- Email: [email protected]
Ectopic pregnancy: histopathology and assessment of cell proliferation with and without methotrexate treatment*†
Vol. 65, No.4, April 1996
FERTILITY AND STERILITY Copyright
i)
Printed on acid-free paper in U. S. A.
1996 American Society for Reproductive Medicine
Ectopic pregnancy: histopathology and assessment of cell proliferation with and without methotrexate treatment*t
Charlotte Floridon, M.D.:j:§ Ole Nielsen, H.T.§ Carsten Byrjalsen, M.D.II Berit H!Illund, M.D.§
Gitte Kerndrup, M.D.§ Sten Grove Thomsen, M.D. II Johan Adolph Andersen, M.D.§
Herning Central Hospital and Odense University Hospital, Odense, Denmark
Objective: To evaluate tubal morphology, trophoblast proliferation, and inflammatory reaction in response to methotrexate (MTX) treatment of ectopic pregnancy (EP). Design: Nonrandomized controlled study. Setting: Academic hospital. Patients: Archival specimens from 10 EP unsuccessfully treated with MTX and 10 cases primarily treated by surgery. Interventions: Ki67/hCG and Ki67/human placental lactogen double immunohistochemical methods were used to examine trophoblastic spread, placentation, hormone production, decidualization, vascular invasion, hemorrhage, rupture, and proliferative index ofthe cytotrophoblast. Band T-lymphocyte responses were evaluated by CD3 and CD20. Results: Trophoblastic spread and placentation were confined to the tubal mucosa after MTX treatment, whereas invasion of the muscularis and subserosa was common in the controls. The proliferative index was reduced (19% versus 93%), although a high proliferative index was found in two of three cases complicated by rupture. Polar proliferation of Ki67-positive cytotrophoblast toward the implantation site was abolished in MTX-treated cases. Decidual reaction was not observed. No correlation was observed between the above-mentioned findings and gestational age, level of j3-hCG, dose of MTX, or interval to surgery. Conclusion: Trophoblastic spread, differentiation, and invasion were compromised by MTX treatment. Methotrexate seems to decrease cytotrophoblast proliferation. Whether a missing decrease in proliferation index reflects treatment failure awaits a larger population-based study. Fertil Steril 1996;65:730-8 Key Words: Ectopic pregnancy, histopathology, methotrexate
During the past two decades, there has been a vast increase in the incidence of ectopic pregnancies (EP) (1). In former days, surgical intervention was the treatment of choice, but today an increasing number
Received June 16, 1995; revised and accepted October 12, 1995. * Supported by WyethlLederle, Copenhagen, Denmark. t Presented in part at the Fifth Conference of Endocrinology and Metabolism in Human Reproduction, Academic Unit of Obstetrics and Gynaecology, London, United Kingdom, March 27 to 29,1995. :j: Reprint requests: Charlotte Floridon, M.D., Odense University Hospital, KhilVervrenget 20 A, DK-5000 Odense C, Denmark (FAX: 45-66-12-27-83). § Department of Pathology, Odense University Hospital. II Department of Obstetrics and Gynecology, Herning Central Hospital and Odense University Hospital. 730
Floridon et at.
of reports indicate that EP can be successfully and safely treated with methotrexate (MTX). As many as 45% of a population of women with EP has been treated with MTX with success rates of 82% to 95% and preservation of patency in 79% to 91% of the affected tubes (2). Experience concerning later reproductive performance still is limited to a few studies (3, 4), showing that 47% to 69% may obtain an intrauterine pregnancy. Methotrexate is a folic acid antagonist acting through inhibition of dihydrofolic acid reductase, blocking purine synthesis and thymidine binding and thus interfering with DNA synthesis. Consequently, proliferation of the cytotrophoblast and, to a lesser degree, the intermediate trophoblast is inhibited, because these trophoblastic cells are the
Trophoblast proliferation after methotrexate
Fertility and Sterility
Table 1 Clinical and Histopathological Results From MTX-Treated Patients EP case no.
GA*
MTXdose
d
Interval until surgery
Rupture
Intraluminal trophoblastic spread
Intramural trophoblastic spread
Subserosal or serosal trophoblastic spread
Placentation Intraluminal
Intramural
Serosal
d
Proliferation index %
1
37
75mg transvaginal injection
4
2
37
25mgx51M
12
3
45
50mg transvaginal injection
8
4
48
25mgx51M
5
5
50
25mgx51M
12
6
54
25 mg x 5 1M
12
7
55
54mgx31M
14
8
60
25 mg x 51M
19
9
68
50mgx2 transvaginal injection
7
+
10
74
50mgx2 transvaginal injection
32
+
Median Range
52 37 to 74
+
Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
+
35
+
17
Intermediate trophoblast
+
19
Intermediate trophoblast
+
60
Intermediate trophoblast
+
3
Intermediate trophoblast
+
28
Intermediate trophoblast
+
2
+
18
+
9
+
68
Intermediate trophoblast
Intermediate trophoblast
19 2 to 68
* GA, gestational age. only constituents capable of mitotic activity. The effect of MTX on trophoblastic tissue has been studied by Schafer et al. (5), who, based on in vitro tissue cultures, found that trophoblasts from EP compared with those from intrauterine pregnancies required a concentration of MTX 10 times higher to induce an equivalent reduction ofhCG levels. Normally, >8 days were required for hCG to decline after application of MTX, and some cases showed even no reduction in hCG values at all. From this study it may be concluded that, if low and nonresponding patients exist in vivo, treatment with MTX could fail in some patients. Further, Sand et al. (6) demonstrated that MTX inhibited growth of trophoblast cell cultures in vitro at concentrations 1,000 times higher than those required to inhibit choriocarcinoma cell proliferation. The histopathology of EP after MTX treatment Vol. 65, No.4, April 1996
has been presented in only two studies. Kooi et al. (7) studied five cases and concluded that MTX applied locally did not seem to compromise tubal response to the insult of invading chorionic tissue or to interfere with the ultimate tubal repair. On the other hand, Klinkert et al. (8) reported a case of persisting EP that seriously hindered tubal repair 6 months after MTX treatment. The literature concerning the general morphology ofEP after surgical intervention raises controversial questions. Some authors (9-11) report a predominantly intraluminal trophoblastic spread, i.e., the major feature of tubal pregnancy is that chorionic villi and villus-associated cytotrophoplast remain superficially implanted without invasion of the muscular wall and that rupture is due to vascular thrombosis and intraluminal hemorrhage with dilatation of the tube. Others (12-15) describe a predomi-
Floridon et al.
Trophoblast proliferation after methotrexate
731
Figure 1 Polar proliferation of Ki67-positive cytotrophoblast is abolished after MTX treatment (Ki67/human placental lactogen X120).
nantly extraluminal trophoblastic spread, i.e., the main feature of the growing gestation is a rapid penetration of the tubal wall with development of the placenta in the muscularis, subserosa, and serosa, leading to disruption of all layers in combination with interstitial hemorrhage. Methotrexate treatment ofEP still carries the risk of later surgery in approximately 18% of cases, despite careful selection of patients (2). Increased risk of treatment failure has often been associated with high ,8-hCG levels and a living embryo. These low or nonresponders might demonstrate a high proliferative index of the cytotrophoblast as demonstrated by the Ki67-antigen expressed in nuclei in active cell cycle (excluding G0 ). To evaluate the effect of MTX, we studied tubal morphology and synthesis ofhCG and human placental lactogen by trophoblasts from 10 MTX-treated EPs and compared the results to a control group treated primarily by surgery. We further examined the proliferative index by the application of the monoclonal antibody MIB1, detecting the Ki67-antigen and the composition of the inflammatory response. MATERIALS AND METHODS
Histopathology of 10 cases ofEP treated unsuccessfully with MTX were reviewed and compared with a control group of 10 EP primarily treated by surgery. Formalin-fixed paraffin-embedded archival material, each comprising 5 to 10 tissue blocks with the implan732
Floridon et al.
tation site represented, were stained by hematoxylin and eosin, CD 3 (pan T-lymphocyte), CD20 (pan Blymphocyte), and double-stained with monoclonal MIB 1 and hCG or human placental lactogen. For immunostaining, 5-J.Lm thick sections were cut and mounted on SuperFrost/Plus glass slides (Menzel-Glaser, Braunschweig, Germany) and dried for 1 hour in an oven at 60°C. Sections were dewaxed in xylene and rinsed in alcohol. Endogenous peroxidase was blocked using 0.5% hydrogen peroxide in methanol. Mter rinsing in water, antigen unmasking was performed. Microwave heating for 20 minutes in a 10 mM citrate buffer (pH 6) as described by Cattoretti et al. (16) was used for the Ki67 antigen, human placental lactogen, hCG, and CD20, whereas a 15minute pretreatment of 0.05% protease (Type 14; Sigma, St. Louis, U.S.A) in Tris-buffer (pH 7.4) at 37°C was applied for CD3. Mter antigen unmasking, sections were pretreated with 2% bovine serum albumin (BSA). Both standard and double immunohistochemical staining were performed. Standard immunostaining was used to identify CD3 (Dako A452, Glostrup, Denmark) and CD20 (clone L26, Dako M755). Both antibodies was diluted 1:200 and applied for 30 minutes followed by a peroxidase-labeled streptavidin biotin method (Dako, Glostrup, Denmark) using 3-amino-9-ethylcarbazol as chromogen. Double immunohistochemical staining for Ki67 /human placental lactogen and Ki67/hCG was performed using a modified sequential method. The Ki67 antigen was identified using the monoclonal antibody MIB1 (Immunotech, Marseille, France) followed by the peroxidase-labelled streptavidin biotin method (Dako) using DAB-Ni as chromogen. Antihuman placental lactogen (Dako A137) or anti-hCG (Dako A231) were used to identify human placental lactogen and hCG in a indirect immunoalkaline phosphatase method using Fast Red TR as chromogen (Dako). In brief, sections were incubated with MIB1 diluted 1:100. Mter rinse in Tris-buffered saline (TBS), biotinylated goat anti-mouse immunoglobulin (Ig) antibody (Dako E433) was applied at dilution 1:200 followed by TBS rinse and incubation in peroxidase-conjugated streptavidin (Dako P397) at dilution 1:300. All incubations were done at room temperature for 30 minutes. Slides were developed for 10 minutes using a 3.3' -diaminobenzidine tetrahydrochloride (0.05%), hydrogenperoxide (0.01 %), nickel sulfate (2%) solution in a 0.1 M acetate buffer at pH 6 (17). After color development, sections were rinsed in water and pretreated with 2% BSA. Sections then were incubated for 30 minutes with antihuman placental lactogen or anti-hCG at a dilution 1:500. Mter rinse in TBS, 30 minutes incubation with alkaline phosphatase-conjugated swine antirabbit Ig (Dako D306) were done at dilution 1:50 and
Trophoblast proliferation after methotrexate
Fertility and Sterility
Table 2
Clinical and Histopathological Results From Surgery-Treated Patients Placentation
EPno.
GA*
Rupture
Intraluminal trophoblastic spread
Intramural trophoblastic spread
Subserosal or serosal trophoblastic spread
Intraluminal
Intramural
Serosal
d
Proliferation index %
11
33
+
12
33
+
13
36
+
14
44
+
15
44
16
45
17
50
18
53
19
54
20
62
Median Range
45 33 to 62
+
+
Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
+
9
Intermediate trophoblast, syncytiotrophoblast Intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
+
+
+
+
92
+
+
Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
95
93
Intermediate trophoblast
+
Intermediate trophoblast
+
92
+
90
Intermediate trophoblast
+
89
+
+
94
+
+
95
93 9 to 97
sections were developed for 20 minutes in Fast Red TR. Finally, sections were counterstained with Mayers Hematoxylin for 5 seconds and mounted with Aquamount (BDH, Poole, United Kingdom). Negative controls were performed, including omitting the primary antibodies from the last sequence (anti-human placentallactogen/anti-hCG), and no unspecific staining or cross-reactivity was observed. All stainings were carried out during one procedure to minimize differences in staining intensity due to day to day variation or variation due to applying different batches of primary and/or secondary antibodies, buffer composition, or other known sources of errors. The proliferative index, expressed in percent, was estimated by counting the number of MIBl-positive nuclei versus the total number of nuclei in a minimum of 300 cytotrophoblast cells at the implantation site, using an objective lens x40. The nuclei were counted as positive or negative without regard to Vol. 65, No.4, April 1996
97
staining intensity. Intermediate trophoblast, because of a low proliferation rate, syncytiotrophoblast, decidua, maternal, and fetal tissue were all excluded from the analysis of the proliferative index. Statistical analysis of the proliferative index was performed using the Mann-Whitney U-test. The morphology of the implantation site and the mode of trophoblastic spread (trophoblastic spread), placentation, decidualization, vascular invasion, and hemorrhage as well as the inflammatory response were evaluated by conventional histopathological methods.
RESULTS Methotrexate treatment
In the MTX group (Table 1), 10 patients were treated with MTX ranging from 50 to 162 mg either 1M or by ultrasound-guided transvaginal injection
Floridon et al.
Trophoblast proliferation after methotrexate
733
Proliferative index at the implantation site varied from 2% to 68% (median 19%), and nuclear reaction in the cytotrophoblast was strongly positive if proliferation was present. Protein hormones (hCG and human placental lactogen) could not be demonstrated. If positive, the intermediate trophoblast demonstrated a low focal to diffuse degree of immunoreactivity with MIB1 and a diffuse reactivity for hCG and human placental lactogen. The syncytiotrophoblast were totally MIB 1 unreactive, but stained strongly positive for hCG and human placental lactogen. None of the placental chorionic villi showed polar proliferation (Fig. 1) characterized by clusters of MIB1-positive cytotrophoblast emanating from only one pole of the placental villi towards the implantation site, as in normal intrauterine pregnancies. Invasion of only small luminal tubal vessels was found by intermediate trophoblast, none of these showing invasion by cytotrophoblast. The intermeFigure 2 Cytotrophoblastic cell columns showing normal polar proliferation after surgical treatment (Ki67 /human placental lactogen X 120).
into the gestational sac. The gestational age (GA) varied from 37 to 74 days (median 52 days), and the interval between MTX treatment and surgery ranged from 4 to 32 days. In three cases (cases 4, 9, and 10), tubal rupture made surgical removal necessary. Four cases (cases 3, 6, 7, and 8) underwent surgery because of abdominal pain and/or local defence, so did two cases (cases 2 and 5) in combination with rising ,8-hCG levels. One patient (case 1) was salpingectomized because of noncompliance after only 4 days where increasing levels of ,8-hCG is found normally. The maximal ,8-hCG before treatment with MTX varied from 550 to 21,500 IU/L (International Reference Preparation [IRP]), during treatment from 850 to 27,000 IU/L, and presurgically from 260 to 27,000 IU/L. Seven patients had declining ,8-hCG during treatment. All10 tubes showed trophoblastic spread comprising all three trophoblast types (cytotrophoblast, intermediate trophoblast, syncytiotrophoblast) with a luminal superficial placentation at the implantation site. Intramural spread, only expressed by intermediate trophoblast, was found in seven cases and distributed both vertically and longitudinally. None of {pe intermediate trophoblast reached the subserosa or serosa. The morphology of the intermediate trophoblast cells varied according to its location from round to polygonal or ovoid to spindle shaped. Only one case (case 1) had a patchy mucosal decidual reaction. 734
Floridon et al.
Figure 3 Intramural placentation in a tubal pregnancy (-). Chorionic villi with polar proliferation towards the implantation site is seen in the lumen (---t) (H&E x 40).
Trophoblast proliferation after methotrexate
Fertility and Sterility
Figure 4 Intermediate trophoblast can be confused with decidual cells (H&E x 300).
diate trophoblast destructed the vessel wall, replaced the endothelial cells, as well as filled the lumen with subsequent formation of hyalinized vascular channels. The tubes were dilated by free blood filling the lumen of the tube. Larger intramural or subserosa! vessels at a distance from the implantation site were not invaded by trophoblast, and hemorrhage into the maternal soft tissue was not marked. The inflammatory response consisted of a diffuse
and/or perivascular infiltration of acute and chronic inflammatory cells, but only to a low degree. The number ofT-lymphocytes was decreased compared with the controls. As seen in Table 1, there was no correlation between GA, dose of MTX, level of ,8-hCG, interval to surgery, or rupture on one side and trophoblastic spread, placentation, or proliferative index on the other, although a relatively high proliferative index was found in two of three cases complicated by rupture.
Figure 5 Immunohistochemical analysis aids in distinguishing the intermediate trophoblast (human placental lactogen-positive) from decidual cells (human placental lactogen-negative) (KI67I human placental lactogen X300).
Vol. 65, No. 4, April 1996
Floridon et al.
Trophoblast proliferation after methotrexate
735
Surgical Treatment
The control group of 10 patients (Table 2) had their tubes removed either by laparotomy or laparoscopy. ,B-hCG was not measured in any of these cases. In six patients, tubal rupture was observed. The GA varied from 33 to 62 days (median 45 days). Nine fallopian tubes showed mucosal trophoblastic spread with all trophoblastic components (cytotrophoblast, intermediate trophoblast, syncytiotrophoblast) present and luminal superficial placentation. One tube (case 12) was without chorionic villi but a few cytotrophoblasts, intermediate trophoblasts, and syncytiotrophoblasts were observed superficially in the mucosa. Intramural spread was seen in seven cases, with all the different trophoblastic components (cytotrophoblast, intermediate trophoblast, syncytiotrophoblast) represented. Intramural placental chorionic villi were seen in five cases. In the subserosa, trophoblastic spread was present in five cases and placental chorionic villi in was present in one case. Decidual reaction was not observed. The proliferative index at the implantation site varied from 9% to 97% (median 93%). Ifpositive, the MIBl, hCG, and human placental lactogen staining intensity of all tree trophoblastic cell types was identical to the one seen after MTX treatment. The cytotrophoblastic cell columns of the villous stems showed polarization of growth toward the implantation site, visualized by clusters of MIBl-positive cytotrophoblasts (Fig. 2). Intravascular invasion of both small luminal and larger distant tubal vessels by intermediate trophoblasts were found in all cases. In two cases, syncytiotrophoblast was seen in the lumina of vessels. Hemorrhage was common into the lumen as well as into the muscular and subserosal parts of the tube as a result of trophoblastic invasion. The inflammatory response was more pronounced compared with the MTX group, and the number of T-lymphocytes was higher. Methotrexate Versus Surgery
A statistically significant difference in proliferative index between the two groups was found (19% versus 93%, P < 0.01) (Tables 1 and 2). Chorionic villous invasion of the muscular wall never was seen in the MTX group but was a common feature in the surgically treated cases (Fig. 3). In both groups, the inflammatory response consisted of a diffuse and/or perivascular infiltration of acute and chronic inflammatory cells, but to a lesser degree in the MTX group where the number of Tlymphocytes was decreased as well. There was no difference in the number of B-lymphocytes between 736
Floridon et al.
the two groups. It was not possible to detect any relation between tubal rupture and the microscopical location of trophoblastic spread, placentation, or hemorrhage. DISCUSSION
The last decade has witnessed an epidemic rise in the incidence of EP. Although success rates of 82% to 95% (2) are reported after MTX treatment, experience still is limited as far as later reproductive performance is concerned. Intrauterine pregnancies (3, 4) have been achieved in 47% to 69% of cases, which seems to compare favorably with the 44% to 52% reported after surgery (18). It is still largely unknown how MTX affects the invasive property of the trophoblast and the growth of placental chorionic villi and why some EPs fail to respond to MTX, why tubal ruptures occur, and whether the inflammatory response is influenced by MTX. An answer to these questions may help to optimize MTX therapy and increase the fertility rates. Former conclusions by Kooi et a1. (7) from 1992 concerning the morphology of tubes after treatment with MTX may be misleading because these authors seem unaware of the existence of the intermediate trophoblast studied in detail by Kurman et a1. (19). This cell type, occurring in all instances of pregnancy, is thought to represent an intermediate stage of differentiation of the trophoblast. The intermediate trophoblast is capable of mitotic activity and hormone synthesis (hCG and human placental lactogen), it is generally mononuclear and expresses many of the characteristics of both cytotrophoblast and syncytiotrophoblast. The intermediate trophoblast plays an important role in implantation-invasion and establishment of the maternal-fetal circulation (19). The results by Kooi et a1. (7) were based on routine stainings of only five cases of tubal EP. Stainings visualizing cells in proliferation and/or hormone-producing trophoblastic tissue were not carried out. Case 1 exhibited predominantly intraluminal trophoblastic spread and a marked decidual response adjacent to the tubal rupture, but the evidence for a tubal decidual reaction is not convincing (7). It may be difficult to distinguish between a decidual reaction and the intermediate trophoblast (Figs. 4 and 5). Decidual cells have small, uniform, round nuclei and homogenous cytoplasm surrounded by a well-defined cell membrane, whereas intermediate trophoblastic cells show larger, hyperchromatic nuclei and amphophilic cytoplasm with indistinct borders and stain positive for hCG human placental lactogen (19). This is in contrast to decidual cells. A predominantly extraluminal trophoblastic spread was seen in case 2. The remaining three cases were removed surgically > 1
Trophoblast proliferation after methotrexate
Fertility and Sterility
year after MTX treatment for reasons not related to the clinical course of the EP, and the ,8-hCG was negative before surgery. In these cases, the ultimate tubal repair was found to be normaL Seriously hindered tubal repair and persistent EP was casuistically reported by Klinkert et aL (8) and based on injection of 5 mg MTX only. The patient underwent surgery 6 months after MTX treatment. We agree with Fernandez (20) and Kock and Kooi (21) that such a treatment regimen does not seem to be superior to expectant management, former reports indicating failure rates of 31 % to 36% (22, 23). The EP was located in the isthmic part of the tube (8) and experience with this location still is limited to a few cases. Finally, a persistent trophoblastic disease cannot be discerned from a new EP because no follow-up of ,8-hCG levels was reported. Our treatment regimen might be insufficient, too, because five cases were treated with 0.5 mg/kg 1M MTX for 5 days. A recent study (24) seems to indicate that patients given such a dose were undertreated and required additional therapy to obtain a decrease in,8hCG levels. Outcome of treatment might depend on peak serum levels ofMTX, more than on the cumulative dose. The main achievements of our study were to demonstrate that trophoblastic spread and placentation always were confined to the tubal mucosa after MTX treatment; invasion of placental chorionic villi and polar proliferation toward the implantation site were observed only in the surgery group; a significant change in proliferative capacity of the cytotrophoblast in the MTX group, although a relatively high proliferative index was found in two of three cases complicated by rupture. Our findings in the surgery group support that EP is associated with a component of extraluminal trophoblastic spread (12-15). We observed both intraluminal, intramural, and subserosal trophoblastic spread as well as placental chorionic villi within the same locations. Invasiveness of the trophoblast after the formation of chorionic villi appears to rely on intermediate trophoblast only (19). The mechanisms responsible for trophoblastic invasion with subsequent intramural or subserosal placentation in the fallopian tube still are unclear. It has been proposed by Randall et aL (13) that these processes are entirely a function of trophoblastic tissue, because the blastocyst appears to be able to implant at any location, the maternal tissues playing a purely passive role. Feinberg et aL (25) have suggested that an interplay between urokinase-type plasminogen activator (PA) and two different plasminogen activator inhibitors (PAl 1 and PAl 2) is responsible for trophoblastic invasion in the intrauterine location. Their findings of PAl 1 in intermediate trophoblast Vol. 65, No.4, April 1996
invading the uterine lining including the arteries raise the possibility that PAl 1 could either function to promote trophoblastic invasion or be required for cessation of the invasive process. A study concerning this issue regarding EP is presently under consideration at our laboratory. The statistically significant difference in proliferative activity (19% versus 93%, P < 0.01) between the two groups in our study supports the in vitro demonstration by Schafer et aL (5) concerning the inhibitory effect of MTX on the trophoblast. They showed that the longer the interval between MTX treatment and hCG measurement, the more pronounced the effect. The decreased number ofT-lymphocytes might indicate that MTX is capable of inhibiting the inflammatory process in the tubal wall by reducing activation and/or recruitment of inflammatory cells, which could explain a better preservation of the tube, but it may be a result as well of the abolished trophoblastic spread and deep placentation. From the present study, it can be concluded that trophoblastic spread in MTX-treated cases seems to be limited to the mucosal part of the tube and that a high proliferative index of cytotrophoblast is associated with an increased risk of treatment failure. Our results might be influenced by selection bias because only archival material with sufficient tissue blocks were included and/or the small study groups. Future studies concerning morphology and the assessment of cell proliferation in EP would benefit from a prospective randomized design with careful preparation of the specimens to avoid distortion or other damage. . REFERENCES 1. Tuomivaara L. Ectopic pregnancy. An analysis of etiological
2. 3.
4. 5.
6.
7.
8.
risk factors and subsequent fertility [thesisl. Oulu, Finland: University of Oulu, 1990. Floridon e, Thomsen SG. Methotrexate treatment of ectopic pregnancy. Acta Obstet Gynecol Scand 1994;73:746-52. Fernandez H, Benifla J-L, Leladier e, Baton e, Frydman R. Methotrexate treatment of ectopic pregnancy: 100 cases treated by primary transvaginal injection under sonographic control. Fertil Steril 1993;59:773-7. Stovall TG, Ling FW. Single dose methotrexate: an expanded clinical trial. Am J Obstet Gynecol 1993; 168:1759-65. Schafer D, Pfuhl JP, Baumann R, Neubert S, Bender HG, Naujoks H. Trophoblast tissue culture of human intrauterine and ectopic pregnancies and treatment with methotrexate. Hum Reprod 1992;7:311-9. Sand PK, Stubblefield PA, Ory SJ. Methotrexate inhibition of normal trophoblasts in vitro. Am J Obstet Gynecol 1986; 155:324-9. Kooi S, van Etten FHPM, Kock HeLV. Histopathology offive tubes after treatment with methotrexate for a tubal pregnancy. Fertil SteriI1992;57:341-5. Klinkert J, van Geldorp HJ, ehadha-Ajwani S, Huikeshoven F JM. Tubal damage after intratubal methotrexate treatment. Fertil Steril 1993;59:926-7.
Floridon et al.
Trophoblast proliferation after methotrexate
737
9. Pauerstein CJ, Croxatto HB, Eddy CA, Ramzy I, Walters MD. Anatomy and pathology of tubal pregnancy. Obstet Gynecol 1986; 67 :30 1-8. 10. Senterman M, Jibodh R, Tulandi T. Histopathologic study of ampullary and isthmic tubal ectopic pregnancy. Am J Obstet Gynecol 1988; 159:939-41. 11. Stock RJ. Tubal pregnancy. Obstet Gynecol Clin North Am 1991; 18:73-94. 12. Budowick M, Johnson TRB, Jr, Genadry R, Parmley TH, Woodruff JD. The histopathology of the developing tubal ectopic pregnancy. Fertil Steril1980;34:169-71. 13. Randall S, Buckley CH, Fox H. Placentation in the fallopian tube. Int J Gynecol Pathol 1987;6:132-9. 14. Parmley TH. The histopathology of tubal pregnancy. Clin Obstet Gynecol 1987;30:119-28. 15. Dietl J, Buchholz F, Kindler P A. Histopathology oftubal pregnancy. Int J Gynecol Obstet 1988;27:385-8. 16. Cattoretti G, Pileri S, Parravicini C, Beckers MHG, Poggi S, Key G, et al. Antigen unmasking on formalin-fixed, paraffinembedded tissue sections. J Pathol 1993; 171:83-98. 17. Scopsi L, Larsson LI. Increased sensitivity in peroxidase immunocytochemistry. Histochemistry 1986;84:221-30. 18. LundorffP, Thorburn J, Lindblom B. Fertility outcome after
738
Floridon et al.
19.
20. 21. 22.
23.
24.
25.
Trophoblast proliferation after methotrexate
conservative surgical treatment of ectopic pregnancy evaluated in a randomized trial. Fertil Steril 1992;57:998-1002. Kurman RJ, Main CS, Chen HC. Intermediate trophoblast: a distinctive form of trophoblast with specific morphological, biochemical and functional features. Placenta 1984;5:34969. Fernandez H. Risks of intratubal methotrexate [letterl. Fertil Steril 1993;60:929. Kock I, Kooi S. Risks of intratubal methotrexate [letter]. Fertil Steril 1993; 60:930. Ylostalo P, Cacciatore B, Sjoberg J, Kiiiiriiiinen M, Tenhunen A, Stenman UH. Expectant management of ectopic pregnancy. Obstet Gynecol 1992;80:345-8. Fernandez H, Rainhorn JD, Papiernik E, Bellet D, Frydman R. Spontaneous resolution of ectopic pregnancy. Obstet Gynecol 1988:71:171-4. Fernandez H, Bourget P, Ville Y, Lelaider C, Frydman R. Treatment of unruptured tubal pregnancy with methotrexate: pharmacokinetic analysis of local versus intramuscular administration. Fertil Steril 1994;62:943-7. Feinberg RF, Kao LC, Haimowitz JE, Queenan JT, Wun TC, Strauss JF, et al. Plasminogen activator inhibitor types 1 and 2 in human trophoblasts. PAI-1 is an immunocytochemical marker of invading trophoblasts. Lab Invest 1989; 61:20-6.
Fertility and Sterility
FERTILITY AND STERILITY Copyright
i)
Printed on acid-free paper in U. S. A.
1996 American Society for Reproductive Medicine
Ectopic pregnancy: histopathology and assessment of cell proliferation with and without methotrexate treatment*t
Charlotte Floridon, M.D.:j:§ Ole Nielsen, H.T.§ Carsten Byrjalsen, M.D.II Berit H!Illund, M.D.§
Gitte Kerndrup, M.D.§ Sten Grove Thomsen, M.D. II Johan Adolph Andersen, M.D.§
Herning Central Hospital and Odense University Hospital, Odense, Denmark
Objective: To evaluate tubal morphology, trophoblast proliferation, and inflammatory reaction in response to methotrexate (MTX) treatment of ectopic pregnancy (EP). Design: Nonrandomized controlled study. Setting: Academic hospital. Patients: Archival specimens from 10 EP unsuccessfully treated with MTX and 10 cases primarily treated by surgery. Interventions: Ki67/hCG and Ki67/human placental lactogen double immunohistochemical methods were used to examine trophoblastic spread, placentation, hormone production, decidualization, vascular invasion, hemorrhage, rupture, and proliferative index ofthe cytotrophoblast. Band T-lymphocyte responses were evaluated by CD3 and CD20. Results: Trophoblastic spread and placentation were confined to the tubal mucosa after MTX treatment, whereas invasion of the muscularis and subserosa was common in the controls. The proliferative index was reduced (19% versus 93%), although a high proliferative index was found in two of three cases complicated by rupture. Polar proliferation of Ki67-positive cytotrophoblast toward the implantation site was abolished in MTX-treated cases. Decidual reaction was not observed. No correlation was observed between the above-mentioned findings and gestational age, level of j3-hCG, dose of MTX, or interval to surgery. Conclusion: Trophoblastic spread, differentiation, and invasion were compromised by MTX treatment. Methotrexate seems to decrease cytotrophoblast proliferation. Whether a missing decrease in proliferation index reflects treatment failure awaits a larger population-based study. Fertil Steril 1996;65:730-8 Key Words: Ectopic pregnancy, histopathology, methotrexate
During the past two decades, there has been a vast increase in the incidence of ectopic pregnancies (EP) (1). In former days, surgical intervention was the treatment of choice, but today an increasing number
Received June 16, 1995; revised and accepted October 12, 1995. * Supported by WyethlLederle, Copenhagen, Denmark. t Presented in part at the Fifth Conference of Endocrinology and Metabolism in Human Reproduction, Academic Unit of Obstetrics and Gynaecology, London, United Kingdom, March 27 to 29,1995. :j: Reprint requests: Charlotte Floridon, M.D., Odense University Hospital, KhilVervrenget 20 A, DK-5000 Odense C, Denmark (FAX: 45-66-12-27-83). § Department of Pathology, Odense University Hospital. II Department of Obstetrics and Gynecology, Herning Central Hospital and Odense University Hospital. 730
Floridon et at.
of reports indicate that EP can be successfully and safely treated with methotrexate (MTX). As many as 45% of a population of women with EP has been treated with MTX with success rates of 82% to 95% and preservation of patency in 79% to 91% of the affected tubes (2). Experience concerning later reproductive performance still is limited to a few studies (3, 4), showing that 47% to 69% may obtain an intrauterine pregnancy. Methotrexate is a folic acid antagonist acting through inhibition of dihydrofolic acid reductase, blocking purine synthesis and thymidine binding and thus interfering with DNA synthesis. Consequently, proliferation of the cytotrophoblast and, to a lesser degree, the intermediate trophoblast is inhibited, because these trophoblastic cells are the
Trophoblast proliferation after methotrexate
Fertility and Sterility
Table 1 Clinical and Histopathological Results From MTX-Treated Patients EP case no.
GA*
MTXdose
d
Interval until surgery
Rupture
Intraluminal trophoblastic spread
Intramural trophoblastic spread
Subserosal or serosal trophoblastic spread
Placentation Intraluminal
Intramural
Serosal
d
Proliferation index %
1
37
75mg transvaginal injection
4
2
37
25mgx51M
12
3
45
50mg transvaginal injection
8
4
48
25mgx51M
5
5
50
25mgx51M
12
6
54
25 mg x 5 1M
12
7
55
54mgx31M
14
8
60
25 mg x 51M
19
9
68
50mgx2 transvaginal injection
7
+
10
74
50mgx2 transvaginal injection
32
+
Median Range
52 37 to 74
+
Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
+
35
+
17
Intermediate trophoblast
+
19
Intermediate trophoblast
+
60
Intermediate trophoblast
+
3
Intermediate trophoblast
+
28
Intermediate trophoblast
+
2
+
18
+
9
+
68
Intermediate trophoblast
Intermediate trophoblast
19 2 to 68
* GA, gestational age. only constituents capable of mitotic activity. The effect of MTX on trophoblastic tissue has been studied by Schafer et al. (5), who, based on in vitro tissue cultures, found that trophoblasts from EP compared with those from intrauterine pregnancies required a concentration of MTX 10 times higher to induce an equivalent reduction ofhCG levels. Normally, >8 days were required for hCG to decline after application of MTX, and some cases showed even no reduction in hCG values at all. From this study it may be concluded that, if low and nonresponding patients exist in vivo, treatment with MTX could fail in some patients. Further, Sand et al. (6) demonstrated that MTX inhibited growth of trophoblast cell cultures in vitro at concentrations 1,000 times higher than those required to inhibit choriocarcinoma cell proliferation. The histopathology of EP after MTX treatment Vol. 65, No.4, April 1996
has been presented in only two studies. Kooi et al. (7) studied five cases and concluded that MTX applied locally did not seem to compromise tubal response to the insult of invading chorionic tissue or to interfere with the ultimate tubal repair. On the other hand, Klinkert et al. (8) reported a case of persisting EP that seriously hindered tubal repair 6 months after MTX treatment. The literature concerning the general morphology ofEP after surgical intervention raises controversial questions. Some authors (9-11) report a predominantly intraluminal trophoblastic spread, i.e., the major feature of tubal pregnancy is that chorionic villi and villus-associated cytotrophoplast remain superficially implanted without invasion of the muscular wall and that rupture is due to vascular thrombosis and intraluminal hemorrhage with dilatation of the tube. Others (12-15) describe a predomi-
Floridon et al.
Trophoblast proliferation after methotrexate
731
Figure 1 Polar proliferation of Ki67-positive cytotrophoblast is abolished after MTX treatment (Ki67/human placental lactogen X120).
nantly extraluminal trophoblastic spread, i.e., the main feature of the growing gestation is a rapid penetration of the tubal wall with development of the placenta in the muscularis, subserosa, and serosa, leading to disruption of all layers in combination with interstitial hemorrhage. Methotrexate treatment ofEP still carries the risk of later surgery in approximately 18% of cases, despite careful selection of patients (2). Increased risk of treatment failure has often been associated with high ,8-hCG levels and a living embryo. These low or nonresponders might demonstrate a high proliferative index of the cytotrophoblast as demonstrated by the Ki67-antigen expressed in nuclei in active cell cycle (excluding G0 ). To evaluate the effect of MTX, we studied tubal morphology and synthesis ofhCG and human placental lactogen by trophoblasts from 10 MTX-treated EPs and compared the results to a control group treated primarily by surgery. We further examined the proliferative index by the application of the monoclonal antibody MIB1, detecting the Ki67-antigen and the composition of the inflammatory response. MATERIALS AND METHODS
Histopathology of 10 cases ofEP treated unsuccessfully with MTX were reviewed and compared with a control group of 10 EP primarily treated by surgery. Formalin-fixed paraffin-embedded archival material, each comprising 5 to 10 tissue blocks with the implan732
Floridon et al.
tation site represented, were stained by hematoxylin and eosin, CD 3 (pan T-lymphocyte), CD20 (pan Blymphocyte), and double-stained with monoclonal MIB 1 and hCG or human placental lactogen. For immunostaining, 5-J.Lm thick sections were cut and mounted on SuperFrost/Plus glass slides (Menzel-Glaser, Braunschweig, Germany) and dried for 1 hour in an oven at 60°C. Sections were dewaxed in xylene and rinsed in alcohol. Endogenous peroxidase was blocked using 0.5% hydrogen peroxide in methanol. Mter rinsing in water, antigen unmasking was performed. Microwave heating for 20 minutes in a 10 mM citrate buffer (pH 6) as described by Cattoretti et al. (16) was used for the Ki67 antigen, human placental lactogen, hCG, and CD20, whereas a 15minute pretreatment of 0.05% protease (Type 14; Sigma, St. Louis, U.S.A) in Tris-buffer (pH 7.4) at 37°C was applied for CD3. Mter antigen unmasking, sections were pretreated with 2% bovine serum albumin (BSA). Both standard and double immunohistochemical staining were performed. Standard immunostaining was used to identify CD3 (Dako A452, Glostrup, Denmark) and CD20 (clone L26, Dako M755). Both antibodies was diluted 1:200 and applied for 30 minutes followed by a peroxidase-labeled streptavidin biotin method (Dako, Glostrup, Denmark) using 3-amino-9-ethylcarbazol as chromogen. Double immunohistochemical staining for Ki67 /human placental lactogen and Ki67/hCG was performed using a modified sequential method. The Ki67 antigen was identified using the monoclonal antibody MIB1 (Immunotech, Marseille, France) followed by the peroxidase-labelled streptavidin biotin method (Dako) using DAB-Ni as chromogen. Antihuman placental lactogen (Dako A137) or anti-hCG (Dako A231) were used to identify human placental lactogen and hCG in a indirect immunoalkaline phosphatase method using Fast Red TR as chromogen (Dako). In brief, sections were incubated with MIB1 diluted 1:100. Mter rinse in Tris-buffered saline (TBS), biotinylated goat anti-mouse immunoglobulin (Ig) antibody (Dako E433) was applied at dilution 1:200 followed by TBS rinse and incubation in peroxidase-conjugated streptavidin (Dako P397) at dilution 1:300. All incubations were done at room temperature for 30 minutes. Slides were developed for 10 minutes using a 3.3' -diaminobenzidine tetrahydrochloride (0.05%), hydrogenperoxide (0.01 %), nickel sulfate (2%) solution in a 0.1 M acetate buffer at pH 6 (17). After color development, sections were rinsed in water and pretreated with 2% BSA. Sections then were incubated for 30 minutes with antihuman placental lactogen or anti-hCG at a dilution 1:500. Mter rinse in TBS, 30 minutes incubation with alkaline phosphatase-conjugated swine antirabbit Ig (Dako D306) were done at dilution 1:50 and
Trophoblast proliferation after methotrexate
Fertility and Sterility
Table 2
Clinical and Histopathological Results From Surgery-Treated Patients Placentation
EPno.
GA*
Rupture
Intraluminal trophoblastic spread
Intramural trophoblastic spread
Subserosal or serosal trophoblastic spread
Intraluminal
Intramural
Serosal
d
Proliferation index %
11
33
+
12
33
+
13
36
+
14
44
+
15
44
16
45
17
50
18
53
19
54
20
62
Median Range
45 33 to 62
+
+
Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
+
9
Intermediate trophoblast, syncytiotrophoblast Intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
+
+
+
+
92
+
+
Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast Cytotrophoblast, intermediate trophoblast, syncytiotrophoblast
95
93
Intermediate trophoblast
+
Intermediate trophoblast
+
92
+
90
Intermediate trophoblast
+
89
+
+
94
+
+
95
93 9 to 97
sections were developed for 20 minutes in Fast Red TR. Finally, sections were counterstained with Mayers Hematoxylin for 5 seconds and mounted with Aquamount (BDH, Poole, United Kingdom). Negative controls were performed, including omitting the primary antibodies from the last sequence (anti-human placentallactogen/anti-hCG), and no unspecific staining or cross-reactivity was observed. All stainings were carried out during one procedure to minimize differences in staining intensity due to day to day variation or variation due to applying different batches of primary and/or secondary antibodies, buffer composition, or other known sources of errors. The proliferative index, expressed in percent, was estimated by counting the number of MIBl-positive nuclei versus the total number of nuclei in a minimum of 300 cytotrophoblast cells at the implantation site, using an objective lens x40. The nuclei were counted as positive or negative without regard to Vol. 65, No.4, April 1996
97
staining intensity. Intermediate trophoblast, because of a low proliferation rate, syncytiotrophoblast, decidua, maternal, and fetal tissue were all excluded from the analysis of the proliferative index. Statistical analysis of the proliferative index was performed using the Mann-Whitney U-test. The morphology of the implantation site and the mode of trophoblastic spread (trophoblastic spread), placentation, decidualization, vascular invasion, and hemorrhage as well as the inflammatory response were evaluated by conventional histopathological methods.
RESULTS Methotrexate treatment
In the MTX group (Table 1), 10 patients were treated with MTX ranging from 50 to 162 mg either 1M or by ultrasound-guided transvaginal injection
Floridon et al.
Trophoblast proliferation after methotrexate
733
Proliferative index at the implantation site varied from 2% to 68% (median 19%), and nuclear reaction in the cytotrophoblast was strongly positive if proliferation was present. Protein hormones (hCG and human placental lactogen) could not be demonstrated. If positive, the intermediate trophoblast demonstrated a low focal to diffuse degree of immunoreactivity with MIB1 and a diffuse reactivity for hCG and human placental lactogen. The syncytiotrophoblast were totally MIB 1 unreactive, but stained strongly positive for hCG and human placental lactogen. None of the placental chorionic villi showed polar proliferation (Fig. 1) characterized by clusters of MIB1-positive cytotrophoblast emanating from only one pole of the placental villi towards the implantation site, as in normal intrauterine pregnancies. Invasion of only small luminal tubal vessels was found by intermediate trophoblast, none of these showing invasion by cytotrophoblast. The intermeFigure 2 Cytotrophoblastic cell columns showing normal polar proliferation after surgical treatment (Ki67 /human placental lactogen X 120).
into the gestational sac. The gestational age (GA) varied from 37 to 74 days (median 52 days), and the interval between MTX treatment and surgery ranged from 4 to 32 days. In three cases (cases 4, 9, and 10), tubal rupture made surgical removal necessary. Four cases (cases 3, 6, 7, and 8) underwent surgery because of abdominal pain and/or local defence, so did two cases (cases 2 and 5) in combination with rising ,8-hCG levels. One patient (case 1) was salpingectomized because of noncompliance after only 4 days where increasing levels of ,8-hCG is found normally. The maximal ,8-hCG before treatment with MTX varied from 550 to 21,500 IU/L (International Reference Preparation [IRP]), during treatment from 850 to 27,000 IU/L, and presurgically from 260 to 27,000 IU/L. Seven patients had declining ,8-hCG during treatment. All10 tubes showed trophoblastic spread comprising all three trophoblast types (cytotrophoblast, intermediate trophoblast, syncytiotrophoblast) with a luminal superficial placentation at the implantation site. Intramural spread, only expressed by intermediate trophoblast, was found in seven cases and distributed both vertically and longitudinally. None of {pe intermediate trophoblast reached the subserosa or serosa. The morphology of the intermediate trophoblast cells varied according to its location from round to polygonal or ovoid to spindle shaped. Only one case (case 1) had a patchy mucosal decidual reaction. 734
Floridon et al.
Figure 3 Intramural placentation in a tubal pregnancy (-). Chorionic villi with polar proliferation towards the implantation site is seen in the lumen (---t) (H&E x 40).
Trophoblast proliferation after methotrexate
Fertility and Sterility
Figure 4 Intermediate trophoblast can be confused with decidual cells (H&E x 300).
diate trophoblast destructed the vessel wall, replaced the endothelial cells, as well as filled the lumen with subsequent formation of hyalinized vascular channels. The tubes were dilated by free blood filling the lumen of the tube. Larger intramural or subserosa! vessels at a distance from the implantation site were not invaded by trophoblast, and hemorrhage into the maternal soft tissue was not marked. The inflammatory response consisted of a diffuse
and/or perivascular infiltration of acute and chronic inflammatory cells, but only to a low degree. The number ofT-lymphocytes was decreased compared with the controls. As seen in Table 1, there was no correlation between GA, dose of MTX, level of ,8-hCG, interval to surgery, or rupture on one side and trophoblastic spread, placentation, or proliferative index on the other, although a relatively high proliferative index was found in two of three cases complicated by rupture.
Figure 5 Immunohistochemical analysis aids in distinguishing the intermediate trophoblast (human placental lactogen-positive) from decidual cells (human placental lactogen-negative) (KI67I human placental lactogen X300).
Vol. 65, No. 4, April 1996
Floridon et al.
Trophoblast proliferation after methotrexate
735
Surgical Treatment
The control group of 10 patients (Table 2) had their tubes removed either by laparotomy or laparoscopy. ,B-hCG was not measured in any of these cases. In six patients, tubal rupture was observed. The GA varied from 33 to 62 days (median 45 days). Nine fallopian tubes showed mucosal trophoblastic spread with all trophoblastic components (cytotrophoblast, intermediate trophoblast, syncytiotrophoblast) present and luminal superficial placentation. One tube (case 12) was without chorionic villi but a few cytotrophoblasts, intermediate trophoblasts, and syncytiotrophoblasts were observed superficially in the mucosa. Intramural spread was seen in seven cases, with all the different trophoblastic components (cytotrophoblast, intermediate trophoblast, syncytiotrophoblast) represented. Intramural placental chorionic villi were seen in five cases. In the subserosa, trophoblastic spread was present in five cases and placental chorionic villi in was present in one case. Decidual reaction was not observed. The proliferative index at the implantation site varied from 9% to 97% (median 93%). Ifpositive, the MIBl, hCG, and human placental lactogen staining intensity of all tree trophoblastic cell types was identical to the one seen after MTX treatment. The cytotrophoblastic cell columns of the villous stems showed polarization of growth toward the implantation site, visualized by clusters of MIBl-positive cytotrophoblasts (Fig. 2). Intravascular invasion of both small luminal and larger distant tubal vessels by intermediate trophoblasts were found in all cases. In two cases, syncytiotrophoblast was seen in the lumina of vessels. Hemorrhage was common into the lumen as well as into the muscular and subserosal parts of the tube as a result of trophoblastic invasion. The inflammatory response was more pronounced compared with the MTX group, and the number of T-lymphocytes was higher. Methotrexate Versus Surgery
A statistically significant difference in proliferative index between the two groups was found (19% versus 93%, P < 0.01) (Tables 1 and 2). Chorionic villous invasion of the muscular wall never was seen in the MTX group but was a common feature in the surgically treated cases (Fig. 3). In both groups, the inflammatory response consisted of a diffuse and/or perivascular infiltration of acute and chronic inflammatory cells, but to a lesser degree in the MTX group where the number of Tlymphocytes was decreased as well. There was no difference in the number of B-lymphocytes between 736
Floridon et al.
the two groups. It was not possible to detect any relation between tubal rupture and the microscopical location of trophoblastic spread, placentation, or hemorrhage. DISCUSSION
The last decade has witnessed an epidemic rise in the incidence of EP. Although success rates of 82% to 95% (2) are reported after MTX treatment, experience still is limited as far as later reproductive performance is concerned. Intrauterine pregnancies (3, 4) have been achieved in 47% to 69% of cases, which seems to compare favorably with the 44% to 52% reported after surgery (18). It is still largely unknown how MTX affects the invasive property of the trophoblast and the growth of placental chorionic villi and why some EPs fail to respond to MTX, why tubal ruptures occur, and whether the inflammatory response is influenced by MTX. An answer to these questions may help to optimize MTX therapy and increase the fertility rates. Former conclusions by Kooi et a1. (7) from 1992 concerning the morphology of tubes after treatment with MTX may be misleading because these authors seem unaware of the existence of the intermediate trophoblast studied in detail by Kurman et a1. (19). This cell type, occurring in all instances of pregnancy, is thought to represent an intermediate stage of differentiation of the trophoblast. The intermediate trophoblast is capable of mitotic activity and hormone synthesis (hCG and human placental lactogen), it is generally mononuclear and expresses many of the characteristics of both cytotrophoblast and syncytiotrophoblast. The intermediate trophoblast plays an important role in implantation-invasion and establishment of the maternal-fetal circulation (19). The results by Kooi et a1. (7) were based on routine stainings of only five cases of tubal EP. Stainings visualizing cells in proliferation and/or hormone-producing trophoblastic tissue were not carried out. Case 1 exhibited predominantly intraluminal trophoblastic spread and a marked decidual response adjacent to the tubal rupture, but the evidence for a tubal decidual reaction is not convincing (7). It may be difficult to distinguish between a decidual reaction and the intermediate trophoblast (Figs. 4 and 5). Decidual cells have small, uniform, round nuclei and homogenous cytoplasm surrounded by a well-defined cell membrane, whereas intermediate trophoblastic cells show larger, hyperchromatic nuclei and amphophilic cytoplasm with indistinct borders and stain positive for hCG human placental lactogen (19). This is in contrast to decidual cells. A predominantly extraluminal trophoblastic spread was seen in case 2. The remaining three cases were removed surgically > 1
Trophoblast proliferation after methotrexate
Fertility and Sterility
year after MTX treatment for reasons not related to the clinical course of the EP, and the ,8-hCG was negative before surgery. In these cases, the ultimate tubal repair was found to be normaL Seriously hindered tubal repair and persistent EP was casuistically reported by Klinkert et aL (8) and based on injection of 5 mg MTX only. The patient underwent surgery 6 months after MTX treatment. We agree with Fernandez (20) and Kock and Kooi (21) that such a treatment regimen does not seem to be superior to expectant management, former reports indicating failure rates of 31 % to 36% (22, 23). The EP was located in the isthmic part of the tube (8) and experience with this location still is limited to a few cases. Finally, a persistent trophoblastic disease cannot be discerned from a new EP because no follow-up of ,8-hCG levels was reported. Our treatment regimen might be insufficient, too, because five cases were treated with 0.5 mg/kg 1M MTX for 5 days. A recent study (24) seems to indicate that patients given such a dose were undertreated and required additional therapy to obtain a decrease in,8hCG levels. Outcome of treatment might depend on peak serum levels ofMTX, more than on the cumulative dose. The main achievements of our study were to demonstrate that trophoblastic spread and placentation always were confined to the tubal mucosa after MTX treatment; invasion of placental chorionic villi and polar proliferation toward the implantation site were observed only in the surgery group; a significant change in proliferative capacity of the cytotrophoblast in the MTX group, although a relatively high proliferative index was found in two of three cases complicated by rupture. Our findings in the surgery group support that EP is associated with a component of extraluminal trophoblastic spread (12-15). We observed both intraluminal, intramural, and subserosal trophoblastic spread as well as placental chorionic villi within the same locations. Invasiveness of the trophoblast after the formation of chorionic villi appears to rely on intermediate trophoblast only (19). The mechanisms responsible for trophoblastic invasion with subsequent intramural or subserosal placentation in the fallopian tube still are unclear. It has been proposed by Randall et aL (13) that these processes are entirely a function of trophoblastic tissue, because the blastocyst appears to be able to implant at any location, the maternal tissues playing a purely passive role. Feinberg et aL (25) have suggested that an interplay between urokinase-type plasminogen activator (PA) and two different plasminogen activator inhibitors (PAl 1 and PAl 2) is responsible for trophoblastic invasion in the intrauterine location. Their findings of PAl 1 in intermediate trophoblast Vol. 65, No.4, April 1996
invading the uterine lining including the arteries raise the possibility that PAl 1 could either function to promote trophoblastic invasion or be required for cessation of the invasive process. A study concerning this issue regarding EP is presently under consideration at our laboratory. The statistically significant difference in proliferative activity (19% versus 93%, P < 0.01) between the two groups in our study supports the in vitro demonstration by Schafer et aL (5) concerning the inhibitory effect of MTX on the trophoblast. They showed that the longer the interval between MTX treatment and hCG measurement, the more pronounced the effect. The decreased number ofT-lymphocytes might indicate that MTX is capable of inhibiting the inflammatory process in the tubal wall by reducing activation and/or recruitment of inflammatory cells, which could explain a better preservation of the tube, but it may be a result as well of the abolished trophoblastic spread and deep placentation. From the present study, it can be concluded that trophoblastic spread in MTX-treated cases seems to be limited to the mucosal part of the tube and that a high proliferative index of cytotrophoblast is associated with an increased risk of treatment failure. Our results might be influenced by selection bias because only archival material with sufficient tissue blocks were included and/or the small study groups. Future studies concerning morphology and the assessment of cell proliferation in EP would benefit from a prospective randomized design with careful preparation of the specimens to avoid distortion or other damage. . REFERENCES 1. Tuomivaara L. Ectopic pregnancy. An analysis of etiological
2. 3.
4. 5.
6.
7.
8.
risk factors and subsequent fertility [thesisl. Oulu, Finland: University of Oulu, 1990. Floridon e, Thomsen SG. Methotrexate treatment of ectopic pregnancy. Acta Obstet Gynecol Scand 1994;73:746-52. Fernandez H, Benifla J-L, Leladier e, Baton e, Frydman R. Methotrexate treatment of ectopic pregnancy: 100 cases treated by primary transvaginal injection under sonographic control. Fertil Steril 1993;59:773-7. Stovall TG, Ling FW. Single dose methotrexate: an expanded clinical trial. Am J Obstet Gynecol 1993; 168:1759-65. Schafer D, Pfuhl JP, Baumann R, Neubert S, Bender HG, Naujoks H. Trophoblast tissue culture of human intrauterine and ectopic pregnancies and treatment with methotrexate. Hum Reprod 1992;7:311-9. Sand PK, Stubblefield PA, Ory SJ. Methotrexate inhibition of normal trophoblasts in vitro. Am J Obstet Gynecol 1986; 155:324-9. Kooi S, van Etten FHPM, Kock HeLV. Histopathology offive tubes after treatment with methotrexate for a tubal pregnancy. Fertil SteriI1992;57:341-5. Klinkert J, van Geldorp HJ, ehadha-Ajwani S, Huikeshoven F JM. Tubal damage after intratubal methotrexate treatment. Fertil Steril 1993;59:926-7.
Floridon et al.
Trophoblast proliferation after methotrexate
737
9. Pauerstein CJ, Croxatto HB, Eddy CA, Ramzy I, Walters MD. Anatomy and pathology of tubal pregnancy. Obstet Gynecol 1986; 67 :30 1-8. 10. Senterman M, Jibodh R, Tulandi T. Histopathologic study of ampullary and isthmic tubal ectopic pregnancy. Am J Obstet Gynecol 1988; 159:939-41. 11. Stock RJ. Tubal pregnancy. Obstet Gynecol Clin North Am 1991; 18:73-94. 12. Budowick M, Johnson TRB, Jr, Genadry R, Parmley TH, Woodruff JD. The histopathology of the developing tubal ectopic pregnancy. Fertil Steril1980;34:169-71. 13. Randall S, Buckley CH, Fox H. Placentation in the fallopian tube. Int J Gynecol Pathol 1987;6:132-9. 14. Parmley TH. The histopathology of tubal pregnancy. Clin Obstet Gynecol 1987;30:119-28. 15. Dietl J, Buchholz F, Kindler P A. Histopathology oftubal pregnancy. Int J Gynecol Obstet 1988;27:385-8. 16. Cattoretti G, Pileri S, Parravicini C, Beckers MHG, Poggi S, Key G, et al. Antigen unmasking on formalin-fixed, paraffinembedded tissue sections. J Pathol 1993; 171:83-98. 17. Scopsi L, Larsson LI. Increased sensitivity in peroxidase immunocytochemistry. Histochemistry 1986;84:221-30. 18. LundorffP, Thorburn J, Lindblom B. Fertility outcome after
738
Floridon et al.
19.
20. 21. 22.
23.
24.
25.
Trophoblast proliferation after methotrexate
conservative surgical treatment of ectopic pregnancy evaluated in a randomized trial. Fertil Steril 1992;57:998-1002. Kurman RJ, Main CS, Chen HC. Intermediate trophoblast: a distinctive form of trophoblast with specific morphological, biochemical and functional features. Placenta 1984;5:34969. Fernandez H. Risks of intratubal methotrexate [letterl. Fertil Steril 1993;60:929. Kock I, Kooi S. Risks of intratubal methotrexate [letter]. Fertil Steril 1993; 60:930. Ylostalo P, Cacciatore B, Sjoberg J, Kiiiiriiiinen M, Tenhunen A, Stenman UH. Expectant management of ectopic pregnancy. Obstet Gynecol 1992;80:345-8. Fernandez H, Rainhorn JD, Papiernik E, Bellet D, Frydman R. Spontaneous resolution of ectopic pregnancy. Obstet Gynecol 1988:71:171-4. Fernandez H, Bourget P, Ville Y, Lelaider C, Frydman R. Treatment of unruptured tubal pregnancy with methotrexate: pharmacokinetic analysis of local versus intramuscular administration. Fertil Steril 1994;62:943-7. Feinberg RF, Kao LC, Haimowitz JE, Queenan JT, Wun TC, Strauss JF, et al. Plasminogen activator inhibitor types 1 and 2 in human trophoblasts. PAI-1 is an immunocytochemical marker of invading trophoblasts. Lab Invest 1989; 61:20-6.
Fertility and Sterility