Uterine large granular lymphocytes: A possible role in embryonic implantation? Ashley King, MA, MRCP, and Y. W. Loke, MA, MD Cambridge, England Leukocytes in endometrium and decidua are phenotypically and morphologically similar to granulated lymphocytes at other mucosal surfaces. Although their exact functions are not known, we suggest they may have a role in the control of implantation and the transformation of the uterine vasculature by trophoblast on which the blood supply to the fetoplacental unit depends. (AM J OBSTET GVNECOl
1990;162:308·10.)
Key words: Large granular lymphocytes, natural killer (NK) cells, mucosal immunity, trophoblast, embryonic implantation The presence of abundant leukocytes in both pregnant and nonpregnant endometrium and decidua has long been recognized. Boyd,1 in his classic monograph on the placenta, described "lymphocyte-like cells" in the uterine lining and European histopathologists of the early twentieth century coined the term endometrial granulocytes for these cells.' The availability of monoclonal antibodies to leukocyte surface markers has made more detailed characterization possible. It is now generally agreed that the predominant leukocyte population in nonpregnant endometrium and decidua belong to the family of large granular lymphocytes, accounting for about 70% of the total, with the remainder being HLA-DR-positive macro phages (-20%) and mature CD3-positive T cells (-10%)."· Very few B cells are found. These uterine large granular lymphocytes express pre thymic T-lineage markers such as CD2 and CD7 but not the more mature ones like that of the T cell receptor-associated CD3 or the peripheral T cell marker CDS. Therefore, they appear not to have been processed by the thymus and would not be expected to have the ability to distinguish between self and nonself or to have rearrangement of the T cell antigen receptor a- and ~-genes. An additional feature of these cells is that they are stained strongly by Leu 1g, a marker used for the identification of peripheral blood large granular lymphocytes, but they do not express other natural killer (NK) markers such as CD16 or Leu 7." 5 Thus uterine large granular lymphocytes appear to correspond to a small subset (-2%) of peripheral blood large From the DIVisIOn of Cellular and Genetic Pathology, Cambridge University Department of Pathology. Supported by the American Fnends of Cambndge UniverSIty, East Anglian Regional Health Authority, M edical Research Council, World Health OrganizatIon, and the Wellcome Trust. Repnnt requests: Dr. Ashley King, DlvlSlon of Cellular and GenetIc Pathology, Cambridge University Department of Pathology, Tennis Court Road, Cambridge CB2 IQP. England. 611115876
308
granular lymphocytes that also have the same phenotype being CD3 negative, CD 16 negative, CDS negative, CD2 positive, CD7 positive, and are designated Leu 1gbroghl.6 By histologic study of the endometrium in different stages of the menstrual cycle it has been shown that uterine large granular lymphocytes greatly increase in number around the sixth postovulatory day, the time at which implantation would take place.' It is not known whether this increase is a result of migration or proliferation in situ. The large granular lymphocytes die 1 or 2 days before menstruation but if pregnancy occurs they are present thoughout the first trimester. Because the increase in number of large granular lymphocytes occurs before pregnancy their recruitment does not appear to be induced by the presence of the implanting blastocyst. A hormonal regulatory mechanism would seem more likely. The immunologic system of the endometrium is one of a number of local systems including the liver and the gut in which the traditional model of peripheral lymphoid organization appears not to apply. Mucosal lymphocytes form relatively autonomous immune systems presumably adapted for the specific conditions and functions of the surfaces they populate. There are interesting similarities between uterine lymphocytes and those in other mucosal organs. Granulated lymphoid cells are scarce in lymph nodes but are prominent at many mucosal surfaces, most notably the gut of humans and rodents where the intestinal intraepithelial lymphocytes contain a large proportion of granulated cells. 7 In both the rat and sheep liver granulated cells (pit cells) are abundant in the sinusoids. 8 • 9 Uterine granulated cells have been noted for many years in rodents (granulated metrial gland cells),IO humans (endometrial granulocytes),2 and more recently in sheep. II However, although human uterine granular lymphocytes share morphologic characteristics with
Volume 162 Number 2
mucosal large granular lymphocytes, their tissue distribution in relation to the epithelium is distinctive. Whereas granular lymphocytes are confined to the intraepithelial compartment at other mucosal sites, the uterine large granular lymphocytes occur predominantly in the stroma. Comparison of the immunophenotype of the lymphocytes from different mucosal sites suggests an overall pattern in which both thymic-dependent (CD3 positive) and thymic-independent (CD3 negative) cells are present. I" Some CD3-positive cells also express CDS but, like the human uterine large granular lymphocyte, do not express peripheral blood large granular lymphocyte markers such as CD16. The recent discovery that some gut intraepithelial lymphocytes are T cell receptor 'VB-positive in mice l3 . 14 raises questions about the relationship in local situations of CD3-negative T cell receptor 'VB-negative to CD3-positive T cell receptor 'VB-positive cells. The abundance of these T cell receptor 'VB-positive cells in athymic mice suggests that there is extrathymic production of T cell receptor 'VBpositive T cells possibly in these surface locations such as the gut, skin, and liver. 15 The function of these mucosal lymphocytes is unknown. The large granular lymphocyte subset in peripheral blood appears to be responsible for NK activity and this spontaneous cytotoxic activity is a property of both CD3-negative and T cell receptor 'VB-positive T cells. 16 NK activity has been demonstrated in lymphocyte preparations from the gut l7 and liver8 of many species but not in human gut intraepithelial lymphocytes. 18 However, Leu 19-positive large granular lymphocytes from early human decidua do exhibit considerable NK activity, 19. 20 which can be augmented after interleukin-2 stimulation. The usual targets for NK cells are virally infected cells, tumor cells, or embryonic cells. Although the target ligands for granulated mucosallymphocytes (whether CD3-negative or T cell receptor 'VB-positive cells) have not been identified they are likely to be different from those involved in antigen recognition by classical T cells. T cell receptor 'VBpositive T cells may recognize antigen in the context of a nonclassical Class I HLA antigen particularly as these cells often express CDS."I CD3-negative NK cells seem to preferentially recognize cells with low or aberrant m~or histocompatibility complex expression, thus providing a surveillance mechanism against malignant transformation. 22 Because large granular lymphocytes with NK activity are so abundant in the endometrium around the time of implantation an obvious question is whether they have an important role in pregnancy. In particular, is their interaction with the invading fetally derived trophoblast cells necessary for successful implantation? As trophoblasts are invasive embryonic cells they would be
Uterine large granular lymphocytes and implantation
309
possible targets for NK cells. Significantly the Class I HLA antigen expressed by these cells appears to be nonclassical, nonpolymorphic,23 and only scantly expressed on the cell surface (unpublished observations). In vitro studies with human trophoblast were impossible until recently when culture techniques were developed, yielding pure trophoblast populations. 24 With the use of these cultured trophoblast cells as targets we have shown that trophoblast is resistant to lysis by decidual NK cells. 19 This might be anticipated because trophoblast invasion is essential to the development of a normal blood supply to the fetoplacental unit. At implantation trophoblast invades maternal decidua and vessels, transforming the latter into low-resistance sinusoidal sacs. These large vessels lack the vasoconstrictive response to autonomic stimuli and maintain an adequate blood supply even in physiologic conditions such as maternal exertion, which cause diversion of blood from other viscera. 25 Histologic studies confirm that failure of the transformation of uterine arteries is associated with intrauterine growth retardation, stillbirth, and preeclampsia. Excessive penetration of maternal tissues also has disastrous consequences as is seen when implantation occurs in deficient decidua as in placenta percreta and ectopic pregnancy, implying that decidua provides some sort of barrier to the invasion!5 Therefore, one approach to these unresolved obstetric problems may lie in examination of the process and control of trophoblast invasion and vascular transformation in early pregnancy. Uterine large granular lymphocytes could influence this process in a number of ways. First, although they are not cytolytic to trophoblast they could limit the invasion by merely being cytostatic. This concept is supported by reports that show sex hormones may increase or decrease NK activity with the highest levels of NK activity occurring in the fertile phase of humans and mice!6 Resistance to metastatic spread of experimental tumors has been shown during this phase in mice!6 Human chorionic gonadotropin has also been shown to suppress NK lysis of K562 cells 27 and high levels of human chorionic gonadotropin are obviously present in the uterus in the first trimester. Second, uterine leukocytes could produce cytokines that might influence trophoblast growth and differentiation. Colony-stimulating factor-I and interleukin3 produced from murine lymphoma cell lines are observed to stimulate proliferation of placental cells in vitro. 2 • Most of the latter are shown to stain for cytokeratin microfilaments and therefore are assumed to be trophoblast. However, the human invasive trophoblast populations do not proliferate in vivo but only differentiate as they move into maternal tissues 29 so the relevance of this finding in the mouse to the human situation is not clear. Third, lymphokines produced by
310
King and Loke
large granular lymphocytes may influence other elements in the endometrium. Transforming growth factor-~ is thought to be produced by decidual NK lineage cells in mice. This exerts a negative effect on the generation of cytotoxic T-lymphocytes, thus inhibiting the efferent arm of any local destructive response.30 These activities against migrating trophoblast could account for the predominantly stromal distribution of uterine large granular lymphocytes in contrast with the intraepitheliallocation of gut granulated lymphocytes whose main function is thought to be to monitor the integrity of the epithelium." From the preceding review it is clear that the lymphocytes found in the uterus are not those that are usually associated with classical allogeneic recognition. Instead, they appear to be large granular lymphocytes that exhibit NK activity. Like similar cells elsewhere at mucosal surfaces in the body, the functions of decidual lymphocytes are not known, but it is tempting to think they may exert some influence on placental development. Elucidation of the mechanisms involved should increase our understanding of the process of implantation .
February 1990 Am J Obstet Gyneco1
12.
13. 14. 15.
16. 17.
18.
19.
20.
REFERENCES I. BoydJD, Hamilton WJ. The human placenta. Cambridge,
21.
2.
22.
3.
4.
5.
6.
7. 8.
9. 10. 11 .
England : W Heffer and Sons Ltd. , 1970:201. Hamperl H, Hellweg G. Granular endometrial stromal cells. Obstet Gynecol 1958;2:379-87. King A, Wellings V, Gardner L, Loke YW. Immunocytochemical characterization of the unusual large granular lymphocytes in human endometrium throughout the menstrual cycle. Hum Immunol 1989;24: 195-205. Starkey PM, Sargent IL, Redman CWG. Cell populations in human early pregnancy: characterization and isolation of large granular lymphocytes by flow cytometry. Immunology 1988;65:129-34. Bulmer IN, Ritson A. The decidua in early pregnancy. In : Beard R, Sharp F, eds. Early pregnancy loss, mechanisms, and treatment. Proceedings of the Royal College of Obstetricians and Gynaecologists. London: Royal College of Obstetricians and Gynaecologists, 1988: 171-80. Lanier LL, Le AM, Civin CI , Loken MR, PhillipsJH. The relationship of CD16 (Leu II) and Leu 19 (NKH- l) antigen expression on human peripheral blood NK cells and cytotoxic T lymphocytes. J Immunol 1986; 136: 4480-6. Dobbins WOo Human intestinal intraepithelial lymphocytes. Gut 1986;27:972-85. Bouwens L, Remels L, Baekela nd M, et al. Large granular lymphocytes or "pit cells" from rat liver: isolation, ultrastructural characterization and natural killer activity. Eur J Immunol 1987; 17:37-42. Meensen E, Gorrell MD, Brandon MR. Presence of a distinct CD8 + and CD5 - leukocyte subpopulations in the sheep liver. Immunology 1988;64:615-69. Stewart IJ. A morphological study of granulated metrial gland cells and trophoblast cells in the labyrinthine placenta of the mouse. J Anat 1984 ; 139:627-38. Lee CS, Gogolin-Ewens K, Brandon MR. Identification
23 .
24. 25. 26.
27. 28. 29.
30.
of a unique lymphocyte subpopulation in the sheep uterus. Immunology 1988;63:157-64. Mowat AMcI. The cellular basis of gastrointestinal immunity. In: Marsh MN, ed . Immunopathology of the small intestine. Chichester: Wiley and Sons Ltd. 1984:4172. Goodman T and Lefrancois L. Expression of the '(8 Tcell receptor on intestinal CD8 + intraepithelial lymphocytes. Nature 1988 ;333:855-7. Bonneville M, Janeway CA, Ito K, et al. Intestinal intraepithelial lymphocytes are a distinct set of '(8T cells. Nature 1988;336:479-481. Yoshikai Y, Reis MD, Mak TW. Athymic mice express a high level of functional ,(-chain but greatly reduced levels of Ct- and ~-chain T-cell reeptor messages. Nature 1988;324:482-5. Ortaldo JR, Longo DL. Human natural lymphocyte effector cells: definition, analysis of activity. and clinical effectiveness. JNCI 1988;80:999-1010. Flexman JP. Shellam GR, Mayrhofer G. Natural cytotoxicity. responsiveness to interferon, and morphology of IEL from small intestine of the rat. Immunology 1983;48: 733-41 . Cerf-Bensussan M, Guy-Grand D, Griscelli C. Intraepitheliallymphocytes of human gut: isolation , characterization and study of natural killer activity. Gut 1985;26: 81-8. King A, Birkby C, Loke YW. Early human decidual cells exhibit NK activity against the K562 cell line but not against first trimester trophoblast. Cell Immunol 1989; 118:337-44. Manaseki S. Searle RF. Natural Killer (NK) cell activity of the first trimester human decidua. Cell Immunol 1989; 121 : 166-73. Janeway CA. Frontiers of the immune system. Nature 1988;333 :804-6. Stork us WJ, Howell DN. Salter RD , DawsonJR, Cresswell P. NK susceptibility varies inversely with target cell Class I HLA antigen expression. J Immunol 1987; 138: 1657-9. Ellis SA, Sargent IL. Redman CWG. McMichael AJ. Evidence for a novel HLA antigen found on human extravillous trophoblast and a choriocarcinoma cell line. Immunology 1986;59:595-60 I. Loke YW. Burland K. Human trophoblast cells cultured in modified medium and supported by extracellular matrix. Placenta 1988;9: 173-82. Robertson WG. Pathology of the pregnant uterus. In: Fox H , ed. Obstetrical and gynecological pathology. London : Churchill Livingstone, 1987 : 1149-76. Hrushesky WJM, Gruber SA, Sothern RB. Hoffman RA. et al. Natural killer cell activity: age, estrous- . and circadian-stage dependence a nd inverse correlation with metastatic potential. JNCI 1988;80:1232-7. Sulke AN, Jones DB. Wood PJ. Hormonal modulation of human natural killer cell activity in vitro. J Reprod Fertil 1985;7: 105-10. Wegmann TG. Placental immunotropism: maternal T cells enhance placental growth and function. Am J Reprod Immunol Microbiol 1987;15:67-70. Bulmer IN, Morrison L, Johnson PM. Expression of the proliferation markers Ki67 and transferrin receptor by human trophoblast populations. J Reprod Immunol 1988; 14:291-302. Clark DA, Falbo M, Rowley RB , Banwatt B. StedronskaClark J. Active suppression of host versus graft reaction in pregnant mice. IX. Soluble suppressor activity obtained from allopregnant mouse decidua that blocks the cytolytic effector response to IL-2 is related to transforming growth factor-~. J Immunol 1988;141:3833-40.
1990;162:308·10.)
Key words: Large granular lymphocytes, natural killer (NK) cells, mucosal immunity, trophoblast, embryonic implantation The presence of abundant leukocytes in both pregnant and nonpregnant endometrium and decidua has long been recognized. Boyd,1 in his classic monograph on the placenta, described "lymphocyte-like cells" in the uterine lining and European histopathologists of the early twentieth century coined the term endometrial granulocytes for these cells.' The availability of monoclonal antibodies to leukocyte surface markers has made more detailed characterization possible. It is now generally agreed that the predominant leukocyte population in nonpregnant endometrium and decidua belong to the family of large granular lymphocytes, accounting for about 70% of the total, with the remainder being HLA-DR-positive macro phages (-20%) and mature CD3-positive T cells (-10%)."· Very few B cells are found. These uterine large granular lymphocytes express pre thymic T-lineage markers such as CD2 and CD7 but not the more mature ones like that of the T cell receptor-associated CD3 or the peripheral T cell marker CDS. Therefore, they appear not to have been processed by the thymus and would not be expected to have the ability to distinguish between self and nonself or to have rearrangement of the T cell antigen receptor a- and ~-genes. An additional feature of these cells is that they are stained strongly by Leu 1g, a marker used for the identification of peripheral blood large granular lymphocytes, but they do not express other natural killer (NK) markers such as CD16 or Leu 7." 5 Thus uterine large granular lymphocytes appear to correspond to a small subset (-2%) of peripheral blood large From the DIVisIOn of Cellular and Genetic Pathology, Cambridge University Department of Pathology. Supported by the American Fnends of Cambndge UniverSIty, East Anglian Regional Health Authority, M edical Research Council, World Health OrganizatIon, and the Wellcome Trust. Repnnt requests: Dr. Ashley King, DlvlSlon of Cellular and GenetIc Pathology, Cambridge University Department of Pathology, Tennis Court Road, Cambridge CB2 IQP. England. 611115876
308
granular lymphocytes that also have the same phenotype being CD3 negative, CD 16 negative, CDS negative, CD2 positive, CD7 positive, and are designated Leu 1gbroghl.6 By histologic study of the endometrium in different stages of the menstrual cycle it has been shown that uterine large granular lymphocytes greatly increase in number around the sixth postovulatory day, the time at which implantation would take place.' It is not known whether this increase is a result of migration or proliferation in situ. The large granular lymphocytes die 1 or 2 days before menstruation but if pregnancy occurs they are present thoughout the first trimester. Because the increase in number of large granular lymphocytes occurs before pregnancy their recruitment does not appear to be induced by the presence of the implanting blastocyst. A hormonal regulatory mechanism would seem more likely. The immunologic system of the endometrium is one of a number of local systems including the liver and the gut in which the traditional model of peripheral lymphoid organization appears not to apply. Mucosal lymphocytes form relatively autonomous immune systems presumably adapted for the specific conditions and functions of the surfaces they populate. There are interesting similarities between uterine lymphocytes and those in other mucosal organs. Granulated lymphoid cells are scarce in lymph nodes but are prominent at many mucosal surfaces, most notably the gut of humans and rodents where the intestinal intraepithelial lymphocytes contain a large proportion of granulated cells. 7 In both the rat and sheep liver granulated cells (pit cells) are abundant in the sinusoids. 8 • 9 Uterine granulated cells have been noted for many years in rodents (granulated metrial gland cells),IO humans (endometrial granulocytes),2 and more recently in sheep. II However, although human uterine granular lymphocytes share morphologic characteristics with
Volume 162 Number 2
mucosal large granular lymphocytes, their tissue distribution in relation to the epithelium is distinctive. Whereas granular lymphocytes are confined to the intraepithelial compartment at other mucosal sites, the uterine large granular lymphocytes occur predominantly in the stroma. Comparison of the immunophenotype of the lymphocytes from different mucosal sites suggests an overall pattern in which both thymic-dependent (CD3 positive) and thymic-independent (CD3 negative) cells are present. I" Some CD3-positive cells also express CDS but, like the human uterine large granular lymphocyte, do not express peripheral blood large granular lymphocyte markers such as CD16. The recent discovery that some gut intraepithelial lymphocytes are T cell receptor 'VB-positive in mice l3 . 14 raises questions about the relationship in local situations of CD3-negative T cell receptor 'VB-negative to CD3-positive T cell receptor 'VB-positive cells. The abundance of these T cell receptor 'VB-positive cells in athymic mice suggests that there is extrathymic production of T cell receptor 'VBpositive T cells possibly in these surface locations such as the gut, skin, and liver. 15 The function of these mucosal lymphocytes is unknown. The large granular lymphocyte subset in peripheral blood appears to be responsible for NK activity and this spontaneous cytotoxic activity is a property of both CD3-negative and T cell receptor 'VB-positive T cells. 16 NK activity has been demonstrated in lymphocyte preparations from the gut l7 and liver8 of many species but not in human gut intraepithelial lymphocytes. 18 However, Leu 19-positive large granular lymphocytes from early human decidua do exhibit considerable NK activity, 19. 20 which can be augmented after interleukin-2 stimulation. The usual targets for NK cells are virally infected cells, tumor cells, or embryonic cells. Although the target ligands for granulated mucosallymphocytes (whether CD3-negative or T cell receptor 'VB-positive cells) have not been identified they are likely to be different from those involved in antigen recognition by classical T cells. T cell receptor 'VBpositive T cells may recognize antigen in the context of a nonclassical Class I HLA antigen particularly as these cells often express CDS."I CD3-negative NK cells seem to preferentially recognize cells with low or aberrant m~or histocompatibility complex expression, thus providing a surveillance mechanism against malignant transformation. 22 Because large granular lymphocytes with NK activity are so abundant in the endometrium around the time of implantation an obvious question is whether they have an important role in pregnancy. In particular, is their interaction with the invading fetally derived trophoblast cells necessary for successful implantation? As trophoblasts are invasive embryonic cells they would be
Uterine large granular lymphocytes and implantation
309
possible targets for NK cells. Significantly the Class I HLA antigen expressed by these cells appears to be nonclassical, nonpolymorphic,23 and only scantly expressed on the cell surface (unpublished observations). In vitro studies with human trophoblast were impossible until recently when culture techniques were developed, yielding pure trophoblast populations. 24 With the use of these cultured trophoblast cells as targets we have shown that trophoblast is resistant to lysis by decidual NK cells. 19 This might be anticipated because trophoblast invasion is essential to the development of a normal blood supply to the fetoplacental unit. At implantation trophoblast invades maternal decidua and vessels, transforming the latter into low-resistance sinusoidal sacs. These large vessels lack the vasoconstrictive response to autonomic stimuli and maintain an adequate blood supply even in physiologic conditions such as maternal exertion, which cause diversion of blood from other viscera. 25 Histologic studies confirm that failure of the transformation of uterine arteries is associated with intrauterine growth retardation, stillbirth, and preeclampsia. Excessive penetration of maternal tissues also has disastrous consequences as is seen when implantation occurs in deficient decidua as in placenta percreta and ectopic pregnancy, implying that decidua provides some sort of barrier to the invasion!5 Therefore, one approach to these unresolved obstetric problems may lie in examination of the process and control of trophoblast invasion and vascular transformation in early pregnancy. Uterine large granular lymphocytes could influence this process in a number of ways. First, although they are not cytolytic to trophoblast they could limit the invasion by merely being cytostatic. This concept is supported by reports that show sex hormones may increase or decrease NK activity with the highest levels of NK activity occurring in the fertile phase of humans and mice!6 Resistance to metastatic spread of experimental tumors has been shown during this phase in mice!6 Human chorionic gonadotropin has also been shown to suppress NK lysis of K562 cells 27 and high levels of human chorionic gonadotropin are obviously present in the uterus in the first trimester. Second, uterine leukocytes could produce cytokines that might influence trophoblast growth and differentiation. Colony-stimulating factor-I and interleukin3 produced from murine lymphoma cell lines are observed to stimulate proliferation of placental cells in vitro. 2 • Most of the latter are shown to stain for cytokeratin microfilaments and therefore are assumed to be trophoblast. However, the human invasive trophoblast populations do not proliferate in vivo but only differentiate as they move into maternal tissues 29 so the relevance of this finding in the mouse to the human situation is not clear. Third, lymphokines produced by
310
King and Loke
large granular lymphocytes may influence other elements in the endometrium. Transforming growth factor-~ is thought to be produced by decidual NK lineage cells in mice. This exerts a negative effect on the generation of cytotoxic T-lymphocytes, thus inhibiting the efferent arm of any local destructive response.30 These activities against migrating trophoblast could account for the predominantly stromal distribution of uterine large granular lymphocytes in contrast with the intraepitheliallocation of gut granulated lymphocytes whose main function is thought to be to monitor the integrity of the epithelium." From the preceding review it is clear that the lymphocytes found in the uterus are not those that are usually associated with classical allogeneic recognition. Instead, they appear to be large granular lymphocytes that exhibit NK activity. Like similar cells elsewhere at mucosal surfaces in the body, the functions of decidual lymphocytes are not known, but it is tempting to think they may exert some influence on placental development. Elucidation of the mechanisms involved should increase our understanding of the process of implantation .
February 1990 Am J Obstet Gyneco1
12.
13. 14. 15.
16. 17.
18.
19.
20.
REFERENCES I. BoydJD, Hamilton WJ. The human placenta. Cambridge,
21.
2.
22.
3.
4.
5.
6.
7. 8.
9. 10. 11 .
England : W Heffer and Sons Ltd. , 1970:201. Hamperl H, Hellweg G. Granular endometrial stromal cells. Obstet Gynecol 1958;2:379-87. King A, Wellings V, Gardner L, Loke YW. Immunocytochemical characterization of the unusual large granular lymphocytes in human endometrium throughout the menstrual cycle. Hum Immunol 1989;24: 195-205. Starkey PM, Sargent IL, Redman CWG. Cell populations in human early pregnancy: characterization and isolation of large granular lymphocytes by flow cytometry. Immunology 1988;65:129-34. Bulmer IN, Ritson A. The decidua in early pregnancy. In : Beard R, Sharp F, eds. Early pregnancy loss, mechanisms, and treatment. Proceedings of the Royal College of Obstetricians and Gynaecologists. London: Royal College of Obstetricians and Gynaecologists, 1988: 171-80. Lanier LL, Le AM, Civin CI , Loken MR, PhillipsJH. The relationship of CD16 (Leu II) and Leu 19 (NKH- l) antigen expression on human peripheral blood NK cells and cytotoxic T lymphocytes. J Immunol 1986; 136: 4480-6. Dobbins WOo Human intestinal intraepithelial lymphocytes. Gut 1986;27:972-85. Bouwens L, Remels L, Baekela nd M, et al. Large granular lymphocytes or "pit cells" from rat liver: isolation, ultrastructural characterization and natural killer activity. Eur J Immunol 1987; 17:37-42. Meensen E, Gorrell MD, Brandon MR. Presence of a distinct CD8 + and CD5 - leukocyte subpopulations in the sheep liver. Immunology 1988;64:615-69. Stewart IJ. A morphological study of granulated metrial gland cells and trophoblast cells in the labyrinthine placenta of the mouse. J Anat 1984 ; 139:627-38. Lee CS, Gogolin-Ewens K, Brandon MR. Identification
23 .
24. 25. 26.
27. 28. 29.
30.
of a unique lymphocyte subpopulation in the sheep uterus. Immunology 1988;63:157-64. Mowat AMcI. The cellular basis of gastrointestinal immunity. In: Marsh MN, ed . Immunopathology of the small intestine. Chichester: Wiley and Sons Ltd. 1984:4172. Goodman T and Lefrancois L. Expression of the '(8 Tcell receptor on intestinal CD8 + intraepithelial lymphocytes. Nature 1988 ;333:855-7. Bonneville M, Janeway CA, Ito K, et al. Intestinal intraepithelial lymphocytes are a distinct set of '(8T cells. Nature 1988;336:479-481. Yoshikai Y, Reis MD, Mak TW. Athymic mice express a high level of functional ,(-chain but greatly reduced levels of Ct- and ~-chain T-cell reeptor messages. Nature 1988;324:482-5. Ortaldo JR, Longo DL. Human natural lymphocyte effector cells: definition, analysis of activity. and clinical effectiveness. JNCI 1988;80:999-1010. Flexman JP. Shellam GR, Mayrhofer G. Natural cytotoxicity. responsiveness to interferon, and morphology of IEL from small intestine of the rat. Immunology 1983;48: 733-41 . Cerf-Bensussan M, Guy-Grand D, Griscelli C. Intraepitheliallymphocytes of human gut: isolation , characterization and study of natural killer activity. Gut 1985;26: 81-8. King A, Birkby C, Loke YW. Early human decidual cells exhibit NK activity against the K562 cell line but not against first trimester trophoblast. Cell Immunol 1989; 118:337-44. Manaseki S. Searle RF. Natural Killer (NK) cell activity of the first trimester human decidua. Cell Immunol 1989; 121 : 166-73. Janeway CA. Frontiers of the immune system. Nature 1988;333 :804-6. Stork us WJ, Howell DN. Salter RD , DawsonJR, Cresswell P. NK susceptibility varies inversely with target cell Class I HLA antigen expression. J Immunol 1987; 138: 1657-9. Ellis SA, Sargent IL. Redman CWG. McMichael AJ. Evidence for a novel HLA antigen found on human extravillous trophoblast and a choriocarcinoma cell line. Immunology 1986;59:595-60 I. Loke YW. Burland K. Human trophoblast cells cultured in modified medium and supported by extracellular matrix. Placenta 1988;9: 173-82. Robertson WG. Pathology of the pregnant uterus. In: Fox H , ed. Obstetrical and gynecological pathology. London : Churchill Livingstone, 1987 : 1149-76. Hrushesky WJM, Gruber SA, Sothern RB. Hoffman RA. et al. Natural killer cell activity: age, estrous- . and circadian-stage dependence a nd inverse correlation with metastatic potential. JNCI 1988;80:1232-7. Sulke AN, Jones DB. Wood PJ. Hormonal modulation of human natural killer cell activity in vitro. J Reprod Fertil 1985;7: 105-10. Wegmann TG. Placental immunotropism: maternal T cells enhance placental growth and function. Am J Reprod Immunol Microbiol 1987;15:67-70. Bulmer IN, Morrison L, Johnson PM. Expression of the proliferation markers Ki67 and transferrin receptor by human trophoblast populations. J Reprod Immunol 1988; 14:291-302. Clark DA, Falbo M, Rowley RB , Banwatt B. StedronskaClark J. Active suppression of host versus graft reaction in pregnant mice. IX. Soluble suppressor activity obtained from allopregnant mouse decidua that blocks the cytolytic effector response to IL-2 is related to transforming growth factor-~. J Immunol 1988;141:3833-40.
