heterogeneity of circulating pro-lactin is further suggested by the fact that the ELISA,but not the RIA nor the bioassay, measured significantlyhigher prolactin levels in SLE patients than in normalcontrols. Because the concentration of bioactive prolac-tin showed no difference between SLE patients andcontrols, the biopotency of circulating prolactin ap-peared to be higher in normal individuals than in SLEpatients. This result would be consistent with the lack ofdetection of bioactive prolactin isoforms by ELISA innormal plasma. However, when the mean plasma bioas-say:ELISA or bioassay:RIA ratios were determined asan index for prolactin biopotency, the differences be-tween SLE patients and controls were not statisticallysignificant. Moreover, the 3 assays measured similarprolactin levels in the plasma of active and inactive SLEpatients. The lack of a difference may not relate to thedistribution of patients (more being in the intermediaterange than at the extremes of the SLEDAI score), sinceno difference was detected between active SLE andhealthy control samples. These results suggest that theimmunologic and biologic heterogeneity of plasma pro-lactin is not associated with the activity of SLE.

The molecular heterogeneity of prolactin sug-gested by ELISA was further substantiated byimmunoprecipitation–Western blots, which showedprolactin-like antigens of 130 kd and 23 kd in theplasma. Because of its apparent molecular weight, the23-kd protein could correspond to unmodified prolactin.Likewise, the 130-kd prolactin-like protein has the sizeexpected for the big-big prolactin (.100,000 kd) that hasbeen described in human serum (6,13) and has been

claimed to be a disulfide-bond aggregate of several23-kd prolactin molecules (13). Similarly, high molecularweight disulfide-bond aggregates (150–170 kd) are pro-duced by the binding of prolactin to antiprolactinautoantibodies (14) and antiprolactin autoantibodieshave been reported in the circulation of SLE patients(15). However, it is unlikely that the 130-kd proteinoriginates from such disulfide-bond aggregates, sincethis protein is found under b-mercaptoethanol reduction.In this regard, similar high molecular weight (53–100 kd)prolactin variants resistant to b-mercaptoethanol reduc-tion have been detected in serum and amniotic fluidfrom humans (16). Some of these prolactin isoformsshowed binding affinity for concanavalin A, which re-flected glycosylation as being a part of their molecularchange (16).

In contrast to the immunologic and biologicproperties of prolactin, chemical characteristics of thehormone did correlate, albeit inversely, with the activityof SLE, since the plasma of inactive SLE patientsshowed a higher concentration of the 130-kd prolactin.The association between the chemical heterogeneity ofprolactin and SLE has already been suggested. Forexample, sera from SLE patients with hyperpro-lactinemia yielded novel proteins migrating at 70, 32,and 16.5 kd (17), and a 60-kd immunoreactive prolac-tin protein was preferentially detected in T lympho-cytes from SLE patients compared with those fromnormal subjects (12).

The higher proportion of the 130-kd prolactinversus the 23-kd prolactin in inactive SLE patientssuggests an abnormality in the mechanisms controllingthe balance between prolactin isoforms. Bromocriptine,a potent dopamine receptor agonist that selectivelyinhibits prolactin secretion from the anterior pituitarygland, suppresses SLE activity (4). Although bromocrip-tine is known to inhibit the processing and secretion ofvarious pituitary prolactin isoforms (6), the possibility ofa selective action of dopamine or other regulatory signalon the secretion of prolactin variants remains to bedetermined.

The possible implication of the association of the130-kd prolactin-like protein with inactive lupus is notclear, because the nature and function of this protein areunknown. Of interest is the fact that large molecularweight forms of prolactin generally have reduced bio-logic receptor binding and immunologic activities incomparison with monomeric prolactin (6–8). However,arguing against the 130-kd prolactin having lower bio-logic and/or immunologic activity, neither the bioassaynor the RIA or ELISA detected differences in the level

1334 CRUZ ET AL

of prolactin between active and inactive SLE patients.Thus, molecular aggregation appears not to mask theepitopes and sites of prolactin that are important for itsbiologic activity, at least not in our in vitro assays.However, the possibility may exist that the 130-kdprolactin, although biologically active in vitro, could notexert its full biologic activity in vivo because the activesites were being blocked in the in vivo situation, or itsaccess to target cells through the capillary network mayhave been restricted due to its large molecular size orchanges in net charge. The latter possibility has alreadybeen proposed for autoantibody-bound prolactin andsubstantiated by showing that autoantibody–prolactinaggregates are cleared more slowly from the circulationthan are free prolactins (14).

In summary, our results support the molecularheterogeneity of prolactin in the circulation of SLEpatients. Further characterization of prolactin isoformsin the circulation of SLE patients is required to helpunderstand the role of this hormone in the pathophysi-ology of this disease. Nevertheless, the close associationbetween the 130-kd prolactin-like isotype and inactiveSLE suggests a potential use of this protein as a markerfor the inactive disease and as an index for the efficacy oftreatment.

ACKNOWLEDGMENTS

The authors thank Fernando Lopez Barrera and Gab-riel Nava for their expert technical assistance, and gratefullyacknowledge Fernando Larrea for his valuable training andadvice.

REFERENCES

1. Reber PM. Prolactin and immunomodulation. Am J Med 1993;95:637–44.

2. Weigent DA. Immunoregulatory properties of growth hormoneand prolactin. Pharmacol Ther 1996;69:237–57.

3. Lavalle C, Loyo E, Paniagua R, Bermudez JA, Herrera J, Graef A,

et al. Correlation study between prolactin and androgens in malepatients with systemic lupus erythematosus. J Rheumatol 1987;14:268–72.

4. Walker SE, McMurray RW, Houri JM, Allen SH, Keisler D, SharpGC, et al. Effects of prolactin in stimulating disease activity insystemic lupus erythematosus. Ann N Y Acad Sci 1998;840:762–72.

5. Buskila D, Lorber M, Neumann L, Flusser D, Shoenfeld Y. Nocorrelation between prolactin levels and clinical activity in patientswith systemic lupus erythematosus. J Rheumatol 1996;23:629–32.

6. Sinha YN. Structural variants of prolactin: occurrence and physi-ological significance. Endocr Rev 1995;16:354–69.

7. Farkouh NH, Packer MG, Frantz AG. Large molecular sizeprolactin with reduced receptor activity in human serum: highproportion in basal state and reduction after thyrotropin-releasinghormone. J Clin Endocrinol Metab 1979;48:1026–32.

8. Subramanian MG, Gala RR. Do prolactin levels measured by RIAreflect biologically active prolactin? J Clin Immunoassay 1986;9:42–52.

9. Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, RothfieldNF, et al. The 1982 revised criteria for the classification of systemiclupus erythematosus. Arthritis Rheum 1982;25:1271–7.

10. Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang CH,and the Committee on Prognosis Studies in SLE. Derivation of theSLEDAI: a disease activity index for lupus patients. ArthritisRheum 1992;35:630–40.

11. Corbacho AM, Macotela Y, Nava G, Torner L, Duenas Z, NorisG, et al. Human umbilical vein endothelial cells express multipleprolactin isoforms. J Endocrinol 2000;166:53–62.

12. Larrea F, Martınez-Castillo A, Cabrera V, Alcocer-Varela J,Queipo G, Carino C, et al. A bioactive 60-kilodalton prolactinspecies is preferentially secreted in cultures of mitogen-stimulatedand nonstimulated peripheral blood mononuclear cells from sub-jects with systemic lupus erythematosus. J Clin Endocrinol Metab1997;82:3664–9.

13. Hattory N. The frequency of macroprolactinemia in pregnantwomen and the heterogeneity of its etiologies. J Clin EndocrinolMetab 1996;81:586–90.

14. Hattori N, Inagaki C. Anti-prolactin autoantibodies cause asymp-tomatic hyperprolactinemia: bioassay and clearance studies ofprolactin-immunoglobulin G complex. J Clin Endocrinol Metab1997;82:3107–10.

15. Leanos A, Pascoe D, Fraga A, Blanco-Favela F. Anti-prolactinautoantibodies in systemic lupus erythematosus patients withassociated hyperprolactinemia. Lupus 1998;7:398–403.

16. Fukuoka H, Hamamoto R, Higurashi M. Heterogeneity of serumand amniotic fluid prolactin in humans. Horm Res 1991;35:58–63.

17. Allen SH, Sharp GC, Wang G, Conley C, Takeda Y, Conroy SE,et al. Prolactin levels and antinuclear antibody profiles in womentested for connective tissue disease. Lupus 1996;5:30–7.

MOLECULAR HETEROGENEITY OF CIRCULATING PROLACTIN IN SLE 1335