0022-202X/80/7501-0083$02.00/ 0 THE JouRNAL OF INVESTIGATIVE DERMATOLOGY, 75:83-88, 1980 Copyright © 1980 by The Williams & Wilkins Co.

Vol. 75, No. I Printed in U.S.A.

lntraepithelial Lymphocytes

LEONARD SEELIG, JR., PH.D ., AND RuPERT E. BILLINGHAM, D.Sc.

Department of Cell Biology, The University of Texas Health Science Center at. Dallas, Dallas, T exas U.S.A.

Significant numbers of lymphocytes are normal com­ponents of most epithelia; however, despite intensive study their functional significance remains an enigma and a challenge. As in the gut epithelium, many of its lymphocytes appear to be "tissue-drawn" rather than "antigen-drawn," an indication that the tissue might be a differentiation staging post whence the "graduates" proceed to other places. Mammary glands appear to be primary sites for migration of lymphocytes "schooled" in the intestine. These and other kinds of leukocytes not only enter the mammary interstitium, but pass through its epithelium to become a normal component of milk. Results of ultrastructural studies hint that the transepi­thelial passage ofleukocytes is facilitated by their ability to disrupt apical occluding junctional complexes and thus to produce intercellular gaps extending down to, but not through, the basal lamina. Activity on the part of the myoepithelial cells may also contribute to this diapedetic process.

To address the question of whether or not milk lym­phocytes within the neonate's digestive tract can tra­verse the gut epithelium and enter its tissues, the present investigators inoculated surgically isolated segments of small bowel in adult (FI X DA)F 1 hybrid hosts with suspensions of viable lymph node cells from FI strain rats. Thickening of the mucosa of these segments and hypertrophy of the draining mesenteric lymph nodes, on an immunogenetically specific basis, were attributed to graft-versus-host reactivity-indicative of the ability of some of the "transplanted" lymphocytes to escape from the lumen of the gut segment.

So far in this symposium we have dealt with the cell minor­ities of nonkeratinocyte lineages normally present in the epi­dermis, mainly the branched cells of Langerhans and melano­cytes. To put the epidermal situation into perspective it seems appropriate to consider the extent to which minority cell pop­ulations of adventitious origin, particularly lymphocytes (since we know little about the presence of other cell types), contribute to the normal mass and functional activity of nonepidermal epithelia (Fig 1).

INTRAEPITHELIAL LYMPHOCYTES IN THE SMALL INTESTINE

The normal occurrence of significant numbers of mononu­clear leukocytes (together with occasional eosinophils and mast cells) within the epithelium of the small bowel (Fig 1A) has been recognized for well over a century, and ingenious interpre­tations of their significance have been advanced (1,2]. Many of

The expenses of some of the work described in this article were defrayed by National Institutes of Health Grant Al-10678.

Reprint requests to: Dr. Leonard Seelig, Jr., Department of Cell Biology, The University of Texas Health Science Center at Dallas, 5323 Harry Hines Boulevard, Dallas, Texas 75235.

Abbreviations: GVH reactions: graft-versus-host reactions FI strain: Fischer strain IE lymphocytes: intraepithelial lymphocytes MEM: minimal essential medium

83

these have been discredited by the rapid increase in knowledge of lymphocyte immunobiology.

Intraepithelial (IE) lymphocytes are very similar to lympho­cytes of the lamina propria of the gastrointestinal tract and are slightly larger than the small lymphocytes in blood (3]. Electron microscopy put an end to the longstanding controversy over whether or not IE lymphocytes were present in vacuoles within epithelial cells by establishing unequivocally that they are located between adjacent enterocytes, usually in the basal re­gion of the epithelium [ 4]. A considerable body of evidence has established the following: (a) that they never form specialized adhesions with their neighbors, (b) that they are not a homo­geneous cell population since large, medium, and small lympho­cytes are present, (c) that they appear to cross the basal lamina between the lamina propria and the epithelium in either direc­tion, and (d) that they account for 9 to 23% of the total cell population in normal jejunal epithelium. However, in certain diseases, notably celiac disease, some patients may have more lymphocytes than enterocytes in their epithelium [ 4-7].

Enterocytes can only synthesize DNA and divide in Lieber­kiihn's crypts; their fate is to flow distally as an intact sheet of epithelium along the sides of the villi to be extruded from their tips and they have a functional life span of 2 to 3 days. Intraepithelial lymphocytes, in contrast, can synthesize DNA and divide regardless of their position within the epithelium [5]. Moreover, blastlike IE lymphocytes, some with high rate of DNA synthesis, have been shown to occur in the absence of antigenic stimulation. No convincing evidence has been forth­coming that IE lymphocytes degenerate within the epit~elium , or that significant numbers of them accompany epithelial cells in their migration to extinctive extrusion from the villi tips [3]. Studies involving infusion of mice with 3H-thymidine have indicated that about 70% of the IE lymphocytes are short-lived and .about 30% long-lived [8,9]. That many of these cells are T lymphocytes has been indicated by the considerable reduction in their numbers in thymus-deprived animals, and by functional and other criteria [10]; however, nude mice do have a small number of IE lymphocytes of undetermined natw-e. Viable lymphocytes, exclusive ofPeyer's patch cells, have been isolated from the gut mucosal epithelium and studied in vitro [11,12]. Unlike lymphoid cells from the lamina propria of the gut, a high proportion of which contain IgA, the majority of the epithelially derived lymphocytes are devoid of immunoglobu­lins, which has prompted the suggestion that they are virgin "B cells." Another interesting feature of many IE lymphocytes that differentiates them from lymphocytes of the peripheral blood, thoracic duct lymph, or lamina propria is that they usually contain metachromatic granules (9].

Studies on the proliferative kinetics and fate of IE lympho­cytes have been complicated by the fact that the population is not a homogeneous one. However, most of the available evi­dence is in accord with the conclusion of Ri.ipke and Everett [9], who studied the small intestine of the mouse. These inves­tigators have suggested that both T and B lymphoblasts, being "tissue-drawn" rather than "antigen-drawn," initially invade the epithelium and undergo mitosis. The B cells contribute to the plasma cell population in the superficial layers of the lamina propria. The T lymphoblasts give rise to small, probably long­lived, lymphocytes that enter the circulation, some of them possibly after several weeks in the epithelium.

Although there can be little doubt that lyn1phocytes enter

FIG 1. Micrographs depicting intraepithelialleukocytes in various tissues. A, human small intestine with numerous leukocytes (L) intervening between adjacent epithelial cells. Several leukocytes (arrows) are seen traversing the basement membrane (1 11.m toluidine-blue-stained epoxy section) (X 1,000). B, human tracheal epithelium showing the typical "halo" or "clear" area around the densely stained leukocyte nucleus (arrows) in paraffin sections stained with hematoxylin and eosin (X 1,000). C, portion of a human mammary gland with numerous plasma cells (P) in the connective tissue surrounding the alveoli. Intraepithelialleukocytes (arrows) are apparent in the basal portion of the epithelium (X 1,000). D , electron micrograph of a leukocyte (L) within the alveolar epithelium of a mammary gland of a lactating rat. These cells, usually located in the epithelium near a myoepithelial cell (M), often exhibit pseudopodia directed toward the apical junctional complex (arrows) (X 18,000).

July 1980

the epithelium of the gut via capillaries in the lamina propria, how they leave it is still open to conjecture. Ultrastructural studies provide no evidence that they normally die within the epithelium, that significant numbers pass upward between ad­jacent enterocytes to break or otherwise pass through the terminal tight junctions to enter the lumen of the gut, or that other than a small proportion of them are transported within the epithelium and lost by extrusion at the tips of the villi. Rather, it appears that they usually leave the villi via lymph vessels of the lamina propria.

An important, unresolved question concerning IE lympho­cytes is their relationship to the majority of immunoblasts that are generated in antigen-stimulated lymphoid tissue remote from the small intestine and that home to its lamina propria, where they become plasma cells [13].

To conclude this survey of IE lymphocytes, whose function is still uncertain, we will consider some of the more recent ideas on their functional significance in relation to evidence that bears upon them.

A. The most obvious possibility is that antigenic material in the lumen of the gut attracts these cells to assume their IE location and stimulates them to undergo blast transformation. This hypothesis is supported by evidence of low IE lymphocyte counts in germ-free animals and even lower counts in T-cell­depleted germ-free animals. However, small numbers of IE lymphocytes are present in completely antigen-free intestines. For example, grafts of fetal mouse intestine transplanted be­neath the renal capsule of syngeneic adult hosts acquire a complement of IE lymphocytes 3 wk later, an indication that some factor from the stroma must be responsible for initially attracting these cells [14]. However, the increased incidence of IE lymphocytes in certain disease states must surely be attrib­utable in part to antigenic material within the intestinal lumen.

B. Intraepithelial lymphocytes, possibly like Langerhans cells, are responsible for interacting with antigen that has traversed the epithelial frontier. This explanation has been somewhat weakened by recent evidence suggesting that the important sites of ingress of luminal antigens into the body are specialized areas of epithelium in Peyer's patches [15,16].

C. The epithelium of the gut is the bursa equivalent, respon­sible for the instruction and maturation of B lymphocytes [17]. Evidence that the majority of IE lymphocytes are T cells and that IE lymphocytes are lacking in the intestine of infant mice greatly weakens this hypothesis.

D. The epithelium of the gut is a site for differentiation of immunologically activated lymphoid cells on their way to be­coming effector cells both locally and elsewhere in the body. The presence of the metachromatic granules in many of the cells may reflect their undifferentiated status.

E. Intraepitheliallymphocytes are a special type of mast cell. Many of them do contain metachromatic granules, and they also contain a small amount of histamine. However, when basophils passively sensitized with IgE antibodies are chal­lenged with the conesponding antigen, they release their his­tamine, whereas IE lymphocytes do not [2].

F. Intraepitheliallymphocytes may belong to a special cate­gory ofT cells capable of combining with antigen and in some way preventing it from initiating "unwanted" immunologic responses. Kupffer's cells in the liver play an analogous role, i.e., they trap and ingest antigens, and thereby reduce their immunogenicity.

ABILITY OF VIABLE LYMPH NODE CELLS INTRODUCED INTO THE LUMEN OF THE SMALL

BOWEL TO TRAVERSE ITS EPITHELIUM

Whether in normal individuals lymphocytes can traverse the epithelium of the gut and enter its lumen is still an unresolved issue. We would like to raise a seemingly more academic ques­tion: If such cells can get into the intestinal lumen, can they get back in the tissues again? This question arises from indirect evidence, of a circumstantial nature, suggesting that in infant

INTRAEPITHELIAL LYMPHOCYTES 85

mice and rats, as well as in human infants, viable lymphocytes transferred n·om the mother to the infant's gut via milk may traverse the suckling's intestinal epithelium and gain access to the blood stream of the host [18]. To test this interpretation, we have been using a simple indirect approach to determine whether viable lymph node cells, introduced experimentally into the lumen of a rat's small intestine, can traverse the intact epithelium and establish themselves in its tissues.

Principle of the Experiment

It is known that when lymphoid cells from an adult parental strain donor are injected or caused to enter the parenchyma of some organs, such as skin or kidneys, of an adult F1 hybrid host that is genetically tolerant of them and that confronts them with alien histocompatibility antigens determined by the major histocompatibility complex, some of the cells react locally against the host (local graft-versus-host [GVH] reaction or normal lymphocyte transfer reaction) , and others pass to the regional or draining lymph node(s) where GVH reactions are also incited [19]. The latter reaction, whose severity is dose­dependent, can be quantified on the basis of hypertrophy of the affected organs [20].

Our approach entailed introduction of a suspension of viable lymph node cells from an adult Fischer (FI) strain rat into a surgically isolated but vascularized segment of small bowel [21] in an adult (FI X DA)Ft hybrid host. We anticipated that if sufficient cells managed to leave the lumen and traverse the gut lining, they would pass to the mesenteric node complex and initiate hypertrophy due to GVH reactivity. In practice, isolated ileal segments (Fig 2), about 10 em long and including at least 1 Peyer's patch, were prepared and the associated mesentery was left intact to preserve the blood, lymphatic, and nerve supplies. The proximal end ofthe segment was closed by suture, and the distal end was brought out through a small opening in the abdominal wall and sutured to the skin to maintain patency of its lumen. Functional continuity of the remaining intestine was restored by end-to-end anastomosis. The isolated bowel segment was thoroughly irrigated with minimal essential me­dium (MEM) containing penicillin and streptomycin to remove all mucus, blood, and food residues. Three days post operation the animals were assigned to 5 groups, and their isolated ileal segments were inoculated via the skin opening with cell suspen­sions (see below) dispensed in 0.25 ml of MEM.

Group I: 100 X 106 FI strain lymph node cells, on each of 3 consecutive days

Group II: 100 X 106 (FI X DA)Ft node cells, on each of 3 consecutive days (controls)

FIG. 2. Schematic drawing of the procedure for isolation of the ileal segment in Fischer x dark-agouti F, hosts and route of introduction of a suspension of viable lymph node cells. Abbreviations: L .I., large intestine; S.I., small intestine.

86 SEELIG AND BILLINGHAM

Group III: A single inoculum of 100 X 106 FI strain node cells Group IV: A single inoculum of 100 X 106 (FI X DA)F1 node

cells (controls) Group V: Unoperated controls receiving no lymph node cells The animals were killed 8 days after the 1st (or only) injection

of lymph node cells. Mesenteric node weights were recorded for each animal, and tissues were removed from the isolated ileal segment and the intact ileum for histological examination.

Results of the Experiment

The mesenteric nodes (Table I) were significantly larger in both experimental series (groups I and III) than in their corre­sponding controls (groups II, IV, and V). Furthermore, node hypertrophy was greater in the case of recipients of 3 doses of node cells than in the case of I-dose recipients. Neither the appearance nor the size of the Peyer's patches in the isolated segment was affected by our experimental manipulation.

The fact (Table II) that the mucosa of the isolated ileal segments receiving injected parental strain node cells was thicker than that of the segments receiving injected F1 hybrid cells corroborated our conclusion that sufficient parental strain node cells (T lymphocytes) left the lumen of the ileal segment [22] to instigate GVH reactivity both locally in the wall and more remotely, in the mesenteric nodes.

In another experiment, Fl strain lymph node cells that had been labeled in vitro with 3H-uridine were injected into the lumina of isolated intestinal segments in (FI X DA)F1 hybrid rats. Autoradiographic examination of sections of tissue re­covered from the gut segments 24 and 48 hr post inoculation revealed labeled cells within both the lamina propria and Peyer's patches. This afforded more direct support for the belief that viable lymphocytes in the intestinal lumen have the ca­pacity to traverse its epithelium.

"SECRETION" OF LEUKOCYTES BY THE MAMMARY GLAND

It has long been known that considerable numbers of viable leukocytes, predominantly macrophages, but approximately

Vol. 75, No. 1

10% T and B lymphocytes, are normal ingredients of colostrum and milk. The leukocyte content of human milk is about 1 to 2 x 106 / ml. However, only recently have we begun to recognize milk cells as an important postpartum component of the ma­ternal immunologic endowment. Studies on human milk lym­phocytes have revealed that they can perform immunologic functions in vitro (including mixed lymphocyte culture reac­tions) , respond to mitogens, as well as to bacterial and viral antigens, and synthesize antibodies [23,24].

A particularly interesting finding is that the repertoire of immune responses of a woman's milk lymphocytes to environ­mental antigens is not necessarily the same as that of her peripheral blood lymphocytes [25]. The mammary gland ap" pears to be part of a common mucosal immune system receiving lymphocytes, via the blood stream, derived principally from immunocompetent cells stimulated by antigens in the individ­ual's gastrointestinal and respiratory tracts [26]. The lympho­cytes in milk, like the antibodies in this exosecretion, seem to represent the optimal population for the protection of sucklings.

That the secretion of cells by the mammary gland during lactation is a purposeful process, and not merely an unavoidable consequence of the large aggregate area of involved secretory epithelium, is suggested by the negligible number of mononu­clear cells-about 2 X 106 -we lose each day from our kidneys via urine [27] and by a growing body of evidence of the short­term benefits the suckling derives from these maternal cells [23].

Raux et al [26] have shown in mice that lymphocytes from the mesenteric lymph nodes "home" to the mammary glands during the latter part of pregnancy and during lactation and that these cells are committed to lgA synthesis. During this period the interstitial tissue surrounding the ductal and alveolar epithelium is heavily infiltrated by mononuclear cells of the lymphocytic series (Fig lC). Prior to lactation, during alveolar development, the IE leukocyte population is relatively constant but it increases rapidly with the onset of lactation [28]. '

In labeling studies using 3H-thyrnidine-injected female rats despite the increase in IE lymphocytes evident during earl; lactation, the percentage of labeled lymphocytes was un-

TABLE I. Mesenteric lymph node weights from hybrid rats with isolated ileal segments

Group" Inoculum in isolated segment Mean MLN weights (mg ± SEM) p-Value

(5) 300 X 106 Fl 304 ± 15 I versus II lymphocytes < 0.001

II (5) 300 X 106 (FIX DA)F 1 159 ± 10 lymphocytes

Ill (5) 100 X 10" FI 209 ± 19 III versus IV lymphocytes <0.05

IV (5) 100 x lOr; (FI X DA)F, 150 ± 9 lymphocytes

v (6) unoperated 141 ± 7 V versus II and IV controls NS

"Number in parentheses is the number of animals per group. Abbreviations: MLN, mesenteric lymph node; FI strain, Fischer strain; NS, not significant.

TABLE II. Ileal mucosal heights from hybrid rats with isolated ileal segments

Mucosal heighls (mm)'' Gro upo Inocu lum in isolated segment

Intact il ea l Isola ted ileal segment (+ SEM) p-Value

segment(+ SEM)

(5) 300 X 106 FI lymphocytes

II (5) 300 X 106 (Fl >:: DA)F 1

lymphocytes Ill (5) 100 x 106 FI

lymphocytes IV (5) 100 X 106 (FI x DA)F 1

lymphocytes

"Number in parentheses is the number of animals per group. & Measured from the muscularis mucosa to the tips of the villi. ,. Intact ileum versus isolated ileum. Abbreviations: FI strain, Fischer strain; NS, not significant.

0.60 ± 0.01

0.46 ± 0.03

0.52 ± 0.02

0.45 ± 0.01

I versus II < 0.005

III versus IV < 0.05

0.56 ± 0.02

0.52 ± 0.02

0.56 ± 0.01

0.53 ± 0.02

p-Value

< 0.05c

< 0.05

NS

<0.05

July 1980

changed and actually decreased between 2 and 6 days post partum; perhaps. this cell populatio~ had previous!~ underg.one antigen stimulatwn and blastogenesis at a remote site (possibly in the gastrointestinal tract) and was redistributed to the mam­mary gland in response to the initiation of lactation.

With Beer, one of us (LS) [29] recently conducted a light and electron microscopic study on the mechanism whereby infil­trating leukocytes traverse the alveolar epithelium dming lac­tation in rats. Like the epithelial cells of the gut, adjacent alveolar epithelial cells are joined by uninterrupted apical oc­cluding junctional complexes, but are only loosely attached by desmosomes at their basal ends. Therefore, there appears to be no major anatomical barrier to penetration of the intercellular clefts by lymphocytes, nor does the mobility of IE leukocytes appear to be restricted by formation of junctional complexes with contiguous epithelial cells. The observation of pseudopo­dia! projections (Fig 1D) from IE leukocytes extending toward the luminal swface of the epithelium suggested migration in this direction. Intraepithelialleukocytes were often observed in association with alveolar myoepithelial cells. Perhaps activity on the part of these contractile cells facilitated the passage of leukocytes into the epithelium. How IE leukocytes disrupted the apical occluding junctions between epithelial cells to gain entrance to the alveolar lumina was not determined. A hint toward resolution of this question was the observation that in certain regions of alveoli the apical junctional complex between adjacent epithelial cells was missing and the resulting gap between epithelial cells was filled with an alveolar secretion in direct contact with a denuded basal lamina. The latter mem­brane appeared to prevent alveolar secretions from entering the extraalveolar connective tissue since secreted material was never observed in this compartment. Of particular in1portance was the observation that leukocytes were frequently present within these gaps, as well as in the ductal system of the mammary gland.

LYMPHOCYTES IN EPITHELIA OF THE MALE REPRODUCTIVE SYSTEM

Apart from being normal minority components of the epithe­liwn of the skin and gut and of alveolar epithelia of lactating mammary glands, lymphocytes appear to be present in practi­cally all other epithelia of surfaces that directly or indirectly open onto the surface of the body. However, IE lymphocytes in epithelia of the thyroid, biliary tract, urinary bladder, and submucosal glands of the bronchi have received little attention [30].

In the epithelia of the male reproductive system, lymphocytes occur in the terminal portions of the seminiferous tubules, where they join the tubuli recti, and in the tubuli recti them­selves, the rete testis, the ductuli efferentes, the epididymis, and the ductus deferens [30,31]. Of these sites epididymal epithelium has the most lymphocytes; they make up about 10% of its cell population. Unlike typical migratory lymphocytes, or those in the blood, many of these cells have granules containing multivesicular bodies and an abundant endoplasmic reticulum.

Lymphocytes have never been demonstrated in normal sem­iniferous epithelium, but monocytes and mast cells do occur among the myoid cells surrounding seminiferous tubules and their incidence appears to be influenced by the cyclical activity of the epitheliwn [32,33].

Most authorities attribute this absence of lymphocytes from seminiferous epithelium to the "blood-testis barrier," whose principal morphological basis is a series of junctional complexes between adjacent Sertoli cells. The functional significance of this barrier is generally held to be containment of potentially autoantigenic material associated with postmeiotic germ cells. However, it is difficult to reconcile this reasoning with the presence of IE lymphocytes in excwTent ducts that lack an effective permeability barrier and in which absorption of lu­minal contents, including antigenic material, does occm.

Lymphocytes also occm in increasing numbers within the

INTRAEPITHELIAL LYMPHOCYTES 87

epithelium of seminal vesicles and the prostate in aging rats [34).

LYMPHOCYTES IN EPITHELIA OF THE FEMALE REPRODUCTIVE TRACT

Like the respiratory and gastrointestinal tracts, the female reproductive tract is chronically exposed to antigenic material in the form of microorganisms and, episodically, to allogeneic semen.

It is well established that varying numbers of neutrophils are normally present in the female reproductive tract in most species and that emigration of neutrophils with mononuclear cells, which occms in response to the presence of sperm, is responsible for the phagocytosis and digestion of gametic re­mains [35,36]. In rodents it appears that leukocytic invasion of the uterine lumen from the mucosa results more from distension of this organ with fluid than from the presence of spermatozoa. No convincing evidence has been forthcoming that leukocytes retmn to tissues of the uterus from its lumen; however, there is indirect evidence that lymphocytes may have the capacity to escape from the uterine lumen and traverse its mucosa [37]; in rats intrauterine inoculation of viable lymphoid cells from al­logeneic donors elicited transplantation immunity and hyper­trophy of the draining lymph nodes, and inoculation of parental strain lymphoid cells per uterum in F, hybrid hosts initiated GVH reactions in the mucosa as well as in the draining nodes.

The maternal component of the placenta comprises decidual tissue that arises early in gestation from epithelioid transfor­mation of fibroblast-like connective tissue cells of the uterine stroma lying close to the invading trophoblast. In the mouse decidual cells from the mature placenta have been shown to have Ia antigens on their surface membranes and to be similar, morphologically, to Langerhans cells [38]. Do these cells play a role in antigen processing during implantation and placenta­tion? Could they process antigenic material of fetal origin in a manner that would prevent it from inciting unwanted immune responses?

These so-called HLA-DR or Ia alloantigens, unlike the fa­miliar transplantation antigens, have a restricted tissue distri­bution, i.e., they are limited to cells of the immune system. They appear to play a role in cooperative events among cells. As we have seen, some cells of the immune system, particularly lymphocytes, normally associate on a relatively large scale with epithelial cells of various types, notably in the intestine and lactating mammary glands. The report that Ia antigens are demonstrable by an indirect immunofluorescence method on a wide variety of epithelial cells [39] has raised the interesting possibility that these epithelial/lymphocytic associations (and possibly interactions) may be controlled by Ia antigens. IgA­secreting plasma cells are scarce in the mammary glands of virgin animals but abundant in glands under the influence of lactogenic hormone, and Ia antigen is expressed on mammary epithelial cells.

REFERENCES 1. Elves MW: The Lymphocytes. London, Lloyd-Luke, 1966 2. Ferguson A: Intraepitheliallyrnphocytes of the small intestine. Part

1. Morphology and experimental immunology of intraepithelial lymphocytes. Gut 18:921-937, 1977

3. Marsh MN: Studies of intestinal lymphoid tissue. I. Electron mi­croscopic evidence of "blast transformation" in epithelial lym­phocytes of small intestinal mucosa. Gut 16:665-682, 1975

4. Meader RD, Landers DF: Electron and light microscopic observa­tions on relationships between lymphocytes and intestinal epi­thelium. Am J Anat 121:763-774, 1967

5. Darlington D, Rogers AW: Epithelial lymphocytes in the small intestine of the mouse. J Anat 100:813-830, 1966

6. Ferguson A, Murray D: Quantitation of intraepithelia l lympho­cytes. Gut 12:988-994, 1971

7. Toner PG, Ferguson A: Intraepithelial cells in the human intestinal mucosa. J Ultrastruct Res 34:329-344 , 1971

8. Lemme! EM Fichtelius KE: Life span of lymphocytes within internal epithelium, Peyer's patch epithelium, epidermis and liver of mice. Int Arch Allergy Appl Immunol41:716-728, 1971

88 SEELIG AND BILLINGHAM

9. Ropke C, Everett NB: Proliferative kinetics of large and small intraepithelial lymphocytes in the small intestine of t he mouse. Am J Anat 145:395-408, 1976

10. Ropke C, Everett NB: Kinetics of intraepithelial lymphocytes in the small intestine of thymus-deprived mice and antigen-de­prived mice. Anat Rec 185:101-108, 1976

11. Rudzik 0, Bienenstock J: Isolation and characteristics of gut mu­cosal lymphocytes. Lab Invest 30:260-266, 1974

12. Rudzik 0, Clancy RL, Perey DYE, Bienenstock J, Singal DP: The distribution of a rabbit thymic antigen and membrane immuno­globulins in lymphoid tissue with special reference to mucosal lymphocytes. J lmmunol 114:1-4, 1975

13. Hall JG, Smith ME: Homing of lymph-borne immunoblasts to the gut. Nature 226:262-263, 1970

14. Ferguson A, Parrott DMV: The effect of antigen deprivation on thymus-dependent and thymus-independent lymphocytes in the small intestine of the mouse. Clin Exp Immunol12:477-488, 1972

15. Bockman DE, Cooper MD: Pinocytosis by epithelium associated with lymphoid follicles in the bursa of Fabricius, appendix, and Peyer's patches. An electron microscopic study. Am J Anat 136: 455-478, 1973

16. Owen RL, Jones AL: Epithelial cell specialization within human Peyer's patches: An ultrastructural study of intestinal lymphoid follicles. Gastroenterology 66:189-203, 1974

17. Fichtelius KE: The mammalian equivalent to bursa Fabricii of birds. Exp Cell Res 46:231-234, 1967

18. Beer AE, Billingham RE, Head JR: Natural transplantation of leukocytes during suckling. Transplant Proc 7:399-402, 1975

19. Billingham RE: The biology of graft-versus-host reactions. Harvey Lect 62:21-78, 1968

20. Levine S: Local and regional forms of graft versus host disease in lymph nodes. Transplantation 6:799-802, 1968

21. Husband AJ, Gowans JL: The origin and antigen-dependent distri­bution of IgA-containing cells in the intestine. J Exp Med 148: 1146-1160, 1978

22. Ferguson A, MacDonald TT: Effects of local delayed hypersensi­tivity on the small intestine. Ciba Found Symp 46:305-319, 1979

23. Head JR: Immunobiology of lactation. Semin Perinatol 1:195-210, 1977

24. Head JR, Beer AE, Billingham RE: Significance of the cellular component of the maternal immunologic endowment in milk. Transplant Proc 9:1465-1471, 1977

25. Parmely MJ, Reath DB, Beer AE, Billingham RE: Cellular immune

Vol. 75, No. 1

responses of human milk T lymphocytes to certain environmental antigens. Transplant Proc 9:1477-1483, 1977

26. Roux ME, McWilliams M, Phillips-Quagliata JM, Weisz-Carring­ton P, Lamm ME: Origin of lgA-secreting plasma cells in the mammary gland. J Exp Med 146:1311-1322, 1977

27. Rolf P: The cells of normal human urine. J Clin Pathol 8:25-31, 1955

28. Seelig LL Jr, Holt RG, Beer AE: Kinetics and transepithelial migration of leukocytes in rat mammary epithelium during preg­nancy and lactation, The Immunology of the Breast Milk. Edited by DH Dayton. New York, Raven Press, 1979, pp 154-172

29. Seelig LL Jr, Beer AE: Transepithelial migration of leukocytes in the mammary gland of lactating rats. Bioi Reprod 18:736-744, 1978

30. Dyrn M, Romrell LJ: Intraepithelial lymphocytes in the male reproductive tract of rats and rhesus monkeys. J Reprod Fertil 42:1-7, 1975

31. Hoffer AP, Hamilton DW, Fawcett DW: The ultrastructure of the principal cells and intraepithelialleukocytes in the initial segment of the rat epididymis. Am J Anat 175:169-202, 1973

32. Hermo L, Clermont Y: Light cells within the limiting membrane of rat seminiferous tubules. Am J Anat 145:467-484, 1976

33. Hermo L, Lalli M: Monocytes and mast cells in the limiting mem­brane of human seminiferous tubules. Bioi Reprod 19:92-100, 1978

34. Allison VF, Clearley GW: Electron microscopic study of cells within the epithelium of the seminal vesicle of aging rats. A nat Rec 172: 262, 1972 (abstract)

35. Austin CR: Fate of spermatozoa in the uterus of the mouse and rat. J Endocrinol 14:335-342, 1957

36. Moyer DL, Rimdusit S, Mishell DR: Sperm distribution and deg­radation in the human female reproductive tract. Obstet Gynecol 35:831-840, 1970

37. Beer AE, Billingham RE: Host responses to intrauterine tissue, cellular and fetal allografts. J Reprod Fertil [Suppl]21:59-83, 1974

38. Jenkinso.n EJ, Searle RF: Ia antigen expression on the developing mouse embryo and placenta. J Reprod Immunol1:1-10, 1979

39. Wiman K, Curman B, Forsum U, K.lareskog L, Malmaniis-Tjern­lund U, Rask L, Triigii.rdh L, Peterson P: Occurrence of Ia antigens of tissues on non-lymphoid origin. Nature 276:711-713, 1978