BIOLOGY OF REPRODUCTION 53, 312-320 (1995)

Mast Cell and Eosinophil Distribution and Activation in Human Endometrium throughout the Menstrual Cycle1 Maria Jeziorska, 3 Lois A. Salamonsen, 4 and David E. Woolley2 ' 3 Departmentof Medicine,3 University Hospital of South Manchester, ManchesterM20 8LR, United Kingdom PrinceHenry's Institute of Medical Research,4 Clayton, Victoria 3168, Australia ABSTRACT Tryptase and chymase immunolocalization techniques have been used to examine the distribution, activation, and tryptase/chymase phenotype of mast cells (MCs) in 107 endometrial specimens that represented every day of the human menstrual cycle. MCs were identified in the endometrium in all stages of the menstrual cycle; similar MC numbers were observed for the functionalis, basalis, and muscularis. Extensive MC activation/degranulation, as judged by extracellular tryptase, was a common feature of the functionalis in specimens sampled just prior to and during menstruation. MC activation was also prominent in the functionalis at times coincident with recognized stromal edema. MCs of the functionalis did not contain chymase; all stained for tryptase and represent the MCT phenotype. By contrast, the basalis and muscularis showed a proportion of MCs containing both tryptase and chymase, MCTC. One important function for extracellular MC tryptase and chymase is their ability to activate precursor forms of the matrix metalloproteinases, enzymes recognized as instrumental in stromal degradation. Quantitative analysis of MC numbers, expressed relative to stromal cell numbers/mm 2, indicated no major changes during the menstrual cycle, although changes in MC morphology, granule content, and activation/degranulation were recognized for specific stages. Eosinophils, detected with monoclonal antibodies EG1 and EG2, were absent from extravascular sites between Days 5 and 26 but showed local accumulations just prior to and during menstruation. Since MCs and eosinophils between them contain avariety of potent mediators, it seems likelythat both cell types assume importantfunctional roles in relation to tissue and vascular remodeling associated with endometrial physiology.

INTRODUCTION

ology, it would seem essential to first obtain an accurate histological analysis of their distribution throughout the normal cycle. Earlier studies have rightly highlighted the problems encountered in histological staining of uterine MCs; most studies having used formalin-fixed tissues and metachromatic staining of the heparin-containing MC granules with conventional stains such as toluidine blue, methylene blue, or azure B [4-8]. These traditional methods have given little indication of MC activation/degranulation and have now been shown to be inferior to the more sensitive immunological techniques that employ antibodies either to the specific MC proteinases tryptase and chymase [16] or to MC histamine [171. Since both enzymes are components of MC granules, their tissue distribution provides an assessment of MCs in both nonactivated and degranulated states. Although variation in tryptase and chymase content is a major feature of MC heterogeneity in human tissues, tryptase is the dominant enzyme of virtually all human MCs [16-18]. We have used tryptase and chymase immunolocalization techniques to examine the distribution, activation, and tryptase/chymase phenotypes of MCs in 107 endometrial specimens that represent every day of the normal human menstrual cycle. Moreover, since MCs are known to release chemotactic factors for eosinophils, we have also used immunotechniques to examine the distribution of eosinophils in these same specimens. Our MC and eosinophil observations are presented and discussed in relation to endometrial remodeling, especially the association of MC activation with times of stromal edema and menstruation.

The human endometrium undergoes monthly cycles of proliferation, cellular differentiation, and secretory activity; in the absence of blastocyst implantation, the functionalis undergoes breakdown and is discharged at menstruation. This extensive tissue remodeling is regulated overall by responses to changes in ovarian steroid hormones [1]. Whereas stromal, epithelial, and endothelial cells represent the major cellular components of the endometrium, other cells such as macrophages, granulated T cells, neutrophils, and mast cells (MCs) have all been described in detail [2, 3]. Recently the functional roles of these latter "inflammatory" cells in remodeling processes during the normal menstrual cycle have been receiving attention [3]. The MC in particular has been the subject of several histological reports [4-9], and MC numbers have been related to various clinical conditions [10-12]. The MC is now considered to play a pivotal role in a variety of biological responses including angiogenesis, fibrosis, wound healing, inflammation, and tissue remodeling [13-15]. Since MCs contain a variety of potent mediators such as histamine, heparin, proteinases, leukotrienes, cytokines, and growth factors [13, 15], it is possible that the MC could profoundly affect the various morphological changes associated with the menstrual cycle. However, in order to investigate the precise functional role(s) of MCs in endometrial physiAccepted April 6, 1995. Received January 18, 1995. 'The work was supported by the North West Regional Health Authority (D.E.W.) and the NH &MRC of Australia (L.A.S.). 2Correspondence. FAX: 44 161 434 5194.

312

MAST CELLS AND EOSINOPHILS IN HUMAN ENDOMETRIUM

MATERIALS AND METHODS Endometrial tissue was obtained at curettage from 107 women who had regular menstrual cycles (28-29 days) and no apparent endometrial dysfunction and who gave their informed consent for collection of the tissue (approved by the Human Ethics Committees at the Monash Medical Centre and the Epworth Hospital, Melbourne). The women either had proven fertility and were scheduled for tubal ligation or were undergoing testing for patency of the fallopian tubes. Patients with leiomyomas or endometriosis and those who had received steroid treatment of any kind over the past year were excluded. Thirty-four specimens were fixed in 10% buffered formalin at 4C overnight, and 83 were fixed in Carnoy's fixative at room temperature for 2 h; all were processed routinely to paraffin blocks and showed similar reactivity with the tryptase antibody. Tissue sections were cut at 5 .m, hydrated, and stained with hematoxylin and eosin for histological dating of the menstrual cycle according to the method of Noyes et al. [19]. Specimens were classified according to an idealized 28-day reproductive cycle and then grouped into early, mid-, and late proliferative phases; early, mid-, and late secretory phases; and menstrual phase. A proportion (n = 10) of the curetted specimens contained superficial fragments of myometrium. Mouse monoclonal antibody to human tryptase (Chemicon International Inc., London, UK), used for Carnoy'sfixed tissue, was diluted 1:300 in Tris-buffered saline (TBS, ph 7.6) and applied to the tissue sections for 2 h at 20 0C. After three 10-min washes in TBS, biotinylated rabbit antimouse antibody (Dako Ltd., Glostrup, Denmark) diluted 1:100 in TBS was applied to the sections for 45 min. After further washing, the StreptABComplex (Dako Ltd) conjugated with alkaline phosphatase was applied for 45 min, sections were washed in TBS, and the alkaline phosphatase was developed through use of New Fuchsin substrate as previously described [20]. For formalin-fixed tissue, mouse anti-human tryptase antibody AA1 from Dako was used. Mouse monoclonal antibody to human MC chymase (Chemicon International Inc.) was diluted 1:100 in TBS and applied to the tissue sections for 2 h at 20°C; subsequent procedures for the localization of tryptase were performed as described above. Mouse monoclonal antibodies against human eosinophil cationic protein EG1 and EG2 (Kabi Pharmacia, Milton Keynes, UK.; recognizing resting/activated and activated eosinophils, respectively) were diluted 1:100 in TBS and applied to the tissue sections for 2 h at 20 0C. After washing with TBS, rabbit anti-mouse IgG in TBS (1:20) was applied for 45 min; this was followed, after further washing, by mouse APAAP in TBS (1:50) for 45 min. The sequence of secondary antibody and mouse APAAP was repeated (15 min each step), and alkaline phosphatase was developed as for tryptase.

313

TBS or normal mouse IgG (Dako), at IgG concentrations similar to those for the primary antibodies, was substituted for the primary antibody on control tissue sections of each specimen, and negative results were consistently produced. ComparativeStaining Techniquesfor MCs The tryptase immunolocalization technique was compared to other conventional staining using four formalinfixed and four Carnoy's-fixed endometrial specimens from early- and mid-secretory phase. Consecutive tissue sections from one specimen were each stained with the following procedures: 1) acidified toluidine blue; 2) long toluidine blue [21]; 3) Alcian blue/safranin; 4) chloroacetate esterase [22]; and 5) tryptase immunolocalization. Each staining procedure was evaluated on three different sections from each specimen. Assessment of MC Distribution Since the endometrium is a complex tissue that undergoes cyclical changes in the gland/stroma ratio, and stromal changes in relation to visible edema, the assessment of MC numbers per unit area was recognized as problematical. In this study the MC as well as the stromal cell distribution in functional endometrium is expressed in relation to unit area (mm2 ), thereby allowing an assessment of the ratio of MC to stromal cells/mm 2 throughout the cycle. This was evaluated for 14 specimens across the cycle; cell counts were made on three tissue sections for each specimen by use of 25 graticule fields of a 40 X objective, of necessity excluding glandular structures and large blood vessels. All specimens were examined by means of a Zeiss (Thornwood, NY) photomicroscope III with 16 X, 40x, and 100 X oil immersion objectives; photography was performed with Ektachrome 160 ASA tungsten film and TMAX 100 pro film (Eastman Kodak, Rochester, NY). RESULTS Assessment of MC StainingMethods Analysis of the different staining procedures showed that tryptase immunolocalization was far superior to all other methods. In terms of MC counts, approximately twice as many endometrial MCs were visualized by tryptase staining as by toluidine blue methods, the relative sensitivities for MC detection being tryptase > chloroacetate esterase > long toluidine blue > toluidine blue > Alcian blue/safranin and azure B (data not shown). This superiority of tryptase staining was retained regardless of whether the tissues were fixed in formalin or Carnoy's, a factor that affected both the esterase and metachromatic staining techniques. Interestingly, the chloroacetate esterase method was also better than the metachromatic staining procedures. The better sensitivity of the tryptase method was effectively illustrated by

314

JEZIORSKA ET AL.

the identification of relatively small MCs containing very few granules in the early proliferative and early secretory phases-MCs that the other stains often failed to recognize. Moreover, the extracellular distribution of MC tryptase also provided evidence of MC activation/degranulation, another feature that made tryptase immunolocalization the method of choice. MC Distributionin EndometrialSpecimens MCs were observed in all but 4 of 107 specimens examined. The endometrium (functionalis and basalis), muscularis, and endometrial/myometrial interface all contained MCs, but the top superficial layer of the functionalis contained relatively fewer MCs. MCs were identified in the endometrium at all stages of the menstrual cycle, but their morphology and states of activation/degranulation reflected specific stages of the cycle. For example, in the early proliferative phase the MCs were small and intact with no extracellular tryptase; however, in mid-proliferative phase, and again in mid-secretory phase, clear evidence of extracellular tryptase appeared as halos around MCs, indicative of MC degranulation (Fig. 1). These two phases correlated well with the recognized times of stromal edema, and observations of stromal disruption and lysis were commonly associated with this MC degranulation and tryptase release as visualized by Nomarski interference microscopy (data not shown). As to whether some observations of extracellular tryptase might have been artificially induced by the physical trauma of sampling the tissues, our concerns were alleviated to some extent by the reproducibility afforded by having several specimens for each day of the cycle. Moreover, the consistent observations of little or negligible extracellular tryptase observed for endometrial MCs during the late proliferative and early secretory phases, and the lack of significant extracellular tryptase for MCs of the basalis and muscularis throughout the cycle, provided some reassurance that MC integrity was retained during sampling processes. All photomicrographs and analytical observations excluded tissue locations close to the edges of each tissue specimen, since these sites might have been physically traumatized by sampling procedures. MC Activation/Degranulation Extensive MC activation, as judged by extracellular tryptase, was a common feature of specimens sampled 2 days prior to, and during, menstruation (Fig. f). These observations were consistently associated with local observations of connective tissue disruption and edema. High-power magnification of individual MCs of the functionalis at specific phases of the cycle demonstrated different morphological appearances-from MCs showing densely stained intracellular tryptase (e.g., Days 5-7 and 11-18) to those

showing extensive extracellular release of granule tryptase, and others showing partial degranulation (Fig. 1, g-j). Such observations indicate that MCs undergo activation/ degranulation at specific stages of the menstrual cycle and that they are intimately involved with stromal connective tissue changes such as edema and breakdown of tissue architecture at menstruation. Our assessment of 103 specimens indicated that MC degranulation was most pronounced just prior to and during menstruation, with significant tryptase release also being observed for the midproliferative and mid-secretory phases, as summarized diagrammatically in Figure 4. MC Tryptase/Chymase Phenotype MC heterogeneity, indicated by the differential content of MC tryptase and chymase, is an established feature of MC biology, and specific tissues usually illustrate one particular MC phenotype. MCs of the functionalis did not contain chymase; all stained for tryptase and represent the MCT phenotype (Fig. 2, a and b). By contrast, the basalis and muscularis showed some MCs positive for chymase, indicating the MCTC phenotype (Fig. 2, c and d). All MCs contained tryptase, the ratio of MCT:MCTC in the basalis and muscularis approximating 4:1 and 1:1, respectively. Such observations appeared consistent throughout the menstrual cycle. Eosinophils Since MCs are known to produce chemoattractants for eosinophils, the extensive MC activation observed at specific phases of the cycle suggested that eosinophils might also contribute to endometrial physiology. Application of the eosinophil monoclonal antibody markers, EG1 and EG2, showed that the endometrium contained no eosinophils in extravascular sites between Days 5 and 26. Eosinophils confined to blood vessels were occasionally visualized, but extravascular eosinophils became evident just prior to (Days 26-28) and during menstruation (Fig. 3, a and b). When observed, the eosinophils usually appeared

FIG. 1. Distribution and morphology of MCs in human endometrium throughout the normal menstrual cycle. a) Early proliferative phase. Small MCs scattered in dense stroma. Red staining for tryptase confined to MCs. x 400. b) Mid-proliferative phase. Numerous MCs with halos of extracellular tryptase indicating MC degranulation/activation. x160. c) Late secretory phase, Day 28 (POD 14 [postovulatory day]). MCs scattered in the stroma with pronounced extracellular tryptase. x 160. d) Mid-proliferative phase. MCs with some extracellular tryptase associated with localized stromal disruption. x 400. e) Mid-secretory phase. MCs and extracellular tryptase inedematous stroma (arrow). x 400. f) Late secretory phase, Day 28 (POD 14). Perivascular MCs with marked extracellular tryptase associated with areas of stromal lysis. x 400. g-j) High-magnification photomicrographs of MCs at different stages of degranulation/activation. g) Early proliferative phase-MC with intense staining for intracellular tryptase; (h)late proliferative phase-MC with weak intracellular staining for tryptase; (i) mid-secretory phase-MC with intense intra- and extracellular staining for tryptase; (j)late secretory phase--degranulated MC with mainly extracellular tryptase. g,h,j: x 1300; i: x 1000.

MAST CELLS AND EOSINOPHILS IN HUMAN ENDOMETRIUM

315

316

JEZIORSKA ET AL.

FIG. 2. Identification of MC tryptase/chymase phenotypes in human uterus. a) Functionalis, menstruation phase. Photomicrograph shows distribution of MCs as demonstrated by staining for MC tryptase. x 160. b) Tissue section consecutive with (a), stained for MC chymase. Note absence of chymase indicating an MCT phenotype for the phenotype. x 400. d) Muscularis, functionalis. x 160. c) Basalis, menstruation phase. Photomicrograph stained for MC chymase showing a few MCs (arrows) of MCTC menstruation phase. Chymase-containing MCs (arrows) indicating the presence of the MCTC phenotype in the muscularis. x400.

as local extravascular accumulations, often with a localized release of eosinophil cationic protein indicative of eosinophil activation (Fig. 3c). MC Distribution/Numbersthroughoutthe Menstrual Cycle Because the endometrium is a complex tissue, undergoing almost constant remodeling and changes in stromal edema during the cycle, the presentation of MC numbers per unit area requires comparison with stromal cell numbers per unit area to provide more meaningful data. Figure 4 presents the distribution of both MCs and stromal cells/mm 2 of stromal tissue obtained from careful numerical analysis of 14 endometrial specimens across the cycle as described in Materials and Methods. Although the numbers for MC/ mm2 showed some variation between specimens, the values for each specimen showed some proportionality with the numbers for stromal cells/mm2 . This ratio of MC:stromal cells/mm 2 of approximately 1:100 was averaged for the majority of specimens. The data therefore suggest that it is un-

likely that total MC numbers show major changes at different stages of the cycle, since their relative proportion to stromal cells appeared reasonably consistent (see Fig. 4). However, changes in MC morphology, granule content, and activation/degranulation are recognized in endometrial specimens at specific phases of the menstrual cycle, especially the relationship of stromal edema with MC degranulation, which is illustrated in Figure 4. DISCUSSION Several studies have examined the distribution of MCs in the human uterus through use of a range of metachromatic staining techniques [5-8, 23]. All have emphasized problems encountered in histological staining and MC quantification, mainly for reasons of low granule content due to early MC maturation stage or postdegranulation [4-24]. The tryptase immunolocalization has undoubtedly resolved many of the earlier reservations, providing more accurate assessments of MC numbers and evidence of degranulation.

MAST CELLS AND EOSINOPHILS IN HUMAN ENDOMETRIUM

FIG. 3. Demonstration of eosinophils in human endometrium. No eosinophils were observed in extravascular sites between Days 5 and 26. a) Late secretory phase, Day 28. Photomicrograph shows extravascular accumulations of eosinophils (EG1 and EG2 markers). x 160. b) Late secretory phase, Day 28. Extravascular accumulation of eosinophils (top center) and single eosinophils within blood vessels (arrows). x 400. c)Menstruation, Day 1.Eosinophils (arrows) and areas of red blood cells associated with extracellular eosinophil cationic protein (asterisk) indicative of eosinophil degranulation/activation. x 400.

317

Whereas most earlier MC studies focused on abnormal clinical conditions of the uterus [5-8], the present study has been restricted to an examination of endometrial MCs throughout the normal menstrual cycle of fertile women. In contrast to earlier reports that MCs were rarely seen in the functional endometrium [7] or were observed in very low numbers (1-3 MCs/mm 2) [231, our study has shown a prominent distribution of MCs in 103 of 107 specimens examined. Edematous changes and increasing gland diameter during the cycle represent the major difficulties in the assessment of MC numbers per unit area, and we are doubtful that any attempt to relate MC numbers to clinical or normal conditions will prove meaningful. However, it is our opinion that the proportional relationship of MC and stromal cell numbers for each specimen supports the conclusion that there is no major influx or loss of MCs prior to menstruation. Similar numbers of MCs were noted in the basalis and myometrium; but whereas these tissues rarely showed any evidence of extracellular tryptase, this was a common observation for the functionalis at specific stages of the cycle, suggesting that MC degranulation represents an important functional feature of endometrial biology. Since the demonstration of extracellular tryptase implies that other MC granule components such as histamine, heparin, serine proteases, and multifunctional cytokines/growth factors are released at the same time, it was to be expected that such sites would often show localized edema and disruption of the stromal matrix. MC heterogeneity is a recognized feature of different tissues; detailed biochemical, ultrastructural, and pharmacological studies of isolated human uterine MCs have suggested a distinctive phenotype in comparison to the well-characterized skin and lung MCs [25-271. Variations in tryptase and chymase content is a commonly cited parameter of MC heterogeneity [16, 28], and our immunolocalization studies of these enzymes have shown that different MC phenotypes exist within different layers of the human uterus. The MCT phenotype contains only tryptase of these two enzymes and predominates in lung and intestinal mucosal tissues, while the MCTr type contains tryptase, chymase, carboxypeptidase, and cathepsin G-like protease and predominates in normal skin and interstitial submucosa [13, 28]. Whereas only the MCT was observed in the functionalis, the MCTc phenotype became more prominent in the deeper tissues (namely, the basalis and muscularis). Although the biological substrates of the MC proteinases remain obscure, recent reports have implicated both MC tryptase and chymase in the activation of precursor forms of the matrix metalloproteinases. Tryptase directly activates prostromelysin [29, 30], whereas chymase activates both prostromelysin and procollagenase [30, 31]. Both stromelysin and collagenase are reportedly involved in the degradative processes associated with menstruation [32, 331. Their codistribution at menstruation with extracellular MC proteinases resulting from MC degranulation would

318

JEZIORSKA ET AL.

FIG. 4. Diagrammatic representation of the interrelationships between MC activation, stromal edema, and endometrial remodeling, and the proportionality of MCs to stromal cells during the normal menstrual cycle.

suggest a concerted and integrated proteolytic action prior to and during menstruation. Endometrial interstitial stroma contains a fibrillar matrix composed of collagens type I, III, V, and VI and fibronectin, whereas basement membrane structures are composed of laminin, type IV collagen, and heparan sulphate proteoglycan [34-36]. These matrix components may undergo compositional changes during the menstrual cycle [37, 381, especially the depletion of type VI collagen during the periimplantation period and at times of edema [39]. The prime function of type VI collagen microfibrils is to link the major elements of the extracellular matrix via their celladhesion and collagen-binding properties; thus the fact that MC tryptase or chymase effectively degrades type VI collagen [40] seems of immediate relevance to the stromal disruption and edematous changes that apparently coincide with endometrial MC degranulation. The abilities of MC tryptase and chymase to degrade other components of the stromal matrix or basement membranes are uncertain, and these are currently under investigation. Earlier studies showed that soluble MC products stimulated collagenase and prostaglandin E production by fibroblasts [41] and stimulated interleukin (IL)-1 production by

monocyte-macrophages [42]. These responses might well be explained by recent reports that human MCs express a range of multifunctional cytokines including IL-4, IL-5, IL-6 [43], and especially tumor necrosis factor a [13, 44]-a factor known to stimulate metalloproteinase and IL-1 expression by specific target cells [45, 46] including human endometrial stromal cells [47]. Thus MC activation not only has the potential to stimulate the production of matrix-degrading metalloproteinases by neighboring cells but also can facilitate their activation via MC tryptase or chymase. Histamine is a major product of MC activation and is one of several mediators responsible for inducing changes in vasopermeability and edema [48]. Similarly, heparin has numerous effects on specific target cells [49]; its recognized anticoagulant actions may be relevant to endometrial bleeding at menstruation, and its interactions with growth factors and its proposed role in angiogenesis [15] may be important for endometrial vascularization. Thus release of histamine and that of heparin associated with MC degranulation are likely to induce a variety of behavioral responses by neighboring cells in microenvironmental locations, possibly playing a pivotal role in the balance between metabolic and catabolic behavior.

MAST CELLS AND EOSINOPHILS IN HUMAN ENDOMETRIUM

MC activation/degranulation in vivo may be induced by soluble factors derived from a variety of cell types, including T cells, monocyte-macrophages, neutrophils, eosinophils, and platelets [13, 28]. Thus MC activation need not necessarily represent an IgE-mediated "allergic" response in the endometrium, which is the traditional view of MC activation in atopy. The observations of MC degranulation demonstrated in this study may therefore have several explanations, including the possibility of hormonal control previously described in studies of rodent uteri [50-55]. However, since most of the animal uterine studies relate to MC numbers or histamine release in relation to hormonal treatments or at times of implantation, it has proved difficult to compare results from these animal studies with our MC findings throughout the normal human menstrual cycle. Microenvironmental MC activation in vivo is often associated with the influx of inflammatory cells such as neutrophils and eosinophils [141. Such a role is consistent with the appearance of both these cell types in the endometrium at the time of menstruation. Although many aspects of eosinophil function remain to be elucidated, eosinophils are known to elaborate inflammatory mediators that can potentially bring about tissue damage. Such factors include the major basic protein, which is a strong platelet agonist that neutralizes heparin and stimulates MC histamine release; three cationic proteins with multifunctional properties including neurotoxic, cytotoxic, and antiproliferative activities; and the production of platelet-activating factor (PAF), leukotriene C4 (LTC 4), and toxic oxygen metabolites [56]. These eosinophil-derived mediators are known to have profound effects on vasoactivity, macrophage and endothelial cell activation, and histamine release by MCs as well as on the production of proteolytic enzymes such as eosinophil collagenase and gelatinase B [56, 57]. Many factors released by activated MCs may explain the accumulation of eosinophils and neutrophils in the endometrium at menstruation; PAF, LTC4, LTB 4, prostaglandin D2, IL-5, and GM-CSF may all function as chemotactic factors [15]. Whatever the specific MC-eosinophil-neutrophil interactions may be, it seems clear from this histological study that these three granulocytes assume important functional roles in relation to the tissue and vascular remodeling associated with menstruation. Moreover, the apparent relationship between MC activation and tissue edema in mid-proliferative and mid-secretory phases suggests that MC mediators may contribute to normal endometrial physiology, for example by the provision of histamine and various growth factors during the proliferative phase, and may possibly contribute to endometrial preparation for blastocyst implantation. Such ideas require further study, but it seems likely that the observations of MCs and eosinophils reported here for the normal menstrual cycle indicate that these cells probably have major influences on endometrial function and tissue remodeling.

319

ACKNOWLEDGMENTS We thank Anna Butt for excellent technical assistance, Marion Marsh and Fiona Hammond for tissue collection, and Prof. Gabor Kovacs and Dr. Malcolm Bamett for providing tissue from their patients.

REFERENCES 1. Fox H.The normal and abnormal endometrium. In: Fox H,Langley FA (eds.), Postgraduate Obstetrical and Gynaecological Pathology. Oxford: Pergamon Press; 1973: 115145. 2. Finn CA. Implantation, menstruation and inflammation. Biol Rev 1986; 61:313-328. 3. Hunt J. Immunologically relevant cells in the uterus. Biol Reprod 1994; 1994:461-466. 4. Sheppard BL, BonnarJ. Mast cells in the human uterus. In: PepysJ, Edwards AM (eds.), The Mast Cell: Its Role in Health and Disease. Tunbridge Wells: Pitman Medical; 1979: 142-148. 5. CrowJ, More L,Howe S.The mast cells of the human uterus. APMIS 1988; 96:921-926. 6. Crow J, Wilkins M, Howe S,More L,Helliwell P. Mast cells in the female genital tract. IntJ Gynecol Pathol 1991; 10:230-237. 7. Drudy L,Sheppard B, Bonnar J. Mast cells in the normal uterus and in dysfunctional uterine bleeding. EurJ Obstet Gynecol Reprod Biol 1991; 39:1193-1201. 8. Drudy L,Sheppard BL, Bonnar J. Mast cell and histamine concentration of the human post-menopausal uterus. EurJ Obstet Gynecol Reprod Biol 1991; 42:39-42. 9. Drudy L,Sheppard BL, BonnarJ. The ultrastructure of mast cells in the uterus throughout the normal menstrual cycle and the postmenopause. J Anat 1991; 175:51-63. 10. Graham RM. Mast cells and cancer of the cervix. Surg Gynecol Obstet 1966; 123:33-39. 11. Mehra U,Devi PK, Chakravati RTV, Chaundry RP. The relationship between endometrial mast cell count and bleeding in women following insertion of an uterine device. Am J Obstet Gynecol 1970; 107:852-856. 12. Tursi A,Mastrorilli A,Ribatti D, Louidice L,Contino R,Claudatus J. Possible role of mast cells in the mechanism of action of intrauterine contraceptive devices. Am J Obstet Gynecol 1984; 148:1064-1066. 13. Galli SJ. New concepts about the mast cell. New EnglJ Med 1993; 328:257-265. 14. Church MK, Caulfield JP. Mast cell and basophil functions. In: Holgate ST, Church MK (eds.), Allergy. London: Gower Medical Publishing; 1993: 5.1-5.12. 15. Norrby K, Woolley D. Role of mast cells in mitogenesis and angiogenesis in normal tissue and tumour tissue. Adv Biosci 1993; 89:71-116. 16. Irani A-M, Bradford TR, Kepley CL, Schechter NM. Detection of MCT and MCTc types of

human mast cells by immunohistochemistry using new monoclonal anti-tryptase and anti-chymase antibodies. J Histochem Cytochem 1989; 1989:1509-1515.

17. Johansson O, Virtanen M,Hilliges M,Yang Q. Histamine immunohistochemistry is superior to the conventional heparin-based routine staining methodology for investigations of human skin mast cells. Histochem J 1994; 26:424-430.

18. Weidner N,Austen KF. Heterogeneity of mast cells at multiple body sites. Fluorescent determination of avidin binding and immunofluorescent determination of chymase, tryptase, and carboxypeptidase content. Pathol Res Pract 1993; 189:156-162. 19. Noyes RW, Hertig AT, Rock J. Dating the endometrial biopsy. Fertil Steril 1950; 1:3-25. 20. Robinson G, Ellis IO,MacLennan KA. Immunocytochemistry. In: BancroftJD, Stevens A (eds.), Theory and Practice of Histological Techniques. Edinburgh: Churchill Livingstone; 1990: 430-431. 21. Wingren U, Enerback L.Mucosal mast cells of the rat intestine: a reevaluation of fixation

and staining properties with the special reference to protein blocking and solubility of the granular glycosaminoglycan. HistochemJ 1983; 15:571-582. 22. Bromley M, Woolley DE. Histopathology of the rheumatoid lesion. Identification of cell types at sites of cartilage erosion. Arthritis Rheum 1984; 27:857-863. 23. Hourihan HM, Sheppard BL, Belsey EM, Brosens IA. Endometrial vascular features prior to and following exposure to levonorgestrel. Contraception 1991; 43:375-385. 24. Friedman MM, Metcalfe DD, KalinerM. Electron microscopic comparison of human nasal and lung mast cell degranulation. In: Befus AD, Bienstock J, Denburg JA (eds.), Mast Cell Differentiation and Heterogeneity. New York: Raven Press; 1986: 367-377. 25. Massey WA, Guo C-B, Dvorak AM, Hubbard WC, Bhagavan BS, Cohan VL, Warner JA, Kagey-Sobotka A, Lichtenstein LM.Human uterine mast cells. Isolation, purification, characterisation, ultrastructure and pharmacology. J Immunol 1991; 147:1621-1627. 26. Tainsh KR, Lau HYA, Liu WL, Pearce FL. The human skin mast cell: a comparison with the human lung cell and a novel mast cell type, the uterine mast cell. Agents Actions 1991; 33:16-19. 27. Guo C-B, Kagey-Sobotka A, Lichtenstein LM, Bochner BS. Immunophenotyping and functional analysis of purified human uterine mast cells. Blood 1992; 79:708-712. 28. Schwartz LB. Mast cells: function and contents. Curr Opin Immunol 1994; 6:91-97. 29. Gruber BL, Marchese MJ, Suzuki K, Schwartz LB, Okada Y, Nagase H, Ramamurthy NS.

320

30.

31.

32.

33.

34. 35.

36. 37.

38.

39.

40.

41.

42.

JEZIORSKA ET AL.

Synovial procollagenase activation by human mast cell tryptase dependence upon matrix metalloproteinase activation. J Clin Invest 1989; 84:1657-1662. Lees M,Taylor DJ, Wooley DE. Mast cell proteinases activate precursor forms of collagenase and stromelysin, but not of gelatinases A and B. EurJ Biochem 1994; 223:171177. Saarinen J, Kalkinen N, Welgus HG, Kovaaren PT. Activation of human interstitial procollagenase through direct cleavage of the leu 83-thr 84 bond by mast cell chymase. J Biol Chem 1994; 269:18134-18140. Marbaix E, Donnez J, Courtoy PJ, Eeckhout Y. Progesterone regulates the activity of collagenase and related gelatinases A and B in human endometrial explants. Proc Natl Acad Sci USA 1992; 89:11789-11793. Hampton AL, Salamonsen LA. Expression of messenger ribonucleic acid encoding matrix metalloproteinases and their tissue inhibitors is related to menstruation. J Endocrinol 1994; 141:R1-R3. Kislaus U, Herr JC, Little CD. Immunolocalization of extracellular matrix proteins ans collagen synthesis in first trimester human decidua. Anat Rec 1987; 218:402-415. Bulletti C, Galassi A, Jasonni VM, Martinelli G, Tabanelli S, Flamigni C. Basement membrane components in normal hyperplastic and neoplastic endometrium. Cancer 1988; 62:142-149. Aplin JD. Cellular biochemistry of the endometrium. In: Wynn RM, Jollie WP (eds.), Biology of the Uterus, 2nd ed. New York: Plenum Med; 1989: 89-129. Aplin JD, Charlton AK, Ayad S. An immunohistochemical study of human extracellular matrix during the menstrual cycle and first trimester of pregnancy. Cell Tissue Res 1988; 231-240. Schatz F, Papp C, Toth-Pal E, Cudemo V, Hausknecht V, Krikun G, Markiewicz L, Gavi B, Wang E-Y, Feygin N, Papp Z, Lockwood CJ. Protease and protease inhibitor expression during in vitro decidualization of human endometrial stromal cells. Ann NY Acad Sci 1994; 734:33-42. Mylona P, Kieltry CM, Hoyland JA, Aplin JD. Expression of type VI collagen mRNAs in human endometrium during the menstrual cycle and first trimester of pregnancy. J Reprod Fertil 1995; (in press). Kielty CM, Lees M,Shuttleworth CA, Woolley D. Catabolism of the intact type VI collagen microfibrils: susceptibility to degradation by serine proteases. Biochem Biophys Res Common 1993; 191:1230-1236. Yoffe JR, Taylor DJ, Woolley DE. Mast cell products stimulate collagenase and prostaglandin E production by cultures of adherent rheumatoid synovial cells. Biochem Biophys Res Commun 1984; 122:270-276. Yoffe JR, Taylor DJ, Woolley DE. Mast cell products and heparin stimulate the production of mononuclear cell factor by cultured human monocyte-macrophages. BiochemJ 1985; 230:83-88.

43. Bradding P, Feather IH, Wilson S, Bardin PG, Heusser CH, Holgate ST, Howarth PH. The mast cells as a source of IL-4, IL-5 and IL-6 in human allergic mucosal inflammation. J Immunol 1993; 151:3853-3865. 44. Gordon JR, Burd PR, Galli SJ. Mast cells as a source of multifunctional cytokines. Immunol Today 1990; 11:458-464. 45. DayerJM, Beutler B, Cerami A. Cachectin/tumour necrosis factor stimulates collagenase and prostaglandin E2 production by human synovial cells and dermal fibroblasts. J Exp Med 1985; 162:2163-2168. 46. Brennan FM, Maini RN, Feldmann M. TNF--a pivotal role in rheumatoid arthritis? Br J Rheumatol 1992; 31:293-298. 47. Rawdanowicz TJ, Hampton L, Nagase H, Woolley DE, Salamonsen LA. Matrix metalloproteinase production by cultured human endometrial stromal cells: identification of interstitial collagenase, gelatinase-A, gelatinase-B, and stromelysin- and their differential regulation by interleukin- la and tumour necrosis factor-a. J Clin Endocrinol &Metab 1994; 79:530-536. 48. White MV, Kaliner MA. Histamine. In: Gallin JI, Goldstein IM, Snyderman R (eds.), Inflammation; Basic Principles and Clinical Correlates. New York: Raven Press; 1988: 169193. 49. Norrby K.Heparin and angiogenesis: a low-molecular-weight fraction stimulates angi2 ogenesis systematically. Haemostasis 1993; 3(suppl 1):141-149. 50. Brandon J, Bibby MC. A study of changes in uterine mast cells during early pregnancy in the rat. Biol Reprod 1979; 20:977-980. 51. Cocchiara R, Albeggiani G, Di Trapani G, Azzolina A, Lampiasi N, Cervello G, Geraci D. Dispersal of rat uterine mast cells and their functional response to an embryo-derived histamine releasing factor: a possible model for embryo implantation. J Reprod Immunol 1988; 14:191-201. 52. Ionov ID. The mast cell system in reproductive toxicology. Reprod Toxicol 1990; 4:22512255. 53. Mohanty N, Chainy GBN. Effect of estradiol valerate on the uterus of the musk shrew (Suncuss murinusI.). Gen Comp Endocrinol 1992; 88:91-99. 54. Padilla L,Reinicke K, Montesino H, Villena F, Ascencio H, Cruz M, Rudolph MI, Histamine content and mast cell distribution in mouse uterus: the effect of sexual hormones, gestation and labor. Cell Mol Biol 1990; 36:93-100. 55. Spaziani E. Accessory reproductive organs in mammals: control of cell and tissue transport by sex hormones. Pharmacol Rev 1975; 27:207-286. 56. Adolphson CR, Gleich GJ. Eosinophils. In: Holgate ST, Church MK (eds.), Allergy. London: Gower Medical Publishing; 1993: 6.1-6.12. 57. Stahle-Bickdahl M, Parks WC. 92 kDa gelatinase is actively expressed by eosinophils and secreted by neutrophils in invasive squamous cell carcinoma. Am J Pathol 1993; 142:995-1000.