Variation during the menstrual cycle of immune cell populations in human endometrium

European Journal of Obsfefrics & Gynecology and Reproducfive Biology, 39 (1991) 203-207 0 1991 Elsevier science Publishers B.V. OOZB-2243/91/%03.50

EUROBS

203

01114

Variation during the menstrual cycle of immune cell populations in human endometrium P.M. Starkey,

L.M. Clover and M.C.P. Rees

Nujfield Deparfmenf of Obstetrics and Gynaecology, John Radcliffe Hospifal. Headingfon, Oxford, U.K. Accepted for publication

28 August 1990

Morphometric analysis and immunohistology of tissue sections have been used to assess variation, during the normal menstrual cycle, of the bone marrow-derived cell populations in human endometrium. Levels of T cells and macrophages were found to befelatively constant throughout the cycle. In contrast, numbers of large granular lymphocytes, identified as being CD56-positive, were generally low between days 10 and 19, but increased sharply in the latter part of the luteal phase, decreasing again after menstruation. This LGL population is known to be abundant in first trimester pregnancy decidua, and is presumed to play a role in early pregnancy success. Endometrium;

Implantation;

Menstrual cycle

Introduction In early human pregnancy the decidualized endometrium contains numerous cells identified as of bone marrow origin, by their surface expression of the leucocyte common antigen, CD45 [1,2]. Flow cytometric analysis of de&dual cell dispersions prepared from first trimester tissue has shown that the most abundant cell type is the large granular lymphocyte (LGL), with smaller numbers of macrophages and a few classical T cells [3]. The decidual LGL express the CD56 (NKHl) antigen and about half are also positive for CD2, the E rosette receptor. They are negative for CD3 and unlike most peripheral blood LGL

Correspondence: DI’. P.M. Starkey, Nuffield Department of Obstetrics and Gynaecology, John Radcliffe Hospital, Headington, Oxford OX3 9DU, U.K.

[4], they are also negative for CD16, the FcyRIII receptor [3]. The function of the bone marrow-derived cells in early pregnancy decidua is not known, though there is evidence from animal studies that they may play a role in pregnancy success [5]. It has been suggested that they could regulate placentation either by suppressing the maternal immune response, preventing rejection of the semi-allogeneic placenta [6-81, or by secreting cytokines which stimulate the growth and development of placental trophoblast cells 191. Immunohistological studies have demonstrated that the cell types present in early pregnancy decidua are also found in non-pregnant endometrium [2], but no quantitative assessment of the various endometrial cell populations was made. In order to better understand the role of these cells in early pregnancy, and determine which cell types are present at or near the time of implantation, we have used morphometric techniques to quantitate

204

the immune cell populations in non-pregnant endometrium throughout the menstrual cycle. While this study was nearing completion, the results of two similar investigations measuring only the decidual LGL were published [lO,ll]. Materials and methods

Samples of non-pregnant endometrium were obtained from 16 women of reproductive age, undergoing hysterectomy for a subjective complaint of excessive menstrual bleeding. No pelvic pathology. was found at operation and histological examination of the resected uteri by an independent histopathologist, showed no abnormality. All the patients had a history of regular (26-30 day) menstrual cycles and had not used either oral or intrauterine contraception nor taken any hormones for at least 6 months prior to surgery. The stage of the cycle at which the tissue was obtained was determined from the patient’s menstrual history and endometrial histology [12]. At operation the condition of the ovaries was observed and the presence or absence of a corpus luteum noted; 6 patients were in the follicular phase and 10 in the luteal phase. Immediately after removal the uterus was opened and blocks of endometrium/ myometrium taken, snap-frozen in liquid nitrogen and stored at -70”. Cryostat sections (7-8 pm) were cut, acetone fixed, and stained as previously described [13] with mouse monoclonal antibodies followed by peroxidase-conjugated rabbit anti (mouse immunoglobulin) IgG obtained from Dako Ltd, High Wycombe, Bucks, U.K. Monoclonal antibodies (see Table I for details) were obtained as follows: F10/89/4 from Prof. J.W. Fabre, Blond McIndoe Centre, Queen Victoria Hospital, East Grinstead, Sussex, U.K.; 3ClO from Prof. R.M. Steinman, The Rockefeller University, New York, NY 10021, U.S.A.; KB90 and CR3/43 from Dr. D.Y. Mason, John Radcliffe Hospital, Oxford, U.K., all these were gifts; and T3 and NKHl from Coulter Electronics Ltd., Luton, U.K.; 0KT4 and 0KT8 from Ortho Diagnostic Systems, High Wycombe, U.K.; T2, Tl and B from Dako Ltd., High Wycombe, U.K.

TABLE I Details of monoclonal

antibodies

used in this study

Antibody

Antigen cluster designation

Specificity

F10/89/4

CD45

3ClO KB90 CR3/43 T3

CDllc CD3

OKT4 0KT8 T2 Tl NKHl B

CD4 CD8 CD7 CD5 CD56 CD22

Leucocyte common antigen Macrophage Macrophage HLA-DR, -DP, -DQ T cell receptor associated T cell(macrophage) T cell T cell, NK cell T cell NK cell B cell

Dilution

l/loo l/10 I/I I/4

l/2 l/10 I/I l/10 I/I I/4 l/I

Microscopy was with a Leitz Laborlux D light microscope, using a graticule of 100 X 1 mm squares. Morphometric counting was done at X 400 magnification, counting the number of positive cells in 250 squares. Preliminary measurements showed that at least 100 squares needed to be counted for the number of positive cells per 10 squares to approach a constant value. Squares which contained no cells, that is those over the lumens of glands or vessels, were not counted. Counting was done in a line at right angles to the endometrial surface, using the graticule at x 100 magnification to define the start point for each 250 squares. Depending on the thickness of the endometrium, two to four 250 square areas were counted, and the arithmetic average number of positive cells per 250 squares was calculated. ReSUlti

The antibodies used in this study showed that T cells, identified as positive for CD3, CD5 and CD7, were most often found in dense lymphocyte aggregates, particularly (though not exclusively) within the deeper layers of the endometrium. In contrast, decidual LGL, which are CD5dpositive and mostly CD7-positive, when present were found scattered throughout the endometrial stroma. Macrophages were identified using the 3ClO antibody which detects an antigen confined to macro-

phages and absent from dendritic cells [14] and KB90 which reacts with CDllc, the p150,95 member of the LFA protein family [15]. Both antibodies showed macrophages to be scattered throughout the endometrium. The reaction of the CR3/43 antibody with endometrial tissue showed that HLA Class II was expressed on endometrial macrophages, in agreement with previous studies [2]. Serial sectians of the same tissue demonstrated that the T cell aggregates were also strongly class II-positive. The reactivity of glandular endothelial cells with CR31/43 varied between samples, in a way which bone no relation to the stage of the cycle. In some tissues all the glands were strongly HLA Class II-positive, some had a mixture of positive and negative glands, and in some all the

250

glands were negative. Vascular endothelial cells were Class Il-positive in all samples. The morphometric data from 16 endometrial samples is summarised in Fig. 1, a-d. The number of bone marrow-derived cells, identified as CD45 positive, is high during menstruation, decreases after day 6 when proliferation of endometrial cells is increasing, and increases again in the early luteal phase. Similarly LGL, detected as CD56positive cells, are also high during menstruation, decreasing after day 6, then gradually increasing until a sharp increase in late luteal phase. Numbers of CD7-positive cells largely follow changes in LGL numbers. CD56 and CD7 show the largest variation during the cycle of any of the markers tested; the maximum level of CD5dpositive cells

a

FOllbuIar

Early luteal

Mid Iukal

Late luta

0

Y.nltrual

FOlltDUtU

Early luteal

Mid lutal

Late lutNl

Fig. 1. Bone marrow-derived cell populations in non-pregnant endometrium. The number of positive cells in 1562.5 pm2 of tissue was determined morphometricahy as described in the text. Tissues were grouped according to the stage of the menstrual cycle, and the median values for each group calculated. All antibodies were used on all tissue samples, except for anti-CD3, which was not used on the early luteal or late luteal samples. Antibodies were (a) anti-CD45 3ClO and KB90, (b) anti-CDS6 and anti-CD7, (c) CR3/43, anti-CD5 and antiCD3, and (d) anti-CD22, anti-CD4 and anti-CD8.

206

is twenty times as high as the minimum level, while that of CD7-positive cells shows an 11 fold increase. The classical T cell markers, CD3 and CD5, are more or less constant throughout the cycle, as are the CD4 and CD8 T cell markers. Macrophages, as detected by 3ClO and KB90, decrease at the beginning of the cycle to a minimum during the follicular phase; numbers then increase slightly through the latter half of the cycle. The number of CDZZ-positive B cells detected at any stage in the cycle is very low. The number of cells expressing HLA Class II antigens in the endometrium was often higher than the number of bone marrow-derived cells, as it includes glandular epithelium and vascular endothelium. Class II expression was highest during menstruation, decreasing in the follicular phase, and rising again thereafter through the luteal phase. Discussion

Morphometric analysis @f tissue sections provides an objective assessment of the frequency of positive cells in a tissue. The agreement in this study between the results given with the two macrophage markers 3C10 and KB90 supports this. Inaccuracies arise where cells are not scattered evenly through the tissue, as applied to T cells in endometrium, or where cells are so numerous that it is difficult to score each separately. This may be responsible for the number of CD45-positive cells being lower than the sum of the various bone marrow-derived subsets. In this study, cells were counted throughout the whole of the endometrium, including both the basalis and functionalis. It has been suggested that the variation in cell populations in the basalis is less than in the functionalis. While we did find a trend, in the late luteal phase, for the increases in CD56-positive cells to be smaller near the endometrial/myometrial junction than at the top of the endometrium, this was not statistically significant. This study confirms that the bone marrow-derived populations in the non-pregnant endometrium are the same as those in early pregnancy

decidua, though the proportions of each vary. In first trimester decidua, the most abundant cell type is the CD56-positive LGL; of decidual bone marrow-derived cells, 48% are LGL, 25% macrophages and only 11% classical T cells [3]. In nonpregnant endometrium, the CDM-positive LGL are present in only small numbers during most of the cycle, but levels increase sharply after day 19 when implantation would be expected to occur. Similar results have recently been described by others [lo]. LGL density is still high during menstruation, and the decline in numbers after day 6 suggests cell death. This is consistent with reports that prior to menstruation the LGL nuclei become pyknotic, in contrast to the normal nuclear morphology of LGL in the early part of the cycle or in pregnancy decidua [lo]. The increased LGL numbers in late luteal phase probably reflect a combination of proliferation in situ and recruitment from peripheral blood, as has been shown for the analogous metrial gland cells in mouse [16]. Immunohistology of endometrial samples using Ki67, a marker of cell proliferation, shows some CD56-positive cells to be actively dividing particularly in secretory phase [ll] though the absolute numbers of dividing cells were low. It has been suggested that the endometrial LGL arise by differentiation from decidual T cells [17], but this would seem to be in contradiction to the evidence that LGL and T cells are distinct [18,19]. Decidual LGL may also express CD2 and CD7, on average about 50% of decidual CDSB-positive cells are also positive for CD2 [3] and CD7 shows a similar distribution. In endometrium we found numbers of CD7-positive cells to be very similar to that of CD56-positive cells. Since CD7 is also expressed on T cells, these results would be consistent with most but not all endometrial LGL being positive for CD7. Macrophage numbers were found to be fairly constant throughout the cycle, though there was a slight decrease after day 6 with numbers subsequently increasing steadily. This may reflect a role for endometrial macrophages in the extensive tissue remodelling that occurs during the menstrual cycle, as macrophages are known to play an important role in wound repair for example [20]. T

201

cell numbers were also found to be relatively invariant, though there were increases at the beginning and end of the cycle. Endometrial T cells appeared to be positive for HLA Class II, suggesting activation, in contrast to the T cells of early pregnancy decidua which are largely HLA Class II-negative [3]. Both CDCpositive and CDS-positive populations were detected. However since no double labelling experiments were included, and CD4 is also expressed on some macrophages [21] it is not certain that all the CDCpositive cells were T cells. Our results suggest that T cells and macrophages are present in endometrium throughout the menstrual cycle but that the LGL population, that is so prominent in early pregnancy decidua, increases sharply only in the late luteal phase of the normal cycle. The signal for their recruitment into the endometrium is clearly under ovarian hormonal control and independent of the presence of a fertilised blastocyst. This is consistent with the finding that decidual cell populations in the uterus in ectopic pregnancy are similar if not identical to those found in a normal uterine pregnancy (P.M. Starkey and K.A. McCallum, unpublished data). The results suggest that these LGL must play a role in early pregnancy from implantation onwards, though the precise nature of that role is still a matter for speculation. References 1 Bulmer JN, Sunderland CA. Bone-marrow origin of endometrial granulocytes in the early human placental bed. J Reprod Immunol 1983;5:383-387. 2 Bulmer JN, Johnson PM. Immunohistological characterisation of the decidkral leucocytic infiltrate related to endometrial gland epjthelium in early human pregnancy. Immunology 1985;55:35-44. 3 Starkey PM, Sargent IL, Redman CWG. Cell populations in human early pregnancy decidua: characterization and isolation of large granular lymphocytes by flow cytometry. Immunology 1988;65:129-134. 4 Lamer LL, Le kM, Civin CI et al. The reIationship of CD16(Leu-11) and Leu-19 (NKH-1) antigen expression on human peripheral blood NK cells and cytotoxic T lymphocytes. J Immunol 1986;136:4480-4486.

5 Clark DA, Slapsys RM, Croy BA et al. Suppressor cell activity in uterine decidua correlates with success or failure of murine pregnancies. J Immunol 1983;131:540-542. 6 Clark DA, Slapsys R, Chaput A et al. Immunoregulatory molecules of trophoblast and decidual suppressor cell origin at the matemofetal interface. Am J Reprod Irmnunol Microbiol 1986;10:100-104. 7 Daya S, Clark DA, Devlin C et al. Preliminary characterisation of two types of suppressor cells in the human uterus. Fertil Steril 1985;44:778-785. 8 Daya S, Clark DA, Devlin C et al. Suppressor cells in human decidua. Am J Obstet Gynecol 1985;151:267-270. 9 Athanassakis I, Bleackley RC, Paetkau V et al. The immunostimulatory effect of T cells and T cell lymphokines on murine fetally derived placental cells. J Immunol 1987;138:37-44. 10 King A, Wellings V, Gardner L et al. Immunocytochemical characterisation of the unusual large granular lymphocytes in human endometrium throughout the menstrual cycle. Human Immunol 1989;24:195-205. 11 Pace D, Morrison L, Bulmer JN. Proliferative activity in endometrial stromal granulocytes throughout menstrual cycle and early pregnancy. J Clin Path01 1989;42:35-39. 12 Noyes RW, Hertig AI, Rock J. Dating the endometrial biopsy. Fertil Steril 1950;1:3-25. 13 Sutton L, Gadd M, Mason DY et al. Cells bearing class II MHC antigens in the human placenta and amniochorion. Immunology 1986;58:23-29. 14 Witmer MD, Koide S, Cohn ZA. Specific anti-mononuclear phagocyte monoclonal antibodies. Application to the purification of dendritic cells and the tissue localization of macrophages. J Exp Med 1983;158:126-145. 15 MacDonald SM, Pulford K, Falini B, et al. A monoclonal antibody recognizing the p150/95 leucocyte differentiation antigen. Immunology 1986;59:427-431. 16 Peel S, Stewart I, Bulmer D. The morphology of granulated metrial gland cells in lethally-irradiated, bone marrow-reconstituted mice. J Anat 1982;135:849. 17 Bulmer JN, Hollings D, Ritson A. Immunocytochemical evidence that endometrial stromal granulocytes are granulated lymphocytes. J Path01 1987;153:281-288. 18 Herberman RB, Ortaldo JR. Natural killer cells: their role in defenses against disease. Science 1981;214:24-30. 19 Hercend T, Schmidt RE. Characteristics and uses of natural killer cells. Imrnunol Today 1988;9:291-293. 20 Cohn ZA. The macrophage - versatile element of inflammation. Harvey Lectures 1983;77:63-80. 21 Wood GS, Warner NL, War&e RA. Anti-Leu3/T4 antibodies react with cells of monocyte/macrophage and Langerhans lineage. J Immunol 1983;131:212-216.