Human uterine natural killer cells: a reappraisal

Share Embed


Descrição do Produto

Molecular Immunology 42 (2005) 511–521

Review

Human uterine natural killer cells: a reappraisal Judith N. Bulmera,∗ , Gendie E. Lashb a b

School of Clinical and Laboratory Sciences, University of Newcastle upon Tyne, Framlington Place, Newcastle upon Tyne NE2 4HH, UK School of Surgical and Reproductive Sciences, University of Newcastle upon Tyne, Framlington Place, Newcastle upon Tyne NE2 4HH, UK Available online 1 October 2004

Abstract The presence of granulated cells within the uterus of many species has been recognised for many years but only recently have these been recognised to be a type of NK cell. Various terms have been applied to the cells, including endometrial granulocyte, K cell and, in mouse and rat, granulated metrial gland cell. Although early studies are often based on histology and electron microscopy, these often include important information for current studies. In vitro studies of purified cells have focused particularly on cytotoxicity and cytokine production and roles in the control of trophoblast invasion and spiral artery remodelling in human pregnancy have been proposed. Evidence in mouse has implicated uNK cell production of IFN-␥ in vascular remodelling but evidence for such a role for human uNK cells remains to be established. Investigation of uNK cells in human pregnancy is hampered by the lack of availability of tissues from the first half of the second trimester of pregnancy when vascular remodelling occurs and also by possible differences between cells from different regions of decidua. The presence of similar cells in species with no trophoblast invasion into the uterus and epitheliochorial placentation raises the question of whether control of trophoblast invasion by human uNK cells is important in vivo and raises the possibility of another function which is conserved between species. © 2004 Elsevier Ltd. All rights reserved. Keywords: Uterine NK cell; Endometrium; Decidua; Endometrial granulated lymphocyte; Pregnancy; Trophoblast invasion; Spiral artery

1. Introduction The presence of granulated cells within human endometrium has been recognised since the 1920s (Weill, 1921) and several later reports described their distribution and staining characteristics. The cells were characterised by the presence of phloxinophilic cytoplasmic granules, in the phloxine tartrazine stain (Hamperl and Hellweg, 1958; Kazzaz, 1972; Dallenbach-Hellweg, 1987). They were noted only in endometrium, predominantly in the late secretory phase of the menstrual cycle and early pregnancy and were scanty in proliferative endometrium and at the end of normal pregnancy (Hamperl and Hellweg, 1958; Dallenbach-Hellweg and Nette, 1964; Kazzaz, 1972). An origin by dual differentiation of endometrial stromal cells was proposed (DallenbachHellweg, 1987) and immunofluorescence studies suggested ∗ Corresponding author. Present address: Department of Pathology, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP, UK. Fax: +44 191 222 8100.

0161-5890/$ – see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.molimm.2004.07.035

that the cells secreted relaxin (Dallenbach and DallenbachHellweg, 1964), although the antiserum used failed to stain rat ovary, a potent relaxin source. Cells with cytoplasmic granules were also recognised in the endometrium of several animal species, including the socalled ‘granulated metrial gland (GMG) cells’, in rat and mouse which localise to decidua and to an area in the mesometrial triangle, termed the metrial gland. A stromal origin for these was generally accepted, despite suggestions in humans (Weill, 1921), mouse (Smith, 1966) and monkey (Bartelemez et al., 1951) of a lymphocytic origin. More conclusive evidence of a lymphocyte derivation came from electron microscope studies of pseudodecidua in rat–mouse chimaeras which exploited ultrastructural differences between rat and mouse metrial gland cells (Peel et al., 1983). In their fully developed form rat and mouse GMG cells bear little resemblance to lymphocytes; mouse GMG cells are up to 50 ␮m in diameter with PAS-positive cytoplasmic granules up to 5 ␮m in diameter. Immunohistochemical studies of human endometrium from normal early pregnancy demon-

512

J.N. Bulmer, G.E. Lash / Molecular Immunology 42 (2005) 511–521

3. Localisation and distribution

ers of lower numbers of CD56+ cells in term decidua (Haller et al., 1993; Vargas et al., 1993) may be due to the use of different tissues, some studies being based on decidua attached to the placental membranes and delivered placenta, while others have studied decidua in placental bed biopsies. In late secretory phase endometrium and early pregnancy decidua uNK cells are detected in the stratum functionalis, often forming aggregates around spiral arteries and glands (Bulmer et al., 1991). This perivascular distribution was considered to represent the effect of diffusion of progesterone from blood into perivascular tissues but more recently has been considered to reflect trafficking of uNK cells or their precursors from the circulation (Trundley and Moffett, 2004). An alternative explanation is that the perivascular distribution may reflect a role in the remodelling of spiral arteries which is an essential feature of pregnancy (see below). There is little information regarding the distribution of uNK cells in the second trimester, tissues from this stage of normal pregnancy being unavailable. We have analysed leucocyte numbers in the first half of pregnancy in a series of placental bed biopsies and failed to note significant differences CD56+ cell numbers between first (8–12 weeks’ gestation) and early second (13–20 weeks’ gestation) trimester decidua, with uNK cells still forming the majority leucocyte population in the early second trimester (Scaife, Bulmer, Robson, Searle, Innes, unpublished data). In contrast, in histological studies of pregnancy hysterectomy samples from 8–19 weeks’ gestation reduced numbers of granulated cells were noted after 13 weeks’ gestation (Fig. 1). It is possible that higher numbers of CD56+ cells in early second trimester decidua than granulated cells, identified in phloxine tartrazine stained sections, may reflect loss of granules, degranulation after the first trimester having been suggested (Spornitz, 1992). Numbers of granulated cells in phloxine tartrazine stained sections and CD56+ cells were comparable in a study of non-pregnant and early pregnancy endometrium (Bulmer et al., 1991) but there have been no studies of granule content of CD56+ cells after the first trimester. Immunohistochemical studies are complicated by the fact that the cytoplasmic granules do not reliably survive freezing but this could be achieved using paraffin embedded samples with histochem-

In early studies endometrial granulated cells were identified by the presence of cytoplasmic granules. The cells were absent in proliferative endometrium, increasing in numbers premenstrually and in pregnancy until the third month of gestation, thereafter declining to be virtually absent at term (Hamperl and Hellweg, 1958; Dallenbach-Hellweg and Nette, 1964). Immunohistochemical studies for CD56 have largely confirmed this distribution, although CD56+ cells are present in the proliferative and early secretory phase, albeit in small numbers (Bulmer et al., 1991; King et al., 1989a). Several studies have recorded a reduction in CD56+ cells in term decidua, but substantial numbers of CD56+ remain in both decidua basalis and decidua parietalis in third trimester placental bed biopsies (Scaife et al., 2003). Reports from oth-

Fig. 1. Relationship between uNK cell numbers detected by phloxine tartrazine staining in decidua basalis and gestational age. Cells were counted in 5 × 400 fields of the central part of the placental bed in pregnancy hysterectomy samples (n = 35).

strated expression of the leucocyte common antigen, CD45, and CD2 but not CD3, CD4 or CD8, nor of the natural killer (NK) cell antigen CD57 (Bulmer and Sunderland, 1984). Their identification as an NK-type cell (suggested by the large granular lymphocyte morphology) was confirmed by expression of CD56 (Ritson and Bulmer, 1987a), which was noted to be particularly intense. It was soon clear that the phenotype of uterine granulated lymphocytes differed from ‘usual’ peripheral blood NK cells, being CD56bright CD16− CD57−. This phenotype was similar to a small subgroup of peripheral blood NK cells.

2. Terminology Granulated cells have been identified in the endometrium of many species but many terms have been used to describe these and the terminology is confusing and remains inconsistent. There has been some reluctance to adopt the term NK cells, probably at least partly to ensure that their distinctive features are appreciated. The different terms used to describe the cells have been well summarised by Stewart (1998). In humans, names have included ‘granular endometrial stromal cell’ (Hamperl and Hellweg, 1958), ‘endometrial granulocytes’, a particularly unfortunate term given their lymphocyte-like morphology, ‘K¨ornchenzellen’ or ‘K’ cells (Dallenbach-Hellweg, 1987) and, more recently ‘endometrial (or decidual) granulated (or granular) lymphocyte (or leucocyte)’ (Bulmer et al., 1991). In mouse and rat terminology is less complex with ‘granulated metrial gland cell’ being used until recently. With recognition of the diversity of NK populations in general, and demonstration of perforin and serine proteases (King et al., 1993; Gudelj et al., 1997) and expression of NK receptors it seems reasonable to adopt the term ‘uterine NK cell’, which emphasises that these represent a special NK cell subgroup. It should be remembered, however, that at least in humans uNK cells are not really ‘uterine’ but ‘endometrial’, being absent in myometrium.

J.N. Bulmer, G.E. Lash / Molecular Immunology 42 (2005) 511–521

ical staining of cytoplasmic granules or double immunolabelling for perforin or granzyme.

4. Uterine NK cell relationship with trophoblast Results of studies which have compared the distribution of uNK cells between decidua basalis, which underlies the placenta and is infiltrated by trophoblast, and decidua parietalis which lines the remainder of the uterine cavity, have been inconsistent. Some have reported increased numbers of uNK cells in decidua basalis and considered this to support a role in the control of trophoblast invasion (Sindram-Trujillo et al., 2003) but others, including ourselves, have failed to detect any difference in the different decidual areas (Khong, 1987; Haller et al., 1995). We analysed decidua in placental bed and non-placental bed biopsies from 8 to 20 weeks’ gestation and at term and did not detect any significant differences in numbers of any of the major decidual leucocyte populations (Scaife, Bulmer, Robson, Searle, Innes, unpublished data). The discrepancies in results of different groups may reflect the different tissues used, with some using decidua basalis and parietalis attached to the delivered term placenta and membranes, respectively, while others have examined much deeper decidua in placental bed biopsies. Decidua basalis is usually defined by the presence of extravillous trophoblast but central areas of the placental bed may differ in composition from more lateral areas. There are, however, several indications that macrophages are often more closely associated with trophoblast than uNK cells, with close associations between macrophages and extravillous trophoblast in decidua basalis (Bulmer et al., 1988) and a recent study of rhesus monkey decidua also highlights the close association of macrophages, rather than uNK cells with trophoblast (Slukvin et al., 2004). Although uNK cell numbers may not vary between decidua basalis and parietalis, the possibility of a functional difference between the two sites has not been considered in detail. Since uNK cells express receptors for the non-classical HLA antigens expressed by extravillous trophoblast (King et al., 2000), contact with trophoblast in decidua basalis could lead to altered function such as a differing cytokine profiles and increased interferon-␥ levels have been reported in decidua basalis (von Rango et al., 2003). It is also possible that uNK cell function varies in different areas of decidua basalis, those cells which are in a perivascular position being exposed to a different microenvronment than cells distant from the vessels. The possibility of differences in uNK cell function in distinct areas of decidua is important for in vitro studies of purified uNK cells. Initial selection of decidua is usually based on its macroscopic appearance and inevitably decidua parietalis predominates in these samples. It would be extremely difficult to select only decidua basalis from macroscopic or even dissecting microscopic examination of pregnancy samples for in vitro studies. Hence the effect of trophoblast on uNK cell function is usually investigated by

513

exposing purified uNK cells to trophoblast or to HLA antigens expressed on transfected cells. The use of laser capture microscopy (Torres et al., 2002; Phillips et al., 2003) may provide an alternative way to compare the function of uNK cells in different areas of decidua and could even be used to investigate perivascular uNK cells.

5. Uterine NK cell relationship with decidual cells Although there has been considerable interest in the role of uNK cells in pregnancy, their presence in endometrium does not depend on trophoblast; uNK cells are detected in non-pregnant endometrium throughout the menstrual cycle and are also seen in decidualised uterine endometrium associated with an ectopic pregnancy at an extrauterine site. Large numbers of uNK cells are also seen after progesterone treatment, and also in ‘decidua’ at ectopic sites, such as the focal decidualisation which can occur at several sites in pregnancy, including cervix and ovarian serosa and also in foci of endometriosis. This distribution suggests regulation by progesterone but uNK cells do not themselves express progesterone receptor, nor is their function altered by progesterone (King et al., 1996; Stewart et al., 1998). The cells have, however, recently been shown to express oestrogen receptor beta, prolactin receptor and glucocorticoid receptor (Gubbay et al., 2002; Henderson et al., 2003). An alternative explanation for the perivasular distribution of uNK cells is that the cells are closely associated with endometrial stromal cell decidualisation and it is in the perivascular areas that the endometrial stromal cells first undergo predecidual change in late secretory phase endometrium (Robertson, 1981; Dallenbach-Hellweg, 1987). The co-localisation of uNK cells with decidualised endometrial stromal cells is striking and is entirely independent of trophoblast, with comparable uNK cell numbers being seen in intrauterine decidua associated with ectopic pregnancy (Vassiliadou and Bulmer, 1998a). Uterine NK cells are also present in foci of endometriosis which have undergone decidual change, even within the omentum and other extragenital sites, emphasising the close relationship with decidual stromal cells. This has also been illustrated in in vitro studies in both humans (King et al., 1999) and mouse (Stewart, 2000). Since uNK cells do not themselves express progesterone receptor, the influence of progesterone is likely to be indirect via the stromal cells. Various products of decidualised endometrial stromal cells could potentially stimulate the recruitment and differentation of uNK cells from peripheral blood NK cells or precursors (see below). However, the presence of uNK cells in small foci of decidualisation such as in the cervix, ovarian serosa and endometriotic foci, without a prominent vascular supply, raises the possibility that the decidual stromal cells could have a role in differentiation of uNK cells, similar to the role of bone marrow stromal cells in NK cell differentiation (Rosmaraki et al., 2001). IL15 is essential for generation of normal NK cell numbers in

514

J.N. Bulmer, G.E. Lash / Molecular Immunology 42 (2005) 511–521

vivo (Rosmaraki et al., 2001) and decidualised endometrial stromal cells produce IL-15, at least in part apparently under progesterone control (Okada et al., 2000; Kitaya et al., 2000). Much work has focused on the relationship between uNK cells and peripheral blood NK cell subgroups. There are, however, differences between the CD56bright NK cells in peripheral blood and uNK cells; in peripheral blood CD56bright NK cells are less granular and less cytotoxic than CD56dim NK cells (Cooper et al., 2001), whereas uNK cells are characterised by abundant cytoplasmic granules and at least in some studies are able to efficiently lyse NK targets (Jones et al., 1997). It is possible that uNK cells represent an NK cell population which is unique to the uterus and which differentiates from precursor cells under the influence of endometrial stromal cells, which themselves are influenced by progesterone. The close association of the endometrial stromal cells with uNK cells merits further study. Investigation of uNK cell function in progesterone treated endometrium and in intrauterine decidua associated with ectopic pregnancy would also be worthwhile, but these samples are rarely predicted preoperatively.

6. Trafficking or in situ proliferation? The variation with menstrual cycle phase of the uNK cells is well established but the mechanism for the marked increase in the cells which is seen in the mid and late secretory phase of the menstrual cycle continuing into early pregnancy is uncertain. It is important to remember that CD56+ cells are present in proliferative and early secretory phase endometrium, albeit in small numbers. Immunohistochemical studies of frozen sections and purified cells have established that up to 40% of CD56-positive uNK cells isolated from late secretory phase endometrium express Ki67 (Jones et al., 1998a) and, indeed stromal mitoses are a recognised feature of premenstrual endometrium. It is, therefore, possible that local proliferation of uNK cells could account for their dramatic increase in number. At no time during the menstrual cycle are there substantial numbers of CD16-positive NK cells in endometrium and expansion by local proliferation would have to be from precursors within the endometrium remaining after menstruation or from residual CD56+ cells in the stratum basalis which is not shed at menstruation. Others favour influx of peripheral blood NK cells from the circulation with modification to the specialised uNK cell type within the uterine microenvironment and there have been several studies of adhesion molecules in endometrium which could play a role in adhesion of lymphocytes to endometrium (Marzusch et al., 1993; Ruck et al., 1994; Burrows et al., 1995). Expression of various adhesion molecules by uNK cells and by endometrial endothelium could explain homing of uNK cell to the endometrium and may account for their distribution in a perivasular position. Since CD16-positive cells are not seen in endometrium recruitment from blood would suggest recruitment of CD56bright cells or their pre-

cursors within the circulation with modification within the uterus. Various mechanisms could play a role in the recruitment of uNK cells or their precursors to the uterus, including expression of CXCR3, CCR5 and CCR7 by CD56bright NK cells (Campbell et al., 2001), CXCL12 by endovascular trophoblast (Hanna et al., 2003) and MIP-1␤ by endometrial stromal cells (Kitaya et al., 2003). Recruitment of uNK cells cannot be dependent on trophoblast and a stromal cell product appears more likely, candidates including IL-15, prolactin and glucocorticoids (Kitaya et al., 2000; Gubbay et al., 2002; Trundley and Moffett, 2004). Recruitment of CD56bright CD16-negative cells from the peripheral circulation has also been suggested and an increase in circulating CD56bright NK cells has been reported in the peripheral blood of women of reproductive age compared with males (King et al., 1991). Evidence from mouse pregnancy supports trafficking of precursor cells to the uterus in early pregnancy in mouse (Chantakru et al., 2002).

7. Where do uNK cells go? The reduction of uNK cells in later pregnancy and also at the end of the menstrual cycle also remains unexplained. Death by both apoptosis and necrosis has been proposed for the loss of uNK cells in the late stages of mouse pregnancy (Kusakabe et al., 1999) but no information is available regarding the fate of the cells in human pregnancy. We have failed to detect apoptosis of decidual lymphocytes in early human pregnancy using TUNEL (Pongcharoen et al., in press), although others have reported apoptosis of decidual leucocytes (Hammer and Dohr, 1999). The possibility that uNK cells degranulate in late pregnancy has been raised (Spornitz, 1992) and this could account for discrepancies in the number of CD56+ cells noted in immunohistochemical studies and the number of cells recognised in older studies by their cytoplasmic granules. This could be addressed by analysis of perforin in uNK cells at different stages of pregnancy. The rounded hyperchromatic nuclei observed in uNK cells in premenstrual endometrium and also strikingly in progesterone treated endometrium has led to the suggestion that the cells undergo apoptosis premenstrually. There is, however, no evidence for this from TUNEL studies (Jones et al., 1998b). Furthermore, CD56+ cells separated from late secretory phase and premenstrual endometrium are still capable of proliferative activity and cytotoxicity, retain expression of bcl-2 and up to 40% express Ki67 indicating proliferation rather than apoptosis (Jones et al., 1998a; Searle et al., 1999).

8. Functional investigations Granulated lymphocytes are the predominant leucocyte population in human endometrium in early pregnancy and functional studies of leucocyte populations in human decidua have not surprisingly focused on these cells. Other major

J.N. Bulmer, G.E. Lash / Molecular Immunology 42 (2005) 511–521

515

leucocyte populations in human endometrium are T lymphocytes, predominantly of the CD8 subtype, and macrophages; unlike other mucosal sites, B lymphocytes are very scanty, as are CD16+ CD56+ NK cells (Bulmer et al., 1991) and plasma cells are not a normal feature, their presence indicating endometritis. Despite considerable effort, however, the in vivo role of uterine NK cells remains unknown.

sive. We have noted different cytokine production by CD8+ cells purified from early pregnancy decidua between 8 and 13 weeks’ gestation (Scaife et al., 2004) but to date the possibility that uNK cell function differs at different gestational ages has not been addressed.

8.1. Preparation of uNK cells for functional investigations

Following recognition as possible NK cells there were several reports of cytotoxic activity by CD56-positive cells in early pregnancy decidua against the NK cell target K562 (King et al., 1989b; Ritson and Bulmer, 1989; Ferry et al., 1990). Reports of their cytotoxic capability varied, however, and the general consensus was that, despite their prominent perforin granules the cytotoxic activity of uNK cells was reduced compared with peripheral blood NK cells. In contrast, Christmas et al. (1990) produced CD3-negative lymphocyte clones from human early pregnancy decidua and failed to detect any cytotoxic activity, proposing that reports of cytotoxic activity by freshly isolated CD56-positive cells could be due to the inclusion of small numbers of peripheral blood NK cells. Because of their abundance most studies of uNK cells have purified cells from early pregnancy samples. There is only one report of the cytotoxic activity of primary cultures of uNK cells from non-pregnant endometrium; this demonstrated lysis of K562 cells by CD56-positive cells purified by MiniMACS immunomagnetic positive selection at all stages of the menstrual cycle, apart from the early proliferative phase (Jones et al., 1997). Interestingly in this study the lytic activity of uterine CD56-positive cells did not differ from that of peripheral blood CD56-positive cells. Reports of cytotoxic activity by uNK cells have inevitably simulated interest in the possibility that these cells can control the invasion into the uterus by extravillous trophoblast which is an essential feature of normal pregnancy (see below).

Comparison of data from different groups is complicated by the varying approaches to uNK cell separation and purification, which is likely to lead to the inclusion of different cell types. Initial dispersal of decidua may be by mechanical or enzymatic methods, leading to major differences in the composition of unseparated decidual cell suspensions (Ritson and Bulmer, 1987b). Different durations of enzyme digestion are likely to lead to release into the cell suspension of varying numbers of peripheral blood cells contained within decidual vessels and this is indicated by the inclusion in different studies of varying proportions of CD16-positive cells. Furthermore, CD56 can also be expressed by activated T cells and NK T cells and the inclusion of these may alter results, particularly in analysis of potential cytokine production. The phenotype of the ‘uterine NK cell’ population itself also varies, with subpopulations expressing CD2, CD8, CD49a, CD49d, CD103 and CD11a, amongst others (Bulmer et al., 1991; Gudelj et al., 1996; Searle et al., 1999). Different subsets of uNK cells have also been reported in mouse, based on lectin reactivity (Paffaro et al., 2003). The possibility that tissue disaggregation and cell purification may lead to cell activation should also be considered. Flow cytometry studies have reported expression of CD69 by uterine NK cells (King et al., 1991; Nishikawa et al., 1991) but immunohistochemical studies of tissue sections indicated that in situ CD69 expression was predominantly by endometrial T cells, with expression of CD69 by CD56 decidual lymphocytes appearing during the isolation procedure (Vassiliadou and Bulmer, 1998b). Whether there are phenotypic differences between uNK cells in decidua basalis and decidua parietalis has not been reported and it is likely that exposure to HLA-C, E and Gpositive trophoblast would alter function and therefore phenotype. The possibility that small gestational age differences could have a profound effect on cell function has also not been considered. Most researchers purify uNK cells from ‘first trimester’ or ‘early pregnancy’ decidua ranging from 8 to 13 weeks’ gestation. Profound changes occur in the structure of the placenta during this period, including the entrance of blood into the intervillous space at around 10 weeks’ gestation (Jauniaux et al., 2000). It seems feasible therefore that the function of cells in decidua at 8 weeks gestation when there is no maternal blood flow into the intervillous space would differ from that at 13 weeks’ gestation when blood flow is well established and trophoblast invasion into decidua is exten-

8.2. Cytotoxicity

8.3. Immunoregulatory activity There has been considerable interest in the possibility of local intrauterine immunosuppression in pregnancy and studies in mouse pregnancy had implicated TGF-␤2 production by a non-T non-B granulated lymphocyte as a local intrauterine suppressor cell population (Clark et al., 1990). Uterine NK cells were attractive candidates for such a role in human pregnancy and TGB-␤2 was reported in CD56-positive cells from human decidua (Clark et al., 1994). However, numerous cell types in human decidua, including decidual stromal cells, macrophages and endometrial glands have been implicated in local immunosuppressive activity and the relative importance of any of these in vivo cannot be assessed with any confidence. Investigation of pathological pregnancy may provide clues to in vivo function, with abnormalities in immunosuppressive activity being reported in a proportion of miscarriages (Lea et al., 1995; Vassiliadou et al., 1999). In studies of human pregnancy, however, it is difficult to assess

516

J.N. Bulmer, G.E. Lash / Molecular Immunology 42 (2005) 511–521

cause or effect since tissues are often retained for several days until clinical presentation leads to removal. Thus, although CD56+ cells from human decidua may mediate immunosuppression within the uterus, the importance of any in vivo effect remains to be established.

rent effort is focused on a potential role in the control of trophoblast invasion and uterine vascular remodelling in pregnancy.

9. Control of trophoblast invasion 8.4. Cytokine production There is considerable interest in the role of cytokines in pregnancy, a bias towards type 2 cytokines favouring successful pregnancy, while a type 1 cytokines are considered to be detrimental (Wegmann et al., 1993). Although the original hypothesis was based on T helper cell cytokine production, these cells are a minor leucocyte population in human decidua and attention has focused on uNK cell cytokine production. Uterine NK cells are a rich source of many different cytokines and growth factors including TNF-␣, IL-10, GMCSF, IL-1␤, TGF-␤1, CSF-1, LIF and IFN-␥ (Jokhi et al., 1994, 1997; Rieger et al., 2001). It was suggested that an unfavourable type 1 cytokine balance with local IL-2 production could lead to generation of LAK cells within decidua capable of trophoblast lysis and hence pregnancy loss (Drake and Head, 1989). However, investigation of cytotoxic activity of CD56+ uNK cells in sporadic abortion failed to show any evidence of LAK activity, around half of the samples showing reduced cytotoxicity of K562 cells by uNK cells compared with gestationally matched normal pregnancy (Vassiliadou and Bulmer, 1998c). Undoubtedly uNK cells can produce a range of cytokines but at present the functional importance of these in normal pregnancy is not known. Current interest is directed towards the role of uNK cell derived cytokines, particularly IFN-␥, in the control of trophoblast invasion by a non-cytotoxic mechanism and remodelling of uterine spiral arteries in the first half of normal pregnancy (see below). 8.5. Innate immunity Although the proposal that uNK cells have a fundamental role in pregnancy is attractive, a role in innate immunity within the uterus is also possible. The endometrium is unusual as a mucosal site in being devoid of plasma cells and containing very few B lymphocytes. Apart from pregnancy with exposure to the semi-allogeneic fetoplacental unit, endometrium is potentially exposed even in the non-pregnant state to a range of foreign antigens, including spermatozoa. Given that granulated lymphocytes occur in the endometrium of several species which do not undergo an invasive type of placentation (see below), it is possible that the role of uNK cells in pregnancy, in particular, but also in non-pregnant endometrium is to protect the fetoplacental unit from infection. 8.6. Summary Despite intensive research activity, the role of the uNK cells remains unknown. The emphasis in research studies has been on a potential role in pregnancy and much cur-

Invasion of uterine tissues and spiral arteries is an essential feature of successful pregnancy in humans (Pijnenborg et al., 1983). This invasion is tightly controlled, excessive invasion occurring in placenta accreta (which is associated with lack of decidua), and inadequate invasion being well recognised in pre-eclampsia, some cases of fetal growth restriction, preterm labour and late miscarriage (Pijnenborg et al., 1983). Invasion of trophoblast into the uterus occurs from implantation onwards and extends by the end of the first half of pregnancy into the inner third of the myometrium. Trophoblast cells also invade the uterine spiral arteries, with cells migrating within the lumen as far as the inner third of the myometrium. Whether due solely to the effect of trophoblast, the end result of the invasion of uterine tissues and arteries by trophoblast is the loss of smooth muscle and elastin from the arterial media and their conversion into dilated vascular channels able to supply a greatly increased blood supply required to the fetoplacental unit and unable to respond to vasomotor influences. Due to the high abundance of uNK cells in the decidua in the first and second trimesters of pregnancy, and their association with extravillous trophoblast cells it has been proposed that they play an active role in the regulation of trophoblast invasion. Several mechanisms have been suggested for this regulation including cytotoxicity, local cytokine production and induction of trophoblast apoptosis. While peripheral blood CD56+ CD16+ NK cells are highly cytotoxic to several different cell types, experimental evidence from several laboratories suggests that CD56bright CD16-negative uNK cells possess very little cytotoxic activity against trophoblast cells without prior activation with IL-2 (King et al., 1989b; Christmas et al., 1990; Ferry et al., 1991; Rouas-Freiss et al., 1997). This lack of cytotoxicty against trophoblast has been attributed to the expression of non-classical HLA-class 1 antigens by extravillous trophoblast (Chumbley et al., 1994; Rouas-Freiss et al., 1997). Decidual NK cells express a range of receptors which could allow recognition of invasive trophoblast cells which express HLA-G, HLA-E and HLA-C (King et al., 2000). In mouse and rat, time lapse video studies of co-cultures revealed lysis of occasional labyrinthine trophoblast cells (Stewart and Mukhtar, 1988) and we have noted lysis of rare choriocarcinoma cells when co-cultured with positively selected CD56-positive lymphocytes from early pregnancy decidua in time lapse video studies (Bulmer, unpublished data). Nevertheless, there is no in situ evidence of trophoblast lysis by uNK cells in early pregnancy decidua, and indeed in humans and other primates, macrophages are equally if not more closely associated with the invasive extravillous trophoblast within decidua basalis (Bulmer et al., 1988; Slukvin et al., 2004).

J.N. Bulmer, G.E. Lash / Molecular Immunology 42 (2005) 511–521

As already noted uNK cells are a rich source of many different cytokines and growth factors including VEGF-C, PlGF, Ang-2, TNF-␣, IL-10, GM-CSF, IL-1␤, TGF-␤1, CSF1, LIF and IFN-␥ (Jokhi et al., 1994, 1997; Rieger et al., 2001; Li et al., 2001). The cytokine profile of uNK cells does, however, appear to differ dependent on the phenotype of the surrounding cells, such as HLA-G expressing cells. Altered production of IFN-␥ has been reported in decidua basalis compared with decidua parietalis (von Rango et al., 2003) and van der Meer et al. (2004) have reported increased production of IFN-␥ and VEGF by uNK cells in response to HLA-G. The role of several of the cytokines and growth factors produced by uNK cells on the control of extravillous trophoblast invasion has been studied using in vitro invasion assays. Ang2 (Dunk et al., 2000), TNF-␣ (Bauer et al., 2004; Otun et al., 2003), IFN-␥ (Otun et al., 2003) and TGF-␤1 (Graham et al., 1994; Otun et al., 2004) have all been shown to inhibit trophoblast invasion. However, the overall effect of uNK cells on trophoblast invasion, examined by use of uNK cell conditioned media or co-culture experiments in in vitro invasion assays, has not yet been investigated in detail. In the placental bed of uncomplicated pregnancies during early gestation up to 30% of the extravillous trophoblast cells are undergoing apoptosis as determined by immunostaining for M30, which detects a neoepitope of cytokeratin 18 revealed after caspase mediated cleavage (Disep et al., 2003). Double immunohistochemical labelling has shown that many of these apoptosing trophoblast cells are surrounded by uNK cells. However, the stimulus for the induction of apoptosis in the trophoblast cells remains unclear. Indeed, the ability of uNK cells to induce cellular apoptosis has been very poorly studied. Li et al. (2001) demonstrated no difference in HUVEC cell apoptosis with or without culture in uNK cell conditioned media. In addition, in our own hands TNF-␣, TGF␤1 and IFN-␥ were all unable to alter the level of M30 immunopositive cells in placental explants cultured for 6 days in the presence of cytokine (Otun et al., 2003, 2004). Another possible explanation for the observed association between apoptosing trophoblast cells and uNK cells, may be that the trophoblast cells which undergo apoptosis recruit uNK cells after apoptosis has been initiated. If uNK cells regulate trophoblast invasion then the most likely mechanism by which this occurs is mediated by uNK cell cytokine secretion. The study of uNK cell cytokine production in vivo is problematic for several reasons including; the short half life of many cytokines, difficulty in differentiating between bound exogenous cytokine immunostaining and endogenous production and potential differences in the activation state of the uNK cells dependent on their microenvironment. Therefore, many researchers (including ourselves) have used purified uNK cell culture systems, although it has still not been established whether isolation leads to activation of this cell type. In addition, there is a growing literature suggesting that the uNK cell cytokine profile is altered on contact with extravillous trophoblast cells, and in particular with HLA-G,

517

and therefore the source (decidua basalis versus decidua parietalis) of uNK cells for isolation or in vivo studies may become important. One novel approach to answer these questions will be to use laser microdissection of placental bed and non-placental bed biopsies coupled with real-time RT-PCR to examine cytokine mRNA levels.

10. Control of vascular remodelling Uterine NK cells have also been proposed to play a role in spiral artery transformation. The majority of evidence for uNK cell involvement in spiral artery transformation has come from transgenic mouse models. For example; tg␧26 (NK− T− B− ) female mice do not have fertility problems but experience >50% fetal loss between 10 and 14 days post conception, the surviving fetuses have very small placentae and are small throughout the rest of their adult lives compared to controls. In addition, there is a lack of remodelling of the uterine spiral arteries in these mice. However, transplantation of bone marrow from scid donors results in a reconstitution of NK (and uNK) cells without a reconstitution of T cells reverses the reproductive deficiencies of the tg␧26 (Guimond et al., 1998). In an elegant set of experiments the same group has also demonstrated that the major uNK cell product responsible for the spiral artery remodelling defects is IFN-␥ (Ashkar et al., 2000). Mice deficient in uNK cells (RAG2−/−) or IFN-␥ signalling (IFN-␥−/−) have implantation site abnormalities and failure of decidual artery remodelling (Ashkar et al., 2000). These defects are not reversed when RAG2−/− mice are engrafted with bone marrow from IFN-␥−/−, strongly suggesting the functions of uNK cells in mouse are mediated by IFN-␥ (Ashkar et al., 2000). Much less is known about the mechanisms involved in spiral artery remodelling in humans but it is likely that they are, at least in part, different from those controlling trophoblast invasion. Early structural changes in decidual spiral arteries, including dilatation and medial disorganization, occur before cellular interaction with trophoblast (Craven et al., 1998), but at a time when uNK cells are present (Bulmer et al., 1991). Uterine NK cells reduce in number, at least as evidenced by their granule content, after 20 weeks’ gestation when vascular changes are generally complete (Fig. 1). In the placental bed uNK cells are seen to be very closely associated with non-transformed, transformed and transforming spiral arteries. In addition to IFN-␥, Ang-2 may be an important uNK cell derived mediator of spiral artery transformation. On binding to Tie-2 Ang-2 acts to counteract the effects of VEGF-A and Ang-1 to destabilise vessel structure and in the adult is only found to be expressed in tissues where vascular remodelling is taking place. uNK cells are a rich source of Ang-2 (Li et al., 2001) and in the placental bed are found predominantly in association with vessels undergoing remodelling. The mechanisms responsible for spiral artery transformation in the human are poorly understood due to a lack

518

J.N. Bulmer, G.E. Lash / Molecular Immunology 42 (2005) 511–521

of appropriate models for the study of this important phenomenon. In addition, the sequential phenotypic and morphometric changes that the remodelled vessels undergo are very poorly defined. At present, several groups including our own are involved in establishing appropriate models to answer these questions (Cartwright et al., 2002).

11. Granulated lymphocytes in the endometrium of other species The existence of granulated cells within the uterus of species other than human has been recognised for some time and raises important questions about their function. Most studies in species other than humans have focused on mouse uNK cells, formerly known as granulated metrial gland cells, and there have been many studies utilising mouse knockouts to investigate the role of these cells in vivo, studies which are impossible to replicate in the human system. Although mouse placentation is haemochorial in type there are important differences between mouse and human and the rat may be a better model of intrauterine events in pregnancy for human pregnancy than mouse, since the latter species does show invasion of the uterus by trophoblast in pregnancy. The presence of granulated cells in the endometrium of non-human primates was recognised in early histological studies (Bartelemez et al., 1951; Cardell et al., 1969) and recent immmunohistochemical studies have analysed the distribution of CD56+ cells in the endometrium of the Rhesus monkey at the time of implantation and in early pregnancy (Slukvin et al., 2004) and functional studies of isolated decidual lymphocytes have confirmed NK cytotoxicity (Slukvin et al., 2001). The cells are also closely associated with vessels invaded by trophoblast (Slukvin et al., 2000), although macrophages appeared to have a stronger association with trophoblast than uNK cells. Primates may therefore be a good model for further analysis of the function of uNK cells. Granulated cells have also been reported in the decidua and endometrium of a wide range of other species, termed minor species in a recent review (Stewart, 1998). These include species which do not have haemochorial placentation and include those species with the least invasive epitheliochorial type of placentation (Engelhardt and King, 1996/1997). Indeed trophoblast invasion is not a major feature in many species which do have a haemochorial placenta and trophoblast invasion is not a major feature of mouse pregnancy. This raises questions about the proposed role of the cells in the control of trophoblast invasion since the cells are present in species where trophoblast invasion does not occur. The possibility of other roles must therefore be considered. In human pregnancy the remodelling which occurs in the spiral arteries has been largely attributed to the effects of trophoblast (Kam et al., 1999), although priming prior to trophoblast invasion has been reported (Craven et al., 1998). It is possible that in species which do not undergo trophoblast invasion, there

are vascular changes in the uterus in pregnancy which are mediated by the uNK cell population. NK cells are highly conserved between species and their role may be conserved in pregnancy in the different species. It is possible that the uNK cells have a role in the various species in protection of the uterus and the fetoplacental unit from infection and other foreign antigens, particularly in pregnancy when the immune system may be compromised. Alternatively, the role in the different species may have evolved, with additional or alternative roles being acquired in humans, including a role in the control of normal trophoblast invasion. Studies of pathological pregnancy in humans provide a possible approach to study the importance of the uNK cells in successful human pregnancy.

12. Conclusions The existence of granulated cells in the endometrium of several species has been known for many years but their recognition as a type of NK cell has been relatively recent. Despite many in vitro studies of uterine NK cells from mouse and human, the in vivo role is not known. Current research effort is focused on the possible role of uNK cells in the control of trophoblast invasion and remodelling of the uterine spiral arteries in the first half of pregnancy. After earlier studies focused on the potential cytotoxicity of uNK cells against extravillous trophoblast, more recently there has been increasing interest in cytokine production by these cells. There is persuasive evidence in mouse that the uNK cells play a role in mediating vascular changes in pregnancy, mediated by IFN-␥. To date no similar evidence has emerged in humans, although uNK cell cytokine production is altered in the presence of HLA-G and uNK cells are closely associated with endometrial spiral arteries in both non-pregnant and pregnant endometrium. Although a role in relation to pregnancy and in particular trophoblast invasion and vascular remodelling is attractive, there are problems which require to be considered. Uterine NK cells are confined to the endometrium, whereas trophoblast invasion and spiral artery transformation extends into the inner third of the myometrium where there are no uNK cells. Invasion by trophoblast occurs in extrauterine ectopic pregnancy in the absence of uNK cells (Vassiliadou and Bulmer, 1998a; von Rango et al., 2001); it is difficult to assess whether such invasion is excessive since the wall of the fallopian tube is not comparable to that of the uterus and tube rupture may be due to mechanical reasons rather than excessive trophoblast invasion. Uterine NK cells are also present in many species which do not undergo an invasive form of placentation and an alternative role would have to be proposed for these. Uterine NK cells are present in large numbers in nonpregnant endometrium and are particularly prominent in progesterone treated endometrium. Although some studies in mouse have supported trafficking of precursors from blood,

J.N. Bulmer, G.E. Lash / Molecular Immunology 42 (2005) 511–521

it is uncertain whether the increase in uNK cells in the late secretory phase of the menstrual cycle can be attributed to local proliferation or influx from the circulation. The very close relationship between uNK cells and decidualised endometrial stromal cells, suggests a fundamental role for these cells in the differentiation of uNK cells within the uterus.

References Ashkar, A.A., Di Santo, J.P., Croy, B.A., 2000. Interferon gamma contributes to initiation of uterine vascular modification, decidual integrity, and uterine natural killer cell maturation during normal murine pregnancy. J. Exp. Med. 192, 259–270. Bartelemez, G.W., Corner, G.W., Hartman, G.G., 1951. Cyclic changes in the endometrium of the rhesus monkey (Macaca mulatto). Contrib. Embryol. Carnegie Inst. Wash. 34, 99–144. Bauer, S., Pollheimer, J., Hartmann, J., Husslein, P., Aplin, J.D., Knofler, M., 2004. Tumor necrosis factor-alpha inhibits trophoblast migration through elevation of plasminogen activator inhibitor-1 in first-trimester villous explant cultures. J. Clin. Endocrinol. Metab. 89, 812–822. Bulmer, J.N., Sunderland, C.A., 1984. Immunohistological characterization of lymphoid cell populations in the early human placental bed. Immunology 52, 349–357. Bulmer, J.N., Smith, J., Morrison, L., Wells, M., 1988. Maternal and fetal cellular relationships in the human placental basal plate. Placenta 9, 237–246. Bulmer, J.N., Morrison, L., Longfellow, M., Ritson, A., Pace, D., 1991. Granulated lymphocytes in human endometrium: histochemical and immunohistochemical studies. Hum. Reprod. 6, 791–798. Burrows, T.D., King, A., Loke, Y.W., 1995. The role of integrins in adhesion of decidual NK cells to extracellular matrix and decidual stromal cells. Cell. Immunol. 166, 53–61. Campbell, J.J., Qin, S., Unutmaz, D., Soler, D., Murphy, K.E., Hodge, M.R., Wu, L., Butcher, E.C., 2001. Unique subpopulations of CD56+ NK and NK-T peripheral blood lymphocytes identified by chemokine receptor expression repertoire. J. Immunol. 166, 6477–6482. Cardell, R.R., Hisaw, F.L., Dawson, A.B., 1969. The fine structure of granular cells in the uterine endometrium of the rhesus monkey (Macaca mulatto) with a discussion of the possible function of these cells in relaxin secretion. Am. J. Anat. 124, 307–340. Cartwright, J.E., Kenny, L.C., Dash, P.R., Crocker, I.P., Aplin, J.D., Baker, P.N., Whitley, G.S., 2002. Trophoblast invasion of spiral arteries: a novel in vitro model. Placenta 23, 232–235. Chantakru, S., Miller, C., Roach, L.E., Kuziel, W.A., Maeda, N., Wang, W.C., Evans, S.S., Croy, B.A., 2002. Contributions from self-renewal and trafficking to the uterine NK cell population of early pregnancy. J. Immunol. 168, 22–28. Christmas, S.E., Bulmer, J.N., Meager, A., Johnson, P.M., 1990. Phenotypic and functional analysis of human CD3− decidual leucocyte clones. Immunology 71, 182–189. Chumbley, G., King, A., Robertson, K., Holmes, N., Loke, Y.W., 1994. Resistance of HLA-G and HLA-A2 transfectants to lysis by decidual NK cells. Cell. Immunol. 155, 312–322. Clark, D.A., Flanders, K.C., Banwatt, D., Millar-Book, W., Manuel, J., Stedronska-Clark, J., Rowley, B., 1990. Murine pregnancy decidua produces a unique immunosuppressive molecule related to transforming growth factor beta-2. J. Immunol. 144, 3008–3014. Clark, D.A., Vince, G., Flanders, K.C., Hirte, H., Starkey, P., 1994. CD56+ lymphoid cells in human first trimester pregnancy decidua as a source of novel transforming growth factor-beta 2-related immunosuppressive factors. Hum. Reprod. 9, 2270–2277. Cooper, M.A., Fehniger, T.A., Caligiuri, M.A., 2001. The biology of human natural killer-cell subsets. Trends Immunol. 22, 633–640.

519

Craven, C.M., Morgan, T., Ward, K., 1998. Decidual spiral artery remodelling begins before cellular interaction with cytotrophoblasts. Placenta 19, 241–252. Dallenbach, F.D., Dallenbach-Hellweg, G., 1964. Immunohistolgische untersuchungen zur lakalisation des relaxius in menschlicher placenta und decidua. Virch. Arch. Path. Anat. 337, 301–316. Dallenbach-Hellweg, G., 1987. Histopathology of the Endometrium. Springer-Verlag, Berlin. Dallenbach-Hellweg, G., Nette, G., 1964. Morphological and histochemical observations on trophoblast and decidua of the basal plate of the human placenta at term. J. Anat. 115, 309–326. Disep, B., Innes, B.A., Ball, E., Robson, S.C., Bulmer, J.N., 2003. Quantitative assessment of apoptosis in the placenta bed of normal and abnormal pregnancies. Placenta 24, A.16. Drake, B.L., Head, J.R., 1989. Murine trophoblast can be killed by lymphokine-activated killer cells. J. Immunol. 143, 9–14. Dunk, C., Shams, M., Nijjar, S., Rhaman, M., Qiu, Y., Bussolati, B., Ahmed, A., 2000. Angiopoietin-1 and angiopoietin-2 activate trophoblast Tie-2 to promote growth and migration during placental development. Am. J. Pathol. 156, 2185–2199. Engelhardt, H., King, G.J., 1996/1997. Uterine natural killer cells in species with epitheliochorial placentation. Nat. Immun. 15, 53– 69. Ferry, B.L., Starkey, P.M., Sargent, I.L., Watt, G.M., Jackson, M., Redman, C.W., 1990. Cell populations in the human early pregnancy decidua: natural killer activity and response to interleukin-2 of CD56positive large granular lymphocytes. Immunology 70, 446–452. Ferry, B.L., Sargent, I.L., Starkey, P.M., Redman, C.W., 1991. Cytotoxic activity against trophoblast and choriocarcinoma cells of large granular lymphocytes from human early pregnancy decidua. Cell. Immunol. 132, 140–149. Graham, C.H., Connelly, I., MacDougall, J.R., Kerbel, R.S., StetlerStevenson, W.G., Lala, P.K., 1994. Resistance of malignant trophoblast cells to both the anti-proliferative and anti-invasive effects of transforming growth factor-beta. Exp. Cell. Res. 214, 93–99. Gubbay, O., Critchley, H.O., Bowen, J.M., King, A., Jabbour, H.N., 2002. Prolactin induces ERK phosphorylation in epithelial and CD56(+) natural killer cells of the human endometrium. J. Clin. Endocrinol. Metab. 87, 2329–2335. Gudelj, L., Deniz, G., Rukavina, D., Johnson, P.M., Christmas, S.E., 1996. Expression of functional molecules by human CD3− decidual granular leucocyte clones. Immunology 87, 609–615. Gudelj, L., Christmas, S.E., Laskarin, G., Johnson, P.M., Podack, E.R., Rukavina, D., 1997. Membrane phenotype and expression of perforin and serine esterases by CD3− peripheral blood and decidual granular lymphocyte-derived clones. Am. J. Reprod. Immunol. 38, 162– 167. Guimond, M.J., Wang, B., Croy, B.A., 1998. Engraftment of bone marrow from severe combined immunodeficient (SCID) mice reverses the reproductive deficits in natural killer cell-deficient tg epsilon 26 mice. J. Exp. Med. 187, 217–223. Haller, H., Radillo, O., Rukavina, D., Tedesco, F., Candussi, G., Petrovic, O., Randic, L., 1993. An immunohistochemical study of leucocytes in human endometrium, first and third trimester basal decidua. J. Reprod. Immunol. 23, 41–49. Haller, H., Tedesco, F., Rukavina, D., Radillo, O., Gudelj, L., Beer, A.E., 1995. Decidual–trophoblast interactions: decidual lymphoid cell populations in basal and parietal decidua. J. Reprod. Immunol. 28, 165–171. Hammer, A., Dohr, G., 1999. Apoptotic nuclei within the uterine decidua of first trimester pregnancy arise from CD45 positive leukocytes. Am. J. Reprod. Immunol. 42, 88–94. Hamperl, H., Hellweg, G., 1958. Granular endometrial stroma cells. Obstet. Gynecol. 11, 379–387. Hanna, J., Wald, O., Goldman-Wohl, D., Prus, D., Markel, G., Gazit, R., Katz, G., Haimov-Kochman, R., Fujii, N., Yagel, S., Peled, A., Mandelboim, O., 2003. CXCL12 expression by invasive trophoblasts

520

J.N. Bulmer, G.E. Lash / Molecular Immunology 42 (2005) 511–521

induces the specific migration of CD16− human natural killer cells. Blood 102, 1569–1577. Henderson, T.A., Saunders, P.T., Moffett-King, A., Groome, N.P., Critchley, H.O., 2003. Steroid receptor expression in uterine natural killer cells. J. Clin. Endocrinol. Metab. 88, 440–449. Jauniaux, E., Watson, A.L., Hempstock, J., Bao, Y.P., Skepper, J.N., Burton, G.J., 2000. Onset of maternal arterial blood flow and placental oxidative stress. A possible factor in human early pregnancy failure. Am. J. Pathol. 157, 2111–2122. Jokhi, P.P., King, A., Sharkey, A.M., Smith, S.K., Loke, Y.W., 1994. Screening for cytokine messenger ribonucleic acids in purified human decidual lymphocyte populations by the reversetranscriptase polymerase chain reaction. J. Immunol. 153, 4427– 4435. Jokhi, P.P., King, A., Loke, Y.W., 1997. Cytokine production and cytokine receptor expression by cells of the human first trimester placental–uterine interface. Cytokine 9, 126–137. Jones, R.K., Bulmer, J.N., Searle, R.F., 1997. Cytotoxic activity of endometrial granulated lymphocytes during the menstrual cycle in humans. Biol. Reprod. 57, 1217–1222. Jones, R.K., Searle, R.F., Stewart, J.A., Turner, S., Bulmer, J.N., 1998a. Apoptosis, bcl-2 expression, and proliferative activity in human endometrial stroma and endometrial granulated lymphocytes. Biol. Reprod. 58, 995–1002. Jones, R.K., Searle, R.F., Bulmer, J.N., 1998b. Apoptosis and bcl-2 expression in normal human endometrium, endometriosis and adenomyosis. Hum. Reprod. 13, 3496–3502. Kam, E.P., Gardner, L., Loke, Y.W., King, A., 1999. The role of trophoblast in the physiological change in decidual spiral arteries. Hum. Reprod. 14, 2131–2138. Kazzaz, B.A., 1972. Specific endometrial granular cells. A semiquantitative study. Eur. J. Obstet. Gynecol. 3, 77–84. Khong, T.Y., 1987. Immunohistologic study of the leukocytic infiltrate in maternal uterine tissues in normal and preeclamptic pregnancies at term. Am. J. Reprod. Immunol. Microbiol. 15, 1–8. King, A., Wellings, V., Gardner, L., Loke, Y.W., 1989a. Immunocytochemical characterization of the unusual large granular lymphocytes in human endometrium throughout the menstrual cycle. Hum. Immunol. 24, 195–205. King, A., Birkby, C., Loke, Y.W., 1989b. Early human decidual cells exhibit NK activity against the K562 cell line but not against first trimester trophoblast. Cell. Immunol. 118, 337– 344. King, A., Balendran, N., Wooding, P., Carter, N.P., Loke, Y.W., 1991. CD3− leukocytes present in the human uterus during early placentation: phenotypic and morphologic characterization of the CD56++ population. Dev. Immunol. 1, 169–190. King, A., Wooding, P., Gardner, L., Loke, Y.W., 1993. Expression of perforin, granzyme A and TIA-1 by human uterine CD56+ NK cells implies they are activated and capable of effector functions. Hum. Reprod. 8, 2061–2067. King, A., Gardner, L., Loke, Y.W., 1996. Evaluation of oestrogen and progesterone receptor expression in uterine mucosal lymphocytes. Hum. Reprod. 11, 1079–1082. King, A., Gardner, L., Loke, Y.W., 1999. Co-stimulation of human decidual natural killer cells by interleukin-2 and stromal cells. Hum. Reprod. 14, 656–663. King, A., Hiby, S.E., Gardner, L., Joseph, S., Bowen, J.M., Verma, S., Burrows, T.D., Loke, Y.W., 2000. Recognition of trophoblast HLA class I molecules by decidual NK cell receptors—a review. Placenta 21, S81–S85. Kitaya, K., Yasuda, J., Yagi, I., Tada, Y., Fushiki, S., Honjo, H., 2000. IL-15 expression at human endometrium and decidua. Biol. Reprod. 63, 683–687. Kitaya, K., Yasuda, J., Nakayama, T., Fushiki, S., Honjo, H., 2003. Effect of female sex steroids on human endometrial CD16neg CD56bright natural killer cells. Fertil. Steril. 79, 730–734.

Kusakabe, K., Okada, T., Sasaki, F., Kiso, Y., 1999. Cell death of uterine natural killer cells in murine placenta during placentation and preterm periods. J. Vet. Med. Sci. 61, 1093–1100. Lea, R.G., Underwood, J., Flanders, K.C., Hirte, H., Banwatt, D., Finotto, S., Ohno, I., Daya, S., Harley, C., Michel, M., et al., 1995. A subset of patients with recurrent spontaneous abortion is deficient in transforming growth factor beta-2-producing suppressor cells in uterine tissue near the placental attachment site. Am. J. Reprod. Immunol. 34, 52–64. Li, X.F., Charnock-Jones, D.S., Zhang, E., Hiby, S., Malik, S., Day, K., Licence, D., Bowen, J.M., Gardner, L., King, A., Loke, Y.W., Smith, S.K., 2001. Angiogenic growth factor messenger ribonucleic acids in uterine natural killer cells. J. Clin. Endocrinol. Metab. 86, 1823–1834. Marzusch, K., Ruck, P., Geiselhart, A., Handgretinger, R., Dietl, J.A., Kaiserling, E., Horny, H.P., Vince, G., Redman, C.W., 1993. Distribution of cell adhesion molecules on CD56++, CD3−, CD16− large granular lymphocytes and endothelial cells in first-trimester human decidua. Hum. Reprod. 8, 1203–1208. Nishikawa, K., Saito, S., Morii, T., Hamada, K., Ako, H., Narita, N., Ichijo, M., Kurahayashi, M., Sugamura, K., 1991. Accumulation of CD16–CD56+ natural killer cells with high affinity interleukin 2 receptors in human early pregnancy decidua. Int. Immunol. 3, 743– 750. Okada, S., Okada, H., Sanezumi, M., Nakajima, T., Yasuda, K., Kanzaki, H., 2000. Expression of interleukin-15 in human endometrium and decidua. Mol. Hum. Reprod. 6, 75–80. Otun, H.A., Lash, G.E., Bulmer, J.N., Searle, R.F., Robson, S.C., 2003. Inhibition of first trimester human extravillous trophoblast invasion by TNF-␣ and IFN-␥. Placenta 24, A.67. Otun, H.A., Lash, G.E., Innes, B., Bulmer, J.N., Searle, R.F., Robson, S.C., 2004. Regulation of trophoblast invasion by TGF-␤: role of apoptosis and proliferation. J. Soc. Gynecol. Investig. 11, 221a. Paffaro Jr., V.A., Bizinotto, M.C., Joazeiro, P.P., Yamada, A.T., 2003. Subset classification of mouse uterine natural killer cells by DBA lectin reactivity. Placenta 24, 479–488. Peel, S., Stewart, I.J., Bulmer, D., 1983. Experimental evidence for the bone marrow origin of granulated metrial gland cells of the mouse uterus. Cell Tissue Res. 233, 647–656. Phillips, R.J., Innes, B.A., Searle, R.F., Bulmer, J.N., Robson, S.C., 2003. Use of laser capture microdissection and RT-PCR to study gene expression in heterogeneous, differentiating trophoblast populations. J. Soc. Gynecol. Investig. 10, 398a. Pijnenborg, R., Bland, J.M., Robertson, W.B., Brosens, I., 1983. Uteroplacental arterial changes related to interstitial trophoblast migration in early human pregnancy. Placenta 4, 397–413. Pongcharoen, S., Bulmer, J.N., Searle, R.F., in press. No evidence for apoptosis of decidual leucocytes in normal and molar pregnancy; implications for immune privilege. Clin. Exp. Immunol. Rieger, L., Kammerer, U., Hofmann, J., Sutterlin, M., Dietl, J., 2001. Choriocarcinoma cells modulate the cytokine production of decidual large granular lymphocytes in coculture. Am. J. Reprod. Immunol. 46, 137–143. Ritson, A., Bulmer, J.N., 1987a. Endometrial granulocytes in human decidua react with a natural-killer (NK) cell marker, NKH1. Immunology 62, 329–331. Ritson, A., Bulmer, J.N., 1987b. Extraction of leucocytes from human deciduas. A comparison of dispersal techniques. J. Immunol. Meth. 104, 231–236. Ritson, A., Bulmer, J.N., 1989. Isolation and functional studies of granulated lymphocytes in first trimester human decidua. Clin. Exp. Immunol. 77, 263–268. Robertson, W.B., 1981. The Endometrium. Butterworths Press, London. Rosmaraki, E.E., Douagi, I., Roth, C., Colucci, F., Cumano, A., Di Santo, J.P., 2001. Identification of committed NK cell progenitors in adult murine bone marrow. Eur. J. Immunol. 31, 1900–1909. Rouas-Freiss, N., Goncalves, R.M., Menier, C., Dausset, J., Carosella, E.D., 1997. Direct evidence to support the role of HLA-G in protecting

J.N. Bulmer, G.E. Lash / Molecular Immunology 42 (2005) 511–521 the fetus from maternal uterine natural killer cytolysis. Proc. Natl. Acad. Sci. U.S.A. 94, 11520–11525. Ruck, P., Marzusch, K., Kaiserling, E., Horny, H.P., Dietl, J., Geiselhart, A., Handgretinger, R., Redman, C.W., 1994. Distribution of cell adhesion molecules in decidua of early human pregnancy. An immunohistochemical study. Lab. Invest. 71, 94–101. Scaife, P.J., Searle, R.F., Robson, S.C., Innes, B.A., Bulmer, J.N., 2003. A comparison of cytotoxic decidual leucocyte populations in early and late pregnancy. J. Soc. Gynecol. Investig. 10, 360. Scaife, P.J., Searle, R.F., Robson, S.C., Innes, B.A., Bulmer, J.N., 2004. cytokine production by CD8+ decidual leucocytes in early pregnancy. J. Soc. Gynecol. Investig. 11, 285A. Searle, R.F., Jones, R.K., Bulmer, J.N., 1999. Phenotypic analysis and proliferative responses of human endometrial granulated lymphocytes during the menstrual cycle. Biol. Reprod. 60, 871–878. Sindram-Trujillo, A.P., Scherjon, S.A., van Hulst-van Miert, P.P., van Schip, J.J., Kanhai, H.H., Roelen, D.L., Claas, F.H., 2003. Differential distribution of NK cells in decidua basalis compared with decidua parietalis after uncomplicated human term pregnancy. Hum. Immunol. 64, 921–929. Slukvin, I.I., Lunn, D.P., Watkins, D.I., Golos, T.G., 2000. Placental expression of the nonclassical MHC class I molecule Mamu-AG at implantation in the rhesus monkey. Proc. Natl. Acad. Sci. U.S.A. 97, 9104–9109. Slukvin, I.I., Watkins, D.I., Golos, T.G., 2001. Phenotypic and functional characterization of rhesus monkey decidual lymphocytes: rhesus decidual large granular lymphocytes express CD56 and have cytolytic activity. J. Reprod. Immunol. 50, 57–79. Slukvin, I.I., Breburda, E.E., Golos, T.G., 2004. Dynamic changes in primate endometrial leukocyte populations: differential distribution of macrophages and natural killer cells at the rhesus monkey implantation site and in early pregnancy. Placenta 25, 297–307. Smith, L.J., 1966. Metrial gland and other glycogen containing cells in the mouse uterus following mating and through implantation of the embryo. Am. J. Anat. 119, 15–23. Spornitz, U.M., 1992. The functional morphology of the human endometrium and decidua. Adv. Anat. Embryol. Cell Biol. 124, 1–99. Stewart, I., Mukhtar, D.D., 1988. The killing of mouse trophoblast cells by granulated metrial gland cells in vitro. Placenta 9, 417–425. Stewart, I.J., 1998. Granulated metrial gland cells in minor species. J. Reprod. Immunol. 40, 129–146. Stewart, I.J., 2000. Mouse granulated metrial gland cells require contact with stromal cells to maintain viability. J. Anat. 197, 495–502.

521

Stewart, J.A., Bulmer, J.N., Murdoch, A.P., 1998. Endometrial leucocytes: expression of steroid hormone receptors. J. Clin. Pathol. 51, 121– 126. Torres, M.S., Ace, C.I., Okulicz, W.C., 2002. Assessment and application of laser microdissection for analysis of gene expression in the rhesus monkey endometrium. Biol. Reprod. 67, 1067–1072. Trundley, A., Moffett, A., 2004. Human uterine leukocytes and pregnancy. Tissue Antigens 63, 1–12. van der Meer, A., Lukassen, H.G., van Lierop, M.J., Wijnands, F., Mosselman, S., Braat, D.D., Joosten, I., 2004. Membrane-bound HLAG activates proliferation and interferon-gamma production by uterine natural killer cells. Mol. Hum. Reprod. 10, 189–195. Vargas, M.L., Santos, J.L., Ruiz, C., Montes, M.J., Aleman, P., GarciaTortosa, C., Garcia-Olivares, E., 1993. Comparison of the proportions of leukocytes in early and term human decidua. Am. J. Reprod. Immunol. 29, 135–140. Vassiliadou, N., Bulmer, J.N., 1998a. Characterization of tubal and decidual leukocyte populations in ectopic pregnancy: evidence that endometrial granulated lymphocytes are absent from the tubal implantation site. Fertil. Steril. 69, 760–767. Vassiliadou, N., Bulmer, J.N., 1998b. Expression of CD69 activation marker by endometrial granulated lymphocytes throughout the menstrual cycle and in early pregnancy. Immunology 94, 368–375. Vassiliadou, N., Bulmer, J.N., 1998c. 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. Vassiliadou, N., Searle, R.F., Bulmer, J.N., 1999. Immunoregulatory activity of decidua in spontaneous early pregnancy loss. Hum. Reprod. 14, 2252–2256. von Rango, U., Classen-Linke, I., Kertschanska, S., Kemp, B., Beier, H.M., 2001. Effects of trophoblast invasion on the distribution of leukocytes in uterine and tubal implantation sites. Fertil. Steril. 76, 116–124. von Rango, U., Classen-Linke, I., Raven, G., Bocken, F., Beier, H.M., 2003. Cytokine microenvironments in human first trimester decidua are dependent on trophoblast cells. Fertil. Steril. 79, 1176–1186. Wegmann, T.G., Lin, H., Guilbert, L., Mosmann, T.R., 1993. Bidirectional cytokine interactions in the maternal-fetal relationship: is successful pregnancy a TH2 phenomenon? Immunol. Today 14, 353–356. Weill, P., 1921. Les cellules granuleuses des muqueuses intestinale et uterine. Arch. Anat. Microsc. 17, 77–82.

Lihat lebih banyak...

Comentários

Copyright © 2017 DADOSPDF Inc.