Cellular Mechanisms Involved in Cyclic Stromal Renewal of the Uterus 111. CELLS OF THE IMMUNE RESPONSE''2 HELEN A. PADYKULA Laboratory of Electron Microscopy, Wellesley College, Wellesley, Massachusetts 02181
ABSTRACT The principal cell types associated with the humoral immune response (monocyte-macrophages, lymphocytes, and plasma cells) are numerous in the endometrial stroma of the uterus during the first four postpartum days in two types of mammals, the marsupial North American opossum and the eutherian albino rat. This transient cellular differentiation coincides with the physiologic period of rapid uterine regression which includes massive reduction in the amount of extracellular stromal material. In addition, heterophils and eosinophils, cell types also known to be associated with phagocytic and immunologic activity, appear in the stroma during the first two postpartum days; their presence may, however, be associated more directly with the postpartum estrus that occurs on day 1 postpartum than with endometrial regression. Thus, the five cell types, which are known in pathologic conditions to be components of an inflammatory response to a foreign antigen, are conspicuously present in the normal regressing endometrium. Furthermore, there is ample ultrastructural evidence of frequent macrophagic-lymphocytic interaction, transformation of lymphocytes, and active secretion by plasma cells during this early postpartum period. An hypothesis has been derived by uniting this new description of endometrial stromal cell differentiation with the existing literature on uterine collagenase activity, an important feature of postpartum regression (reviews by Gross, '74; Harris and Krane, '74). It is based on the assumption that during regression the extracellular action of neutral collagenase (and possibly other extracellular proteases) release new antigenic sites in proteins located in the ground substance. In the case of collagenase, these transient antigenic sites would arise a t the locus of enzymic cleavage as well as from the subsequent denaturation of the fragments of the collagen molecule. This endogenous antigenic stimulus would be strong and temporary, and would lead to the cellular manifestations of the transient humoral immunologic response which are evident in the regressing stroma of these two mammals. This humoral immune reaction may be one of the regulatory mechanisms involved in the cyclic renewal of the extracellular compartment of the uterine stroma. The transient accumulation and subsesequent disappearance of macrophages in the postpartum endometrial stroma has received considerable attention (Fluhmann, '28; Deno, '37; Lobe1 and Deane, '62; Banon et al., '64; Smith and Henzl, '69; Brandes and Anton, '69). These morphological and histochemical studies have considered the macrophages in isolation as phagocytes involved in uterine involution. Phagocytic involvement in collagen degradation was indicated by ultrastructural ANAT. IIEc., 184: 49-72.
analysis (Schwarz and Guldner, '67; Parakkal '69, '72). Recent biochemical studies on the enzymic basis of collagen degradation in the postpartum rat uterus provide a better framework for interpreting the probable function of uterine macrophages during regression (Gross and Lapiere, '62; Jeffrey and Gross, '70; Jeffrey et d.,'71; Received Mav 21. '75. AcceDted Aue. - 18.. '15. I This study is dedicated t i the memory of the late Dr. Helen Wendler Deane. zThis work was supported by U. S. Public Health Service Grant HD 01056.
49
50
HELEN A. PADYKULA
Woessner and Ryan, '73; Ryan and Woessner, '74; Koob and Jeffrey, '74). It has been clearly established that the degradation of collagen fibrils is initiated by a neutral collagenase that functions extracellularly ; the resulting polypeptide fragments of the collagen molecule and residual portions of the fibrils may be further degraded within the macrophages (reviews by Gross, '74; Harris and Krane, '74). The uterine macrophages may also be the producers of neutral collagenase, since it has been recently demonstrated that peritoneal macrophages will synthesize this collagenase after stimulation by lymphocytes sensitized by antigen or mitogen (Wahl et al., '75). Another important view of macrophagic function emerges from strong evidence accumulated from various experimental systems that places the macrophage into a position of central involvement in cellular and humoral immunity (review by Silverman, '70). I n these immunologic functions macrophages are closely associated with lymphocytes. The present report describes uterine macrophages as occurring in association with lymphocytes and plasma cells in the postpartum endometrium of the albino rat and North American opossum. The new definitions of macrophagic function, along with the evidence presented here as well a s in the companion studies (Padykula and Taylor, '76; Padykula and Campbell, '76) provide a broader basis for interpreting the possible activities of the uterine macrophage. The presence of plasma cells with lymphocytes and macrophages signals the occurrence of a humoral immune response (Weiss, '72). This suggests that a n immunologic event occurs in the involuting uterus during the normal course of physiologic events. The stimulus for antibody production presumably may arise either from the placental-fetal complex or from a n endogenous uterine antigenic stimulus. The discussion centers on the possibility that new transient antigenic sites are revealed in extracellular macromolecular components, such as collagen fibrils, during the enzymic degradation and denaturation that occurs during regression. In addition, morphologic evidence is reviewed which suggests that immunologic informa-
tion is cyclically eliminated from the uterus through physical segregation of macrophages and lymphocytes and then emigration across the uterine epithelia and/or into lymphatic vessels (also Padykula and Taylor, '76; Padykula and Campbell, '76). MATERIALS AND METHODS
This cytological analysis was derived from semi-thin and ultrathin sections of postpartum uten of adult opossums and albino rats (CD strain). The animals as well as the procedures for light and electron microscopic study are described in the companion papers (Padykula and Taylor, '76; Padykula and Campbell, '76). RESULTS
During the first three postpartum days of the opossum and rat, numerous monocyte-macrophages and lymphocytes are present i n the endometrial stroma. Transforming lymphocytes are also present. Furthermore, plasma cells are common, although they appear to be fewer in number than the lymphocytes and macrophages. Heterophils are present particularly in the superficial stroma; they are conspicuous during day 1 postpartum but decline steadily in number thereafter. These various transient cells were observed both in the opossum and albino rat. In addition, eosinophils are abundant in the deep stroma of rat endometrium during day 1 postpartum but decline steadily in number thereafter. The illustrations presented here are taken entirely from the opossum because this species was studied more intensively. Also a n unusual process of macrophagic-lymphocytic isolation was conspicuous i n the opossum endometrial stroma, which may be important in eventual interpretation of the apparent immune response in the involuting stroma. The presence of intermediary forms of phagocytic cells suggests that tissue monocytes are being converted into full-blown macrophages through progressive enlargement of cell size and of the phagolysosomd system. This morphological evidence indicates that the postpartum flock of endometrial macrophages may originate from blood monocytes, as has been experimentally demonstrated in other systems (Volkmann and Gowans, '65A,B). It is generally assumed that during the first four post-
IMMUNE RESPONSE I N REGRESSING ENDOMETRIUM
partum days these macrophages are engaged in the removal of the extracellular macromolecules located in the endometrial stroma. These materials are chiefly collagen fibrils and the macromolecular constituents of the ground substance. Figure 3 illustrates the intracellular occurrence of collagen fibrils within phagosomes (also Schwarz and Guldner, '67; Parakkal '69, '72). At this time, the macrophages are frequently observed in close association and alignment with lymphocytes (figs. 1-5, 7 ) . The apposed surfaces of these two cell types are frequently flattened (fig. 2). This lymphocytic-macrophagic association may also involve specialization of the plasma membranes at certain sites, particularly of the lymphocyte (figs. 4A, 4B); whether this particular junctional specialization represents attachment sites or intercellular communication is unknown. However, the structural features suggest this association may be comparable to the attachment points between neurons at the synapse (Peters et al., '70). Transforming lymphocytes (Biberfeld, '71 ) with conspicuous cisternae of rough endoplasmic reticulum, distributed more or less singly in a cytoplasmic matrix rich in ribosomes and polysomes, are present (fig. 5). The occurrence of plasma cells manifesting various stages in the synthetic state (figs. 1, 2 , 6) suggests active secretion of antibody. Thus these cytological appearances are comparable to other systems in which foreign antigen is processed by macrophages that interact with lymphocytes to trigger antibody production by plasma cells. A problem of interpretation, however, presents itself here because involution is a normal process, and thus the antigen may originate from postpartum modification of the macromolecules normally present in the endometrial stroma rather than from a foreign source. The interaction of lymphocytes and macrophages is of immunologic significance and thus it is of considerable interest to find that these cells in the opossum endometrium may acquire at some point during this postpartum differentiation an inner extracellular coat as well as an outer cellular vesture that isolates them from the rest of the stroma (figs, 7, 8). The morphologi-
51
cal appearance suggests that the macrophage is immobilized within the enclosure since its surface is smooth except at several locations where uncoated cytoplasmic processes extend through the extracellular coat (figs. 7, 8). However, it is also possible that the morphologic relationships in figures 7 and 8 (and in figs. 10, 11, Padykula and Taylor, '76) may instead represent stages in the enclosure of macrophages and lymphocytes within the basal laminae of the uterine epithelia, Analysis of serial sections would be required to define this complex relationship further. The isolation of a lymphocyte and a macrophage as a pair (fig. 8 ) was quite often observed (also fig. 9 and text figure 1 in Padykula and Taylor, '76). This remarkable niorphological event may represent isolation of immunologic information. Furthermore, a companion investigation (Padykula and Taylor, '76) demonstrates that macrophages and lymphocytes are eliminated from the postpartum stroma of the opossum by transepithelial emigration to the lumen of the uterine glands and to the uterine cavity (text fig. 1 in Padykula and Taylor '76). I n the albino rat the lymphatic vessels offer a n additional mode of egress (Padykula and Campbell, '76). This cyclic appearance of lymphocytes and macrophages, their physical association free in the stroma as well as in the isolation chamber illustrated in figures 7 and 8, may reflect a n immune mechanism which is differentiated for the purpose of eliminating extracellular stromal macromolecules in a cyclic manner. This possibility is discussed below. DISCUSSION
The uteriis and its capacity to respond immunologically The embryo-fetus with its placental membranes may be viewed as a successful allograft on the uterine endometrium. Current evidence suggests that this success is derived more from the low antigenicity of the trophoblast than from inability of the endometrium to respond immunologically (Beer and Billingham, '71, Beer et al., '75). It has long been known that the injection of vital dyes or carbon particles will result in the appearance of endome-
52
HELEN A. PADYKULA
trial macrophages, especially during estrogenic influence (Fluhmann, ’28; Nicol and Vernon-Roberts, ’65; Vernon-Roberts, ’69). Experimental evidence indicates that skin allografts to the rat uterus cause a 3-fold increase in weight of the draining lymph nodes whereas autografts do not (Beer and Billingham, ’71 ). Hypertrophy of draining lymph nodes occurs also during hetercspecific pregnancies (Beer and Billingham, ’71). Recently Beer et al. (’75) in a study of interstrain (allogeneic) matings demonstrated that the maternal response against fetal alloantigens leads to heavier placental-fetal units than after syngeneic matings. Surgical removal of the lymph nodes (“para-aortic nodes,” according to the authors) that drain the uterine horns interferes with this enlargement of the placental-fetal units, a n observation which suggests that humoral immunity is involved. This varied experimental evidence along with the present report of the coexistence of macrophages, lymphocytes, and plasma cells in the postpartum endometrium, indicates that the uterus possesses the capacity to respond immunologically. The distribution of the lymphatic vessels in the uterus may contribute to its ability to accept grafts, since in the rabbit (McLean and Scothorne, ’70), and most likely in the rat, lymphatic vessels are absent from the endometrium. They occur primarily as a plexus located between the two muscle layers of the myometrium (Hoggan and Hoggan, 1881). In these species, lymphatic vessels occur also at the endometrial-myornetrial junction and in the serosa. This absence of endometrial lymphatic vessels may be the basis of the so-called “privileged site” status of the uterus in these species, since Barker and Billingham (’68), by experimentally eliminating lymphatic drainage from a region in other organs, created “privileged sites” which do not reject grafts. It should be added that the amount of lymphatic drainage in the endometrium may be related, a s a consequence of evolutionary process, to the degree of penetration by the placenta into the endometrium (Hoggan and Hoggan, 1881). The relationship of this special spatial arrangement of uterine lymphatic vessels to the immunologic activity
of the endometrium must await further investigation.
Possible significance of the occurrence in the regressing endometrium of the cells associated with the humoral immune response Heterophils (neutrophils) invade the regressing endometrium early in the f i s t day after delivery. They are conspicuous in the superficial endometrium during day 1 and decrease in number during day 2. Although this cell type in abnormal states usually represents the first stage of inflammation, the heterophils in the normal regressing endometrium seem to be programmed into the events of estrus (Yanigamachi and Chang, ’67) whether estrus occurs postpartum or during a cycle. Our histological observations on day 1 postpartum in the rat and opossum (Padykula and Campbell, ’76) indicate that heterophils migrate from superficial endometrial blood vessels, pass through the superficial stroma, and migrate across the luminal epithelium into the uterine cavity. Here they function in the phagocytosis of sperm (Austin ‘57; Reid, ’64). Eosinophils also appear transiently at the time of the postpartum estrus: since this immigration is known to reflect rising estrogen levels (Ross and Klebanoff, ’SS), it may also be an event in the endometrial cellular program of estrus. Thus the transient appearance of heterophils and eosinophils may not be directly related to the subsequent accumulation of macrophages, lymphocytes, and plasma cells in the postpartum uterus. However, the possibility of the interrelationship of all these cell types should remain open since these two granulocytes are known to be cellular components of inflammatory processes. The occurrence of plasma cells in concert with lymphocytes and macrophages indicates that a humoral immune response is taking place (Weiss, ’72). Thus it is likely that antibody production is occurring in the regressing endometrium. This antibody (or antibodies) might be released into the circulation as IgG or into the uterine cavity as IgA. The presence of plasma cells indicates that B lymphocytes have been present. The frequent apposition of the surfaces of lymphocytes and macrophages
IMMUNE RESPONSE I N REGRESSING ENDOMETRIUM
suggests cellular interaction or communication, as occurs in the presence of foreign antigen (e.g., Nielsen et al.,'74). The presence of transforming lymphocytes signals the possible activation of lymphocytes to become blast cells capable of proliferation (e.g., Biberfeld, '71). The nature of the antigenic material that stimulates the appearance of this transient postpartum cellular representation of a humoral immune response must next be considered. There are three possible sources of foreign antigens: (1) sperm if the female was mated during the day 1 postpartum estrus; ( 2 ) the placental-fetal complex that was intimately apposed and/or fused to portions of the endometrium until parturition; and ( 3 ) normal macromolecules of the stroma, such as collagen, which are altered during extracelluIar enzymic degradation in a manner that transiently reveals new antigenic sites. Since our postpartum opossums and rats were isolated from males, the antigenic source cannot be spermatozoa. The natural allograft represented by the placental-fetal complex is a possibility as an antigenic source, especially in the opossum, a wild animal in which genetic strain variation is more apt to occur. However, at diestrus of a non-pregnant cycle in the opossum, influx of macrophages, lymphocytes, and plasma cells occurs in a degree comparable to that occurring after parturition. This similarity in the cellular manifestations of regression in the non-pregnant and pregnant cycles, to a considerable degree, lessens the possibility of antigenic stimulation by the marsupial feto-placental complex. It should be mentioned that this similarity is a special situation because endometrial expansion during the luteal phase in the opossum is comparable whether or not pregnancy occurs; this led early investigators to designate the luteal phase of a non-pregnant cycle as "pseudopregnancy" (Hartman, '23). Genetic variation due to strain differences in the albino rat is less probable than in the opossum because we used highly inbred members of the Charles River CD strain. As an example of the macromoleci~lar alterations that may occur during uterine involution, it is appropriate to consider the role of neutral collagenase in the produc-
53
tion of new antigenic sites. The purified enzyme attacks the collagen molecule in its polymeric position within the insoluble collagen fibril and cleaves through the triple helical conformation at a specific site to produce two unequal fragments (usually one 75% fragment and the other a 25% piece of the collagen molecule) (review by Gross, '74). In the case of rat uterine collagenase, further cleavage of two small peptide fragments occurs in vitro; however, it is not known whether or not this additional enzymic action arises from contamination (Jeffrey and Gross, '70). Nevertheless, the cleaved collagen molecule would present new antigenic sites in an extracellular maternal location. Furthermore, according to proposed mechanisms for collagenolytic degradation of insoluble fibrils (Harris and Krane, '74: fig. 5, p. 606; Gross, '74: fig. 18, p. 410), the two fragments of the collagen molecule would move away from the fibril into the ground substance where they would most likely denature at body temperature. Thus denaturation would reveal antigenic sites additional to those arising from the cleaved triple helix. Since collagen degradation proceeds very rapidly (Harkness and Moralee, '56; Ryan and Woessner, '74), cleavage by collagenase and the subsequent denaturation would constitute a powerful transient local antigenic stimulus to the maternal organism. The solubilized denatured collagen fragments might be further degraded by extracellular proteases or phagocytosed by uterine macrophages (Hams and Krane, '74; Gross, '74) to be digested within phagosomes by cathepsin B, (Burleigh et al., '74). To account for Schwarz and Guldner's ('67) and Parakkal's ('69, '72) demonstrations of intact collagen fibrils within phagosomes of uterine phagocytes, Harris and Krane ('74) suggested in diagrammatic representation (their fig. 7, p. 607) that multimolecular fragments of collagen fibrils may be released by collagenase cleavage and ingested by macrophages for further digestion within phagosomes. Thus it seems likely that the macrophage is involved in the terminal degradation of collagen fibrils as well as of other extracellular macromolecules. Furthermore, the possibility exists that neutral collagenase
54
HELEN A. PADYKULA
may be secreted by the macrophages (Wahl et al., '75), although the possibility of secretory involvement of fibroblasts (Werb and Burleigh, '74) or epithelia (Eisen and Gross, '65) should remain open. Hypothesis involving a humoral immune reaction in the regulation of cyclic renewal of extracellular macromolecules in the stroma of the normal endometrium The hormonal environment during the first few postpartum days favors the emigration of heterophils, monocytes, eosinophils, and lymphocytes from the blood into the endometrial stroma. In addition, this milieu favors the synthesis and release of neutral collagenase (and possibly other enzymes capable of extracellular cleavage of macromolecules) into the ground substance of the stroma. The polymeric collagen fibril is dismantled initially by collagenolytic cleavage which releases fragments of monomeric molecules as well as larger cross-striated pieces of the collagen fibril. Numerous antigenic sites are transiently revealed as a result of rapid enzymic cleavage of the collagen triple helix and the subsequent denaturation of fragments in the ground substance. The various fragments of the enzymically cleaved collagen fibrils are ingested by monocytemacrophages that respond to them as unnatural proteins or polypeptides and process them as antigens. At this time, macrophagic-lymphocytic interaction is conspicuous, as the two cell surfaces become closely apposed and may even acquire special differentiations. B lymphocytes differentiate into plasma cells to produce antibody to the collagen-derived antigens. This antibody would most likely enter the vascular and lymphatic vessels and thus be recognizable systemically. At the close of this episode, lymphocytes, carrying immunologic information, and macrophages, engorged with the remnants of extracellular stromal macromolecules, have been largely eliminated from the endometrium by transepithelial emigration into the uterine cavity, glandular lumens, or into lymphatic vessels. Thus, immunologic information contained in lymphocytes as well as extracellular debris sequestered in macrophages, both materials related to the pre-
vious pregnancy or cycle, have been largely removed from the uterine stroma, and a new ovarian-uterine cycle proceeds in which stromal macromolecules are again synthesized and secreted into the extracellular compartment of the uterus. The cellular mechanism proposed in this hypothesis suggests involvement of an autoimmune response in the normal removal of extracellular stromal macromolecules. This hypothesis may be applicable to the interpretation of the so-called collagen vascular diseases, such as rheumatoid arthritis, in which neutral collagenase is known to be released and aggregations of lymphocytes, macrophages, and plasma cells occur at the sites of lesions. ACKNOWLEDGMENTS
Richard V. T. Stearns prepared the p h e tomicrographs and the final electronmicrographs. The author assumes full responsibility for the interpretations presented here but gratefully acknowledges generous discussion of the observations with Dr. Leon P. Weiss, Dr. Myron S . Silverman, Dr. Miles Crenshaw, Dr. Joachim Sommers, and Dr. Edward H. Bossen. LITERATURE CITED Austin, C. R. 1957 Fate of spermatozoa in the uterus of the mouse and rat. J. Endocrinol., 14: 335-342. Banon, P., D. Brandes and J. K. Frost 1964 Lysosomal enzymes in the rat ovary and endometrium during the estrous cycle. Acta Cytol., 8: 4 1 6 4 2 5 . Barker, C. F., and R. E. Billingham 1968 The role of afferent lymphatics in the rejection of skin homografts. J. Exp. Med., 128: 197-221. Beer, A. E., and R. E. Billingham 1972 Immunobiology of mammalian reproduction. In: Advances in Immunology. Vol. 14. R. J. Dixon, Jr. and H. G. Kunkel, eds. Academic Press, New York, pp. 1-84. Beer, A. E., J. R. Scott and R. E. Billingham 1975 Histoincompatibility and maternal immunological status as determinants of f e t e placental weight and litter size in rodents. J. Exp. Med., 142: 180-196. Biberfeld, P. 1971 Morphogenesis in blood lymphocytes stimulated with phytohaemoagglutinin (PHA). Acta Path. Microbiol., Scand. Section A, Supplement No. 223. 3This hypothesis has been derived from the observations reported here and in the two accompany-
ing studies (Padykula and Taylor, '76; Padykula and
Campbell, '76).
IMMUNE RESPONSE IN REGRESSING ENDOMETRIUM B r a n d s , D., and E. Anton 1969 A n electron microscopic cytochemical study of macrophages during uterine involution. J. Cell Biol., 41: 450-461. Burleigh, M. C., A. J. Barrett and G. S. Lazarus 1974 Cathepsin B1. A lysosomal enzyme that degrades native collagen. Biochem. J., 137: 387-398. Deno, R. A. 1937 Uterine macrophages in the mouse and their relation to involution. Am. J. Anat., 60: 433-469. Eisen, A. Z., and J. Gross 1965 The role Of epithelium and mesenchyme in the production of a collagenolytic enzyme and a hyaluronidase in the anuran tadpole. Dev. Biol., 12: 408418. Fluhmann, C. F. 1928 The reticubendothelid cells of the uterus: A n experimental study. Am. J. Obst. Gynec., 15: 783-796. Gross, J. 1974 Collagen biology: structure, degradation, and disease. In: The Harvey Lectures, 1972-1973, Academic Press, New York, pp. 351-432. Gross, J., and C. M. Lapiere 1962 Collagenolytic activity in amphibian tissues; a tissue culture assay. Proc. Nat. Acad. Sci. (U.S.A.), 48: 1014-1022. Harkness, R. D., and B. E. Moralee 1956 The time course and route of loss of collagen from the rat's uterus during postpartum involution. J. Physiol., 132: 502-508. Harris, E. D., Jr., and S. M. Krane 1974 Collagenase. New Eng. J. Med. (Part l), 291: 557-563; (Part 2), 291: 605-609; (Part 3), 201: 652-661. Hartman, C. 1923 The oestrous cycle in the opossum, Didelphis virginiana. Am. J. Anat., 32: 353-421. Hoggan, G., and F. E. Hoggan 1881 On the comparative anatomy of the lymphatics of the uterus. J. Anat., 16: 50-89. Jeffrey, J. J., R. J. Coffey and A. Z. Eisen 1971 Studies on uterine collagenase. I. Relationship of enzyme production to collagen metabolism. Biochim. Biophys. Acta, 252: 136-142. Jeffrey, J. J., and J. Gross 1970 Collagenase from the rat uterus. Isolation and partial characterization. Biochemistry, 9: 268-274. Jeffrey, J. J., and T. J. Koob 1973 Hormonal regulation of collagen catabolism in the uterus. Excerpta Med.: Endocrinology, Int. Congr. Series, 273: 1115-1123. Koob, T. J., and J. J. Jeffrey 1974 Hormonal regulation of collagen degradation in the uterus: Inhibition of collagenase expression by progesterone and cycIic AMP. Biochim. Biophys. Acta, 354: 61-70. Lobel, B., and H. W. Deane 1962 Enzymic activity associated with postpartum involution of the uterus and with its regression after hormone withdrawal in the rat. Endocrinology, 70: 567-578. McLean, J. M., and R. J. Scothorne 1970 The lymphatics of the endometrium in the rabbit. J. Anat., 107: 39-48.
55
Nicol, T., and 8 . Vernon-Roberts 1965 The influence of the estrous cycle, pregnancy, and ovariectomy in RES activity. J. Reticuloendothel. SOC.,2: 15-29. Nielsen, M. H., H. Jensen, 0. Braendstrup and 0. Werdelin 1974 Macrophagelymphocyte clusters in the immune response to soluble protein antigen in vitro. 11. Ultrastmcture of clusters formed during the early response. J. Exp. Med., 140: 1260-1272. Padykula, H. A., and A. Campbell 1976 Cellu&ar mechanisms involved in cyclic stromal renewal of the uterus. 11. The albino rat. Anat. Rec., 184: 27-48. Padykula, H. A., and J. M. Taylor 1976 Cellular mechanisms involved in cyclic stromal renewal of the uterus. 1. The North American Opossum. Anat. Rec., 184: 5-26. Parakkal, P. F. 1969 Involvement of macrophages in collagen resorption. J. Cell Biol., 41: 345-354. 1972 Macrophages: The time course and sequence of their distribution in the postpartum uterus. J. Ultrastr. Res., 40: 284-291. Peters, A., S . L.Palay and H. deF. Webster 1970 The Fine Structure of the Nervous System. Hoeber, Med. Div., Harper & Row, New York. Reid, B. L. 1964 Fate of residual sperm in the mouse uterus. J . Anat., 98: 492 (abstract). Ross, R., and S. J. Klebanoff 1966 The eosinophilic leukocyte. Fine structure studies of changes i n the uterus during the estrous cycle. J. Exp. Med., 124: 653-660. Ryan, J. N., and J. F. Woessner, Jr. 1974 Oestradiol inhibition of collagenase role in uterine involution. Nature, 248: 526-528. Schwarz, W., and R. H. Giildner 1967 Elektronenmikroskopische Untersuchungen der Kollagenabbaus im Uterus der Ratte nmach der Schwangerschaft. Z. Zellforsch., 83: 416-426. Silverman, M. S. 1970 The macrophage in c d lular and humoral immunity. J. Reticuloendo. SOC.,8: 105-123. Smith, R. E., and M. R. Henzl 1969 Role of mucopolysaccharides and lysosomal hydrolases i n endometrial regression following withdrawal of esttadiol and chlormadinone acetate. I. Epithelium and stroma. Endocrinology, 85: 50-66. Vernon-Roberts, B. 1969 The effects of steroid hormones on macrophagic activity. Int. Rev. Cytol., 25: 131-159. Volkmann, A., and J. L. Gowans 1965A The production of macrophages in the rat. Brit. J. Exp. Path., 46: 50-61. 1965B The origin of macrophages from bone marrow in the rat. Brit. J. Exp. Path., 46: 62-70. Wahl, L. M., S. M. Wahl, S. E. Mergenhagen and G. R. Martin 1975 Collagenase production by lymphokine-activated macrophages. Science, 187: 261-263. Weis% L. p. 1972 The Cells and Tissues of the
56
HELEN A. PADYKULA
Immune System. Prentice-Hall, Inc., Englewood, N. J. Werb, Z., and M. C. Burleigh 1974 A specific collagenase from rabbit fibroblasts in monolayer culture. Biochem. J., 137: 373-376. Woessner, J. F., Jr., and J. N. Ryan 1973 Collagenase activity in homogenates of the in-
voluting rat uterus. Biochim. Biophys. Acta, 309: 397-405. Yanigamachi, R., and M. C. Chang 1963 Infiltration of leucocytes into the uterine lumen of the golden hamster during the estrous cycle and following mating. J. Reprod. Fert., 5 : 389-396.
All of these illustrations were obtained from preparations of opossum endometrium o n days 1 and 2 postpartum.
PLATE 1 EXPLANATION OF FIGURE
1
Endometrial glands with intervening stroma, semi-thin section, toluidine blue. x 1,950. Macrophages ( M ) , lymphocytes ( L ) , and a plasma cell ( P ) occur in the stroma that intervenes between two glands ( G ) . Cell division (arrows) is evident in the glandular epithelium. In two locations ( * ), lymphocytes are closely apposed to the surface of macrophages (only a portion of the cytoplasm of these macrophages is included here). The ultrastructure of the boxed area is shown in figure 2.
IMMUNE RESPONSE IN REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 1
57
PLATE 2 EXPLANATION O F FIGURE
2
58
Electron micrograph of a portion of the field shown i n figure 1. X 7,500. The plasma cell ( P ) is in active synthetic state: cisternae of rough endoplasmic reticulum are rounded and filled with secretory product. A lymphocyte ( L ) and a macrophage ( M ) are flattened and in close alignment along the region of their apposition (arrows).
IMMUNE RESPONSE IN REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 2
59
PLATE 3 EXPLANATION OF FIGURE
3
60
Phasocyte and lymphocyte. x 25,000. This phagocyte ( M ) conta’ns phagosomes ( P ) that contain fibrils ( F ) which appear to be collagen. Note also the specialized sites (arrows) along the phagocytic cell surface. The cytoplasm of the lymphocyte ( L ) is rich i n ribosomes and has relatively little rough endoplasmic reticulum. The extracellular compartment contains much ground substance ( G S ) and relatively few collagen fibrils (cf).
IMMUNE RESPONSE I N REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 3
61
PLATE 4 EXPLANATION O F FIGURES
4A,B
62
Specialization in the lymphocytic-macrophagic association. A, X 28,700; B, 84,500. In figure 4A, two cytoplasmic processes (CP) extend from the macrophage (MAC), one of which is closely associated with the surface of the lymphocyte (LYM). These processes contain numerous membrane-limited, electron opaque inclusions of various shapes. In one area (arrow), the apposed surfaces show a junctional specialization, which is enlarged in figure 4B. At the differentiated site, the plasma membranes of the lymphocyte and macrophage remain separate, but show junctional specialization in the form of greater cytoplasmic density ( d ) along the inner surface of the lymphocytic plasma membrane.
IMMUNE RESPONSE I N REGRESSING ENDOMETRIUM
PLATE 4
Helen A. Padykula
63
PLATE 5 EXPLANATION OF FIGURE
5
64
Transforming lymphocyte. x 27,500. Single cisternae of rough endoplasmic reticulum ( R E R ) are conspicuous; some cisternae are arranged at right angles to the surface of the cell (arrows). Note the prevalence of polysomes ( p ) and single ribosomes in the cytoplasmic matrix. G S , ground substance; cf, collagen fibrils.
IMMUNE RESPONSE I N REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 5
65
PLATE 6 EXPLANATION O F FIGURE
6
66
Endometrial plasma cell. x 12,500. Active synthetic state is reflected by elaborate oriented cisternae of rough endoplasmic reticulum (RER) which are filled with secretory product as well as by presence of the large Golgi apparatus. Packaging of secretory product appears to be occurring along the inner or concave surface of the Golgi (G) apparatus ( * ) . Coated vesicles (arrows) are numerous in the Golgi region. A n aggregate of smooth membraned tubules ( S M ) is present.
IMMUNE RESPONSE IN REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 6
67
PLATE 7 EXPLANATION O F FIGURE
7
68
Enclosure of a macrophage i n the subepithelial stroma. x 7,500. The enclosure of macrophages ( M ) within an inner extracellular coat (arrowheads) and an outer fibroblast-like cellular vesture ( X ) is a regular feature of the stromal differentiation. The extracellular coat is discontinuous a t points where macrophagic pseudopodial processes ( m p ) of irregular shape extend free into the ground substance of the stroma. The processes of the fibroblast-like cell ( x p ) enclose the coated macrophage as well as some uncoated pseudopodial processes (mp). Note that the processes of the investing cell are very thin in some regions. This process of enclosure is most frequently observed in regions close to the epithelial basal lamina (EL) which is slack. Another example of lymphocytic-macrophagic apposition is evident a t the bottom of this figure (*'). E, epithelium.
lMMUNE RESPONSE IN REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 7
69
PLATE 8 EXPLANATION OF FIGURE
8
70
Enclosure of a lymphocyte and a macrophage in the stroma. x 13,000. A lymphocyte (LYM) and a macrophage (MAC) are isolated together within an extracellular coat ( e c ) , except for an uncoated macrophagic process (mp). The processes of a fibroblast-like cell ( X ) follow the contours of this lymphocytic-macrophagic complex.
IMMUNE RESPONSE IN REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 8
71
ABSTRACT The principal cell types associated with the humoral immune response (monocyte-macrophages, lymphocytes, and plasma cells) are numerous in the endometrial stroma of the uterus during the first four postpartum days in two types of mammals, the marsupial North American opossum and the eutherian albino rat. This transient cellular differentiation coincides with the physiologic period of rapid uterine regression which includes massive reduction in the amount of extracellular stromal material. In addition, heterophils and eosinophils, cell types also known to be associated with phagocytic and immunologic activity, appear in the stroma during the first two postpartum days; their presence may, however, be associated more directly with the postpartum estrus that occurs on day 1 postpartum than with endometrial regression. Thus, the five cell types, which are known in pathologic conditions to be components of an inflammatory response to a foreign antigen, are conspicuously present in the normal regressing endometrium. Furthermore, there is ample ultrastructural evidence of frequent macrophagic-lymphocytic interaction, transformation of lymphocytes, and active secretion by plasma cells during this early postpartum period. An hypothesis has been derived by uniting this new description of endometrial stromal cell differentiation with the existing literature on uterine collagenase activity, an important feature of postpartum regression (reviews by Gross, '74; Harris and Krane, '74). It is based on the assumption that during regression the extracellular action of neutral collagenase (and possibly other extracellular proteases) release new antigenic sites in proteins located in the ground substance. In the case of collagenase, these transient antigenic sites would arise a t the locus of enzymic cleavage as well as from the subsequent denaturation of the fragments of the collagen molecule. This endogenous antigenic stimulus would be strong and temporary, and would lead to the cellular manifestations of the transient humoral immunologic response which are evident in the regressing stroma of these two mammals. This humoral immune reaction may be one of the regulatory mechanisms involved in the cyclic renewal of the extracellular compartment of the uterine stroma. The transient accumulation and subsesequent disappearance of macrophages in the postpartum endometrial stroma has received considerable attention (Fluhmann, '28; Deno, '37; Lobe1 and Deane, '62; Banon et al., '64; Smith and Henzl, '69; Brandes and Anton, '69). These morphological and histochemical studies have considered the macrophages in isolation as phagocytes involved in uterine involution. Phagocytic involvement in collagen degradation was indicated by ultrastructural ANAT. IIEc., 184: 49-72.
analysis (Schwarz and Guldner, '67; Parakkal '69, '72). Recent biochemical studies on the enzymic basis of collagen degradation in the postpartum rat uterus provide a better framework for interpreting the probable function of uterine macrophages during regression (Gross and Lapiere, '62; Jeffrey and Gross, '70; Jeffrey et d.,'71; Received Mav 21. '75. AcceDted Aue. - 18.. '15. I This study is dedicated t i the memory of the late Dr. Helen Wendler Deane. zThis work was supported by U. S. Public Health Service Grant HD 01056.
49
50
HELEN A. PADYKULA
Woessner and Ryan, '73; Ryan and Woessner, '74; Koob and Jeffrey, '74). It has been clearly established that the degradation of collagen fibrils is initiated by a neutral collagenase that functions extracellularly ; the resulting polypeptide fragments of the collagen molecule and residual portions of the fibrils may be further degraded within the macrophages (reviews by Gross, '74; Harris and Krane, '74). The uterine macrophages may also be the producers of neutral collagenase, since it has been recently demonstrated that peritoneal macrophages will synthesize this collagenase after stimulation by lymphocytes sensitized by antigen or mitogen (Wahl et al., '75). Another important view of macrophagic function emerges from strong evidence accumulated from various experimental systems that places the macrophage into a position of central involvement in cellular and humoral immunity (review by Silverman, '70). I n these immunologic functions macrophages are closely associated with lymphocytes. The present report describes uterine macrophages as occurring in association with lymphocytes and plasma cells in the postpartum endometrium of the albino rat and North American opossum. The new definitions of macrophagic function, along with the evidence presented here as well a s in the companion studies (Padykula and Taylor, '76; Padykula and Campbell, '76) provide a broader basis for interpreting the possible activities of the uterine macrophage. The presence of plasma cells with lymphocytes and macrophages signals the occurrence of a humoral immune response (Weiss, '72). This suggests that a n immunologic event occurs in the involuting uterus during the normal course of physiologic events. The stimulus for antibody production presumably may arise either from the placental-fetal complex or from a n endogenous uterine antigenic stimulus. The discussion centers on the possibility that new transient antigenic sites are revealed in extracellular macromolecular components, such as collagen fibrils, during the enzymic degradation and denaturation that occurs during regression. In addition, morphologic evidence is reviewed which suggests that immunologic informa-
tion is cyclically eliminated from the uterus through physical segregation of macrophages and lymphocytes and then emigration across the uterine epithelia and/or into lymphatic vessels (also Padykula and Taylor, '76; Padykula and Campbell, '76). MATERIALS AND METHODS
This cytological analysis was derived from semi-thin and ultrathin sections of postpartum uten of adult opossums and albino rats (CD strain). The animals as well as the procedures for light and electron microscopic study are described in the companion papers (Padykula and Taylor, '76; Padykula and Campbell, '76). RESULTS
During the first three postpartum days of the opossum and rat, numerous monocyte-macrophages and lymphocytes are present i n the endometrial stroma. Transforming lymphocytes are also present. Furthermore, plasma cells are common, although they appear to be fewer in number than the lymphocytes and macrophages. Heterophils are present particularly in the superficial stroma; they are conspicuous during day 1 postpartum but decline steadily in number thereafter. These various transient cells were observed both in the opossum and albino rat. In addition, eosinophils are abundant in the deep stroma of rat endometrium during day 1 postpartum but decline steadily in number thereafter. The illustrations presented here are taken entirely from the opossum because this species was studied more intensively. Also a n unusual process of macrophagic-lymphocytic isolation was conspicuous i n the opossum endometrial stroma, which may be important in eventual interpretation of the apparent immune response in the involuting stroma. The presence of intermediary forms of phagocytic cells suggests that tissue monocytes are being converted into full-blown macrophages through progressive enlargement of cell size and of the phagolysosomd system. This morphological evidence indicates that the postpartum flock of endometrial macrophages may originate from blood monocytes, as has been experimentally demonstrated in other systems (Volkmann and Gowans, '65A,B). It is generally assumed that during the first four post-
IMMUNE RESPONSE I N REGRESSING ENDOMETRIUM
partum days these macrophages are engaged in the removal of the extracellular macromolecules located in the endometrial stroma. These materials are chiefly collagen fibrils and the macromolecular constituents of the ground substance. Figure 3 illustrates the intracellular occurrence of collagen fibrils within phagosomes (also Schwarz and Guldner, '67; Parakkal '69, '72). At this time, the macrophages are frequently observed in close association and alignment with lymphocytes (figs. 1-5, 7 ) . The apposed surfaces of these two cell types are frequently flattened (fig. 2). This lymphocytic-macrophagic association may also involve specialization of the plasma membranes at certain sites, particularly of the lymphocyte (figs. 4A, 4B); whether this particular junctional specialization represents attachment sites or intercellular communication is unknown. However, the structural features suggest this association may be comparable to the attachment points between neurons at the synapse (Peters et al., '70). Transforming lymphocytes (Biberfeld, '71 ) with conspicuous cisternae of rough endoplasmic reticulum, distributed more or less singly in a cytoplasmic matrix rich in ribosomes and polysomes, are present (fig. 5). The occurrence of plasma cells manifesting various stages in the synthetic state (figs. 1, 2 , 6) suggests active secretion of antibody. Thus these cytological appearances are comparable to other systems in which foreign antigen is processed by macrophages that interact with lymphocytes to trigger antibody production by plasma cells. A problem of interpretation, however, presents itself here because involution is a normal process, and thus the antigen may originate from postpartum modification of the macromolecules normally present in the endometrial stroma rather than from a foreign source. The interaction of lymphocytes and macrophages is of immunologic significance and thus it is of considerable interest to find that these cells in the opossum endometrium may acquire at some point during this postpartum differentiation an inner extracellular coat as well as an outer cellular vesture that isolates them from the rest of the stroma (figs, 7, 8). The morphologi-
51
cal appearance suggests that the macrophage is immobilized within the enclosure since its surface is smooth except at several locations where uncoated cytoplasmic processes extend through the extracellular coat (figs. 7, 8). However, it is also possible that the morphologic relationships in figures 7 and 8 (and in figs. 10, 11, Padykula and Taylor, '76) may instead represent stages in the enclosure of macrophages and lymphocytes within the basal laminae of the uterine epithelia, Analysis of serial sections would be required to define this complex relationship further. The isolation of a lymphocyte and a macrophage as a pair (fig. 8 ) was quite often observed (also fig. 9 and text figure 1 in Padykula and Taylor, '76). This remarkable niorphological event may represent isolation of immunologic information. Furthermore, a companion investigation (Padykula and Taylor, '76) demonstrates that macrophages and lymphocytes are eliminated from the postpartum stroma of the opossum by transepithelial emigration to the lumen of the uterine glands and to the uterine cavity (text fig. 1 in Padykula and Taylor '76). I n the albino rat the lymphatic vessels offer a n additional mode of egress (Padykula and Campbell, '76). This cyclic appearance of lymphocytes and macrophages, their physical association free in the stroma as well as in the isolation chamber illustrated in figures 7 and 8, may reflect a n immune mechanism which is differentiated for the purpose of eliminating extracellular stromal macromolecules in a cyclic manner. This possibility is discussed below. DISCUSSION
The uteriis and its capacity to respond immunologically The embryo-fetus with its placental membranes may be viewed as a successful allograft on the uterine endometrium. Current evidence suggests that this success is derived more from the low antigenicity of the trophoblast than from inability of the endometrium to respond immunologically (Beer and Billingham, '71, Beer et al., '75). It has long been known that the injection of vital dyes or carbon particles will result in the appearance of endome-
52
HELEN A. PADYKULA
trial macrophages, especially during estrogenic influence (Fluhmann, ’28; Nicol and Vernon-Roberts, ’65; Vernon-Roberts, ’69). Experimental evidence indicates that skin allografts to the rat uterus cause a 3-fold increase in weight of the draining lymph nodes whereas autografts do not (Beer and Billingham, ’71 ). Hypertrophy of draining lymph nodes occurs also during hetercspecific pregnancies (Beer and Billingham, ’71). Recently Beer et al. (’75) in a study of interstrain (allogeneic) matings demonstrated that the maternal response against fetal alloantigens leads to heavier placental-fetal units than after syngeneic matings. Surgical removal of the lymph nodes (“para-aortic nodes,” according to the authors) that drain the uterine horns interferes with this enlargement of the placental-fetal units, a n observation which suggests that humoral immunity is involved. This varied experimental evidence along with the present report of the coexistence of macrophages, lymphocytes, and plasma cells in the postpartum endometrium, indicates that the uterus possesses the capacity to respond immunologically. The distribution of the lymphatic vessels in the uterus may contribute to its ability to accept grafts, since in the rabbit (McLean and Scothorne, ’70), and most likely in the rat, lymphatic vessels are absent from the endometrium. They occur primarily as a plexus located between the two muscle layers of the myometrium (Hoggan and Hoggan, 1881). In these species, lymphatic vessels occur also at the endometrial-myornetrial junction and in the serosa. This absence of endometrial lymphatic vessels may be the basis of the so-called “privileged site” status of the uterus in these species, since Barker and Billingham (’68), by experimentally eliminating lymphatic drainage from a region in other organs, created “privileged sites” which do not reject grafts. It should be added that the amount of lymphatic drainage in the endometrium may be related, a s a consequence of evolutionary process, to the degree of penetration by the placenta into the endometrium (Hoggan and Hoggan, 1881). The relationship of this special spatial arrangement of uterine lymphatic vessels to the immunologic activity
of the endometrium must await further investigation.
Possible significance of the occurrence in the regressing endometrium of the cells associated with the humoral immune response Heterophils (neutrophils) invade the regressing endometrium early in the f i s t day after delivery. They are conspicuous in the superficial endometrium during day 1 and decrease in number during day 2. Although this cell type in abnormal states usually represents the first stage of inflammation, the heterophils in the normal regressing endometrium seem to be programmed into the events of estrus (Yanigamachi and Chang, ’67) whether estrus occurs postpartum or during a cycle. Our histological observations on day 1 postpartum in the rat and opossum (Padykula and Campbell, ’76) indicate that heterophils migrate from superficial endometrial blood vessels, pass through the superficial stroma, and migrate across the luminal epithelium into the uterine cavity. Here they function in the phagocytosis of sperm (Austin ‘57; Reid, ’64). Eosinophils also appear transiently at the time of the postpartum estrus: since this immigration is known to reflect rising estrogen levels (Ross and Klebanoff, ’SS), it may also be an event in the endometrial cellular program of estrus. Thus the transient appearance of heterophils and eosinophils may not be directly related to the subsequent accumulation of macrophages, lymphocytes, and plasma cells in the postpartum uterus. However, the possibility of the interrelationship of all these cell types should remain open since these two granulocytes are known to be cellular components of inflammatory processes. The occurrence of plasma cells in concert with lymphocytes and macrophages indicates that a humoral immune response is taking place (Weiss, ’72). Thus it is likely that antibody production is occurring in the regressing endometrium. This antibody (or antibodies) might be released into the circulation as IgG or into the uterine cavity as IgA. The presence of plasma cells indicates that B lymphocytes have been present. The frequent apposition of the surfaces of lymphocytes and macrophages
IMMUNE RESPONSE I N REGRESSING ENDOMETRIUM
suggests cellular interaction or communication, as occurs in the presence of foreign antigen (e.g., Nielsen et al.,'74). The presence of transforming lymphocytes signals the possible activation of lymphocytes to become blast cells capable of proliferation (e.g., Biberfeld, '71). The nature of the antigenic material that stimulates the appearance of this transient postpartum cellular representation of a humoral immune response must next be considered. There are three possible sources of foreign antigens: (1) sperm if the female was mated during the day 1 postpartum estrus; ( 2 ) the placental-fetal complex that was intimately apposed and/or fused to portions of the endometrium until parturition; and ( 3 ) normal macromolecules of the stroma, such as collagen, which are altered during extracelluIar enzymic degradation in a manner that transiently reveals new antigenic sites. Since our postpartum opossums and rats were isolated from males, the antigenic source cannot be spermatozoa. The natural allograft represented by the placental-fetal complex is a possibility as an antigenic source, especially in the opossum, a wild animal in which genetic strain variation is more apt to occur. However, at diestrus of a non-pregnant cycle in the opossum, influx of macrophages, lymphocytes, and plasma cells occurs in a degree comparable to that occurring after parturition. This similarity in the cellular manifestations of regression in the non-pregnant and pregnant cycles, to a considerable degree, lessens the possibility of antigenic stimulation by the marsupial feto-placental complex. It should be mentioned that this similarity is a special situation because endometrial expansion during the luteal phase in the opossum is comparable whether or not pregnancy occurs; this led early investigators to designate the luteal phase of a non-pregnant cycle as "pseudopregnancy" (Hartman, '23). Genetic variation due to strain differences in the albino rat is less probable than in the opossum because we used highly inbred members of the Charles River CD strain. As an example of the macromoleci~lar alterations that may occur during uterine involution, it is appropriate to consider the role of neutral collagenase in the produc-
53
tion of new antigenic sites. The purified enzyme attacks the collagen molecule in its polymeric position within the insoluble collagen fibril and cleaves through the triple helical conformation at a specific site to produce two unequal fragments (usually one 75% fragment and the other a 25% piece of the collagen molecule) (review by Gross, '74). In the case of rat uterine collagenase, further cleavage of two small peptide fragments occurs in vitro; however, it is not known whether or not this additional enzymic action arises from contamination (Jeffrey and Gross, '70). Nevertheless, the cleaved collagen molecule would present new antigenic sites in an extracellular maternal location. Furthermore, according to proposed mechanisms for collagenolytic degradation of insoluble fibrils (Harris and Krane, '74: fig. 5, p. 606; Gross, '74: fig. 18, p. 410), the two fragments of the collagen molecule would move away from the fibril into the ground substance where they would most likely denature at body temperature. Thus denaturation would reveal antigenic sites additional to those arising from the cleaved triple helix. Since collagen degradation proceeds very rapidly (Harkness and Moralee, '56; Ryan and Woessner, '74), cleavage by collagenase and the subsequent denaturation would constitute a powerful transient local antigenic stimulus to the maternal organism. The solubilized denatured collagen fragments might be further degraded by extracellular proteases or phagocytosed by uterine macrophages (Hams and Krane, '74; Gross, '74) to be digested within phagosomes by cathepsin B, (Burleigh et al., '74). To account for Schwarz and Guldner's ('67) and Parakkal's ('69, '72) demonstrations of intact collagen fibrils within phagosomes of uterine phagocytes, Harris and Krane ('74) suggested in diagrammatic representation (their fig. 7, p. 607) that multimolecular fragments of collagen fibrils may be released by collagenase cleavage and ingested by macrophages for further digestion within phagosomes. Thus it seems likely that the macrophage is involved in the terminal degradation of collagen fibrils as well as of other extracellular macromolecules. Furthermore, the possibility exists that neutral collagenase
54
HELEN A. PADYKULA
may be secreted by the macrophages (Wahl et al., '75), although the possibility of secretory involvement of fibroblasts (Werb and Burleigh, '74) or epithelia (Eisen and Gross, '65) should remain open. Hypothesis involving a humoral immune reaction in the regulation of cyclic renewal of extracellular macromolecules in the stroma of the normal endometrium The hormonal environment during the first few postpartum days favors the emigration of heterophils, monocytes, eosinophils, and lymphocytes from the blood into the endometrial stroma. In addition, this milieu favors the synthesis and release of neutral collagenase (and possibly other enzymes capable of extracellular cleavage of macromolecules) into the ground substance of the stroma. The polymeric collagen fibril is dismantled initially by collagenolytic cleavage which releases fragments of monomeric molecules as well as larger cross-striated pieces of the collagen fibril. Numerous antigenic sites are transiently revealed as a result of rapid enzymic cleavage of the collagen triple helix and the subsequent denaturation of fragments in the ground substance. The various fragments of the enzymically cleaved collagen fibrils are ingested by monocytemacrophages that respond to them as unnatural proteins or polypeptides and process them as antigens. At this time, macrophagic-lymphocytic interaction is conspicuous, as the two cell surfaces become closely apposed and may even acquire special differentiations. B lymphocytes differentiate into plasma cells to produce antibody to the collagen-derived antigens. This antibody would most likely enter the vascular and lymphatic vessels and thus be recognizable systemically. At the close of this episode, lymphocytes, carrying immunologic information, and macrophages, engorged with the remnants of extracellular stromal macromolecules, have been largely eliminated from the endometrium by transepithelial emigration into the uterine cavity, glandular lumens, or into lymphatic vessels. Thus, immunologic information contained in lymphocytes as well as extracellular debris sequestered in macrophages, both materials related to the pre-
vious pregnancy or cycle, have been largely removed from the uterine stroma, and a new ovarian-uterine cycle proceeds in which stromal macromolecules are again synthesized and secreted into the extracellular compartment of the uterus. The cellular mechanism proposed in this hypothesis suggests involvement of an autoimmune response in the normal removal of extracellular stromal macromolecules. This hypothesis may be applicable to the interpretation of the so-called collagen vascular diseases, such as rheumatoid arthritis, in which neutral collagenase is known to be released and aggregations of lymphocytes, macrophages, and plasma cells occur at the sites of lesions. ACKNOWLEDGMENTS
Richard V. T. Stearns prepared the p h e tomicrographs and the final electronmicrographs. The author assumes full responsibility for the interpretations presented here but gratefully acknowledges generous discussion of the observations with Dr. Leon P. Weiss, Dr. Myron S . Silverman, Dr. Miles Crenshaw, Dr. Joachim Sommers, and Dr. Edward H. Bossen. LITERATURE CITED Austin, C. R. 1957 Fate of spermatozoa in the uterus of the mouse and rat. J. Endocrinol., 14: 335-342. Banon, P., D. Brandes and J. K. Frost 1964 Lysosomal enzymes in the rat ovary and endometrium during the estrous cycle. Acta Cytol., 8: 4 1 6 4 2 5 . Barker, C. F., and R. E. Billingham 1968 The role of afferent lymphatics in the rejection of skin homografts. J. Exp. Med., 128: 197-221. Beer, A. E., and R. E. Billingham 1972 Immunobiology of mammalian reproduction. In: Advances in Immunology. Vol. 14. R. J. Dixon, Jr. and H. G. Kunkel, eds. Academic Press, New York, pp. 1-84. Beer, A. E., J. R. Scott and R. E. Billingham 1975 Histoincompatibility and maternal immunological status as determinants of f e t e placental weight and litter size in rodents. J. Exp. Med., 142: 180-196. Biberfeld, P. 1971 Morphogenesis in blood lymphocytes stimulated with phytohaemoagglutinin (PHA). Acta Path. Microbiol., Scand. Section A, Supplement No. 223. 3This hypothesis has been derived from the observations reported here and in the two accompany-
ing studies (Padykula and Taylor, '76; Padykula and
Campbell, '76).
IMMUNE RESPONSE IN REGRESSING ENDOMETRIUM B r a n d s , D., and E. Anton 1969 A n electron microscopic cytochemical study of macrophages during uterine involution. J. Cell Biol., 41: 450-461. Burleigh, M. C., A. J. Barrett and G. S. Lazarus 1974 Cathepsin B1. A lysosomal enzyme that degrades native collagen. Biochem. J., 137: 387-398. Deno, R. A. 1937 Uterine macrophages in the mouse and their relation to involution. Am. J. Anat., 60: 433-469. Eisen, A. Z., and J. Gross 1965 The role Of epithelium and mesenchyme in the production of a collagenolytic enzyme and a hyaluronidase in the anuran tadpole. Dev. Biol., 12: 408418. Fluhmann, C. F. 1928 The reticubendothelid cells of the uterus: A n experimental study. Am. J. Obst. Gynec., 15: 783-796. Gross, J. 1974 Collagen biology: structure, degradation, and disease. In: The Harvey Lectures, 1972-1973, Academic Press, New York, pp. 351-432. Gross, J., and C. M. Lapiere 1962 Collagenolytic activity in amphibian tissues; a tissue culture assay. Proc. Nat. Acad. Sci. (U.S.A.), 48: 1014-1022. Harkness, R. D., and B. E. Moralee 1956 The time course and route of loss of collagen from the rat's uterus during postpartum involution. J. Physiol., 132: 502-508. Harris, E. D., Jr., and S. M. Krane 1974 Collagenase. New Eng. J. Med. (Part l), 291: 557-563; (Part 2), 291: 605-609; (Part 3), 201: 652-661. Hartman, C. 1923 The oestrous cycle in the opossum, Didelphis virginiana. Am. J. Anat., 32: 353-421. Hoggan, G., and F. E. Hoggan 1881 On the comparative anatomy of the lymphatics of the uterus. J. Anat., 16: 50-89. Jeffrey, J. J., R. J. Coffey and A. Z. Eisen 1971 Studies on uterine collagenase. I. Relationship of enzyme production to collagen metabolism. Biochim. Biophys. Acta, 252: 136-142. Jeffrey, J. J., and J. Gross 1970 Collagenase from the rat uterus. Isolation and partial characterization. Biochemistry, 9: 268-274. Jeffrey, J. J., and T. J. Koob 1973 Hormonal regulation of collagen catabolism in the uterus. Excerpta Med.: Endocrinology, Int. Congr. Series, 273: 1115-1123. Koob, T. J., and J. J. Jeffrey 1974 Hormonal regulation of collagen degradation in the uterus: Inhibition of collagenase expression by progesterone and cycIic AMP. Biochim. Biophys. Acta, 354: 61-70. Lobel, B., and H. W. Deane 1962 Enzymic activity associated with postpartum involution of the uterus and with its regression after hormone withdrawal in the rat. Endocrinology, 70: 567-578. McLean, J. M., and R. J. Scothorne 1970 The lymphatics of the endometrium in the rabbit. J. Anat., 107: 39-48.
55
Nicol, T., and 8 . Vernon-Roberts 1965 The influence of the estrous cycle, pregnancy, and ovariectomy in RES activity. J. Reticuloendothel. SOC.,2: 15-29. Nielsen, M. H., H. Jensen, 0. Braendstrup and 0. Werdelin 1974 Macrophagelymphocyte clusters in the immune response to soluble protein antigen in vitro. 11. Ultrastmcture of clusters formed during the early response. J. Exp. Med., 140: 1260-1272. Padykula, H. A., and A. Campbell 1976 Cellu&ar mechanisms involved in cyclic stromal renewal of the uterus. 11. The albino rat. Anat. Rec., 184: 27-48. Padykula, H. A., and J. M. Taylor 1976 Cellular mechanisms involved in cyclic stromal renewal of the uterus. 1. The North American Opossum. Anat. Rec., 184: 5-26. Parakkal, P. F. 1969 Involvement of macrophages in collagen resorption. J. Cell Biol., 41: 345-354. 1972 Macrophages: The time course and sequence of their distribution in the postpartum uterus. J. Ultrastr. Res., 40: 284-291. Peters, A., S . L.Palay and H. deF. Webster 1970 The Fine Structure of the Nervous System. Hoeber, Med. Div., Harper & Row, New York. Reid, B. L. 1964 Fate of residual sperm in the mouse uterus. J . Anat., 98: 492 (abstract). Ross, R., and S. J. Klebanoff 1966 The eosinophilic leukocyte. Fine structure studies of changes i n the uterus during the estrous cycle. J. Exp. Med., 124: 653-660. Ryan, J. N., and J. F. Woessner, Jr. 1974 Oestradiol inhibition of collagenase role in uterine involution. Nature, 248: 526-528. Schwarz, W., and R. H. Giildner 1967 Elektronenmikroskopische Untersuchungen der Kollagenabbaus im Uterus der Ratte nmach der Schwangerschaft. Z. Zellforsch., 83: 416-426. Silverman, M. S. 1970 The macrophage in c d lular and humoral immunity. J. Reticuloendo. SOC.,8: 105-123. Smith, R. E., and M. R. Henzl 1969 Role of mucopolysaccharides and lysosomal hydrolases i n endometrial regression following withdrawal of esttadiol and chlormadinone acetate. I. Epithelium and stroma. Endocrinology, 85: 50-66. Vernon-Roberts, B. 1969 The effects of steroid hormones on macrophagic activity. Int. Rev. Cytol., 25: 131-159. Volkmann, A., and J. L. Gowans 1965A The production of macrophages in the rat. Brit. J. Exp. Path., 46: 50-61. 1965B The origin of macrophages from bone marrow in the rat. Brit. J. Exp. Path., 46: 62-70. Wahl, L. M., S. M. Wahl, S. E. Mergenhagen and G. R. Martin 1975 Collagenase production by lymphokine-activated macrophages. Science, 187: 261-263. Weis% L. p. 1972 The Cells and Tissues of the
56
HELEN A. PADYKULA
Immune System. Prentice-Hall, Inc., Englewood, N. J. Werb, Z., and M. C. Burleigh 1974 A specific collagenase from rabbit fibroblasts in monolayer culture. Biochem. J., 137: 373-376. Woessner, J. F., Jr., and J. N. Ryan 1973 Collagenase activity in homogenates of the in-
voluting rat uterus. Biochim. Biophys. Acta, 309: 397-405. Yanigamachi, R., and M. C. Chang 1963 Infiltration of leucocytes into the uterine lumen of the golden hamster during the estrous cycle and following mating. J. Reprod. Fert., 5 : 389-396.
All of these illustrations were obtained from preparations of opossum endometrium o n days 1 and 2 postpartum.
PLATE 1 EXPLANATION OF FIGURE
1
Endometrial glands with intervening stroma, semi-thin section, toluidine blue. x 1,950. Macrophages ( M ) , lymphocytes ( L ) , and a plasma cell ( P ) occur in the stroma that intervenes between two glands ( G ) . Cell division (arrows) is evident in the glandular epithelium. In two locations ( * ), lymphocytes are closely apposed to the surface of macrophages (only a portion of the cytoplasm of these macrophages is included here). The ultrastructure of the boxed area is shown in figure 2.
IMMUNE RESPONSE IN REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 1
57
PLATE 2 EXPLANATION O F FIGURE
2
58
Electron micrograph of a portion of the field shown i n figure 1. X 7,500. The plasma cell ( P ) is in active synthetic state: cisternae of rough endoplasmic reticulum are rounded and filled with secretory product. A lymphocyte ( L ) and a macrophage ( M ) are flattened and in close alignment along the region of their apposition (arrows).
IMMUNE RESPONSE IN REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 2
59
PLATE 3 EXPLANATION OF FIGURE
3
60
Phasocyte and lymphocyte. x 25,000. This phagocyte ( M ) conta’ns phagosomes ( P ) that contain fibrils ( F ) which appear to be collagen. Note also the specialized sites (arrows) along the phagocytic cell surface. The cytoplasm of the lymphocyte ( L ) is rich i n ribosomes and has relatively little rough endoplasmic reticulum. The extracellular compartment contains much ground substance ( G S ) and relatively few collagen fibrils (cf).
IMMUNE RESPONSE I N REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 3
61
PLATE 4 EXPLANATION O F FIGURES
4A,B
62
Specialization in the lymphocytic-macrophagic association. A, X 28,700; B, 84,500. In figure 4A, two cytoplasmic processes (CP) extend from the macrophage (MAC), one of which is closely associated with the surface of the lymphocyte (LYM). These processes contain numerous membrane-limited, electron opaque inclusions of various shapes. In one area (arrow), the apposed surfaces show a junctional specialization, which is enlarged in figure 4B. At the differentiated site, the plasma membranes of the lymphocyte and macrophage remain separate, but show junctional specialization in the form of greater cytoplasmic density ( d ) along the inner surface of the lymphocytic plasma membrane.
IMMUNE RESPONSE I N REGRESSING ENDOMETRIUM
PLATE 4
Helen A. Padykula
63
PLATE 5 EXPLANATION OF FIGURE
5
64
Transforming lymphocyte. x 27,500. Single cisternae of rough endoplasmic reticulum ( R E R ) are conspicuous; some cisternae are arranged at right angles to the surface of the cell (arrows). Note the prevalence of polysomes ( p ) and single ribosomes in the cytoplasmic matrix. G S , ground substance; cf, collagen fibrils.
IMMUNE RESPONSE I N REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 5
65
PLATE 6 EXPLANATION O F FIGURE
6
66
Endometrial plasma cell. x 12,500. Active synthetic state is reflected by elaborate oriented cisternae of rough endoplasmic reticulum (RER) which are filled with secretory product as well as by presence of the large Golgi apparatus. Packaging of secretory product appears to be occurring along the inner or concave surface of the Golgi (G) apparatus ( * ) . Coated vesicles (arrows) are numerous in the Golgi region. A n aggregate of smooth membraned tubules ( S M ) is present.
IMMUNE RESPONSE IN REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 6
67
PLATE 7 EXPLANATION O F FIGURE
7
68
Enclosure of a macrophage i n the subepithelial stroma. x 7,500. The enclosure of macrophages ( M ) within an inner extracellular coat (arrowheads) and an outer fibroblast-like cellular vesture ( X ) is a regular feature of the stromal differentiation. The extracellular coat is discontinuous a t points where macrophagic pseudopodial processes ( m p ) of irregular shape extend free into the ground substance of the stroma. The processes of the fibroblast-like cell ( x p ) enclose the coated macrophage as well as some uncoated pseudopodial processes (mp). Note that the processes of the investing cell are very thin in some regions. This process of enclosure is most frequently observed in regions close to the epithelial basal lamina (EL) which is slack. Another example of lymphocytic-macrophagic apposition is evident a t the bottom of this figure (*'). E, epithelium.
lMMUNE RESPONSE IN REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 7
69
PLATE 8 EXPLANATION OF FIGURE
8
70
Enclosure of a lymphocyte and a macrophage in the stroma. x 13,000. A lymphocyte (LYM) and a macrophage (MAC) are isolated together within an extracellular coat ( e c ) , except for an uncoated macrophagic process (mp). The processes of a fibroblast-like cell ( X ) follow the contours of this lymphocytic-macrophagic complex.
IMMUNE RESPONSE IN REGRESSING ENDOMETRIUM Helen A. Padykula
PLATE 8
71
