J Clin Pathol 1987;40:901-905

Leucocyte populations in ectopic tubal pregnancy

U EARL, D P LUNNY, J N BULMER

From the Department of Pathology, University ofLeeds, Leeds

SUMMARY Leucocytes at the ectopic implantation site in 10 cases of early tubal pregnancy werecharacterised with a series of monoclonal antibodies using an indirect immunoperoxidase techniqueon cryostat sections. Most were HLA-DR positive macrophages, and there were a small number ofmature T lymphocytes (UCHTI and Dako-TI positive cells). These results were compared withthose reported in normal first trimester intrauterine pregnancies, and the contributions of the vari-ous leucocyte types to successful implantation at both the ectopic and intrauterine sites wereassessed.

The developing embryonic and extraembryonic tis-sues of the placenta carry an assortment of tissuespecific and transplantation antigens, some of whichare foreign to the maternal immune system. Despitewidespread infiltration of maternal decidua, myo-metrium and uterine vessels by semiallogenic extra-villous cytotrophoblast, a detrimental immuneresponse between mother and fetus does not occur.

In early intrauterine pregnancy, however, largenumbers of potential maternal immune effector cellsare dispersed throughout the decidua lining theuterus, and leucocytes aggregate particularly aroundspiral arteries, the degenerating glands of the deciduabasalis, and in areas of trophoblastic invasion.`13Immunohistological characterisation of this infiltratehas shown that it mainly consists of macrophages andT lineage lymphocytes.3I-I The distribution and anti-genic phenotype of the T lineage cells suggests thatthey correspond to specific endometrial granulocytes,sometimes referred to as K6rnchenzellen or K cells.34Endometrial granulocytes form a unique populationof endometrial leucocytes that is prominent in the latesecretory phase of the menstrual cycle and in first tri-mester decidua. Morphologically, endometrial granu-locytes differ from circulating polymorphonuclearleucocytes in having phloxinophilic cytoplasmic gran-ules, and an oval, lobulated, pyknotic and hyper-chromatic lymphocyte-like nucleus.The functions of leucocyte populations at intra-

uterine implantation sites in humans have not beenfully determined. This contrasts with the findings inmurine decidua, in which there is evidence that leuco-cyte populations modulate the maternal immuneresponse.6

Successful pregnancy does not depend on an intra-uterine environment. Ectopic pregnancy usually

Accepted for publication 19 February 1987

arises in the fallopian tube, and progresses until tubalrupture or gross tubal distension occurs, necessitatingsurgical removal. Rarely, tubal pregnancy progressesto term with delivery of a normal infant.7 Tropho-blast in ectopic pregnancy is identical with normalfirst trimester intrauterine trophoblastic tissue in thefirst trimester, with regard to the expression of tro-phoblast and major histocompatibility complex anti-gens, and the localisation of pregnancy related hor-monal products.89 Ectopic pregnancy, therefore,provides an excellent opportunity to study early pla-cental development, and immunological interactionsbetween mother and fetus.

In this report we have characterised the leucocyticinfiltrate at the ectopic tubal implantation site andcompared the findings with those reported in normalintrauterine pregnancy.

Material and methods

Ten specimens of tubal pregnancies were receivedfresh within 30 minutes of surgical removal. Tubalrupture had occurred in one case, and a macro-scopically normal fetus was present in two of thespecimens. Five millimetre cubes of tissue were dis-sected to include tube wall adjacent to the pregnancyimplantation site. Tissue blocks were snap frozen inisopentane cooled in liquid nitrogen (BDH, Poole,Dorset), and then stored in liquid nitrogen until use.In seven cases the gestational age of the specimen wasknown, and these ranged from two to 10 weeks afterthe last menstrual period. Further blocks were fixed informalin and routinely processed for diagnostic pur-poses.

Endometrial curettings were taken at the time ofsalpingectomy in four cases and examined after rou-tine fixation in formalin and processing.The table gives details of the nine murine mono-

901

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.8.901 on 1 August 1987. D

ownloaded from

Earl, Lunny, Bulmer902Table Specificities, sources, and dilutions ofmonoclonal antibodies used

Monoclonal antibody Specificity detected Source/reference Dilution

F-10-89-4 Leucocyte common antigen Serotec, Bicester, UK'0 1/100 peritoneal exudateUCHTI Mature T lymphocytes (CD3) Unipath, Bedford, UK" 1/20 purified immunoglobulinDako-TI Mature T lymphocytes (CD5) Dakopatts A/S, Denmark 1/10 culture supernatant

proportion of immature B lymphocytesDako-T2 T lineage cells (CD7) Dakopatts A/S, Denmark 1/10 culture supernatantOKT 11 E rosette receptor (CD2) Ortho Diagnostics, UK 1/20 purified immunoglobulinOKT 8 Peripheral T cells (CD8) Ortho Diagnostics, UK"2 1/20 purified immunoglobulin

cytotoxic/suppressor subsetDako-Macrophage Tissue macrophages Dakopatts A/S, Denmark 1/20 culture supematantAnti-Leu-M3 Tissue macrophages Becton-Dickinson, USA 1/20 purified immunoglobulinM1/70 C3b-receptor Serotec, Bicester, UK 1/1 culture supernatantDako-Pan-B B lymphocytes (CD22) Dakopatts A/S, Denmark 1/10 culture supernatantTGI Granulocytes, 10% of monocytes PCL Beverley, University College 1/1 culture supernatant

Hospital, LondonCR3/43 Non-polymorphic class II MHC 13 1/1O culture supernatant

products (DR, DP, DQ) I chainDA6.231 Non-polymorphic class II MHC 1 1/10 culture supernatant

products (DR, DP, DQ) ,B chain

clonal antibodies and one rat monoclonal antibody(M1/70) used.'° -14

Serial 6 im cryostat sections were cut, dried in airovernight, fixed in acetone for 10 minutes, wrapped infoil, and stored at -20°C. After rehydration in Trisbuffered saline, pH 7-6 (TBS), monoclonal antibodieswere applied using a standard indirect immu-noperoxidase technique.3 An appropriate negativecontrol was included for each case. Sections werelightly counterstained with Harris's haematoxylinand mounted in synthetic resin.

Sections measuring 5 um were fixed in formalin andprocessed routinely before being stained by the stan-dard haematoxylin and eosin method.

Results

FORMALIN FIXED TISSUESChorionic villi were present in all specimens, and insix cases abundant extravillous cytotrophoblastinfiltrated the fallopian tube wall and vascular spaces.Chronic inflammatory cells of lymphocyte and histio-cyte morphology were present at the implantation siteand showed some tendency to aggregate around con-gested blood vessels. No consistent associationbetween chronic inflammatory cells and chorionicvilli or extravillous trophoblast was seen. Plasma cellswere rarely seen. Acute inflammatory cells wereidentified in areas of tubal haemorrhage and necrosis.Convincing evidence of previous salpingitis, repre-sented by tubal diverticulae proximal to the ectopicimplantation or hydrosalpinx, was seen in two cases.Focal decidualisation of the tubal mucosa was seen inthree specimens.

All four endometrial samples showed secretorychange, and in two there was an Arias-Stella reaction.Decidualised stromal fragments were present in only

one specimen. Endometrial granulocytes wereidentified in all cases.

FROZEN TISSUESAll specimens included tube wall with residual epi-thelium. Implanting chorionic villi and infiltratingextravillous trophoblast were present in three cases.Cells bearing leucocyte common antigen (LCA) wereidentified by the antibody F10-89-4. There were sub-stantial numbers of leucocytes in all specimens, eitherdispersed randomly, or forming interstitial aggregates(fig 1). There was no consistent association of leuco-cytes with vascular structures or implanting chorionicvilli.More than 75% of the total interstitial leucocytes

were stellate and labelled with the macrophage mark-ers Dako-Macrophage and anti-Leu-M3 (fig 2). Serialsections labelled with CR3/43 or DA6.231 showedthat these macrophages uniformly expressed non-polymorphic determinants of class II MHC antigens(figs 3a and b).A smaller proportion of interstitial leucocytes, at

most 25% of LCA positive cells, were small roundcells. They were diffusely dispersed and showed noassociation with vascular or trophoblastic tissues.These cells were reactive with the mature pan-T lym-phocyte markers UCHT1 and Dako-Tl, and also invarying proportions with the T lineage markersDako-T2, OKT1 1, and the suppressor/cytotoxic sub-set marker OKT8; T8 positive cells were usually lessabundant than TI- or T2-positive cells (fig 4). Wheresome aggregation of T cells occurred, a variableamount of reactivity was observed with the class IIMHC markers in consecutive sections, and this reac-tivity was independent of macrophages associatedwith the T lymphocyte aggregates. It proved toodifficult, however, to identify individual T lympho-

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.8.901 on 1 August 1987. D

ownloaded from

Leucocyte populations in ectopic tubal pregnancy

A~~~~~e :6.... :j"

1 A,

tiXt

646XMjt ~*,6J

4'14- p

P..,I

#4& .." .0.11

4.

Fig 1 Interstitial leucocytes labelled with lantigen marker F-10-89-4. There are two cecells (small arrows) and larger, irregular, st(arrowhead). (Haematoxylin counterstain.

cytes and assess their HLA-DR state

implantation site in ectopic pregnancy are HLA-DRpositive macrophages; in the normal non-pregnantfallopian tube these form only a small proportion ofthe total.15 7 Extensive numbers of macrophagesdefined by morphology, immunohistology, and histo-chemistry have been reported in normal intrauterine

As tissues in the first trimester, particularly in areas oftrophoblastic invasion, and around spiral arteries andendometrial glands.4 The association of macrophageswith a pronounced lymphocytic infiltrate around firsttrimester intrauterine glands, has led to speculationthat there may be a local immune response to glandu-lar epithelium.5 At the ectopic implantation site, theinflux of macrophages appear to be randomly distrib-

J*¢t uted, with no predeliction for tubal epithelium, vascu-lar structures, or trophoblastic tissues. This may bedue in part to the different responses of tubal mucosaand endometrium to pregnancy. Although derivedfrom a common Mullerian duct origin, tubal mucosadiffers from endometrium in its relative deficiency inprogesterone mediated decidualisation. Endometrialglands show a series of morphological changes

ATr throughout pregnancy; initially they appear hyper-euecommon secretory, but later they involute so that by the third

eucocypes-round trimester they appear only as slit shaped channels.'8tellate cells Furthermore, the epithelial antigenic phenotype

changes, so that early in gestation the glands lose classI MHC determinants and aquire novel antigens

with serial sec- expressed by trophoblastic cells.'9 Fallopian tube epi-tions alone.

Occasional interstitial cells were reactive withM1/70, which is directed against the C3b receptor,carried by monocytes, granulocytes, and a proportionof B cells. A few interstitial cells were reactive withDako-Pan-B (B lymphocyte marker) or TGI (granu-locyte marker).A constant population of leucocytes was present in

the basal layer of the epithelium of the fallopian tube.These intraepithelial leucocytes labelled as mature Tcells and showed variable OKT8 positivity.The stromal Hofbauer cells of fetal chorionic villi

labelled with FO-89-4 and were reactive with macro-phage markers but not with CR3/43 or DA6.231.Tubal epithelium and vascular structures were reac-tive with markers of class II MHC products, as pre-viously described.'5

Discussion

The presence of a chronic inflammatory cell infiltratearound the implantation site in tubal pregnancy hasbeen well documented, and as in the present study,seems to be related to the pregnancy tissues present,and not to pre-existing chronic salpingitis.'6 Using aseries of immunohistological markers, we have shownthat most of the leucocytes at, and adjacent to, the

.j..

N.,*3 _

9i A

T

4t '. S

:A.

v~~~~~~~ .. 4

I

A,,W8.kk̂b. 4t 4g ......... 4'.

It,

Fig 2 Stellate interstitial leucocytes labelled withDako-Macrophage adjacent to infiltrating trophoblast (T).(Haematoxylin counterstain.)

E,s

903

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.8.901 on 1 August 1987. D

ownloaded from

904

B si: ^

i:w; w w

^. s:+%: we * :asP.o:

b} s <g.. 9: . .

t_ .t., 8:t. 45

_ W:4 ^C ^ e

::

S

:J.

>: ''# r

!. 9.

k ^}: s.. w e N :X,i,} Xi. r

*# .58:im.EE dASe e

Z K.: _ w.

tj,-i;...fl

Ai

}s::5 r ..... ff .v

«S 4 r 1.s. ,,fP£.....v ::' dW:r ^. . 8 ... ..

eqk .., S _ ... , a>:- 5GSS - jg5 .,aS.

'#11 WZ | BUl: g,.... St N (,.. .

..aais*-bF: s.

Earl, Lunny, Bulmer

PrI'

44W>i. Ql

4:

A5M

.MW

.:

a I',

'la4 .4

*_s- .;L.

* p

*I....

4..

*:O*

-;

!ikka_r.

:. }. t

Fig 3 Consecutive sections showing interstitial leucocytes reactive with (a) Dako-Macrophage and (b) class IIMHCmarker DA6.231. (Haematoxylin counterstain.)

thelium does not show morphological changes inectopic pregnancy nor is there loss of class I MHCsurface antigens, although changes in expression ofHLA-D locus products have been described.8 15

*t ~~~~~~~t-i':.:9

:

60

.,,

Fig4 Small round interstitial leucocytes labelled with pan-Tlymphocyte marker, UCHTJ. (Haematoxylin counterstain.)

The role of macrophages in normal and ectopicpregnancy may be primarily phagocytic-that is, act-ing to remove tissue debris-or it may also be con-cerned with the immunological properties of antigenprocessing and presentation. In murine decidua thereis evidence that macrophages regulate the activities ofother immunocompetent cells through the release ofprostaglandin E2, which inhibits the production ofcytotoxic T lymphocytes.20

Classical T lymphocytes formed a much smallerproportion of the leucocytes present at the ectopicimplantation site. No T lineage cells bearing the Erosette receptor and immature T cell markers butlacking peripheral T cell markers, were shown. SuchT lineage cells have been reported in first trimesterintrauterine decidua, both distant from and adjacentto the implantation site.3 These cells outnumber clas-sic T cells, expressing the mature T cell phenotype bya ratio of 10:1, and are thought to correspond to theendometrial granulocytes present in routinely pro-cessed tissue.3 Immunohistological characterisationof these cells in frozen tissue is impeded by thedisruption through freezing of their distinctiveeosinophilic and phloxinophilic cytoplasmic granules.

Endometrial granulocytes were once thought to bederived from endometrial stromal cells, and to secreterelaxin in preparation for menstruation.21 Recentstudies, however, have shown that these cells may bederived from bone marrow and undergo further mor-phological and phenotypic differentiation within

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.8.901 on 1 August 1987. D

ownloaded from

Leucocyte populations in ectopic tubal pregnancy 905endometrium.2 3 Endometrial granulocytes are prom-inent in late secretory and menstrual phase endo-metrium and in first trimester decidua. Their functionis unknown, but their absence at the ectopic gestationsite makes it unlikely that they participate in localimmune modulation of materno-fetal cellular inter-actions and blastocyst implantation. A smallgranulated lymphocyte has been shown in murinedecidua; this produces a non-MHC dependent, solu-ble, suppressor factor which inhibits generation ofantipaternal cytotoxic lymphocytes.6 Suppressoractivity has been reported in human decidualexplants22; this suppression seems to originate fromsmall lymphocyte-like cells.23 24 Decidualisation ofendometrial stroma often occurs in ectopic preg-nancy, although as we confirmed, it is usually focaland patchy. 16 25 Soluble suppressor factors producedby granulated suppressor cells within uterine deciduamay act at a distant ectopic implantation site to per-mit blastocyst implantation and development.The presence of leucocyte and macrophage mark-

ers, but the absence of HLA-DR antigens on placen-tal Hofbauer cells at the ectopic gestation site, agreeswith results reported for normal intrauterine tissues inthe first trimester and reflects the early gestational

4 26ages of the ectopic specimens.In conclusion, ectopic pregnancy differs from nor-

mal intrauterine pregnancy in that it has no special-ised T lineage cells at the implantation site. Largenumbers of macrophages are present both in ectopicand intrauterine pregnancy tissues. Local immunemodulation of maternal cytotoxicity, in response toforeign fetal antigens, may revolve around pro-duction of soluble suppressor factors by macrophageswhich down regulate the activities of other immuno-competent cells. Conversely, soluble factors derivedfrom cells within uterine decidua may be capableof acting early in gestation at the distant tubalimplantation site. Isolation of these immune effectorcells from pregnancy tissues and study of theirproperties in functional assays may resolve thesequestions.We thank Dr PCL Beverley for providing the anti-body TG1 and theatre staff at the Clarendon WingTheatres, Leeds General Infirmary, for obtainingfresh specimens.References

I Pijnenborg R, Dixon G, Robertson WB, Brosens I. Tropho-blastic invasion of human decidua from 8-18 weeks ofpregnancy. Placenta 1980;1:3-19.

2 Bulmer JN, Sunderland CA. Bone-marrow origin of endometrialgranulocytes in the early human placental bed. J Reprod Immu-nol 1983;5:383-7.

3 Bulmer JN, Sunderland CA. Immunohistological characterisa-tion of lymphoid populations in the early human placental bed.Immunology 1984;52:349-57.

4 Bulmer JN, Johnson PM. Macrophage populations in the human

placenta and amniochorion. Clin Exp Immunol 1984;S7:393-403.

5 Bulmer JN, Johnson PM. Immunohistological characterizationof the decidual leucocyte infiltrate related to endometrial glandepithelium in early human pregnancy. Immunology 1985;55:35-44.

6 Clark DA. Materno-fetal relations. Immunol Lett 1985;9:239-47.7 Augensen K. Unruptured tubal pregnancy at term with survival

of mother and child. Obstet Gynecol 1983;61:259-61.8 Earl U, Wells M, Bulmer JN. The expression of major histo-

compatibility antigens by trophoblast in ectopic tubal preg-nancy. J Reprod Immunol 1985;8:13-24.

9 Earl U, Wells M, Bulmer JN. Immunohistochemical character-ization of trophoblast antigens and secretory products inectopic tubal pregnancy. Int J Gynecol Pathol 1986;5: 132-42.

10 Dalchau R, Kirkley J, Fabre JW. Monoclonal antibody to ahuman leucocyte-specific membrane glycoprotein probablyhomologous to the leucocyte-common antigen of the rat. Eur JImmunol 1980;10:737-44.

11 Beverley PCL, Callard RE. Distinctive functional characteristicsof human T lymphocytes defined by E rosetting or a mono-clonal anti-T cell antibody. Eur J Immunol 1981;11:329-34.

12 Rheinherz EL, Kung PC, Goldstein G, Schlossman SF. A mono-clonal antibody reactive with the human cytotoxic/suppressorT cell subset previously defined by a heteroantiserum termedTH2. J Immunol 1980;124:1301-7.

13 Sunderland CA, Naiem M, Mason DY, Redman DWG, StirratGM. The expression of major histocompatibility antigens byhuman chorionic villi. J Reprod Immunol 1981;3:323-31.

14 Guy K, Van Heyningen V, Cohen BB, Deane DL, Steel CM.Differential expression and serologically distinct sub-populations of human Ia antigens detected with monoclonalantibodies to Ia alpha and beta chains. Eur J Immunol1982;12:942-80.

15 Bulmer JN, Earl U. The expression of class II MHC gene prod-ucts by fallopian tube epithelium in pregnancy and throughoutthe menstrual cycle. Immunology 1987;61:207-13.

16 Pauerstein CJ, Croxatto HB, Eddy CA, Ramzy I, Walters MD.Anatomy and pathology of tubal pregnancy. Obstet Gynecol1986;67:301-8.

17 Peters WM. Nature of "basal" and "reserve" cells in oviductaland cervical epithelium in man. J Clin Pathol 1986;39:306-12.

18 Bulmer JN, Wells M, Bhabra K, Johnson PM. Immuno-histological characterization of endometrial gland epitheliumand extravillous trophoblast in third trimester human placentalbed tissues. Br J Obstet Gynaecol 1986;93:823-32.

19 Johnson PM, Bulmer JN. Uterine gland epithelium in humanpregnancy often does not express detectable maternal MHCantigens but does express fetal trophoblast antigens. J Immunol1984;132:1608-10.

20 Tawfik OW, Hunt JS, Wood GW. A major soluble factor pro-duced by uterine cells during murine pregnancy is prosta-glandin E2. J Reprod Immunol 1986;suppl: 126.

21 Dallenbach-Hellweg G. Histopathology of the endometrium. 3rded. Berlin: Springer-Verlag, 1981:26.

22 Golander A, Zakuth V, Shechter Y, Spirer Z. Suppression oflymphocyte reactivity in vitro by a soluble factor secreted byexplants of human decidua. Eur J Immunol 1981;11:849-51.

23 Daya S, Clark DA, Devlin C, Jarrell J, Chaput A. Suppressorcells in human decidua. Am J Obstet Gynecol 1985;151:267-70.

24 Daya S, Clark DA. Production of immunosuppressor factor(s) bypreimplantation human embryos. Am J Reprod ImmunolMicrobiol 1986;11:98-101.

25 Breen JN. A 21 year survey of 654 ectopic pregnancies. Am JObstet Gynecol 1970;106:1004-16.

26 Sutton L, Mason DY, Redman CWG. Cells bearing class IIMHC antigens in the human placenta and amniochorion.Immunology 1986;58:23-9.

Requests for reprints to: Dr U Earl, Department of Pathol-ogy, University of Leeds, Leeds LS2 9JT, England.

copyright. on D

ecember 17, 2020 by guest. P

rotected byhttp://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.8.901 on 1 August 1987. D

ownloaded from