Endometrial CD56+ natural killer cells in women with recurrent miscarriage: a histomorphometric study

K. Clifford1,3, A.M. Flanagan2 and L. Regan1

1 Departments of Obstetrics and Gynaecology and 2 Histopathology, Imperial College School of Medicine at St Mary's, London W2 1NY, UK


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Endometrial natural killer (NK) cells were compared in luteal-phase endometrial samples from women with recurrent miscarriage and from normal subjects. Cryostat sections were labelled using a monoclonal antibody to CD56 using an avidin–biotin complex method and a morphometric study performed. Increased mean numbers of CD56+ cells were documented in the endometrium of women with recurrent early miscarriage only. These findings suggest a possible role for NK cells in the pathogenesis of recurrent early pregnancy loss.

Key words: endometrium/natural killer cells/recurrent miscarriage


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
At the time of implantation the endometrium is populated by abundant maternal leukocytes, the majority of which are large granular lymphocytes (LGL) (Bulmer et al., 1987Go; King et al., 1989Go, 1991Go; Klentzeris et al., 1992Go) whose cytoplasmic granules contain cytolytic molecules (King et al., 1993Go). These cells are unusual in that while they stain strongly for the natural killer (NK) marker CD56, most do not express other classical NK markers such as CD16 and CD3 (King et al., 1989Go). The number of endometrial LGL varies with the menstrual cycle with few LGL in proliferative and early secretory phase endometrium and a sharp increase in numbers in the mid to late luteal phase—the time of embryonic implantation (King et al., 1989Go; Bulmer et al., 1991Go; Starkey et al., 1991Go; Klentzeris et al., 1992Go). The population of LGL is maintained in the decidua if pregnancy occurs and persists until after the first trimester when numbers decline.

The function of LGL remains unclear but their abundance at the time of implantation suggests they may be important in the establishment of pregnancy. As the LGL are in intimate contact with invading trophoblast both temporally and spatially, one possible role is that of limiting trophoblast invasion into the decidua (King and Loke, 1991Go). Experimental evidence from murine models supports this notion: NK cell-depleted transgenic mice have an increased incidence of fetal loss compared to immunocompetent mice and this is associated with histological abnormalities of placental vasculature (Guimond et al., 1997Go). Further, when NK cells are replenished by bone marrow transplantation reproductive outcome is improved (Guimond et al., 1998Go). The part played by endometrial leukocytes in the mechanism of infertility and early pregnancy failure is not yet clear. Alterations in endometrial leukocyte populations have been observed in women with unexplained infertility (Klentzeris et al., 1994Go) and in women with infertility associated with autoimmune thyroid disease (Stewart-Akers et al., 1998Go). A recent study reported no changes in endometrial T-cells in a group of women undergoing spontaneous early pregnancy loss (Vassiliadou and Bulmer, 1998aGo). However the same authors report decreased cytotoxic capability of decidual CD56+ lymphocytes from spontaneous aborters indicating a functional deficiency of these cells associated with early pregnancy loss (Vassiliadou and Bulmer, 1998bGo). Alterations in endometrial NK cell subsets have been reported in women with recurrent miscarriage (Lachapelle et al., 1996Go). Increased numbers of peripheral blood CD56+ cells have been demonstrated in women with recurrent miscarriage (Kwak et al., 1995Go) and in women who have experienced multiple in-vitro (IVF) failures (Beer et al., 1996Go). These observations are of interest as they raise the possibility that alterations in the CD56+ NK cell population may result in early pregnancy loss. The purpose of this morphometric study was to quantify the endometrial CD56+ NK cell population in women with a history of recurrent miscarriage and therefore lend further support to this hypothesis.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Subjects
Thirty-nine women were recruited for the study which was approved by Kensington, Chelsea and Westminster ethical committee. Twenty-nine women with a history of recurrent miscarriage were recruited from St Mary's Recurrent Miscarriage Clinic having been investigated according to national guidelines (RCOG Clinical Guideline, 1998). All women were found to have normal parental chromosomes, a normal uterine cavity on ultrasound scanning and were negative for antiphospholipid antibodies. The control group consisted of 10 parous women with no history of miscarriage who attended the Samaritan Hospital for laparoscopic sterilization. No subject had taken hormonal treatment or used an intrauterine contraceptive device (IUCD) for the preceding 3 months, and at least 4 months had elapsed since the last miscarriage. All women collected daily early morning urine specimens for measurement of luteinizing hormone (LH) excretion during the cycle in which the biopsy was obtained in order to ascertain the timing of the LH surge and all had abstained from coitus since the time of the last menstrual period. Written informed consent was gained in all cases.

Endometrial biopsies
The endometrial biopsies were taken in the luteal phase of the menstrual cycle between day 7 and day 10 after the mid-cycle LH surge. This was determined from daily early morning urinary LH measurements as described previously (Clifford et al., 1994Go). Endometrial tissue was obtained as an outpatient procedure, using an endometrial sampler (Z-sampler, BEI Medical Systems International, Gembloux, Belgium). Each specimen was placed onto OCT embedding matrix (CellPath, Hemel Hempstead, UK) on a cork disc and snap-frozen in isopentane in liquid nitrogen. The discs were then stored at –70°C until required.

Serial cryostat sections of 5–7 µm thick were mounted on clean glass slides and air-dried at room temperature. After fixing in acetone for 7 min the slides were wrapped in foil and stored at –20°C until immunolabelling.

Immunocytochemistry
Duplicate sections of all specimens were immunolabelled. The slides were washed in tris-buffered saline (TBS) and blocked with 3% bovine serum albumin in TBS for 15 min. The specimens were incubated with primary mouse anti-human natural killer cell, CD56, monoclonal antibody (Dako, Glostrup, Denmark) at a dilution of 1:20 for 60 min. After three washes in TBS the slides were incubated with secondary biotinylated rabbit anti-mouse antibody (Dako, Glostrup, Denmark) at a dilution of 1:600 for 30 min. The slides were then washed 3 times in TBS and incubated with avidin-biotin complex (ABC) for 30 min. The sections were counterstained with haematoxylin, dehydrated in 100% alcohol, and cleared with xylene prior to mounting. Negative controls were performed by omission of the primary antibody.

Histomorphometric analysis
Morphometry was performed using transmitted light microscopy using an eyepiece graticule. All sections were scored blindly without the observer's (K.C.) knowledge of the subject details. A random selection of slides was scored blindly by a second observer (A.M.F.) and similar numbers of positively staining cells found. The number of positively staining cells in 10 non-overlapping high-powered (x400) fields (h.p.f.) was counted. The procedure was repeated in a second specimen from the same subject and the mean of the absolute number of cells per 10 h.p.f. was calculated.

Statistics
The mean number of positively staining cells per 10 h.p.f. was compared in different groups using Student's two-sample t-test. Pearson's correlation test was used to analyse the relationship between the mean number of cells and various factors in the reproductive history. P = 0.05 level of significance was used throughout.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Twenty-nine women had a history of recurrent pregnancy loss (range 3–7 miscarriages; median = 4) of whom 23 had experienced early pregnancy losses only (<13 weeks gestation) and six had suffered at least one late loss (>13 weeks gestation). The median time between the last miscarriage and the endometrial biopsy was 7 months with no woman undergoing a biopsy less than 4 months from the last miscarriage. Fourteen women had a history of a livebirth prior to their miscarriages. The control group, consisting of 10 parous women with no history of miscarriage or infertility, had a median parity of 2; (range 1–3). The median age of the miscarriage group was 32 years (range 25–42) and that of the controls 33 years (range 27–39). The clinical details and results for each subject are shown in Table IGo.


View this table:
[in this window]
[in a new window]
 
Table I. Patient characteristics and the mean number of CD56+ cells for each subject
 
There were significantly more endometrial CD56+ NK cells in women with a history of recurrent miscarriage (mean = 146 per 10 h.p.f.; SD ± 71) than in the control group (mean = 94 per 10 h.p.f.; SD ± 19) (P = 0.001). When the miscarriage group was separated according to type of miscarriage the increased number of CD56+ NK cells was seen only in the group of women who had a history of early pregnancy loss (mean = 161 per 10 h.p.f.; SD ± 72) compared to the control group (mean = 94 per 10 h.p.f.; SD ± 19) (P < 0.001) and women with a history of at least one late loss (mean = 89 per 10 h.p.f.; SD ± 19) (P < 0.001) (Figure 1Go). There was no difference in the number of endometrial CD56+ cells between the control population and those women who had experienced a late miscarriage. In the group of women who had experienced early pregnancy loss only, there was no correlation between the number of CD56+ cells and the maternal age, the number of previous miscarriages, a past history of a livebirth and the time since last miscarriage.



View larger version (9K):
[in this window]
[in a new window]
 
Figure 1. The number of endometrial CD56+ cells in normal subjects and miscarriers separated according to gestation of previous losses. Women with a history of early miscarriage only had a significantly increased mean number of CD56+ cells per 10 high-power fields (h.p.f.) compared with the control group and with women with a history of a late loss (P < 0.001). Bars represent mean values.

 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The endometrial leukocyte profile in women with recurrent miscarriage has recently been a focus of interest. This observational study of the endometrial CD56+ NK cell population in women with recurrent miscarriage has shown an association between increased numbers of CD56+ NK cells and a history of recurrent early pregnancy loss. We found no correlation between the number of CD56+ cells and the number of previous losses, the number of livebirths, or the time since the last miscarriage. It is noteworthy that in this study there was considerable overlap in CD56+ cell numbers between miscarriers and control subjects, indicating that women with recurrent miscarriage represent a heterogeneous group and no single pathology can underlie all cases. It has also been reported that women with recurrent miscarriage have an altered endometrial lymphocyte population, including an increased number of CD56+ cells (Quenby et al., 1999Go), supporting the findings presented here. Another study employing immunofluoroscopy reported that although there was no significant difference in the overall number of endometrial NK cells between women with recurrent miscarriage and controls there was a higher ratio of CD56bright : CD56dim cells, suggesting that NK subsets may be important (Lachapelle et al., 1996Go). We did not assess the intensity of CD56 expression in our study as identification of NK cell subpopulations by immunohistochemistry does not provide useful information; however, this remains an interesting avenue of research for future studies. Women with recurrent miscarriage have also been shown to have increased circulating peripheral blood CD56+ NK cells (Kwak et al., 1995Go; Beer et al., 1996Go) and furthermore an increase in NK cells may be related to adverse pregnancy outcome in subsequent pregnancies (Coulam et al., 1995Go; Aoki et al., 1995Go). A recent report documented an association between increased numbers of circulating NK cells and heavy metal excretion, which may be a causative factor in some cases of subfertility and recurrent miscarriage (Gerhard et al., 1998Go).

The role played by NK cells in implantation remains unclear. Their presence in pre-implantation endometrium indicates that they are not merely part of an immune response to embryonic demise. Extensive trophoblast invasion of the uterine tissue is necessary for successful implantation and placentation. This process must be finely regulated: excessive invasion results in trophoblastic disease while insufficient invasion may lead to early pregnancy loss. Available evidence supports the theory that endometrial leukocytes play an important role in this process and their recruitment has been shown to be regulated by cytokines. Successful pregnancy depends on the presence of trophoblast growth-promoting cytokines—the Th–2 cytokines—including interleukin (IL)-4, IL-5 and IL-10 (Wegmann et al., 1993Go). In the presence of Th-1 cytokines that include IL-2 and interferon (IFN)-{gamma}, trophoblast cell lysis is favoured (King and Loke, 1990Go). Thus pregnancy appears to be a Th-2 dominant state (Wegmann et al., 1993Go). The LGL population itself, together with endometrial thymus-derived (T)-cells, is a major source of cytokines (Saito et al., 1993Go; Jokhi et al., 1994Go). Thus numerical changes in the LGL population could lead to alterations in the endometrial cytokine profile and hence to a disturbance of the Th-2/Th-1 equilibrium in favour of trophoblast cell lysis. Increased decidual NK cytotoxic activity has been described in women with recurrent miscarriage (Chao et al., 1995Go). Furthermore intravenous immunoglobulin G (IVIG) therapy has been shown to down-regulate elevated peripheral blood CD56+ NK cells in women with recurrent miscarriage (Kwak et al., 1996Go; Ruiz et al., 1996Go) which has been reported to be associated with successful pregnancy outcome. Whether similar changes in CD56+ cell numbers are observed at the endometrial site is not known and thus the mechanism of action of such treatments is not clear. Due to the heterogeneous nature of the problem, it is unlikely that all women with recurrent miscarriage would benefit from such therapy, and clearly further large studies are needed.

The findings presented here support the hypothesis that endometrial NK cells are important for the establishment of pregnancy and alterations in the endometrial NK cell population are associated with early pregnancy failure. However a causal link remains to be established and further studies are needed to determine the precise role of endometrial leukocytes in early pregnancy. The relationship between NK cells and endometrial cytokine production warrants further investigation and how these processes may be defective in recurrent early pregnancy loss.


    Acknowledgments
 
K.C. was funded by a grant from the Medical Research Council. We would also like to thank the patients and controls who made such a valuable contribution to this study.


    Notes
 
3 To whom correspondence should be addressed Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Aoki, K., Kajiura, S., Matsumoto, Y. et al. (1995) Preconceptional natural-killer-cell activity as a predictor of miscarriage. Lancet, 345, 1340–1342.[ISI][Medline]

Beer, A.E., Kwak, J.Y. and Ruiz, J.E. (1996) Immunophenotypic profiles of peripheral blood lymphocytes in women with recurrent pregnancy losses and in infertile women with multiple failed in vitro fertilization cycles. Am. J. Reprod. Immunol., 35, 376–382.[ISI][Medline]

Bulmer, J.N., Hollings, D. and Ritson, A. (1987) Immunocytochemical evidence that endometrial stromal granulocytes are granulated lymphocytes. J. Pathol., 153, 281–288.[ISI][Medline]

Bulmer, J.N., Morrison, L., Longfellow, M. et al. (1991) Granulated lymphocytes in human endometrium: histochemical and immunohistochemical studies. Hum. Reprod., 6, 791–798.[Abstract]

Chao, K.H., Yang, Y.S., Ho, H.N. et al. (1995) Decidual natural killer cytotoxicity decreased in normal pregnancy but not in anembryonic pregnancy and recurrent spontaneous abortion. Am. J. Reprod. Immunol., 34, 274–280.[ISI][Medline]

Clifford, K., Rai, R., Watson, H. and Regan, L. (1994) An informative protocol for the investigation of recurrent miscarriage: preliminary experience of 500 consecutive cases. Hum. Reprod., 9, 1328–1332.[Abstract]

Coulam, C.B., Goodman, C., Roussev, R.G. et al. (1995) Systemic CD56+ cells can predict pregnancy outcome. Am. J. Reprod. Immunol., 33, 40–46.[ISI][Medline]

Gerhard, I., Waibel, S., Daniel, V. and Runnebaum, B. (1998) Impact of heavy metals on hormonal and immunological factors in women with repeated miscarriages. Hum. Reprod. Update, 4, 301–309.[Abstract/Free Full Text]

Guimond, M.J., Luross, J.A., Wang, B. et al. (1997). Absence of natural killer cells during murine pregnancy is associated with reproductive compromise in TgE26 mice. Biol. Reprod., 56, 169–179.[Abstract]

Guimond, M.J., Wang, B. and Croy, B.A. (1998) Engraftment of bone marrow from severe combined immunodeficient (SCID) mice reverses the reproductive deficits in natural killer cell deficient TgE26 mice. J. Exp. Med., 187, 217–223.[Abstract/Free Full Text]

Jokhi, P.P., King, A., Sharkey, A.M. et al. (1994) Screening for cytokine messenger ribonucleic acids in purified human decidual lymphocyte populations by the reverse-transcriptase polymerase chain reaction. J. Immunol., 153, 4427–4434.[Abstract/Free Full Text]

King, A. and Loke, Y.W. (1990) Human trophoblast and JEG choriocarcinoma cells are sensitive lysis by IL-2 stimulated decidual NK cells. Cell. Immunol., 129, 435–448.[ISI][Medline]

King, A. and Loke, Y.W. (1991) On the nature and function of human uterine granular lymphocytes. Immunol. Today, 12, 432–435.[ISI][Medline]

King, A. and Loke, Y.W. (1993) Effect of IFN-{gamma} and IFN-{alpha} on killing of human trophoblast by decidual LAK cells. J. Reprod. Immunol., 23, 51–62.[ISI][Medline]

King, A., Wellings, V., Gardner, L. and Loke, Y.W. (1989) Immunocytochemical characterization of the unusual large granular lymphocytes in human endometrium throughout the menstrual cycle. Hum. Immunol., 24, 195–205.[ISI][Medline]

King, A., Balendran, N., Wooding, P. et al. (1991) CD3– leukocytes present in the human uterus during early placentation: phenotypic and morphologic characterization of the CD56++ population. Dev. Immunol., 1, 169–190.[Medline]

King, A., Wooding, P., Gardner, L. and Loke, Y.W. (1993) Expression of perforin, granenzyne A and TIA-1 by uterine CD56+ NK cells implies they are activated and capable of effector functions. Hum. Reprod., 8, 2061–2067.[Abstract]

Klentzeris, L.D., Bulmer, J.N., Warren, M.A. et al. (1992) Endometrial lymphoid tissue in the timed endometrial biopsy: morphometric and immunohistochemical aspects. Am. J. Obstet. Gynecol., 167, 667–674.[ISI][Medline]

Klentzeris, L.D., Bulmer, J.N., Warren, M. et al. (1994) Lymphoid tissue in the endometrium of women with unexplained infertility: morphometric and immunohistochemical aspects. Hum. Reprod., 9, 646–652.[Abstract]

Kwak, J.Y., Beaman, K.D., Gilman-Sachs, A. et al. (1995) Up-regulated expression of CD56+, CD56+/CD16+ and CD19+ cells in peripheral blood lymphocytes in pregnant women with recurrent pregnancy losses. Am. J. Reprod. Immunol., 34, 93–99.[ISI][Medline]

Kwak, J.Y., Kwak, F.M., Ainbinder, S.W. et al. (1996) Elevated peripheral blood natural killer cells are effectively downregulated by immunoglobulin G infusion in women with recurrent spontaneous abortions. Am. J. Reprod. Immunol., 35, 363–369.[ISI][Medline]

Lachapelle, M.H., Miron, P., Hemmings, R. and Roy, D.C. (1996) Endometrial T, B, and NK cells in patients with recurrent spontaneous abortion. J. Immunol., 156, 4027–4034.[Abstract]

Quenby, S., Bates, M., Vince, G. et al. (1999) Immunohistochemical investigation of leukocytes in pre-implantation endometrium of women with recurrent miscarriage. J. Soc. Gynecol. Invest., 6 (Suppl.), Abstr. 664.

Royal College of Obstetricians and Gynaecologists (1998) The management of recurrent miscarriage. RCOG Clinical Guidelines. RCOG.

Ruiz, J., Kwak, J.Y., Baum, L. et al. (1996) Intravenous immunoglobulin inhibits natural killer cell activity in vivo in women with recurrent spontaneous abortion. Am. J. Reprod. Immunol., 35, 370–375.[ISI][Medline]

Saito, S., Nishikawa, K., Morii, T. et al. (1993) Cytokine production by CD16–CD56bright natural killer cells in the human early pregnancy decidua. Int. Immunol., 5, 559–563.[Abstract]

Starkey, P.M., Clover, L.M. and Rees, M.C.P. (1991) Variation during the menstrual cycle of immune cell populations in human endometrium. Eur. J. Obstet. Gynecol., 39, 203–207.[ISI][Medline]

Stewart-Akers, A.M., Krasnow, J.S., Brekosky, J. and DeLoia, J.A. (1998) Endometrial leukocytes are altered numerically and functionally in women with implantation defects. Am. J. Reprod. Immunol., 39, 1–11.[ISI][Medline]

Vassiliadou, N. and Bulmer, J.N. (1998a) Characterization of endometrial T lymphocyte subpopulations in spontaneous early pregnancy loss. Hum. Reprod., 13, 44–47.[Abstract]

Vassiliadou, N. and Bulmer, J.N. (1998b) Functional studies of human decidua in spontaneous early pregnancy loss: effect of soluble factors and purified CD56+ lymphocytes on killing of natural killer and lymphokine-activated killer-sensitive targets. Biol. Reprod., 58, 982–987.[Abstract]

Wegmann, T.G., Lin, H., Guilbert, J. and Mossmann, T.R. (1993) Bidirectional cytokine interactions in the maternal–fetal relationship: Is successful pregnancy a Th2 phenomenon? Immunol. Today, 14, 353–356.[ISI][Medline]

Submitted on May 14, 1999; accepted on July 14, 1999.