Concentration of interleukin-1{beta} correlates with prostaglandin E2 and F2{alpha} in human pre-ovulatory follicular fluid

To investigate the role(s) of interleukin-1 (IL-1) in humanovarian function, we measured the concentrations of IL–1,prostaglandins (PGs) and steroids in follicular fluid of 90stimulated ovaries, with reference to oocyte maturation. Concentrationsof IL-1 were significantly higher in the follicles from whichmature oocytes were recovered than in follicles from which oocytescould not be recovered (P < 0.05). IL-1 concentrations alsoincreased in association with oocyte maturation. Positive significantcorrelations were seen between IL-1 and prostaglandin E2 (PGE2)(r = 0.47, P < 0.001), and between IL-1 and prostaglandinF₂ (PGF₂) (r = 0.22, P < 0.05) in pre-ovulatory follicularfluid, but not between IL–1 and oestradiol, or betweenIL-1 and oestradiol, or between IL-1 and progesterone. 0Follicularfluid IL-1 might contribute to prostaglandin-induced oocytematuration and ovulation.     

Ovary: Immunolocalization of transforming growth factor-{beta}1 and transforming growth factor-{beta}2 in the mouse ovary during gonadotrophin-induced follicular maturation

An immunohistochemical approach was utilized to evaluate thecellular distribution of transforming growth factor-₁ (TGF₁)and transforming growth factor ₂ (TGF₂) at different stagesof follicle development in the prepubertal mouse ovary underthe following conditions: (i) after pregnant mare's serum gonadotrophin(PMSG) treatment; (ii) after PMSG and human chorionic gonadotrophin(HCG) treatment; (iii) after PMSG and HCG treatment plus mating.In the immature ovary, TGFF₁ and TGF₂ immunoreactivities arelocalized in theca and granulosa cells and in oocytes. AfterPMSG treatment, TGF₁ and TGF₂ immunoreactivities are localizedin granulosa cells; in addition, TGF₂ staining is noted in thematrix surrounding antral cells. Staining for both TGF₁ andTGF₂ drops in the theca but persists in the oocyte. PMSG plusHCG treatment results in a significant increase in TGF₁ andTGF₂ immunoreactivity in the theca and in the maintenance ofTGF₁ staining in both basal granulosa cells and cumulus cellswhereas TGF₂ immunoreactivity is essentially localized in thematrix surrounding cumulus cells. Staining for TGF₁ and TGF₂persists in the oocyte. Following PMSG plus HCG treatment andmating, TGF₁ immunoreactivity is localized in the luteal cellsof corpora lutea and TGF₂ shows a similar localization pattern.This study provides evidence that TGF₁ and TGF₂ peptides areexpressed in specific cell types during induced follicular maturationin the mouse ovary.     

Integrins are not involved in the process of human sperm-oolemmal fusion

BACKGROUND: We investigated whether integrins are required forthe human sperm–oocyte binding and fusion processes. METHODS:The expression of several integrin subunits at the human oocyteplasma membrane was investigated using immunofluorescence microscopy,and the functional role of integrins expressed at the humanoocyte surface in sperm–oocyte interaction was studiedusing a zona-free human oocyte binding and fusion assay. A totalof 144 unfertilized oocytes were stained with anti-integrinantibodies and 147 zona-free unfertilized oocytes were inseminatedin the presence of various anti-integrin antibodies that wereexpressed in oocyte plasma membrane. RESULTS: The antibodiesof six integrin subunits (₂, ₃, ₅, ₆, _V, _M) and six integrinsubunits (₁, ₂, ₃, ₄, ₅, ₆) were bound to the surface of fixedunfertilized oocytes. In contrast, the presence of ₁ and ₄ subunitscould not be verified. The human sperm–oocyte bindingwas only partially inhibited by blocking antibodies of ₂, ₃,₅, ₆, _V, _M, ₁, ₂ and ₃ with a maximum of 55% inhibition, butantibodies of ₄, ₅ and ₆ showed no effect on sperm–oolemmalbinding. A similar reduction of the number of fused sperm wasobserved. However, the ratio of fused sperm to total sperm (boundand fused) was not impaired by all integrin antibodies, suggestingthat integrins had no role in the sperm–oolemmal fusionprocess. CONCLUSIONS: These results suggest that one of thebinding mechanisms can be inhibited by integrin antibodies butthat this mechanism does not play an essential role in the humansperm–oolemmal binding and fusion processes. The othermechanisms, insensitive to integrins, may involve both bindingand fusion processes in human oocytes.     

Molecular aspects of implantation: Human trophoblast adhesion to matrix proteins: inhibition and signal transduction

At the time of implantation, the extracellular matrix proteinslaminin and fibronectin are abundant in the decidua and aredistributed pericellularly around each individual stromal cell.First trimester human trophoblast expresses both laminin andfibronectin receptors, specifically the ₁₁, ₅₁, ₆₁ and ₆₄ integrinheterodimers. In this study we have demonstrated that in-vitroadhesion of first trimester human trophoblast to purified extracellularmatrix proteins and to purified decidual stromal cell monolayerscan be inhibited by monoclonal antibodies directed against appropriateintegrin subunits and by synthetic peptides containing an arginine-glycine-asparticacid sequence. Monoclonal antibodies (mAbs) to the ₅ and ₁ integrinsubunits and a synthetic peptide significantly inhibited adhesionto fibronectin. Binding of trophoblast to laminin was blockedwith mAbs to the ₆ and ₁ but not ₁ and ₄ integrin subunits.Similarly, integrin-mediated adhesion to monolayers of decidualstromal cells could be blocked with mAbs to the ₅, ₆, ₁ and₄ integrin subunits. Integrin-mediated signal transduction innormal and malignant trophoblast was investigated by Westernblotting. A 115 kDa protein was the major tyrosine phosphorylatedprotein detected in trophoblast after binding to laminin orfibronectin. The profile of tyrosine phosphorylated proteinsdiffered for malignant trophoblast.     

Pregnancy: Effect of the vascular endothelium on contractions induced by prostaglandin F 2{alpha}in isolated pregnant guinea pig uterine artery

The purpose of this study was to examine the influence of endotheliumon prostaglandin F₂-mediated contractions in pregnant guineapig uterine artery. Consequently, the effects of prostaglandinF₂ pregnant guinea pig uterine arterial rings with both Intactand denuded endothelium were studied. In vessels with denudedendothelium prostaglandin F₂ (0.1–10 µM) inducedcontraction (pD₂ = 6.17) with greater potency than in vesselswith intact endothelium (pD₂ 5.68). N^G.Monomethyl (10 µM)did not affect the concentration—response curve for prostaglandinF₂ regardless of endothelial condition. In contrast, in bothtypes of preparation, indomethacin (10 µM) increased themaximal response value obtained with prostaglandin F₂ but thiseffect was significantly greater in preparations with intactthan In those with denuded endothelium (128.3 versus 206.5%).Moreover, indomethacin shifted the concentra tion-response curvefor prostaglandin F2 to the left only in preparations with intactendothelium. The PK_A values for prostaglandin F₂ itself didnot differ between preparations: 5.41 and 5.52 for pregnantguinea pig uterine artery with and without endotheliuin, respectively.The receptor reserve expressed as K_A/EC₅₀ was significantlygreater in rings with denuded (4.44) compared to those In ringswith intact endo thelium (1.86). We conclude that prostaglandin-F₂contraction in pregnant guinea pig uterine artery is modu latedby the vascular endothelium. It is probable that cyclo oxygenaseproducts relating to vasodilatation and derived from endotheliummediate this effect, acting as a functional endogenous antagonistand thereby reducing the apparent efficacy and potency of prostaglandinF₂     

Identification, in mouse macrophages and in serum, of a soluble receptor for the Fc portion of IgG (Fc{gamma}R) encoded by an alternatively spliced transcript of the Fc{gamma}RII gene

Low affinity FcR are a heterogeneous group of glycoproteinswhich exist in transmembrane (TM) as well as in soluble forms.Two membrane isoforms of the murine type II FcR, FcRilb1 andFc;Rilb2, have been described. They result from the translationof alternatively spliced premRNA, FcRilb2 lacking sequencesof the first intracytoplasmic domain (IC1). Soluble forms ofFcR (sFcR) have previously been shown to result from proteolysisof membrane receptors. We report here the identification, inmacrophages, of a mRNA derived from the FCRll gene by splicingexons encoding the TM and IC1 domains, i.e. corresponding toa TM-deleted FcRllb2 mRNA. A soluble protein possibly encodedby this mRNA was identified in macrophage supernatants. In accordancewith FcR nomenclature, we propose to name this new FcRll IsoformFcRllb3. It is the most abundant 8FcR present in serum, as comparedwith 8FcR resulting from cleavage of membrane FcR.     

Integrins and adhesion mlecules: Cell adhesion molecules on the oocyte and preimplantation human embryo

The presence of cell adhesion molecules on human oocytes, earlyembryos, and pre-hatched blastocysts was examined by indirectimmunofluorescence and compared to the distribution found onfirst trimester villous placenta with the same antibodies. Sixintegrin subunits (3, V, 1, 3, 4, 5) were observed consistentlythroughout preimplantation development. Evidence was also obtainedfor the presence of integrin subunits 2, 4, L, 2, and 7 on asmall number of oocytes. A more restricted developmental analysisof E-cadherin, ICAM-1, NCAM, and VCAM-1 demonstrated that thesecell adhesion molecules are also present on oocytes and earlyembryos. L-selectin was detected on oocytes but was not foundon 8-cell embryos. The oocyte and early blastomeres have complexsurfaces in which the integrin and CAM families are represented.     

Endocrinology: Substrate dependency of C19 conjugates in hirsute hyperandrogenic women and the influence of adrenal androgen

Serum C₁₉ conjugates, specifically 3-androstanediol glucuronide(3G), reflect peripheral androgen action through the actionof 5-reductase activity. The origin of 5-reduced C₁₉ conjugateshas been controversial and it has been suggested that they arederived primarily from adrenal androgens. We examined concentrationsof 3G, 3-androstanediol sulphate (3S), androsterone glucuronide(AoG) and androsterone sulphate (AoS) in 40 hirsute hyperandrogenicwomen. These patients were divided into four groups based uponindividual, combined or normal concentrations of the adrenalandrogens dehydroepiandrosterone (DHEAS) and 11-hydroxy-androstenedione.Testosterone, unbound testosterone and androstenedione weresimilar in these groups. Serum 3G was equally high in all groupsand was correlated significantly with hirsutism, while the otherconjugates were not. Androsterone glucuronide was raised inall groups but was higher in patients with raised DHEAS. Serum3S was raised in all groups and was higher where both adrenalandrogens were raised. Serum AoS was highly correlated withDHEAS. Serum 3G was correlated with unbound testosterone andandrostenedione but not with the adrenal androgens. The glucuronideconjugates were correlated with one another as were the sulphateconjugates but glucuronides and sulphates were not correlated.These data confirm ovarian and adrenal dependency of C₁₉ conjugates.Serum 3G appears to reflect hirsutism most accurately and isleast dependent on adrenal androgens in patients with mixedhyperandrogenism.     

Interleukin-1{alpha} and -{beta} modulation of luteinized human granulosa cell oestrogen and progesterone biosynthesis

This study was designed to test the hypothesis that inter-leukin-1(IL-1) and directly affect progesterone, and oestradiol productionin cultures of purified human granulosa cells. Luteinized granulosacells were obtained from women during in-vitro fertilizationcycles. Granulosa cells with and without associated white bloodcells were cultured in the presence of IL-1 and IL-1 (0.5–50ng/ml) for 48 h. Media were changed at 24 h intervals and assayedfor progesterone and oestradiol. In separate experiments, granulosacell viability was assessed with the tetrazolium salt reductionassay, haemocytometer cell counts, and Trypan blue dye exclusion.Our results indicate that progesterone synthesis by basal andhuman chorionic gonadotrophin (HCG)-stimulated granulosa cellsco-cultured with white blood cells was inhibited by 5.0 ng/mlof IL-1 and IL-1 at 48 h of culture. In the presence of whiteblood cells, granulosa cell oestradiol synthesis was inhibitedby IL-1 but not IL-1. Oestradiol was inhibited after both 24and 48 h of culture and was maximally affected by 5.0 ng/mlof IL-1. In contrast, basal and HCG-stimulated oestradiol productionby granulosa cells cultured free of white blood cells was inhibitedonly by IL-1. IL-1 at 5.0 ng/ml produced maximal inhibitionof basal oestradiol (57%) and HCG-stimulated oestradiol (41%)production at 48 h of culture. Gonadal steroid inhibition byIL-1 and IL-1 was not mediated through cytotoxic or antiproliferativeeffects on granulosa cells. Specificity of the granulosa cellresponse to IL-1 and IL-1 was demonstrated by abrogation ofsteroid inhibition with anti-IL-1 and IL-1 neutralizing antibodies.In conclusion, IL-1 directly inhibited the production of oestradiolby human ovarian granulosa cells. IL-1 and IL-1 also exertedindirect effects on steroid production via white blood cellsthat are usually present in granulosa cell cultures if stepsare not taken to remove them. These data support the hypothesisthat cytokines play an important role in intra-ovarian regulationof steroid biosynthesis.     

Immunology: Autoimmunity to spermatozoa, asymptomatic Chlamydia trachomatis genital tract infection and {gamma}{delta} T lymphocytes in seminal fluid from the male partners of couples with unexplained infertility

The relationship between an undetected, asymptomatic Chlamydiatrachomatis genital tract infection, the concentration of andb T cells in semen and sperm autoimmunity was examined in 48male partners of couples with unexplained infertility. ImmunoglobulinA (IgA) antibodies to C.trachomatis were detected in seminalfluids from 14 (29.2%) of the men. Only four of these were positivefor circulating anti-chlamydial IgA, suggesting that the stimulusfor antibody production was within the genital tract. In contrast,four men were positive for anti-chlamydial IgG in their semen;all were also seropositive for anti-chlamydial IgG. T lymphocytesbearing the and antigen receptors were present in every semensample. Men with seminal anti-chlamydial IgA, however, had significantly(P = 0.035) elevated semen T cell concentrations (median 3100cells/ml) than did men lacking this antibody (median 1400 cells/ml);concentrations of T cells were comparable in both groups. Genitaltract sperm autoimmunity, as shown by antibodies bound to motileejaculated spermatozoa, was detected in 13 (27.1%) men. Thepresence of these antibodies was associated with elevated concentrationsof both (median 4200 versus 700 cells/ml) and (median 5000versus 850 cells/ml) T cells (P = 0.0002 and 0.0001 respectively).Men with antisperm antibodies only in their serum had seminalT cell concentrations comparable with men testing negative forantisperm antibodies. Anti-chlamydial IgA was identified insemen from four of 10 men with IgA bound to their spermatozoaand in none of the men with only spermatozoabound IgG. Therewas no relationship between sperm quality and the occurrenceof seminal IgA antibodies to either C.trachomatis or spermatozoa.An asymptomatic C.trachomatis infection activates T cells withinthe male genital tract, which may lead to antisperm antibodyformation and immune-mediated infertility.     

Andrology: Evidence of {beta}1 integrins and fibronectin on spermatogenic cells in human testis

Interactions between human spermatozoa and oocytes are an essentialevent in the process of fertilization. Cell —cell andcell — matrix interactions in somatic cells are mediatedby adhesion molecules such as 1 integrins (very late antigens;VLA). Therefore, we investigated the expression of 1 integrinsand the matrix proteins collagen IV, fibronectin and lamininin human testis by immunohistology. Monoclonal antibodies againstthe chain of 1 integrins reacted with the basement membraneof the tubuli seminiferi, spermatocytes, spermatids and testicularspermatozoa. The 3, –5 and –6 chains of 1 integrinsshowed the same pattern, whereas, the 1, –2 and –4chains could not be detected on spermatogenic cells. These VLAsubunits were localized on endothelial cells, leukocytes andbasement membranes. Matrix proteins such as laminin, collagenIV and fibronectin were detectable as components of basementmembranes in human testis. Germinal cells except spermatogoniaexpressed fibronectin only. These results demonstrate that 1integrins and matrix proteins in human testis are normally expressedon somatic tissue and that germinal cells, especially spermatocytes,spermatids and spermatozoa show positive reactions with antibodiesagainst the VLA–3, –5 and –6 complexes andfibronectin. These findings suggest a production of B1 integrinsand fibronectin during the spermatogenesis and a role of theseproteins in adhesive mechanisms of spermatozoa similar to somaticcell—cell or cell—matrix interactions.     

Expression of integrins by human trophoblast and differential adhesion to laminin or fibronectin

The process of placental implantation involves a series of transformationsof trophoblast from a single polarized epithelial layer restingon a basement membrane (villous trophoblast), to cellular aggregates(trophoblast columns) which ultimately disperse to invade uterinedecidua as individual cells (interstitial trophoblast). Suchtissue re-modelling is associated with changes in the constituentsof the extracellular matrix and in the expression of matrixreceptors by the cells, the most relevant being the family ofintegrins which bind to laminin and fibronectin. In this studywe show, by immunohistology and flow cytometry, a gradual lossof laminin receptors with the concomitant acquisition of fibronectinreceptors as trophoblast is transformed from the villous phenotype,through the cell columns, into the extravillous population.The pattern of staining for the ₅, ₆, ₁ and ₄ subunits indicatesthat the integrins expressed by trophoblast are predominantlythe ₅₁ and the ₆₄ heterodimers. We have also shown that isolatedtrophoblast cells assume a flattened, sessile phenotype whencultured on laminin but exhibit a more spreading, motile morphologywhen plated on fibronectin. In addition, numerous multinucleatedgiant cells are observed on a fibronectin substrate. Our datasuggest that the relative expression of laminin and fibronectinreceptors may determine the morphology and behaviour of trophoblastduring the process of implantation.     

Effects of progesterone and oestradiol-17{beta} on 17{beta}-ol-dehydrogenase activity in stromal cells of human endometrium under in-vitro conditions

A 17-ol-dehydrogenase activity could be demonstrated in fibroblastmonolayer cell cultures of proliferative human endometrium.After 24 h incubation 100 nCi [6, 7–3H] oestradiol-17was completely oxidized to oestrone. Progesterone was not ableto enhance the metabolizing velocity. In contrast, progesteroneincubation revealed a decreasing oxidation rate with increasingmolarity. Histological changes after transformation of the endometriumare discussed to explain in-vivo results showing an increased17-ol-dehydrogenase activity in the secretory phase of the cycle     

Endocrinology: Prostaglandin F2{alpha} stimulates release of insulin-like growth factor binding protein-3 from cultured human granulosa-luteal cells

Human ovarian follicular fluid contains a number of insulin-likegrowth factor binding proteins (IGFBP) of which IGFBP-3 is themost abundant. IGFBP-3 synthesis is growth hormone-regulated.We studied the effect of prostaglandin F₂ (PGF₂) on IGFBP-3secretion by cultured human granulosa-luteal cells from follicularaspirates of women participating in an in-vitro fertilizationprogramme. The IGFBP-3 concentration was measured using a specificmonoclonal immunofluorimetric assay. Contrary to a previousreport on unstimulated follicles, this study demonstrated apositive correlation between follicular fluid IGFBP-3 concentrationand follicular size. PGF2a was found to stimulate in a dose-dependentfashion the secretion of IGFBP-3. Significant (p < 0.05)effects were found at PGF₂ concentrations of 10^–8, 10^–7and 10^–6 M. Because IGFBP-3 inhibits progesterone productionstimulated by insulin-like growth factor (IGF)-I, the PGF₂-inducedstimulation of IGFBP-3 production may be one of the mechanismswhereby PGF₂ exerts its luteolytic effect via the IGF system     

Implantation: Tumour necrosis factor {alpha} inhibits in-vitro decidualization of human endometrial stromal cells

Interleukin-1 (IL-1) has been reported previously to inhibitthe in-vitro decidualization of human endometrial stromal cellsas assessed by progesterone-induced prolactin production andmorphological transformation. In this study we examined whetherother cytokines, such as tumour necrosis factor-(TNF), interferon-(IFN), IFN or granulocyte-macrophage colony-stimulating factor(GM-CSF), could affect the decidualization of human endometrialstromal cells in vitro. Of these cytokines, TNF significantlysuppressed prolactin production in a dose-dependent manner,with no apparent effect on cell number. The morphological transformationof endometrial stromal cells was also inhibited by TNF. TNFand IL-1 significantly suppressed cAMP-stimulated prolactinproduction by endometrial stromal cells. Neither the progesteroneconcentration in the supernatant of the endometrial stromalcell culture system nor intracellular calcium concentrationof the endometrial stromal cells were affected by the additionof TNF or IL-1. These results indicated that TNF and IL-1 suppressboth progesterone-induced and cAMP-mediated prolactin productionin endometrial stromal cells, and that this inhibition was notattributable to direct effects on progesterone metabolism orrelated to Ca2+-mediated signal transduction. These experimentssuggested that a local increase of TNF and IL-1 under certainpathological conditions in vivo may disturb blastocyst implantationand/or the maintenance of pregnancy by inhibiting the decidualizationof endometrial stromal cells.     

Pregnancy: Cervical ripening with the cytokines interleukin 8, interleukin 1{beta} and tumour necrosis factor {alpha} in guinea-pigs

It has been suggested that the collagenolytic enzymes releasedfrom white blood cells which infiltrate the pregnant human uterinecervix at term are responsible for connective tissue changeswhich take place during the ripening process. Similarly, aninfiltration of inflammatory cells occurs in pregnant guinea-pigseither spontaneously at term or at preterm after treatment withthe antiprogestin onapristone. The objective of this study wasto evaluate the effects of the inflammatory cytokines interleukin8 (IL-8), interleukin 1 (IL-1), tumour necrosis factor (TNF-)and a combination of IL-1 and TNF- on cervical ripening in guinea-pigsduring advanced pregnancy. The cytokines were applied locally(intracervically) in a gel for 2 days and the effects were assessedon the third day by both extensibility measurements and morphologicalevaluation. IL-8 treatment on days 42 and 43 post coitum (p.c)and on days 48 and 49 p.c. (term: day 67± 3 p.c.) significantly(P < 0.05) increased cervical extensibility at both stagesof pregnancy. Although IL-1 treatment (days 42 and 43 p.c.)led to a slight increase in cervical extensibility, this effectwas not statistically significant. An electron microscope studyperformed on days 48 and 49 p.c. revealed a pronounced cervicalripening accompanied by the dissolution of collagen fibres,stromal oedema and the infiltration of polymorphonuclear leukocytesin all cytokine-treated groups. The morphological effects ofIL-8 and IL-1 were indistinguishable from those observed duringnormal cervical ripening at term. In contrast, TNF-, both aloneand in combination with IL-1, brought about a severe inflammatoryreaction with a massive infiltration of lymphocytes, marcophagesand polymorphonuclear leukocytes at the investigated dose. Weconclude that the local application of the inflammatory cytokinesIL-8, IL-1 and TNF- produces cervical ripening without inducinglabour in pregnant guinea-pigs; the morphological effects ofIL-8 and IL-1 being similar to the physiological cervical ripening.Our data support the view that cytokines, particularly IL-8,may play an important role during physiological, pathologicaland induced cervical ripening and could be clinically usefulas an adjunct to labour and delivery.     

Immunohistological distribution of the secretory endometrial protein, 'pregnancy-associated endometrial {alpha}2-globulin', a glycosylated {beta}-lactoglobulin homologue, in the human fetus and adult employing monoclonal antibodies

We have previously demonstrated that pregnancy-associated endometrial₂-globulin (₂-PEG), the human glycosylated (-lactoglobulin homologue(HG-BLG), is quantitatively the major secretory soluble proteinproduct of the secretory endometrium during the latter halfof the menstrual cycle and decidua spongiosa of the gestationalendometrium during early pregnancy, and is principally localizedto the glandular epithelium. In the present study employingmonoclonal antibodies in immunohistological techniques, thedistribution and localization has been examined in normal andpathological tissues of the adult and first-trimester fetus.No significant staining for ₂-PEG was detected in any non-reproduction-associatedtissue in the normal adult nor any tissue in the fetus. In theadult, most intense staining was associated with the endometrialglandular epithelium in the uterus or in ectopic sites in patientswith endometriosis. During the menstrual cycle and pregnancy,appearance of ₂-PEG in endometriosis was strongly linked withits appearance in uterine endometrial tissue, suggesting thatendometriotic tissue exhibited competence to respond to thesame hormonal milieu required to induce synthesis in the uterineendometrium. Localization to the mucosal epithelium of the Fallopiantube was consistent with synthesis of ₂-PEG, albeit at low levels,and staining at this site reflected fluctuations of stainingwithin the uterus. Of the pathological specimens examined, stainingwas only detected in a proportion of ovarian carcinomas. Nostaining was detected in the mammary gland, a site of -lactoglobulinsynthesis, whether obtained during pregnancy or lactation. Theseobservations support the proposal that during the menstrualcycle and pregnancy, the endometrial glandular epithelium representsthe major source of the glycosylated -lactoglobulin homologueand serum measurements may be employed to assess its functionalpresence. Mechanisms of regulation of its production that wouldaccount for the histological observations are discussed.