Fas-fas ligand system-induced apoptosis in human placenta and gestational trophoblastic disease

PROBLEM

The low frequency of maternal immune responses to paternally inherited fetal antigens raises the following question: What regulates the immunobiology of pregnancy? Data suggest that this state is the result of peripheral immune-tolerance, an active process of immune-regulation in which activated T cells undergo apoptosis. We studied Fas ligand (FasL) expression and apoptosis in normal and pathologic placentas to find out whether the Fas-FasL-induced apoptosis takes place during implantation.

METHOD OF STUDY

FasL expression in paraffin sections was detected using specific antibodies and confirmed with reverse transcriptase-polymerase chain reaction of total RNA from frozen placentas. Apoptosis was detected using the terminal deoxy (d)-UTP nick end-labeling assay.

RESULTS

FasL was found in the normal placenta and in gestational trophoblastic disease. Apoptotic leukocytes were localized to the maternal-fetal interface corresponding in localization with the distribution of FasL.

CONCLUSIONS

We propose that FasL expression in the placenta is a mechanism responsible for the development of maternal immune tolerance specific for paternal alloantigens and operates in pathologic states characterized by trophoblastic invasion/proliferation.

Human lymphocyte antigen expression in hydatidiform mole: androgenesis following fertilization by a haploid sperm

Thirteen hydatidiform moles (complete moles) and lymphocytes from each parent were analyzed for human lymphocyte antigen (HLA-A and HLA-B specificities). It was demonstrated that molar tissues expressed homozygous A and B specificities which were identical to those of the father and not those of the mother. It was concluded that androgenesis was responsible for the pathogenesis of most cases of complete mole. There was homozygous expression of paternal HLA specificities which were heterozygous for A locus and/or B locus in eight of nine cases of complete mole. This suggests that these hydatidiform moles developed from an egg which was fertilized by a haploid sperm which duplicated its own chromosomes after meiosis.

Fc receptors in human placenta

Cryostat sections of placental tissue strongly adsorbed erythrocytes sensitized with IgG antibodies of human, rabbit, and guinea pig origin. No adsorption occurred using erythrocytes sensitized with F(ab')2 fragments. The reaction was strongly inhibited by intact IgG and by Fc fragments, weakly inhibited by pFc' fragments, and not inhibited by Facb and F(ab')2 or albumin. These properties are similar to those of corresponding receptors in normal lymphoid tissues. Results obtained with sections of hydatidiform mole showed that the reaction occurred with the trophoblastic tissue. Porcine placenta had no Fc receptor activity. The presence of an Fc receptor in human placental tissue may therefore be of significance for the selective transfer of IgG from mother to foetus.

Assessment of cell proliferation in hydatidiform mole using monoclonal antibody MIB1 to Ki-67 antigen

AIMS

To assess the role of Ki67 immunoreactivity in predicting the clinical progress of hydatidiform mole.

METHODS

Tissue from 87 hydatidiform moles, 11 normal first trimester placentas, 11 normal term placentas and 17 spontaneous abortions were examined for expression of Ki67 antigen, using the monoclonal antibody MIB1.

RESULTS

Ki67 immunoreactivity was significantly higher in the tissue from normal first trimester placentas than in that from normal term placentas and spontaneous abortions. Among the 87 patients with hydatidiform moles studied, 20 developed persistent gestational trophoblastic disease and required subsequent treatment. There was no statistically significant difference in the Ki67 index between the 20 patients who developed persistent disease and those who did not.

CONCLUSION

Hydatidiform moles which give rise to persistent trophoblastic disease do not have a higher proliferative rate than those which do not. The Ki67 index is not useful for predicting the prognosis of molar pregnancies.

Hyaluronan, CD44 and its variant exons in human trophoblast invasion and placental angiogenesis

Both hyaluronan and one of its receptors, CD44, can be demonstrated in the early human conceptus and in placental stroma. The variants of CD44 resulting from variable exon splicing are found in metastasizing human malignancies and are also involved in hyaluronan uptake and degradation. The resulting hyaluronan fragments are known to be highly angiogenic. We postulated that the self-limited process of trophoblast invasion of the uterine decidua results in part from the strategy of alternative splicing of CD44, similar to that used by invasive cancer cells in the course of metastatic spread and possibly angiogenesis. Monoclonal antibodies specific for CD44s and for an exon expressed during metastatic tumour progression, CD44v7, were used to examine this hypothesis. In this study we found human trophoblasts, for the first time, to express CD44. Intermediate trophoblasts of first and second trimester exhibited the standard form of CD44 while extravillous trophoblasts, which are responsible for the invading characteristics of the placenta, were positive for the alternatively spliced form, the CD44v7-8. Moreover, in the case of placenta accreta there was a prominent membrane staining of the trophoblasts that were embedded in the fibrin layer over the myometrium. The highly metastatic choriocarcinoma cells also expressed CD44v7-8. We propose, therefore, that the invading trophoblasts utilize the alternatively splicing machinery. These cells retain their invasive capabilities through the permissive ECM by carrying the CD44v7-8 isoform, which binds weakly to hyaluronan and thus prevents it from being degraded by intracellular hyaluronidase.

Differential expression of blood-group-related carbohydrate antigens by trophoblast subpopulations

A panel of monoclonal antibodies directed at fucosylated and sialylated carbohydrates on the Type 1 and Type 2 blood group precursor chains was used in an immunohistological study of trophoblast subpopulations. Formalin-fixed paraffin-embedded sections of normal trophoblast throughout pregnancy, hydatidiform moles and gestational choriocarcinoma were examined. Sialyl-Lex was localized in extravillous trophoblast populations in normal and molar pregnancies and choriocarcinoma tumour cells. After neuraminidase treatment, Lex was identified in similar cell populations. Although Lea was identified in syncytial trophoblast cells at the maternofetal interface, no staining was seen for sialyl-Lea. Villous trophoblast did not stain with any of the monoclonal antibodies used. Our results underline the divergent differentiation of trophoblast into villous and invasive extravillous subpopulations. The analogy with tumour cells is discussed.

Immunology of spontaneous abortion and hydatidiform mole

Our study of the sequence of morphological events leading to the formation of hydatidiform mole (HM) led us to this study. Our object in this paper is to present evidence supporting our 1971 hypothesis with results obtained by various immunological methods in our laboratory, and to propose the importance of an immunoregulatory mechanism by blocking antibody (BA). Our hypothesis is that the conceptus, after the early demise of the embryo, may be dropped out in most cases as spontaneous abortion (SpA), and the trophoblast in the remaining cases continues to grow, developing into HM by an "immune selection" mechanism. Our conclusion is that the trophoblast will be damaged by immune cytotoxic mechanism through a decrease of BA in SpA, and continue to grow in HM through continuous production of BA after the demise of the embryo.

The significance of proliferating cell nuclear antigen in human trophoblastic disease: an immunohistochemical study

The expression of proliferating cell nuclear antigen (PCNA) in human trophoblastic disease was assessed immunohistochemically in tissue from 29 spontaneous abortions, 33 partial moles, 40 complete moles and 23 choriocarcinomas using the monoclonal antibody PC10. PCNA immunoreactivity occurred predominantly in the cytotrophoblasts in each of the four types of tissues. Quantitative analysis showed that the choriocarcinoma group gave a statistically significant higher PCNA index than the other three. There was no significant difference between the groups of spontaneous abortion, partial or complete mole. Sixteen of the 73 patients with partial and complete moles developed persistent gestational trophoblastic disease and there was no significant difference between the patients requiring chemotherapy and those who did not. We conclude that choriocarcinoma has a significantly higher PCNA proliferative index whilst hydatidiform moles cannot be distinguished from abortions by such analysis. The PCNA index does not appear to be useful in predicting the progression of molar pregnancies to persistent trophoblastic diseases.

Distribution of Immunocompetent Cells in Decidua of Controlled and Uncontrolled (Choriocarcinoma/Hydatidiform mole) Trophoblast Invasion

PROBLEM

Pregnancy has been considered as a model of successfully controlled tissue invasion where trophoblast cells infiltrate the maternal decidua without being rejected or without destroying the tissue. In choriocarcinoma (CC) and hydatidiform mole (HM), a dysregulation of invasive (malignant/benign) trophoblast cells is present. Immunocompetent cells (IC) are known to be involved in rejection pathways of malignant cells and can also be identified in early pregnancy decidua. The aim of the present study was to identify the phenotype of IC in decidua of women with normal pregnancy (NP), CC and HM.

METHODS

Immunocompetent cells were detected by immunohistochemistry in decidual tissue from first trimester NP (n = 10), CC (n = 12) and HM (n = 11) using antibodies against CD8+, CD3+, CD56+, CD68+ cell surface markers and mast cell tryptase (MCT). A scaled eye piece was used for cell counting to obtain semiquantitative results. Statistical analysis was performed using Wilcoxon rank/Mann-Whitney tests.

RESULTS

We observed a significantly increased number of lymphocytes positive for CD8, CD3 and MCT positive granulocytes in CC and HM compared with the samples from NP (all P < or = 0.001). Lymphocytes positive for natural killer (NK) cell marker CD56 were significantly decreased in CC and HM versus NP (P < or = 0.001). The number of CD68 positive cells (macrophages) were not significantly different among the tissue pools.

CONCLUSION

The increase of CD8/CD3 T cells and mast cells in CC and HM and the decrease of CD56 cells, compared with NP, suggests the necessity of a balance between T and NK cells in controlling trophoblast invasion.

Expressions of proliferation markers (Ki-67, proliferating cell nuclear antigen, and silver-staining nucleolar organizer regions) and of p53 tumor protein in gestational trophoblastic disease

OBJECTIVE

This study was undertaken to determine whether the expressions of 3 proliferation markers (Ki-67, proliferating cell nuclear antigen, and silver-staining nucleolar organizer regions) and of p53 tumor protein could differentiate spontaneous abortions from gestational trophoblastic diseases and also discriminate among gestational trophoblastic disease subgroups.

STUDY DESIGN

Twenty-two partial hydatidiform moles, 17 complete hydatidiform moles, 6 invasive hydatidiform moles, and 20 nonhydropic spontaneous abortions (control group) were evaluated by means of immunohistochemical techniques with antibodies to Ki-67, proliferating cell nuclear antigen, and p53. One-step silver staining was used to detect silver-staining nucleolar organizer regions.

RESULTS

The expressions of Ki-67, proliferating cell nuclear antigen, silver-staining nucleolar organizer regions, and p53 were significantly higher in the gestational trophoblastic disease group than in the control group. The results of linear discriminant analysis showed that silver-staining nucleolar organizer region count had the highest sensitivity and specificity (93.3% and 100%, respectively) for distinguishing gestational trophoblastic disease from spontaneous abortion. Sensitivity and specificity for discriminating gestational trophoblastic disease from spontaneous abortion increased to 100% when all four markers were used together. Proliferating cell nuclear antigen was found to be the best discriminating variable for differentiating among gestational trophoblastic disease subgroups.

CONCLUSION

Our findings suggest that expressions of Ki-67, proliferating cell nuclear antigen, silver-staining nucleolar organizer regions, and p53 may aid in the diagnosis of gestational trophoblastic diseases. These fairly rapid, simple, and economic techniques could serve as a useful adjunct to conventional methods in the diagnosis of gestational trophoblastic diseases.

A study of human leukocyte antigen G expression in hydatidiform moles

OBJECTIVE

Human leukocyte antigen G (HLA-G) is a nonclassic major histocompatibility gene normally expressed only in extravillous trophoblasts throughout pregnancy. It may be responsible in part for the successful evasion of the hemiallogenic trophoblasts from maternal immune surveillance. We investigated whether HLA-G is expressed in molar pregnancies.

STUDY DESIGN

We examined 5 complete hydatidiform mole specimens and 5 partial hydatidiform mole specimens to determine whether HLA-G is expressed by immunohistochemistry and by RNA in situ hybridization analysis.

RESULTS

We found that both the protein and RNA of HLA-G is expressed in complete and partial hydatidiform moles.

CONCLUSION

HLA-G RNA and protein are expressed in molar pregnancies. HLA-G expression is independent of embryonic development and may therefore be an integral part of placental development. Furthermore, expression of HLA-G in the complete hydatidiform mole, a naturally occurring androgenote, confirms expression of the paternal allele of HLA-G. Imprinting of HLA-G is therefore unlikely to play a role in protecting fetal trophoblasts from maternal immune rejection.

Characterization and localization of HLA antigens on hydatidiform mole

Frozen sections of three specimens of hydatidiform mole were stained with monoclonal antibodies to HLA Class I and Class II antigens by means of an indirect immunoperoxidase technique. Class I (HLA A, B, C) antigens were detected on proliferating extravillous trophoblast and on villous stromal cells but not on quiescent villous trophoblast. Trophoblast Class I antigen was detected with four different antibodies to monomorphic determinants but not with antibodies to the appropriate polymorphic HLA A or B type. Stromal cells were reactive with all Class I antibodies. Class II (HLA DR) antigens were not detected on any molar tissue. The expression of HLA antigens by molar tissue is similar to that of the normal first-trimester human placenta.

HLA-G expression in trophoblast cells is independent of embryonic development

HLA-G is a nonclassical major histocompatibility complex class I molecule that is selectively expressed on cytotrophoblasts at the feto-maternal interface where it may play a major role in maternal-fetal tolerance. In this study, we compared HLA-G expression in trophoblasts from normal and pathologic pregnancies by immunohistochemical analysis. First, we found a defective HLA-G expression in miscarriages associated with hypotrophic but normal eggs. Conversely, by studying molar pregnancies, we observed a high HLA-G expression in complete and partial hydatidiform moles. Finally, HLA-G expression could be visualized in extravillous trophoblasts that develop outside of their normal environment, as reported here in ectopic pregnancies. Taken together, these results suggest that HLA-G expression in extravillous trophoblasts is induced in an autonomous manner, independently of embryonic development, and may be an integral part of placental development allowing its tolerance from maternal immune system.

Use of proliferation cell nuclear antigen immunoreactivity for distinguishing hydropic abortions from partial hydatidiform moles

AIMS

To determine whether the expression of proliferating cell nuclear antigen (PCNA) in villous cytotrophoblast could distinguish between placental tissue from a hydropic abortion and that from a partial hydatidiform mole.

METHODS

Tissue from 18 partial hydatidiform moles, 15 hydropic abortions, five normal first trimester placentas and five normal full term placentas were immunostained for expression of PCNA, using the monoclonal antibody PC10.

RESULTS

PCNA immunoreactivity was very much higher in the cytotrophoblast of normal first trimester placentas than in normal term placentas. Villous tissue from partial hydatidiform moles showed, on average, less immunoreactivity for PCNA than did villous tissue from hydropic abortions.

CONCLUSIONS

Immunostaining for PCNA is of no value for differentiating between partial hydatidiform moles and hydropic abortions. The findings indicate that trophoblastic proliferation or hyperplasia is not a feature of partial hydatidiform moles.

The HL-A and ABO antigens in trophoblastic disease

No statistically significant deviations in phenotype frequencies of the 25 HL-A antigens or the ABO antigens were seen when 111 Caucasian patients with trophoblastic disease were compared with 1,259 healthy Caucasian controls. However, an increasein the incidence of HL-A11 was found in 39 patients who currently had the disease, but not in 72 who had recovered from the disease. Further, an increase in the frequency of W18 was observed among 18 patients who currently had 'invasive' disease (choriocarcinoma or invasive mole), but not in 44 who had recovered from such disease. If valid for larger patient population, these increases may suggest association of HL-A11 and W18 with the 'morbidity' of the trophoblastic disease. No increase in histocompatibility was seen in 45 patient-couples over 67 control-couples in terms of decrease in the number of male spouse's HL-A incompatibilities, and no significant difference was seen in the distribution of pregnancies in the two groups. No significant difference was observed in the incidence of different male-female combinations of ABO blood groups between 95 patient-couples and an equal number of control-couples. Lymphocytotoxic antibodies were found in 64 patients (158 sera) or 36% of the 178 patients (413 sera) examined. HL-A specific antibodies were found in 30 or 17% patients (39 sera). Of these 30, 24 patients had molar pregnancies and six had choriocarcinoma. Whether these antibodies have a role in the destruction of neoplastic tissue remains to be determined.

Implantation site in complete molar pregnancy: a study of immunologically competent cells with monoclonal antibodies

The nature and intensity of inflammatory cellular infiltrate in the implantation sites of 10 complete molar pregnancies were evaluated by immunohistochemical staining of frozen tissue sections with monoclonal antibodies. As compared to the implantation site in normal pregnancy, there was an increase in the number of infiltrating inflammatory cells in the molar implantation site. Most of the inflammatory cells were T cells with predominance of T4+ (Leu-3a+) cells over T8+ cells. Inflammatory cells were not detected in the molar vesicles. The possible implications of these findings in molar pregnancy are discussed.

Decidual T lymphocyte activation in hydatidiform mole

AIM

To quantify and compare decidual leucocyte subpopulations in complete and partial hydatidiform molar pregnancy with those in normal early pregnancy.

METHODS

Decidual leucocytes were characterised using an avidin-biotin technique and a panel of monoclonal antibodies in formalin fixed, paraffin embedded decidual tissues from 10 normal first trimester pregnancy terminations and from 13 partial and 13 complete hydatidiform moles. Immunostained cells were fully quantitated and differences between subject groups were analysed using the Mann-Whitney test. T lymphocyte populations were further characterised using double immunohistochemical labelling.

RESULTS

The numbers and percentages of CD3+ and CD4+ T cells were significantly increased in complete hydatidiform moles compared with partial moles and normal early pregnancy decidua. No differences were found in the number or percentage of CD8+ T cells. The CD8+ to CD4+ T cell ratio decreased significantly in complete mole compared with partial mole and normal decidua. The numbers and percentages of CD45RO+ cells increased significantly in both partial and complete hydatidiform mole compared with normal first trimester decidua. Double labelling confirmed that 50% of CD3+ T cells in complete and partial molar pregnancy coexpressed CD45RO, compared with 30% in normal pregnancy. Other leucocyte populations (eGLs, macrophages, B cells, and classical natural killer cells) did not differ between complete and partial mole and normal pregnancy decidua.

CONCLUSIONS

The presence of an increased population of activated decidual CD45RO+ T cells in complete and partial molar pregnancy suggests altered maternal immune responses against molar trophoblast.

Primary chronic abortion, preeclampsia, idiopathic intrauterine growth retardation, hydatidiform mole, and choriocarcinoma: a unifying concept

Different degrees of maternal hyporesponse, as far as blocking activity is concerned, are proposed for primary chronic abortion, preeclampsia, and idiopathic intrauterine growth retardation. On the other hand, a maternal hyperresponse to fetal antigens with a higher production of blocking antibodies may be related to an unusual proliferation of the trophoblast in cases of hydatidiform mole and choriocarcinoma.

Serum cytokines in gestational trophoblastic diseases

Interleukin 1 beta (IL-1 beta), interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-alpha) were assayed by 125I immunoradiometric assay in sera of 42 cases of vesicular mole (VM), 24 cases of choriocarcinoma and 23 normal pregnant women at their first trimester (controls). According to pathologic diagnosis and serial serum hCG beta assays, the cases with VM and choriocarcinoma were subdivided into remission and progressive tumor groups. The progressive tumor groups--both VM and choriocarcinoma--showed marked elevations of serum IL-1 beta, IL-6 and TNF alpha. For choriocarcinoma in remission this elevation was considerably less pronounced. The VM cases in remission had only a slight increase of the mean serum IL-6 value and none of the cases had elevated IL-1 beta or TNF values. These results may indicate that serum IL-1 beta and TNF-alpha assays are valuable biomarkers in the differential diagnosis of gestational trophoblastic disease (GTD). Moreover, normal values of these cytokines may rule out high-risk GTD, whereas markedly elevated values may indicate poor prognosis.