- Email: [email protected]
Thymic reticulum of autoimmune mice
Exp. Pathol. 29, 45-53 (1986) VEE Gustav Fischer Verhtg J ena INSERJ\I U25, Hopital Necker, 75015 Paris, France
Thymic reticulum of autoimmune mice 1. Ultr as tru ctural studi es of the di a betic (db/db) mouse th y mu s By B.
NA BA RRA
and 1.
ANDRIANARISON
With 4 figures (Received March 6, 1985)
Address for correspondence: Dr. B. NABARRA, INSERM U25, Hopital Neck er, 149 rue de Sllvres, F - 75015 Paris, France Key word s: thymic reticulum; diabetes; thymus; epithelial reUs; db/db lUke; ultrastructural study
Summary
The thymus of the db/db mouse, an a utoimmune strain with insulin-dependent diabet es mellitus and manifesting hyperglycemia, j)olyuria, glycosuria and obesity, was observed by electron microscopy. Its comparison with normal thymuses and thymuses from non-diabetic obese mice has revealed two major differences: first, the presence of crystal-like structures in some of the numerous clear vacuol es in cells , and second, a modific ation of the cells forming the border of cystic cavities. These cells contain two t ypes of granules which are quite plentiful. One variety is clear and big, and a second t ype is small and dense, with an aspect similar to that of secretion granules. It is hypothesized that the crystalloid formations found in epithelial cells, and the presenee of small, dense granulations in cells bordering the cystic cavities, are a consequence of the abnormal storage of the thymic hormone, thymulin, which results from a secretory function defect. These observations suggest a possible role for the thymic reticulum in th ymic function deficiencies. Introduction
The th ymus has a complex function that for the most part is mediated by substances secreted by epithelial cells, such as tIle thymic hormone, thymulin (1-5). The level of this hormone (previously called the serum thymic factor FTS) was decrea.sed in some stra.ins of mice with autoimmune pathology (2), and was accompanied by a decreased number of thymulin-containing cells labeled with monoclonal anti-thymulin antibodies and revealed by immunofluorescence (IF), (33, 34). These observations suggest a possible role for the th ymic reti culum in thymic function deficiencies. Different strains of mice with autoimmune disease present these th ymic dysfunctions (2, 12, 33), and it is interesting to look at the morphological aspects of the thymic reticulum in these autoimmune mi ce, since an observed alteration may permit a better comprehension of the development of autoimmune dis ease. In this study, we have examined db/db mice with autoimmune diabetes. Th e mutant diabetic mou se (db/db) appeared in the C56BL/6 strain and was first described by HUMMEL (23, 24). It is eharacterized by an insulin-dependent diabet es mellitus with hyperglycemia, polyuria, glycosuria and obesity. The manifestation of the diabetes is accompanied by his-
46
tological alterations in the endocrine cell populations of the pancreatic islets and other modifications of endocrine functions (6, 7, 8, 27, 37). In these mice, recent studies have shown that immune reactions are altered including the development of autoantibodies (17, 26), and it has been designated as an autoimmune strain. Lastly, M. DEERA.Y-SACHS (13) has shown that db/db mice possess cytotoxic spleen T cells capable of inhibiting insulin release by islet cells. The morphology of the sb/sb mouse thymus has not been extensively examined to date. Only preliminary observations were reported in a previous paper (12). This study presents a more extensive description of the thymic reticulum of db/db mice. J1 aterials and 111ethods Mic e: Homozygous (db/db) aud heterozygous (db+/db) mice, C57BL/6 control mice and obese mice without diabetes aged 15 d, 4 weeks and 2, 4 and 6 months old were purchased from C.S.E.A.L. Orleans-La Source (France). The thymuses were excised after Nembutal anesthesia and prepared for electron microscopic observation. Technical preparation for electron microscopy. Thymic fragments were fixed, immediately after excision, for 1 h in 2 % glutaraldehyde in phosphate buffer, pH 7.4, and after washing in buffer, post fixed for 1 h in 1 % osmium tetroxide. Then these fragments were dehydrated in serial dilutions of alcohol and embedded in Epon. 900 A sections were cnt with a diamond knife, collected on copper grids and stained with uranyl acetate and lead citrate. Four animals were observed in each group for pach series. Some ultrathin sections were collected on glass slides, stained with silver methionate (29) and observed by light microscopy for the topographical study of the tissue.
Fig. 1. General aspect of thymic reticulum in normal mice. Thp micrograph shows ,t type I epithelial cell (EP) with clear vacuoles containing very dense granules (--+). These tells are compressed by numerous lymphocytes. In the upper left hand corner we observe an interdigited reticular cell (IRC). In db/db mice, the same image is observed. x 10,500.
46
Exp. Path. 29 (1986) 1
Resttlts By light microscopy, the topography and morphological aspect of the thymus is normal at an early age (15 d) and at 1-2 months. At 4 and 6 months the thymic cortex is greatly reduced, due, for the most part, to a depletion of the lymphoid population. All observations of the thymic reticulum of autoimmune mice are compared, first with the normal aspect of the thymus, previously described by numerous authors and with our description (B. NABARRA - in preparation) which is summarized briefly in the following text. Second, additional comparisons are made with heterozygote mice of the same strain and non diabetie obese miee. By electron microseopy, two major modifications are found in these thymuses: the first, eoneerning the epithelial eells, and the second, involving the cystic formations.
1. Epithelial cells The mouse thymic retieulum is formed by several types of eells, among which there are three types of epithelial cells which have intracytoplasmic tonofilaments, form desl11osomes and show cytoplasmie specialization. Type I epithelial cells are present in both the cortex and the medulla. In db/db mice, as in normal mice, these cells are stellate in form and they are eharacterized by the presence of dear intracytoplasmic vacuoles with a matrix which seems to be empty. They contain dense granulations, often located next to the membrane, and are sometimes associated with membrane fragments and small empty vesicles (fig. 1). In some of these granules, we have found traces of zinc (an essential eomponent of thymulin) (30). In the homozygous db/db mouse thymus a most important anomaly is observed: in some cases these clear vacuoles contain an enormous crystal-like inclusion. This crystalloid inclusion has a hexagonal or diamond shape. It is electron dense and homogeneous, sometimes having a central area a little less dense with a granular appearance. It is present either
Fig. 2. Aspect of the crystalloid strueturc found in type I epithelial cells in db/db mice. a. Homogeneous crystal-like structure ('t') seen in clear vacuoles containing dense granules (-->- tonofilaments). x 15,500. b. Crystal-like structure with a diamond shape with a central non homogeneous area (-->- tonofilaments). x 23,000. Exp. Path. 29 (1986) 1
47
alone in the vacuole, or associated with the small dense granules which are normal in this type of vacuole, or with a homogeneous electron dense material (figs. 2a, b). Type II and type III epithelial cells are observed only in the medulla. Type II shows, in addition to the epithelial characteristics, a peculiar intracytoplasmic network of empty cavities surrounded by a double membrane which forms very complicated labyrinths. Numerous microvilli are observed in these alveoli. In db/db mice, in contrast to the normal situation, a filamentous substance associated with a finely granular material is often observed in these cavities. Some of these type II cells present large clumps of glycogen in their cytoplasm. This ovservation suggests that these cells are young cells, or early mature cells. Although it is very difficult to count cells in eleetron microscopy, it seems that the number of type II cells is slightly increased in the pathological mouse. Type III epithelial cells, possess a round intracytoplasmic cavity lined by microvilli or by cilia. These cavities are filled with a fine dense material associated with other more gramIlar and filamentous material more often in autoimmune db/db thymuses than in normal ones. 2. Interdigitated reticular cells (IRC) and macrophages In addition to the epithelial cells in the reticulum, two types of non-epithelial cells with phagocytic function, i.e., the interdigitated reticular cell (IRC) and common lllacrophages are found. In normal mice, IRC are sparse and located at the corti co-medullary junction and in the medulla. They are large cells with a clear cytoplasm, a few Iysosomes and intracytoplasmic granulation. The shape of the cell is irregular with a scalloped aspect. In autoimmune db/db mouse thymuses, numerous IRC were observed. Some of these cells, in contrast to normal mice, contain the Birbeck corpuscles noted in other species. The cytoplasm contains various dense granules, and in some cases, the granules present an electron dense core which can easily be distinguished from the granular matrix. The common macrophages, present in the cortex and medulla, contain numerous Iysosomes and their phagolysosomes often contain pyknotic cells. They exhibit no morphological differences between autoimmune db/db mice and normal animals. 3. Cellular associations In the thymic reticulum, in addition to these various types of cells forming the network, it is possible to observe cell associations in which epithelial cells are implicated, i.e., Hassal's corpuscles and cystic formations involving several cells around an irregular lumen. Hassal's corpuscles are small and rare in the mouse. They are formed by a keratinized central core and concentric epithelial cell layers. The cystic cavities are often developed in the mouse. Their lumens, often empty, are bordered by cells whose apical zones are differentiated into microvilli and/or cilia. a) HassaI's corpuscles Under light microscopy, the number of Hassal's corpuscles appears to be increased and they are more developed in diabetic mice than normal ones. By electron microscopy, they are highly developed with several concentric layers and have an intensely keratinized central core. The early stages of Hassal's corpuscles are also present. Indeed, we often observed epithelial cells with dense bundles of tonofilaments arranged in concentric layers around the cellular nucleus without peripheral concentric layers of cells. b) Cystic cavity The most striking feature of these "diabetic" thymuses is the presence of many cystic cavities whose morphological aspect is different from those of the normal monse thymus. These cavities are lined by two types of alternating cells. The first type is large and compJetely filled with moderately electron dense, round, large granulations (fig. 3 Gl, 4). In addition to this cellular element, a highly developed endoplasmic reticulum with numerous
48
Exp. Path. 29 (198G) 1
Fig. 3. The cystic cavity in the db/db mouse thymic reticulum: Around the lumen of the cavity two types of cells, distinguished by the type of granules present in their cytoplasm, are seen. The first type contains big elear granules (G 1 ) whose morphological aspect is similar to that of the mucous sf)cretions and the second type includes small dense granules (G 2 ) which seem to be secretory granules. (_ desmosomes). The apical zone of these cells has microvillosities. x 25,500.
rough lamellae and free ribosomes and a few mitochondria, is observed. Tonofilaments are not found in these cells. This cell type alternates with another type of cell whose cytoplasm contains tonofilaments. These cells form typical desmosomes with the cells described above. Often, the cytoplasm is reduced and completely wedged between the other cells. In some cases, these epithelial cells have a large number of dense, small granules containing a fine homogenous material whose aspect is different from those previously described (fig. 3 G2). These cells are arranged together around the lumen which is often elongated or has an irregular form. The apical area of these cells is lined with short microvilli, or by groups of cilia. It is not surrounded by a membrane in the basal area and is continuous with the reticulum. In SOllle cases, a clear fine material is seen in the lumen and sometimes a few lymphocytes. In one case, we observed some crystal-like inclusions in an epithelial cell bordering the lumen. 4. Controls IV e compared these observations with thymuses from heterozygote mice of the same strain, and with a strain which is obese, but not diabetic. In these two types of thymuses no significant modifications were clearly found in comparison with normal ones. Cells with lipidic inclusions were observed infiltrating the thymuses of obese mice, but uo cristalloid vacuoles, nor granulations in the cystic eavities were found. 4
Exp. Path. 29 (1986) 1
49
Fig. 4. Cells of the first type contain numerous clear granulations (G1 ). Other cells with tonofilments alternate with the first type of cell (~) and form desmosomes. These epithelial cells are completely compressed against the G1 cells. MicrociUosities are observed in the lumen of the cystic cavity. x 10,000.
Discussion
As it has been said previously (12), a progressive lowering of the circulating thymulin level has been noted in the db/db mouse, which is accompanied by a reduced number of positive cells as assessed by a monoclonal anti-thymulin antibody. Under light microscopy, the aspect of the "diabetic" thymus appears inversed, due to the reduction of the cortex. The number of Hassal's corpuscles seems to be increased and more cystic cavities are seen. These phenomena are noted as early as the fourth month, but are more evident at 6 months (12). By electron microscopy, two basic alterations are observed in the thymic reticulum. Firstly, crystal-like structures are often noted in the clear vacuoles of the type I epithelial cells. The existence of these crystals seems to be important, if we take into consideration the secretory function of the cells in question. Crystalloid forms have been observed in other types of cells, in other organs (16,19,36) and have been interpreted as being concentrated proteins. Thus, it can be hypothesized that these crystals observed in epithelial cells of the db/db mouse thymus could be molecules of hormones produced by these cells (with a possible defective conformation] being responsible for the crystalization) concentrated and stored, as a result of a secretory defect. Hence, this could explain the reduced level of circulating hormone. A similar observation has been noted by SCHMITT (36) in the thymus of Swan mice, an autoimmune strain that presents a lupus-like pathology. According to this author as well, a defect exists in the secretory function of the cell. The decreased number of cells producing thymulin can be also explained by this secretory defect. Indeed, it could be thought that the secretory defect and the aceumulation of the
50
Exp. Path. 29 (1986) 1
hormone could become a signal for the cell which, is henceforth programmed to no longer produce the hormone, or to diminish the number of producing cells. This system of retroregulation has recently been demonstrated in mouse (35) and human thymic cultures (10). Concerning the peculiar aspect of type II and III epithelial cells, it is difficult to suggest any hypothesis, since their role as yet is unknown. However, if they participate in the secretory function of the gland, perhaps the material observed in their cavities is a cell product that is not secreted because of a defect in this function. The second important morphological observation that we have described in these db/db mouse thymuses is the increased number of cystic cavities whose cells contain abnormal granules. The existence of cystic cavities, without abnormal granules, is rarely described in animals or man (9, 22, 31, 38). We have observed several of these cavities in normal C57BL/6 and CBA mice. Some authors have reported that their number and size increase after cortisone treatment, depending upon the endocrine status of the animal (14), or following stress or extensive septicemia in man (31). - Furthermore, the aged thymus and thymus of "nude" mice are rich in these cystic formations (11, 28). CORDIER (11) demonstrated in the "nude" mouse that these structures are of endodermic origin, and that during the embryological formation of the thymus the association of endoderm and ectoderm organizes the regulation of the development of the different elements of the reticulum. Thus, when the ectoderm (i.e., for a part the epithelial cells of the thymus) is lacking, the endoderm develops in an uncontrolled manner and forms these cystic cavities. It could be thought that in the db/db mouse, the acquired deficiency of the thymic epithelium, which does not seem to function correctly, leads to an imbalance between the two cOlllPonents of the thymus, and allows the proliferation of the cystic elelllent of the reticulum. The cells bordering these cystic cavities in normal mice do not contain granUlations. On the contrary, in db/db mouse thymuses, we observed two types of granulations. The first type is a big, moderately electro dense granule, and at first sight, it appears to have a morphological aspect similar to the granulations of the mucoid secreting intestinal "goblet" cells. However, granulations of this type are also observed in normal birds (18, 25), in nude mice (11) and in AKR mice infected with leukemia virus (32). Thymuses from this latter strain of mice have been studied by P ANSKY and HOUSE (32), and their suggestion concerning the origin of these granulation appears to be interesting. Indeed, these authors noted the similar morphology of these granular cells and the pancreatic islets of Langerhans. Their explanation for this resemblance is based upon embryological studies suggesting that both of these cell types arise from endoderm. Furthermore, they think that these cells produce a factor directly involved in the metabolism of carbohydrates. This substance could be an insulinase inhibitor or a compound having an activity imitating insulin. From our observations, the significance of these granulations remains unknown. In this respect, it is interesting to note that EISENBARTH et al. (15) have recently discovered two anti-pancreatic islet antibodies which react with subsets of thymic reticulum cells. For these authors, this observation indicates that the autoimmunity of diabetes may be extended to the thymus. The second type of granule found in the epithelial cell bordering the cystic cavities, presents a different morphological aspect. These granules are smaller and denser, and sometimes have a dense central core with a clear halo around its periphery. To the best of oUT knowledge, this type of granulation in the cells forming the cystic cavities has rarely been described previously (20, 21). The morphological aspect of these granules is characteristic of secretion granules, and it could be thought that in this instance as well, we are dealing with a stored and perhaps not secreted factor. Thus, the functional deficit of the epithelial cells would be a defect in the excretion of the factors produced, which would be expressed in two forms: either intravacuolar crystals, or groups of dense small granules spread throughout the cytoplasm of the cells forming the eystic cavity. With this hypothesis, we are in agreement with HIROKAWA (20, 21), who suggested that certain granulations that he observed in aged mouse thymuses could con4*
Exp. Path. 29 (1986) 1
51
sist of stored and denatured thymic hormone. However, we find epithelial cells with dense granules only in the cystic cavities and not strewn throughout the reticulum. How can this phenomenon be explained? ,Ve would like to suggest that the cystic cavities, whose role is unknown, are perhaps the site of regroupment for degenerating cells, different from Hassal's corpuscles. In the same area, the epithelial cells are found to be modified, no longer capable of playing their roles in the microenvironment: the education of T cells, and the self-recognition (3, 4). Thus, this acquired incompetenee may initiate the autoimmune phenomena and the development of lymphoid cells defieient in their funetional capacities or their self control. This relationship is open to speculation, and as lH. DARDENNE (12) suggested, it is possible that the first decrease in the level of thymulin produced, acts upon the loss of T cell suppressive functions, which have, as a consequence, the establishment of anti-pancreatic autoimmunity expressed as antibodies and cytotoxic lymphocytes.
Acknowledgements We thank T. SACKSICK for his excellent photographic assistance, and J. J ACOESEN for reviewing the manuscript.
Literature 1. BACH, J. F., a,nd :JL DAHDENNE, Studies on thymus products. Demonstration and charactprization of a circulating thymic hormone. Immunology 25, 353 (1973). 2. - - J. ]\f. PLEAU and lIL A. BACH, Isolation, biochemical characteristics and biological activity of a circulating thymic hormone in the mouse and in the man. Ann. N.Y. Acad. Sci. 249, 186 (1975). 3. - :XI. A. BACH, C. CARlC,"UD, ]\1. DARDENNE and J. CL. MONIER, Thymic hormones and autoimmunity. In: Autoimlllunity - Genetic, Immunologic, Virologic and Clinical Aspects (ed. N. TALAL). Academic Press, New York 1977, p. 207. 4. BACH, M. A., and J. CHARREIRE, Role of a circulating thymic f11etor in self recognition and self tolerance. Ann. N.Y. Acad. Sci. 332, 55 (1979). 5. BACH, J. F., Thymulin (FTS-Zn). In: Clinics in Immunology and Allergy. W. B. Saunders, New York 1983, p. 133. 6. BAETENS, D., Y. STEFAN, l\I. RAVAZZOLA, F. MALAISSE-LAGEA, D. L. COLEiI'IAN and L. ORCI, Alteration of islet cell populations in spontaneously diabetic mice. Diabetes 27, 1 (1978). 7. BERGLUND, 0., B. J. FRANKEL and B. HELL~IANN, Development of the insulin secretory defect in genetically diabetic (db/db) mouse. Acta Endocrinol. 87, 543 (1978). 8. BOQUIST, L., B. HELLMANN, A. LERNMARK and J. 13. T;\LEDAL, Influence of the mutation "diabetes" on insulin release and islet morphology in mice of different genetic backgrounds. J. Cell BioI. 62, 77 (1974). 9. CHERRY, C., R. EISENSTEIN and A. GLUCKSMAN, Epithelia.l cords and tubules of the rat thymus. Brit. J. Exp. Path. 48, 90 (1967). 10. COHEN, S., S. BEIUUH, l\I. DAHDENNE and J. F. BACH, Regulation in vitro de la secretion de thymuline par les cellules epithBliales thymiques humaines. CR Acad. Sci. (Paris) 297 (serie III) G3 (1983). 11. CORDIER, A., Ultrastructure of the thymus in "nude" mice. J. Ultrastruct. Res. 47, 26 (1974). 12. DARDENNE, l\I., W. SAVINO, L. M. GASTINAL, B. NABARRA and J. F. BACH, Thymic dysfunction in the mutant diabetic (db/db) mouse. J. Immunol. 130, 1195 (1983). 13. DEBRAy-SACHS, M., P. SAl, C. J3OJTARD, R. ASSAN and J. HAMBURGER, Anti-panereatic immunity in genetically diabetic mite. Clin. Exp. Immunol. 51, 1 (1983). 14. EBBESEN, P., and M. NIELSEN, Thymic cysts in oestrogenised Balb/c mice. Acta Pathol. JliIicrobioI. Seand. 80, 211 (1972). 15. EISENBARTH, G., R. JACKSON, S. SHIKANTAS, A. POWERS, J. BusE and H. JliIORr, Utilization of monoelonal antibodies techniques to study type I (insulin-dependent) diabetes mellitus. In: Immunology in Diabetes, ed. ANDllEANI, DI ::I1AHIO and FEDERLIN. Kripton Publications, London 1984, p. 143. 16. FAWCETT, D., The Cell. Saunders Comp., London 1969, p. 319. 17. FEHNAN"DES, G., B. S. HANDWEllGER, E. J. YUNIS 11,nd D. ]VI. BnowN, Immune responses in the mutant diabetic C57BL/Ks-db+ mouse. Discrepancies between in vitro a.ndin vivo immunobiological assays. J. Clin. Invest. 61, 243 (1978). 18. FRAZIEn, J. Ultrastructure of the chick thymus. Z. ZeIlfOTsch.136, 191 (1973).
52
Exp. Path. 29 (1986)
1
19. HAMILTON, D., D. FAWCETT and A. CHRISTENSEN, The liver of the slender salamander. 1. The structure of its crystalline inclusions. Z. Zellforsch. 70, 347 (1966). 20. HIROKAWA, K., Age-related changes of thymus. Morphological and functional aspects. Acta Path. Jap. 28, 843 (1978). 21. - J. MCCLURE and A. GOLDSTEIN, Age-related changes in localization of thymosin in the human thymus. Thymus 4, 19 (1982). 22. HOSHINO, T., The fine structure of ciliated vesicle containing reticular cells in the mouse thymus. Exp. Cell Res. 27, 615 (1962). 23. HUMMEL, K. P., 1\1. M. DICKIE a,nd D. L. COLEMAN, Diabetes, a new mut.
Exp. P
[);3
Thymic reticulum of autoimmune mice 1. Ultr as tru ctural studi es of the di a betic (db/db) mouse th y mu s By B.
NA BA RRA
and 1.
ANDRIANARISON
With 4 figures (Received March 6, 1985)
Address for correspondence: Dr. B. NABARRA, INSERM U25, Hopital Neck er, 149 rue de Sllvres, F - 75015 Paris, France Key word s: thymic reticulum; diabetes; thymus; epithelial reUs; db/db lUke; ultrastructural study
Summary
The thymus of the db/db mouse, an a utoimmune strain with insulin-dependent diabet es mellitus and manifesting hyperglycemia, j)olyuria, glycosuria and obesity, was observed by electron microscopy. Its comparison with normal thymuses and thymuses from non-diabetic obese mice has revealed two major differences: first, the presence of crystal-like structures in some of the numerous clear vacuol es in cells , and second, a modific ation of the cells forming the border of cystic cavities. These cells contain two t ypes of granules which are quite plentiful. One variety is clear and big, and a second t ype is small and dense, with an aspect similar to that of secretion granules. It is hypothesized that the crystalloid formations found in epithelial cells, and the presenee of small, dense granulations in cells bordering the cystic cavities, are a consequence of the abnormal storage of the thymic hormone, thymulin, which results from a secretory function defect. These observations suggest a possible role for the thymic reticulum in th ymic function deficiencies. Introduction
The th ymus has a complex function that for the most part is mediated by substances secreted by epithelial cells, such as tIle thymic hormone, thymulin (1-5). The level of this hormone (previously called the serum thymic factor FTS) was decrea.sed in some stra.ins of mice with autoimmune pathology (2), and was accompanied by a decreased number of thymulin-containing cells labeled with monoclonal anti-thymulin antibodies and revealed by immunofluorescence (IF), (33, 34). These observations suggest a possible role for the th ymic reti culum in thymic function deficiencies. Different strains of mice with autoimmune disease present these th ymic dysfunctions (2, 12, 33), and it is interesting to look at the morphological aspects of the thymic reticulum in these autoimmune mi ce, since an observed alteration may permit a better comprehension of the development of autoimmune dis ease. In this study, we have examined db/db mice with autoimmune diabetes. Th e mutant diabetic mou se (db/db) appeared in the C56BL/6 strain and was first described by HUMMEL (23, 24). It is eharacterized by an insulin-dependent diabet es mellitus with hyperglycemia, polyuria, glycosuria and obesity. The manifestation of the diabetes is accompanied by his-
46
tological alterations in the endocrine cell populations of the pancreatic islets and other modifications of endocrine functions (6, 7, 8, 27, 37). In these mice, recent studies have shown that immune reactions are altered including the development of autoantibodies (17, 26), and it has been designated as an autoimmune strain. Lastly, M. DEERA.Y-SACHS (13) has shown that db/db mice possess cytotoxic spleen T cells capable of inhibiting insulin release by islet cells. The morphology of the sb/sb mouse thymus has not been extensively examined to date. Only preliminary observations were reported in a previous paper (12). This study presents a more extensive description of the thymic reticulum of db/db mice. J1 aterials and 111ethods Mic e: Homozygous (db/db) aud heterozygous (db+/db) mice, C57BL/6 control mice and obese mice without diabetes aged 15 d, 4 weeks and 2, 4 and 6 months old were purchased from C.S.E.A.L. Orleans-La Source (France). The thymuses were excised after Nembutal anesthesia and prepared for electron microscopic observation. Technical preparation for electron microscopy. Thymic fragments were fixed, immediately after excision, for 1 h in 2 % glutaraldehyde in phosphate buffer, pH 7.4, and after washing in buffer, post fixed for 1 h in 1 % osmium tetroxide. Then these fragments were dehydrated in serial dilutions of alcohol and embedded in Epon. 900 A sections were cnt with a diamond knife, collected on copper grids and stained with uranyl acetate and lead citrate. Four animals were observed in each group for pach series. Some ultrathin sections were collected on glass slides, stained with silver methionate (29) and observed by light microscopy for the topographical study of the tissue.
Fig. 1. General aspect of thymic reticulum in normal mice. Thp micrograph shows ,t type I epithelial cell (EP) with clear vacuoles containing very dense granules (--+). These tells are compressed by numerous lymphocytes. In the upper left hand corner we observe an interdigited reticular cell (IRC). In db/db mice, the same image is observed. x 10,500.
46
Exp. Path. 29 (1986) 1
Resttlts By light microscopy, the topography and morphological aspect of the thymus is normal at an early age (15 d) and at 1-2 months. At 4 and 6 months the thymic cortex is greatly reduced, due, for the most part, to a depletion of the lymphoid population. All observations of the thymic reticulum of autoimmune mice are compared, first with the normal aspect of the thymus, previously described by numerous authors and with our description (B. NABARRA - in preparation) which is summarized briefly in the following text. Second, additional comparisons are made with heterozygote mice of the same strain and non diabetie obese miee. By electron microseopy, two major modifications are found in these thymuses: the first, eoneerning the epithelial eells, and the second, involving the cystic formations.
1. Epithelial cells The mouse thymic retieulum is formed by several types of eells, among which there are three types of epithelial cells which have intracytoplasmic tonofilaments, form desl11osomes and show cytoplasmie specialization. Type I epithelial cells are present in both the cortex and the medulla. In db/db mice, as in normal mice, these cells are stellate in form and they are eharacterized by the presence of dear intracytoplasmic vacuoles with a matrix which seems to be empty. They contain dense granulations, often located next to the membrane, and are sometimes associated with membrane fragments and small empty vesicles (fig. 1). In some of these granules, we have found traces of zinc (an essential eomponent of thymulin) (30). In the homozygous db/db mouse thymus a most important anomaly is observed: in some cases these clear vacuoles contain an enormous crystal-like inclusion. This crystalloid inclusion has a hexagonal or diamond shape. It is electron dense and homogeneous, sometimes having a central area a little less dense with a granular appearance. It is present either
Fig. 2. Aspect of the crystalloid strueturc found in type I epithelial cells in db/db mice. a. Homogeneous crystal-like structure ('t') seen in clear vacuoles containing dense granules (-->- tonofilaments). x 15,500. b. Crystal-like structure with a diamond shape with a central non homogeneous area (-->- tonofilaments). x 23,000. Exp. Path. 29 (1986) 1
47
alone in the vacuole, or associated with the small dense granules which are normal in this type of vacuole, or with a homogeneous electron dense material (figs. 2a, b). Type II and type III epithelial cells are observed only in the medulla. Type II shows, in addition to the epithelial characteristics, a peculiar intracytoplasmic network of empty cavities surrounded by a double membrane which forms very complicated labyrinths. Numerous microvilli are observed in these alveoli. In db/db mice, in contrast to the normal situation, a filamentous substance associated with a finely granular material is often observed in these cavities. Some of these type II cells present large clumps of glycogen in their cytoplasm. This ovservation suggests that these cells are young cells, or early mature cells. Although it is very difficult to count cells in eleetron microscopy, it seems that the number of type II cells is slightly increased in the pathological mouse. Type III epithelial cells, possess a round intracytoplasmic cavity lined by microvilli or by cilia. These cavities are filled with a fine dense material associated with other more gramIlar and filamentous material more often in autoimmune db/db thymuses than in normal ones. 2. Interdigitated reticular cells (IRC) and macrophages In addition to the epithelial cells in the reticulum, two types of non-epithelial cells with phagocytic function, i.e., the interdigitated reticular cell (IRC) and common lllacrophages are found. In normal mice, IRC are sparse and located at the corti co-medullary junction and in the medulla. They are large cells with a clear cytoplasm, a few Iysosomes and intracytoplasmic granulation. The shape of the cell is irregular with a scalloped aspect. In autoimmune db/db mouse thymuses, numerous IRC were observed. Some of these cells, in contrast to normal mice, contain the Birbeck corpuscles noted in other species. The cytoplasm contains various dense granules, and in some cases, the granules present an electron dense core which can easily be distinguished from the granular matrix. The common macrophages, present in the cortex and medulla, contain numerous Iysosomes and their phagolysosomes often contain pyknotic cells. They exhibit no morphological differences between autoimmune db/db mice and normal animals. 3. Cellular associations In the thymic reticulum, in addition to these various types of cells forming the network, it is possible to observe cell associations in which epithelial cells are implicated, i.e., Hassal's corpuscles and cystic formations involving several cells around an irregular lumen. Hassal's corpuscles are small and rare in the mouse. They are formed by a keratinized central core and concentric epithelial cell layers. The cystic cavities are often developed in the mouse. Their lumens, often empty, are bordered by cells whose apical zones are differentiated into microvilli and/or cilia. a) HassaI's corpuscles Under light microscopy, the number of Hassal's corpuscles appears to be increased and they are more developed in diabetic mice than normal ones. By electron microscopy, they are highly developed with several concentric layers and have an intensely keratinized central core. The early stages of Hassal's corpuscles are also present. Indeed, we often observed epithelial cells with dense bundles of tonofilaments arranged in concentric layers around the cellular nucleus without peripheral concentric layers of cells. b) Cystic cavity The most striking feature of these "diabetic" thymuses is the presence of many cystic cavities whose morphological aspect is different from those of the normal monse thymus. These cavities are lined by two types of alternating cells. The first type is large and compJetely filled with moderately electron dense, round, large granulations (fig. 3 Gl, 4). In addition to this cellular element, a highly developed endoplasmic reticulum with numerous
48
Exp. Path. 29 (198G) 1
Fig. 3. The cystic cavity in the db/db mouse thymic reticulum: Around the lumen of the cavity two types of cells, distinguished by the type of granules present in their cytoplasm, are seen. The first type contains big elear granules (G 1 ) whose morphological aspect is similar to that of the mucous sf)cretions and the second type includes small dense granules (G 2 ) which seem to be secretory granules. (_ desmosomes). The apical zone of these cells has microvillosities. x 25,500.
rough lamellae and free ribosomes and a few mitochondria, is observed. Tonofilaments are not found in these cells. This cell type alternates with another type of cell whose cytoplasm contains tonofilaments. These cells form typical desmosomes with the cells described above. Often, the cytoplasm is reduced and completely wedged between the other cells. In some cases, these epithelial cells have a large number of dense, small granules containing a fine homogenous material whose aspect is different from those previously described (fig. 3 G2). These cells are arranged together around the lumen which is often elongated or has an irregular form. The apical area of these cells is lined with short microvilli, or by groups of cilia. It is not surrounded by a membrane in the basal area and is continuous with the reticulum. In SOllle cases, a clear fine material is seen in the lumen and sometimes a few lymphocytes. In one case, we observed some crystal-like inclusions in an epithelial cell bordering the lumen. 4. Controls IV e compared these observations with thymuses from heterozygote mice of the same strain, and with a strain which is obese, but not diabetic. In these two types of thymuses no significant modifications were clearly found in comparison with normal ones. Cells with lipidic inclusions were observed infiltrating the thymuses of obese mice, but uo cristalloid vacuoles, nor granulations in the cystic eavities were found. 4
Exp. Path. 29 (1986) 1
49
Fig. 4. Cells of the first type contain numerous clear granulations (G1 ). Other cells with tonofilments alternate with the first type of cell (~) and form desmosomes. These epithelial cells are completely compressed against the G1 cells. MicrociUosities are observed in the lumen of the cystic cavity. x 10,000.
Discussion
As it has been said previously (12), a progressive lowering of the circulating thymulin level has been noted in the db/db mouse, which is accompanied by a reduced number of positive cells as assessed by a monoclonal anti-thymulin antibody. Under light microscopy, the aspect of the "diabetic" thymus appears inversed, due to the reduction of the cortex. The number of Hassal's corpuscles seems to be increased and more cystic cavities are seen. These phenomena are noted as early as the fourth month, but are more evident at 6 months (12). By electron microscopy, two basic alterations are observed in the thymic reticulum. Firstly, crystal-like structures are often noted in the clear vacuoles of the type I epithelial cells. The existence of these crystals seems to be important, if we take into consideration the secretory function of the cells in question. Crystalloid forms have been observed in other types of cells, in other organs (16,19,36) and have been interpreted as being concentrated proteins. Thus, it can be hypothesized that these crystals observed in epithelial cells of the db/db mouse thymus could be molecules of hormones produced by these cells (with a possible defective conformation] being responsible for the crystalization) concentrated and stored, as a result of a secretory defect. Hence, this could explain the reduced level of circulating hormone. A similar observation has been noted by SCHMITT (36) in the thymus of Swan mice, an autoimmune strain that presents a lupus-like pathology. According to this author as well, a defect exists in the secretory function of the cell. The decreased number of cells producing thymulin can be also explained by this secretory defect. Indeed, it could be thought that the secretory defect and the aceumulation of the
50
Exp. Path. 29 (1986) 1
hormone could become a signal for the cell which, is henceforth programmed to no longer produce the hormone, or to diminish the number of producing cells. This system of retroregulation has recently been demonstrated in mouse (35) and human thymic cultures (10). Concerning the peculiar aspect of type II and III epithelial cells, it is difficult to suggest any hypothesis, since their role as yet is unknown. However, if they participate in the secretory function of the gland, perhaps the material observed in their cavities is a cell product that is not secreted because of a defect in this function. The second important morphological observation that we have described in these db/db mouse thymuses is the increased number of cystic cavities whose cells contain abnormal granules. The existence of cystic cavities, without abnormal granules, is rarely described in animals or man (9, 22, 31, 38). We have observed several of these cavities in normal C57BL/6 and CBA mice. Some authors have reported that their number and size increase after cortisone treatment, depending upon the endocrine status of the animal (14), or following stress or extensive septicemia in man (31). - Furthermore, the aged thymus and thymus of "nude" mice are rich in these cystic formations (11, 28). CORDIER (11) demonstrated in the "nude" mouse that these structures are of endodermic origin, and that during the embryological formation of the thymus the association of endoderm and ectoderm organizes the regulation of the development of the different elements of the reticulum. Thus, when the ectoderm (i.e., for a part the epithelial cells of the thymus) is lacking, the endoderm develops in an uncontrolled manner and forms these cystic cavities. It could be thought that in the db/db mouse, the acquired deficiency of the thymic epithelium, which does not seem to function correctly, leads to an imbalance between the two cOlllPonents of the thymus, and allows the proliferation of the cystic elelllent of the reticulum. The cells bordering these cystic cavities in normal mice do not contain granUlations. On the contrary, in db/db mouse thymuses, we observed two types of granulations. The first type is a big, moderately electro dense granule, and at first sight, it appears to have a morphological aspect similar to the granulations of the mucoid secreting intestinal "goblet" cells. However, granulations of this type are also observed in normal birds (18, 25), in nude mice (11) and in AKR mice infected with leukemia virus (32). Thymuses from this latter strain of mice have been studied by P ANSKY and HOUSE (32), and their suggestion concerning the origin of these granulation appears to be interesting. Indeed, these authors noted the similar morphology of these granular cells and the pancreatic islets of Langerhans. Their explanation for this resemblance is based upon embryological studies suggesting that both of these cell types arise from endoderm. Furthermore, they think that these cells produce a factor directly involved in the metabolism of carbohydrates. This substance could be an insulinase inhibitor or a compound having an activity imitating insulin. From our observations, the significance of these granulations remains unknown. In this respect, it is interesting to note that EISENBARTH et al. (15) have recently discovered two anti-pancreatic islet antibodies which react with subsets of thymic reticulum cells. For these authors, this observation indicates that the autoimmunity of diabetes may be extended to the thymus. The second type of granule found in the epithelial cell bordering the cystic cavities, presents a different morphological aspect. These granules are smaller and denser, and sometimes have a dense central core with a clear halo around its periphery. To the best of oUT knowledge, this type of granulation in the cells forming the cystic cavities has rarely been described previously (20, 21). The morphological aspect of these granules is characteristic of secretion granules, and it could be thought that in this instance as well, we are dealing with a stored and perhaps not secreted factor. Thus, the functional deficit of the epithelial cells would be a defect in the excretion of the factors produced, which would be expressed in two forms: either intravacuolar crystals, or groups of dense small granules spread throughout the cytoplasm of the cells forming the eystic cavity. With this hypothesis, we are in agreement with HIROKAWA (20, 21), who suggested that certain granulations that he observed in aged mouse thymuses could con4*
Exp. Path. 29 (1986) 1
51
sist of stored and denatured thymic hormone. However, we find epithelial cells with dense granules only in the cystic cavities and not strewn throughout the reticulum. How can this phenomenon be explained? ,Ve would like to suggest that the cystic cavities, whose role is unknown, are perhaps the site of regroupment for degenerating cells, different from Hassal's corpuscles. In the same area, the epithelial cells are found to be modified, no longer capable of playing their roles in the microenvironment: the education of T cells, and the self-recognition (3, 4). Thus, this acquired incompetenee may initiate the autoimmune phenomena and the development of lymphoid cells defieient in their funetional capacities or their self control. This relationship is open to speculation, and as lH. DARDENNE (12) suggested, it is possible that the first decrease in the level of thymulin produced, acts upon the loss of T cell suppressive functions, which have, as a consequence, the establishment of anti-pancreatic autoimmunity expressed as antibodies and cytotoxic lymphocytes.
Acknowledgements We thank T. SACKSICK for his excellent photographic assistance, and J. J ACOESEN for reviewing the manuscript.
Literature 1. BACH, J. F., a,nd :JL DAHDENNE, Studies on thymus products. Demonstration and charactprization of a circulating thymic hormone. Immunology 25, 353 (1973). 2. - - J. ]\f. PLEAU and lIL A. BACH, Isolation, biochemical characteristics and biological activity of a circulating thymic hormone in the mouse and in the man. Ann. N.Y. Acad. Sci. 249, 186 (1975). 3. - :XI. A. BACH, C. CARlC,"UD, ]\1. DARDENNE and J. CL. MONIER, Thymic hormones and autoimmunity. In: Autoimlllunity - Genetic, Immunologic, Virologic and Clinical Aspects (ed. N. TALAL). Academic Press, New York 1977, p. 207. 4. BACH, M. A., and J. CHARREIRE, Role of a circulating thymic f11etor in self recognition and self tolerance. Ann. N.Y. Acad. Sci. 332, 55 (1979). 5. BACH, J. F., Thymulin (FTS-Zn). In: Clinics in Immunology and Allergy. W. B. Saunders, New York 1983, p. 133. 6. BAETENS, D., Y. STEFAN, l\I. RAVAZZOLA, F. MALAISSE-LAGEA, D. L. COLEiI'IAN and L. ORCI, Alteration of islet cell populations in spontaneously diabetic mice. Diabetes 27, 1 (1978). 7. BERGLUND, 0., B. J. FRANKEL and B. HELL~IANN, Development of the insulin secretory defect in genetically diabetic (db/db) mouse. Acta Endocrinol. 87, 543 (1978). 8. BOQUIST, L., B. HELLMANN, A. LERNMARK and J. 13. T;\LEDAL, Influence of the mutation "diabetes" on insulin release and islet morphology in mice of different genetic backgrounds. J. Cell BioI. 62, 77 (1974). 9. CHERRY, C., R. EISENSTEIN and A. GLUCKSMAN, Epithelia.l cords and tubules of the rat thymus. Brit. J. Exp. Path. 48, 90 (1967). 10. COHEN, S., S. BEIUUH, l\I. DAHDENNE and J. F. BACH, Regulation in vitro de la secretion de thymuline par les cellules epithBliales thymiques humaines. CR Acad. Sci. (Paris) 297 (serie III) G3 (1983). 11. CORDIER, A., Ultrastructure of the thymus in "nude" mice. J. Ultrastruct. Res. 47, 26 (1974). 12. DARDENNE, l\I., W. SAVINO, L. M. GASTINAL, B. NABARRA and J. F. BACH, Thymic dysfunction in the mutant diabetic (db/db) mouse. J. Immunol. 130, 1195 (1983). 13. DEBRAy-SACHS, M., P. SAl, C. J3OJTARD, R. ASSAN and J. HAMBURGER, Anti-panereatic immunity in genetically diabetic mite. Clin. Exp. Immunol. 51, 1 (1983). 14. EBBESEN, P., and M. NIELSEN, Thymic cysts in oestrogenised Balb/c mice. Acta Pathol. JliIicrobioI. Seand. 80, 211 (1972). 15. EISENBARTH, G., R. JACKSON, S. SHIKANTAS, A. POWERS, J. BusE and H. JliIORr, Utilization of monoelonal antibodies techniques to study type I (insulin-dependent) diabetes mellitus. In: Immunology in Diabetes, ed. ANDllEANI, DI ::I1AHIO and FEDERLIN. Kripton Publications, London 1984, p. 143. 16. FAWCETT, D., The Cell. Saunders Comp., London 1969, p. 319. 17. FEHNAN"DES, G., B. S. HANDWEllGER, E. J. YUNIS 11,nd D. ]VI. BnowN, Immune responses in the mutant diabetic C57BL/Ks-db+ mouse. Discrepancies between in vitro a.ndin vivo immunobiological assays. J. Clin. Invest. 61, 243 (1978). 18. FRAZIEn, J. Ultrastructure of the chick thymus. Z. ZeIlfOTsch.136, 191 (1973).
52
Exp. Path. 29 (1986)
1
19. HAMILTON, D., D. FAWCETT and A. CHRISTENSEN, The liver of the slender salamander. 1. The structure of its crystalline inclusions. Z. Zellforsch. 70, 347 (1966). 20. HIROKAWA, K., Age-related changes of thymus. Morphological and functional aspects. Acta Path. Jap. 28, 843 (1978). 21. - J. MCCLURE and A. GOLDSTEIN, Age-related changes in localization of thymosin in the human thymus. Thymus 4, 19 (1982). 22. HOSHINO, T., The fine structure of ciliated vesicle containing reticular cells in the mouse thymus. Exp. Cell Res. 27, 615 (1962). 23. HUMMEL, K. P., 1\1. M. DICKIE a,nd D. L. COLEMAN, Diabetes, a new mut.
Exp. P
[);3