Eur. J. Immunol. 1990. 20: 129-137
Willem Van EwijkO, Pawel KisielowO and Harald Von Boehmer+ Department of Cell Biology I1 and Immunologyo, Erasmus University, Rotterdam, Institute of Immunology and Experimental Therapyn, Polish Academy of Sciences, Wroclaw and Basel Institute for Immunology+, Basel
Immunohistology of TcR
cdp transgenic mice
129
Immunohistology of T cell differentiation in the thymus of H-Y-specificT cell receptor alp transgenic mice We examined the immunohistological aspects of the H-Y specific T cell receptor (TcR) a/p transgene expression in the thymus of male and female transgenic (Tg) mice. Virtually all thymocytes expressed the p transgene in both the male and female thymus. Expression of accessory molecules (co-receptors) in Tg mice deviated from control mice. In the male Tg thymus, CD8 expression was either low or absent on both cortical and medullary thymocytes. In contrast, in the thymus of female mice, CD8+ cells were found both in the cortex and in the medulla.The majority of medullary thymocytes was bright CD8+.This is in clear contrast to the CD8 distribution in control B6 mice, where only a few percent of medullary cells are CD8+. Similarily, the proportion of cells expressing CD4 antigens was reduced in the cortex and medulla of the thymus from maleTg mice, as compared to the thymus of female Tg mice and B6 control mice. Comparative analysis of the stromal cell types of the thymic microenvironments in the three groups of mice revealed that the cortical thymic microenvironment of maleTg mice differed, compared to that of female Tg mice. In particular, the deep cortex showed a closely packed meshwork of epithelial reticular cells. Moreover, H-2Db molecules (which are the restricting elements for the Tg TcR a/p) were abnormally expressed in the thymic cortex of male mice. The cortical microenvironment in female mice, on the other hand, appeared normal. Together, the data indicate that TcR a/p transgene expression in mule mice leads to an aberrant co-receptor expression in both cortical and medullary lymphoid cells as well as an abnormal composition of the cortical microenvironment. Both phenomena may be the consequence of “negative selection” of developing H-Y-specificTcells,as it occurs only in the maleTg thymus.The absence of the H-Y antigen, but presence of the restricting element H-2Db in the thymic cortex of female mice, leads to accumulation of CD8+ in the medulla, a phenomenon interpreted as “positive selection”.
1 Introduction Intrathymic T cell differentiation is a multistep process, whereby committed progenitor cells differentiate under the influence of stromal cell types in various thymic microenvironments [l-31. A t various levels of differentiation, developing T cells acquire differentiation antigens and receptors for growth factors [ a ] . Most importantly, at an early stage of differentiation, T cells acquire the ability to recognize antigenic peptides in the context of MHC class I or MHC class I1 molecules, by virtue of the heterodimeric TcR [7-lo]. Furthermore, in the thymus, T cells become unresponsive to self antigens present in the thymus [ 12-15]. Tcell differentiation proceeds most efficiently in the thymic microenvironment, since thymusless mice are severely deficient in numbers of Tcells [16, 171. It has been shown that within the thymic stroma, both thymic epithelial reticular cells as well as BM-derived MQ, and interdigitating cells influence T cell development [18]. Thus, it has been argued that epithelial reticular cells are involved in “positive selection” of MHC-restricted T cells [ 19-21], whereas
[I 78581 Correspondence: Willem van Ewijk, Dept. of Cell Biology I1 and Immunology, Erasmus University, PO. Box 1738, 3000 DR Rotterdam, The Netherlands 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990
interdigitating cells (IDC) are involved in “negative selection” [23-241. IDC may not be the only cells involved in negative selection, as recently reported evidence suggests that thymic cortical epithelial cells may also be involved in tolerance induction [25]. Together these data indicate that different stromal cell types in the thymus play a crucial inductive role in T cell development. The development of TcR transgenic (Tg) mice has recently provided new insight in mechanisms of positive and negative selection in the process of Tcell development [15, 26-30]. In particular, Tg mice carrying the TcR a@ directed against the male-specific H-Y antigen were shown to delete differentiating H-Y-specificTcells in the thymus at an early stage of development. Such cells were not deleted in the female thymus. Rather, absence of the H-Y antigens, but presence of the restricting element in the female thymus, caused a significant increase in the proportion of the CD8+ TcR+ H-Y-specific T cells. The present study extends on the phenomenon of positive and negative selection in the thymus. By exploiting immunohistological techniques, three questions are addressed: (a) does the expression of the TcR a / p Tg influence the expression of other differentiation markers in developing T cells? (b) does the expression of the TcR a / p Tg influence the architecture of various thymic microenvironments? (c) can the sites of positive and negative selection be allocated to specific thymic microenvironments?
Eur. J. Immunol. 1990. 20: 129-137
W. Van Ewijk, F! Kisielow and H. Von Boehmer
130
100 pl of an appropriately diluted Ig fraction of a rabbitanti-rat Ig antiserum, conjugated to horseradish peroxide 2.1 Tg mice (RaRaHRP; Dako, Copenhagen, Denmark, P162), supplemented with 1% normal mouse serum. After this second Fertilized eggs obtained from a cross of (C57BL/6J x incubation step, sections were rinsed in PBS/HS-Tw for DBA/W) mice were injectedwith genomic DNA containing 30 min (four changes) and developed with 200-pl diaminothe rearranged TcR a and p genes, isolated from the benzidene (25 mg DAB in 5 ml PBS, activated with 50 pl B6.2.16 H-Y-specific cytolyticTcel1 clone. This Tcell clone 30 % HzOz) for 3 min. Finally, the sections were rinsed in recognizes H-Y in the context of the H-2Db-restricting PBS, dehydrated and coverslipped. The sections were element.TheTg founder mouse contained four copies of the examined using a Zeiss photomicroscope (Oberkochen, a gene and 2 copies of the p gene integrated on the same FRG), using 2.5 X, 4 x and 10 x Plan-Apo objectives. For chromosome [31].This founder mouse was backcrossed to photography, a contrastingfilter was used (Kodak, RochesC57L mice. C57L mice express H-2b MHC antigens but ter, NY, Wratten gelatin filter no. 49). Micrographs were lack the Vp8 gene family. For the present experiment, Tg recorded on Ilford Pan-F films. mice were killed at 4-6 weeks of age.
2 Materials and methods
3 Results
2.2 mAb
3.1 T cell differentiation
mAb used in the present experiments were either directed to mouse T cell differentiation antigens or antigens expressed on thymic epithelial and BM-derived cells. The specificitiesand references of those antibodies are listed in Table 1. 2.3 Immunohistology The thymus, mesenteric LN and spleen lymphoid tissues were removed and collectively embedded in Tissue Tek, using teflon specimen molds. Specimens were frozen in a mixture of ethanol and solid COz. After freezing, the teflon mold was removed and 5-pm sections were cut using a Leitz 1720 cryostat (Wetzlar, FRG). Frozen sections were collected on gelatin-coated microscope slides (aqueous solution of 0.1% gelatin and 0.01% KCr(SO&). Sections were allowed to air-dry, briefly dipped in acetone, air-dried using a fan and stored in a dessicator over silicagel for a period no longer than 3 weeks. Before staining with antibodies, sections were soaked in PBS supplemented with 0.5% horse serum (HS) and 0.01% Tween 20 (PBS/HS-Tw).This treatment removes the Tissue Tek and restores the “threedimensional structure”of the tissue sections. Sections were then incubated with 100 pl undiluted culture SN of mAb for at least 30 min. Sections were washed in PBSMS-Tw for 30 min (four changes). Next, sections were incubated with
3.1.1 Introductory remarks
In the first part of this section the expression of theTg TcR chain, as well as the expression of the differentiation antigens Thy-1, CD8, CD4, Ly-1 and Mel-14, will be presented consecutively in B6 control mice, male TcR a$ Tg mice and female a l p Tg mice. In the second part we will deal with changes in microenvironments of the thymus of these mice. 3.1.2 Expression of the H-Y-specific TcR (Fig. la, b, c) Expression of theTgTcR was monitored by the mAb KJ16, which recognizes the Vg8.2 TcR protein. This antibody was chosen because it stains at higher density in frozen tissue sections than the otherwise used F23.1 antibody. As shown in Fig. la, the frequency of KJ16+ cells is low in the control B6 thymus and the intensity of the staining increases from cortex towards medulla. Both the male and female Tg mice show expression of the Tg TcR p chain on almost all lymphoid cells in the thymus. This includes cells in the subcapsular region which in normal mice are mostly TcR[40].Comparison of the staining intensity shows that the cortical thymocytes in male mice generally expressVg8.2 at higher levels than those in female mice. Some individual
Table 1. Specificities and references of antibodies used
Clons code 1. 2. 3. 4.
F23.1 Eu16 SAD22 H129-19
5. 53-6-72
ER-TR3 6. ER-TR4
7. ER-TRS 8. 9. 10. 11.
ER-’7 hael-14 UA-5-215 M5-114
Specificity
Reference
Vp8.2 Vfi.l,V+3.2
Staen et al., 1985 [32] Behlke et al., 1987 [33] Ledbetter and Henenberg, 1979 I341 Pierres et al., 1984 [35] Ledbetter and Henenberg, 1979 [34] Van m e t et al., 1984 [36]
cD 4 cD8 &Ak
Thymic cortical epithelium Thymic medullary epithelium Thymic fibrocytes Homing receptor H-2Db I-Ab
Van Vliet et al., 1984 [36] Van Vliet et al., 1984 [36] Van Wet et al., 1984 [36] Gallatin et al., 1983 [37] Ozato and‘Sachs, 1981 [38] Bhattacharaya et al., 1981 [39]
Eur. J. Immunol. 1990. 20: 129-137
Immunohistology of TcR a/p transgenic mice
131
Figure 1. Frozen sections of thymi of C57BL (a, d), maleTcR a / p Tg mice (b, e), and femaleTcR a@ Tg mice (c, f). Incubation with mAb KJ16 (Vg8.1,V&2) is presented in panels a-c.Thy-1 staining is presented in panels d-f. C = cortex, M = medulla ( X 100).
cells in the medulla appear Vp8.2 dull or negative.This is in clear contrast to the Tg expression in the female thymus, where virtually all medullary thymocytes expressVp8.2 at a high level.
3.1.3 Expression of Thy-1 (Fig. Id, e, f) Comparison of Thy-1 staining intensity of the thymus of B6, maleTg and female Tg mice reveals no significant variations in Thy-1 expression. The slight increase in Thy-1 density in the outer cortex of the female thymus is within the limits of variation of Thy-1 staining on thymic sections. It may be noted, however, that the average size of the cortical thymocytes in the male thymus is slightly larger compared to control and female transgenic mice.
3.1.4 Expression of CD8 (Fig. 2a, b, c) Since the Tg TcR was originally derived from a CD8+ H-Y-specific T cell clone, it was of interest to study CD8 expression in male and female Tg mice. A typical normal CD8 pattern is shown in Fig. 2a, where B6 cortical thymocytes show a high expression of CD8 molecules. From double labeling with CD4 it is known that most of these cells are “double-positive’’ (CD4+8+)cells (previously published data, not shown). “Single-positive’’CD8+cells do occur in the medulla but the majority of medullary thymocytes in control mice is CD8-. CD8 expression is markedly different in both Tg mice. Most striking is the virtual absence of CD8 molecules on cortical and medullary cells in the maleTg thymus (Fig. 2b). By Contrast, the large majority of thymocytes in the female thymus expressed
132
W.Van Ewijk, l? Kisielow and H.Von Boehrner
Eur. J. Immunol. 1990.20: 129-137
Figure 2. Frozen sections of thymi of C57BL (a, d), maleTcR a.@ Tg mice (b, c ) , and femaleTcR dp mice (c, 9. Sections in Fig. a-c show CD8 expression; sections in Fig. d-f show CD4 expression. C = cortex, M = medulla ( X 100).
CD8 (as indicated in Fig. 2c).This figure shows in particular that, in contrast to B6 control mice, now virtually all medullary thymocytes strongly express CD8. Hence, CD8+ cells accumulate in the medulla of female Tg mice.
are even bright CD4+. Expression of CD4 on cortical thymocytes in the female thymus (Fig. 2f) is comparable to CD4 expression in B6 control mice. The medulla of the thymus of femaleTg mice also contains CD4+,though fewer than the B6 thymus, some of which stain brightly.
3.1.5 Expression of CD4 (Fig. 26, e, f) 3.1.6 Expression of Mel-14 (Fig. 3a, b, c )
A typical CD4 staining pattern is observed in the thymus of B6 control mice (Fig. 2d). Again, CD4 is expressed on virtually all cortical thymocytes, and also the majority of medullary thymocytes is CD4+. In the male thymus, CD4 expression is generally low in the cortex and virtually absent on cells at the cortico-medullary junction (Fig. 2e). Some medullary thymocytes clearly express CD4 and some
Mel-14 staining was included in the present study since a high expression of this cell surface marker is present on a subpopulation of immature CD4-8- thymocytes [41, 421. In B6 mice some individual cortical thymocytes express Mel-14 (Fig. 3a). By contrast, both in the male and female Tg mice, a considerable proportion of cortical cells stains
Eur. J. lmmunol. 1990. 20: 129-137
Immunohistology of TcR ./p transgenic mice
133
Figure 3. Frozen sectionsof thymi of C57BL (a), maleTcR a/p Tg (b), male female TcR a l p Tg mice (c). All sections are incubated with mAb Mel-14. C = cortex, M = medulla ( X 100).
brightly with the Mel-14 antibody (Fig. 3b, resp. Fig. 3c). Moreover, taken into account the centripetally oriented migration of cortical thymocytes [43], we argue that these cells accumulate at the cortico-medullary junction. In the medulla of both Tg mice dull Mel-14+ cells are observed. Such cells occur also in the control B6 thymus, albeit at a lower frequency.
3.2.3 Medullary epithelial cells, stained with ER-TR5 (Fig. 4, e, f) The differences in ER-TR5 staining between medullary reticular elements in the B6 control, male, and female Tg mice are minor (c.f. Fig. 4d, e, f). In the thymus of femaleTg mice medullary epithelial cells show a pattern comparable t o control mice; male mice show slightly more and smaller ER-TRS+ cells in the thymus.
3.2 Thymic microenvironments 3.2.1 Introductory remarks In the second part of the results section we focus upon the architecture of the thymic stroma in normal B6 and a l p Tg mice. For this purpose, sections were stained with antibodies directed against cortical or medullary epithelial cells (ER-TR4, resp. ER-TRS), antibodies directed to fibrocytes (ER-TR7) and antibodies directed to class I and class I1 MHC molecules.
3.2.4 Fibrocytes, revealed with ER-TR7 (Fig. 5a, b, c) Comparison of the distribution of ER-TR7+ fibrocytes in B6, male and female Tg mice reveals a different staining pattern in the male Tg thymus. Normally, ER-TR7 reacts with fibrocytes in the capsula, trabecula, and with reticular cells around blood vessels (perivascular spaces) and with some individual fibrocytes in the medullary stroma (Fig. 5a). Such a pattern is also observed in the female Tg mouse thymus (Fig. 5c). In the male thymus, however, an extensive net of fibroblasts is observed, both in the medulla and in the cortex (Fig. 5b).
3.2.2 Cortical epithelial cells stained by ER-TR4 (Fig. 4a, b, c) 3.2.5 Class I MHC expression The ER.TR4 antibody is directed against cortical epithelial stromal cells [36]. Frozen sections stained by this antibody reveal a fine open lacework in the thymic cortex of B6 mice (Fig. 4a, b, c). MaleTg mice show a marked difference in the distribution of ER-TR4+ cells. As shown in Fig. 4b, the staining intensity of cortical stromal cells appears higher and the stromal cells are more closely packed, especially in the deeper part of the cortex. The staining intensity of ER-TR4 in the female cortex is comparable to that in B6 control mice (Fig. 4c), but the cortical epithelial cells are less dispersed compared to the distribution of these cells in the control thymus.
Since the original Tcell clone was H-2Db restricted, it was of obvious interest to study H-2Db expression on stromal cells in the Tg mice. H-2Db molecules are clearly present on cortical epithelial cells in control B6 mice, as shown in Fig. 6a. This distribution pattern is in clear contrast to that of H-2Kb molecules (data not shown). The latter molecules are mostly expressed in the thymic medulla, as previously reported [44, 451. There is also expression of H-2Db on medullary cells, but the “confluent” staining pattern usually observed with anti-H-2K antibodies was not detected. Both cortical and medullary thymocytes show a low, but specific
134
Eur. J. Immunol. 1990. 20: 129-137
W.Van Ewijk, F! Kisielow and H.Von Boehmer
Figure 4. Frozen sections of thymi of C57BL (a, d), maleTcR a / p Tg (b, e), and femaleTcR a/ p mice (c, f). Fig. a-c showsTR4 expression (cortical epithelial-reticular cells), Fig. d-f shows TR5 expression (medullary epithelial-reticular cells). C = cortex, M = medulla ( x loo).
H-2Db expression. Thymic expression of H-2Db in the male mice is grossly aberrant. Here, in the outer cortex, H-2Db staining is confined to a few stubbed reticular cells. As indicated in Fig. 6b, the fine reticular lacework as observed in the B6 control thymus is absent. H-2Db expression in the thymic cortex of female Tg mice is comparable with the H-2Db expression in control mice. Some cortical areas in the female thymus appear free from reticular elements (Fig. 6c, asterisk). Such areas are not specific for the female transgenic thymus since they also occur in the thymi of control B6 mice (data not shown).
staining pattern is “confluent”. Previous studies have shown that in this compartment both thymocytes, epithelial cells and interdigitating cells are MHC+ [45]. Comparison of the staining patterns in B6 and Tg mice shows a marked increase in the intensity of I-Abstaining in both transgenic mice. The I-Abstaining pattern observed in the medulla of Tg mice differs from B6 control mice; now a more flocculent staining pattern is observed indicating that not all medullary stromal cells in Tg mice express MHC class I1 antigens.
3.2.6 Class I1 MHC expression (Fig. 6d, e,
4 Discussion
f)
Class I1 MHC molecules (I-Ab) are clearly expressed both on cortical and medullary stromal cells. In the medulla this
The data described in this report are in good agreement with data obtained by surface staining of thymocytes in
Eur. J. Immunol. 1990. 20: 129-137
Immunohistology of TcR a/p transgenic mice
135
Figure 5. Frozen sections of thymi of C57BL (a), male TcR cdp Q (b), and femaleTcR a/@Tg mice (c). All sections are incubated with mAb ER.TR7 detecting fibrocytes. C = Cortex, M = medulla, c = thymic capsule, p = perivascular lining ( x 100).
suspension.They extend previous results by describing the anatomical location of the various thymocyte subpopulations as well as by analyzing the stromal elements in the various thymuses. The female thymus of the Tg mice resembles in many aspects the thymus of normal B6 mice.There are, however, some important differences which are relevant in considering positive and negative selection of thymocytes as studied in TcR Tg mice. First, it is noted that practically all cells in the thymus express the transgenicTcR p chain. This includes CD4-8- cells in the outer subcapsularcortex. Such cells are among the earliest intrathymic precursors of Tcells and are TcR- in normal mice [7, 401. Second, all of the cortical CD4+8+ express the transgenic receptor, albeit at lower density than their CD4-8- precursors [15]. As can be observed with the light microscope, there is no apparent clustering of the Tg TcR on thymocytes in the outer cortex. Redistribution of TcR was once thought to signify an interaction of lymphoid cells with epithelial cells, a process implicated in positive selection [40]. It appears that a massive redistribution of receptors is not essential for positive selection. Third, in the medulla of female Tg animals we clearly see an over-representation of CD4-8+ thymocytes, which indicates that positive selection occurs in these animals [26-281. We know from previous experiments that most CD4-8+ cells express the transgenic a / p heterodimer, while the medullary population of CD4+8cells expresses the transgenic p but not the a chain. Apparently, the deeper cortex and more specifically the cortico-medullary junction plays an important role in Tcell differentiation: here in normal Tcell development, thymocytes switch off either of the two CD4 and CD8 accessory molecules (as illustrated in Fig. 2a), whereas the level of TcR expression increases (Fig. la, c). Furthermore, in the Tg mice, Mel-14+ cells accumulate in the cortico-medullary microenvironment which may signify a high extent of positive selection in these mice. Overall, it appears that in
the thymus of femaleTg mice most of the outer cortical cells are not affected, and that phenotypic changes related to positive selection are only apparent in the cortico-medullary microenvironment. However, our observations do not exclude the possibility that thymocyte-epithelial cell interaction required for positive selection [26] occurs initially in the outer cortex. In this context we have recently observed that a / p Tg SCZD mice which lack H-2Db molecules do not contain thymocytes with medullary phenotypes at all [28]. This latter observation indicates that the signal for positive selection is not delivered in the microenvironment of the medulla itself. The most obvious change in the thymus from male TcR Tg mice is the virtual absence of CD8 molecules in cortical and medullary cells. Because this staining would also detect intracellular CD8 molecules, the data argue that CD8 expressing cells are absent, rather than having endocytosed surface CD8 molecules. This correlates well with the reduced number of thymocytes in the male thymus and indicates that CD8+ cells expressing the Tg receptor are deleted at an early stage of Tcell maturation. This leaves behind a population of CD4-8- cells expressing relatively high levels of the Tg receptor which are now found all over the cortex of the male mice. Some CD4+8- bright cells are detected in the medulla. From other studies we know that such cells express the p but not the a Tg TcR chain [27]. Finally we have observed rare phenotypes in the periphery of male a/p Tg mice, like cells which have deleted both Tg and express endogenous TcR genes [31]. Cells with this phenotype may account for the Vg8.2 dull cells in the medulla. The strongly reduced number of lymphocytes in the cortex of male mice goes together with a densely packed network of stromal elements. Most likely, the accumulation of cortical epithelial cells represents condensation of epithelial cells and not necessarily an outgrowth. The abnormal
136
W. Van Ewijk, F! Kisielow and H. Von Boehmer
Eur. J. Immunol. 1990.20: 129-137
Figure 6. Frozen sections of thymi of C57BL (a, d), maleTcR a / p Tg (b, e), and femaleTcR a/fi Tg mice (c, f). Sections in panels a-c show MHC class I (H-2Db) expression; sections in panels d-f show MHC class I1 (I-Ab) expression. C = cortex, M = medulla ( x 100).
architecture of the cortex in male a / p Tg mice is also reflected in the abnormal staining with Db MHC antigenspecific antibodies: the fine lacework of reticular cells observed in control C57BL mice is absent and the Db staining is much more intense in the outer compared to the deeper cortex.These data may indicate that the structure of the cortical microenvironment itself is influenced by the presence of developing thymocytes. Similar observations have been described before in the post-irradiation regenerating murine thymus [46], and very recently we have shown that the thymic microenvironment of SCID mice can be altered by repopulation of SCZD mice with BM cells from normal littermates (Van Ewijk, Shores and Singer, unpublished observation).
Earlier experiments on tolerance of class I1 MHC-restricted Tcells in non-Tg mice [12, 141 did not provide any data indicating that CD4+8+ cortical thymocytes could be the target of deletion. On the contrary, the data were interpreted to indicate that immature C D 4 W thymocytes were not affected. We have discussed several reasons for this apparent discrepancy, like different antigen representation in the cortex, different densities of TcR on cortical cells in transgenic and nontransgenic animals and different sensitivity of assays detecting the partial deletion of cortical cells [15]. In the meantime, functional antibody blocking experiments [47-501 and more careful staining analysis [51] indicated that also in those systems the target of deletion are CD4+8+ cortical thymocytes. The reason that not all
Eur. J. Immunol. 1990.20: 129-137
CD4+8+ thymocytes expressing the relevant TcR are deleted in these mice [45] may indicate either that the deletion requires a certain density of TcR expression or that the deleting antigens are differently distributed. Because of an early expression of theTgTcR in our mice, most CD4+8+ thymocytes express the Tg receptor at densities within the top range of that of normal CD4+8+cells. It may be also for this reason that the deletion of CD4+8+is most obvious in our male TcR a/@Tg mice. The authors appreciate secretarial assistance from Ms. Cary Meijerink and Nicole Schoepflin, and photographical artwork from Mr. Tar van 0s. Received August 15, 1989.
5 References 1 Ezine, S., Weissman, I. L. and Rouse, R.V., Nature 1984.309: 629. 2 Kingston, R., Jenkinson, E. J. and Owen, J. J.T., Nature 1985. 317: 811. 3 Von Boehmer, H., Annu. Rev. Immunol. 1988. 6: 309. 4 Kisielow, P.,Leiserson, W. and Von Boehmer, H . , J. Immunol. 1984. 133: 1117. 5 Raulet, D. H., Garman, R. D., Saito, H. and Tonegawa, S., Nature 1985. 314: 103. 6 Jenkinson, E. J., Kingston, R. and Owen, J. J.T., Nature 1987. 329: 160. 7 Crisanti, A., Colantoni, A., Snodgrass, R. and Von Boehmer, H., EMBO J. 1986. 5: 2837. 8 Havran, W. L. and Allison, J. P., Nature 1988. 335: 443. 9 Pardoll, D. M., Fowlkes, B. J., Bluestone, J. A., Kruisbeek, A., Maloy,W. L., Coligan, J. E. and Schwartz, R. H., Nature 1987. 326: 79. 10 Williams, G . T., Kingston, R., Owen, M. J., Jenkinson, E . J., Owen, J. J. T., Nature 1986. 324: 63. 11 Fowlkes, B. J. and Pardoll, D. H., Adv. Immunol. 1989. 44: 207. 12 Kappler, J. W., Roehm, N. and Marrack, P., Cell 1987. 49: 207. 13 Kappler, J. W., Staerz, U., White, J. and Marrack, P., Nature 1988. 332: 35. 14 McDonald, H. R., Schneider, R., Lees, R. K., Howe, R. C., Acha-Orbea, H., Festenstein, H., Zinkernagel, R. M. and Hengartner, H., Nature 1988. 332: 40. 15 Kisielow, F!, Bliithmann, H., Staerz, U. D., Steinmetz, M. and Von Boehmer, H., Nature 1988. 333: 742. 16 Wortis, H. H., Nehlson, S. and Owen, J. J. T., J. Exp. Med. 1971. 134: 681. 17 Jenkinson, E. J. ,Van Ewijk,W. and Owen, J. J.T., J. Exp. Med. 1981. 153: 280. 18 Van Ewijk, W., Lab. Invest. 1988. 59: 579. 19 Bevan, M. J., Nature 1977. 269: 417. 20 Zinkernagel, R. M., Callahan, G. N., Althage, A., Cooper, S., Klein, P. A. and Klein, J., J. Exp. Med. 1978. 147: 882. 21 Von Boehmer, H., Haas,W. and Jerne, N. K., Proc. Natl. Acad. Sci. USA 1978. 75: 2439.
Immunohistology of TcR a/P transgenic mice
137
22 Von Boehmer, H. and Schubiger, K.. Eur. J. Immunol. 1983. 14: 1048. 23 Jenkinson, E. J., Ihittay, P. and Owen, J. J. T., Transplantation 1985. 39: 331. 24 Lo, D. and Sprent, J., Nature 1986. 319: 7672. 25 Van Ewijk,W., Ron,Y., Monaco, J., Kappler, J., Marrack, P.,Le Mew, H., Gerlinger, P., Durrand, B., Benoist, C. and Mathis, D., Cell 1988. 53: 357. 26 Kisielow, P.,Teh, H. S., Bliithmann, H. and Von Boehmer, H., Nature 1988. 335: 730. 27 Teh, H. S., Kisielow, P., Scoll, B., Kishi, H., Uematsu, Y.. Bliithmann, H. and Von Boehmer, H., Nature 1988. 335: 229. 28 Scott, B., Bliithmann, H., Teh, H. S. and Von Boehmer. H.. Nature 1989. 338: 591. 29 Sha, W. C., Nelson, C. A., Newberry, R. D., Kranz, D. M., Russel, J. H. and Loh, D.Y., Nature 1988. 335: 271. 30 Sha, W. C., Nelson, C. A., Newberry, R. D., Kranz. D. M., Russel, J. H. and Loh, D.Y., Nature 1988. 336: 73. 31 Bliithmann, H., Kisielow, P., Kematsu, Y., Malissen, M., Krirnpenfort, P., Berns, A.,Von Boehmer, H. and Steinmctz. M., Nature 1988. 334: 156. 32 Staerz, U., Rammensee, H., Benedetta, J. and Bevan, M. J., J. Immunol. 1985. 134: 3994. 33 Behlke, M. A., Henkel, T. J., Anderson, S. J., Lan, N. C., Hood, L., Braciale,V.. Braciale, T. J. and Loh, D. Y., J. Exp. Med. 1987. 165: 257. 34 Ledbetter, J. A. and Herzenberg, L. A., Irnmunol. Reb: 1979. 47: 63. 35 Pierres, A., Naquet, P.,Van Agthoven. A., Bekkhoucha, F., Benizot, F., Mishal, Z., Schmitt-Verhulst, A . M. and Pierres. M., J. Immunol. 1984. 132: 2775. 36 Van Vliet, E., Melis, M. and Van Ewijk, W., Eur. J. Immunol. 1984. 14: 524. 37 Gallatin, W. M., Weissrnann, I. L. and Butcher, E. C.. Nafure 1983. 304: 30. 38 Ozato, K. and Sachs, D. H., J. Immunol. 1981. 126: 317. 39 Bhattacharaya, A., Dort, M. E. and Springer, T. A., J. Immunol. 1981. 127: 2488. 40 Farr, A. G., Anderson, K. S., Marrack. P.and Kappler. J., Cell 1985. 43: 543. 41 Reichert, R. A., Weissmann, I. L. and Butcher, E. C.. J. Imrnunol. 1986. 136: 3521. 42 Shortman, K., Wilson, A., Van Ewijk, W. and Scollay, R., J. Immunol. 1987. 138: 342. 43 Weissman, I. L., J. Exp. Med. 1973. 137: 504. 44 Rouse, R.V.,Van Ewijk,W., Jones, P. D. and Weissman, I . L..L Immunol. 1979. 122: 2508. 45 VanEwijk,W., Rouse, R.V, Jones, I? D. and Weissman, 1. L.. J. Histochem. Cytochem. 1980. 28: 1089. 46 Huiskamp, R.,Van Vliet, E. and Van Ewijk, W.. J. Immunol. 1985. 143: 2170. 47 Fowlkes, B. J., Schwartz, R. H. and Pardoll, D. M., Nuture 1988. 334: 620. 48 McDonald, H. R., Hengartner, H. and Pedrazzini. P., Nature 1988. 335: 174. 49 McCarthy, S. A., Kruisbeek, A. M., Lippenkamp, J. K.. Sharrow. S. 0. and Singer, A., Nature 1988. 336: 76. 50 Kyewski, B. A., Schirrmacher, V. and Allison, J.. Eur. J. Immunol. 1989. 19: 857. 51 White, J., Herman, A., Pullen, A. M.. Kubo, R., Kappler. J. and Marrack, P., Cell 1989. 56: 27.
Willem Van EwijkO, Pawel KisielowO and Harald Von Boehmer+ Department of Cell Biology I1 and Immunologyo, Erasmus University, Rotterdam, Institute of Immunology and Experimental Therapyn, Polish Academy of Sciences, Wroclaw and Basel Institute for Immunology+, Basel
Immunohistology of TcR
cdp transgenic mice
129
Immunohistology of T cell differentiation in the thymus of H-Y-specificT cell receptor alp transgenic mice We examined the immunohistological aspects of the H-Y specific T cell receptor (TcR) a/p transgene expression in the thymus of male and female transgenic (Tg) mice. Virtually all thymocytes expressed the p transgene in both the male and female thymus. Expression of accessory molecules (co-receptors) in Tg mice deviated from control mice. In the male Tg thymus, CD8 expression was either low or absent on both cortical and medullary thymocytes. In contrast, in the thymus of female mice, CD8+ cells were found both in the cortex and in the medulla.The majority of medullary thymocytes was bright CD8+.This is in clear contrast to the CD8 distribution in control B6 mice, where only a few percent of medullary cells are CD8+. Similarily, the proportion of cells expressing CD4 antigens was reduced in the cortex and medulla of the thymus from maleTg mice, as compared to the thymus of female Tg mice and B6 control mice. Comparative analysis of the stromal cell types of the thymic microenvironments in the three groups of mice revealed that the cortical thymic microenvironment of maleTg mice differed, compared to that of female Tg mice. In particular, the deep cortex showed a closely packed meshwork of epithelial reticular cells. Moreover, H-2Db molecules (which are the restricting elements for the Tg TcR a/p) were abnormally expressed in the thymic cortex of male mice. The cortical microenvironment in female mice, on the other hand, appeared normal. Together, the data indicate that TcR a/p transgene expression in mule mice leads to an aberrant co-receptor expression in both cortical and medullary lymphoid cells as well as an abnormal composition of the cortical microenvironment. Both phenomena may be the consequence of “negative selection” of developing H-Y-specificTcells,as it occurs only in the maleTg thymus.The absence of the H-Y antigen, but presence of the restricting element H-2Db in the thymic cortex of female mice, leads to accumulation of CD8+ in the medulla, a phenomenon interpreted as “positive selection”.
1 Introduction Intrathymic T cell differentiation is a multistep process, whereby committed progenitor cells differentiate under the influence of stromal cell types in various thymic microenvironments [l-31. A t various levels of differentiation, developing T cells acquire differentiation antigens and receptors for growth factors [ a ] . Most importantly, at an early stage of differentiation, T cells acquire the ability to recognize antigenic peptides in the context of MHC class I or MHC class I1 molecules, by virtue of the heterodimeric TcR [7-lo]. Furthermore, in the thymus, T cells become unresponsive to self antigens present in the thymus [ 12-15]. Tcell differentiation proceeds most efficiently in the thymic microenvironment, since thymusless mice are severely deficient in numbers of Tcells [16, 171. It has been shown that within the thymic stroma, both thymic epithelial reticular cells as well as BM-derived MQ, and interdigitating cells influence T cell development [18]. Thus, it has been argued that epithelial reticular cells are involved in “positive selection” of MHC-restricted T cells [ 19-21], whereas
[I 78581 Correspondence: Willem van Ewijk, Dept. of Cell Biology I1 and Immunology, Erasmus University, PO. Box 1738, 3000 DR Rotterdam, The Netherlands 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990
interdigitating cells (IDC) are involved in “negative selection” [23-241. IDC may not be the only cells involved in negative selection, as recently reported evidence suggests that thymic cortical epithelial cells may also be involved in tolerance induction [25]. Together these data indicate that different stromal cell types in the thymus play a crucial inductive role in T cell development. The development of TcR transgenic (Tg) mice has recently provided new insight in mechanisms of positive and negative selection in the process of Tcell development [15, 26-30]. In particular, Tg mice carrying the TcR a@ directed against the male-specific H-Y antigen were shown to delete differentiating H-Y-specificTcells in the thymus at an early stage of development. Such cells were not deleted in the female thymus. Rather, absence of the H-Y antigens, but presence of the restricting element in the female thymus, caused a significant increase in the proportion of the CD8+ TcR+ H-Y-specific T cells. The present study extends on the phenomenon of positive and negative selection in the thymus. By exploiting immunohistological techniques, three questions are addressed: (a) does the expression of the TcR a / p Tg influence the expression of other differentiation markers in developing T cells? (b) does the expression of the TcR a / p Tg influence the architecture of various thymic microenvironments? (c) can the sites of positive and negative selection be allocated to specific thymic microenvironments?
Eur. J. Immunol. 1990. 20: 129-137
W. Van Ewijk, F! Kisielow and H. Von Boehmer
130
100 pl of an appropriately diluted Ig fraction of a rabbitanti-rat Ig antiserum, conjugated to horseradish peroxide 2.1 Tg mice (RaRaHRP; Dako, Copenhagen, Denmark, P162), supplemented with 1% normal mouse serum. After this second Fertilized eggs obtained from a cross of (C57BL/6J x incubation step, sections were rinsed in PBS/HS-Tw for DBA/W) mice were injectedwith genomic DNA containing 30 min (four changes) and developed with 200-pl diaminothe rearranged TcR a and p genes, isolated from the benzidene (25 mg DAB in 5 ml PBS, activated with 50 pl B6.2.16 H-Y-specific cytolyticTcel1 clone. This Tcell clone 30 % HzOz) for 3 min. Finally, the sections were rinsed in recognizes H-Y in the context of the H-2Db-restricting PBS, dehydrated and coverslipped. The sections were element.TheTg founder mouse contained four copies of the examined using a Zeiss photomicroscope (Oberkochen, a gene and 2 copies of the p gene integrated on the same FRG), using 2.5 X, 4 x and 10 x Plan-Apo objectives. For chromosome [31].This founder mouse was backcrossed to photography, a contrastingfilter was used (Kodak, RochesC57L mice. C57L mice express H-2b MHC antigens but ter, NY, Wratten gelatin filter no. 49). Micrographs were lack the Vp8 gene family. For the present experiment, Tg recorded on Ilford Pan-F films. mice were killed at 4-6 weeks of age.
2 Materials and methods
3 Results
2.2 mAb
3.1 T cell differentiation
mAb used in the present experiments were either directed to mouse T cell differentiation antigens or antigens expressed on thymic epithelial and BM-derived cells. The specificitiesand references of those antibodies are listed in Table 1. 2.3 Immunohistology The thymus, mesenteric LN and spleen lymphoid tissues were removed and collectively embedded in Tissue Tek, using teflon specimen molds. Specimens were frozen in a mixture of ethanol and solid COz. After freezing, the teflon mold was removed and 5-pm sections were cut using a Leitz 1720 cryostat (Wetzlar, FRG). Frozen sections were collected on gelatin-coated microscope slides (aqueous solution of 0.1% gelatin and 0.01% KCr(SO&). Sections were allowed to air-dry, briefly dipped in acetone, air-dried using a fan and stored in a dessicator over silicagel for a period no longer than 3 weeks. Before staining with antibodies, sections were soaked in PBS supplemented with 0.5% horse serum (HS) and 0.01% Tween 20 (PBS/HS-Tw).This treatment removes the Tissue Tek and restores the “threedimensional structure”of the tissue sections. Sections were then incubated with 100 pl undiluted culture SN of mAb for at least 30 min. Sections were washed in PBSMS-Tw for 30 min (four changes). Next, sections were incubated with
3.1.1 Introductory remarks
In the first part of this section the expression of theTg TcR chain, as well as the expression of the differentiation antigens Thy-1, CD8, CD4, Ly-1 and Mel-14, will be presented consecutively in B6 control mice, male TcR a$ Tg mice and female a l p Tg mice. In the second part we will deal with changes in microenvironments of the thymus of these mice. 3.1.2 Expression of the H-Y-specific TcR (Fig. la, b, c) Expression of theTgTcR was monitored by the mAb KJ16, which recognizes the Vg8.2 TcR protein. This antibody was chosen because it stains at higher density in frozen tissue sections than the otherwise used F23.1 antibody. As shown in Fig. la, the frequency of KJ16+ cells is low in the control B6 thymus and the intensity of the staining increases from cortex towards medulla. Both the male and female Tg mice show expression of the Tg TcR p chain on almost all lymphoid cells in the thymus. This includes cells in the subcapsular region which in normal mice are mostly TcR[40].Comparison of the staining intensity shows that the cortical thymocytes in male mice generally expressVg8.2 at higher levels than those in female mice. Some individual
Table 1. Specificities and references of antibodies used
Clons code 1. 2. 3. 4.
F23.1 Eu16 SAD22 H129-19
5. 53-6-72
ER-TR3 6. ER-TR4
7. ER-TRS 8. 9. 10. 11.
ER-’7 hael-14 UA-5-215 M5-114
Specificity
Reference
Vp8.2 Vfi.l,V+3.2
Staen et al., 1985 [32] Behlke et al., 1987 [33] Ledbetter and Henenberg, 1979 I341 Pierres et al., 1984 [35] Ledbetter and Henenberg, 1979 [34] Van m e t et al., 1984 [36]
cD 4 cD8 &Ak
Thymic cortical epithelium Thymic medullary epithelium Thymic fibrocytes Homing receptor H-2Db I-Ab
Van Vliet et al., 1984 [36] Van Vliet et al., 1984 [36] Van Wet et al., 1984 [36] Gallatin et al., 1983 [37] Ozato and‘Sachs, 1981 [38] Bhattacharaya et al., 1981 [39]
Eur. J. Immunol. 1990. 20: 129-137
Immunohistology of TcR a/p transgenic mice
131
Figure 1. Frozen sections of thymi of C57BL (a, d), maleTcR a / p Tg mice (b, e), and femaleTcR a@ Tg mice (c, f). Incubation with mAb KJ16 (Vg8.1,V&2) is presented in panels a-c.Thy-1 staining is presented in panels d-f. C = cortex, M = medulla ( X 100).
cells in the medulla appear Vp8.2 dull or negative.This is in clear contrast to the Tg expression in the female thymus, where virtually all medullary thymocytes expressVp8.2 at a high level.
3.1.3 Expression of Thy-1 (Fig. Id, e, f) Comparison of Thy-1 staining intensity of the thymus of B6, maleTg and female Tg mice reveals no significant variations in Thy-1 expression. The slight increase in Thy-1 density in the outer cortex of the female thymus is within the limits of variation of Thy-1 staining on thymic sections. It may be noted, however, that the average size of the cortical thymocytes in the male thymus is slightly larger compared to control and female transgenic mice.
3.1.4 Expression of CD8 (Fig. 2a, b, c) Since the Tg TcR was originally derived from a CD8+ H-Y-specific T cell clone, it was of interest to study CD8 expression in male and female Tg mice. A typical normal CD8 pattern is shown in Fig. 2a, where B6 cortical thymocytes show a high expression of CD8 molecules. From double labeling with CD4 it is known that most of these cells are “double-positive’’ (CD4+8+)cells (previously published data, not shown). “Single-positive’’CD8+cells do occur in the medulla but the majority of medullary thymocytes in control mice is CD8-. CD8 expression is markedly different in both Tg mice. Most striking is the virtual absence of CD8 molecules on cortical and medullary cells in the maleTg thymus (Fig. 2b). By Contrast, the large majority of thymocytes in the female thymus expressed
132
W.Van Ewijk, l? Kisielow and H.Von Boehrner
Eur. J. Immunol. 1990.20: 129-137
Figure 2. Frozen sections of thymi of C57BL (a, d), maleTcR a.@ Tg mice (b, c ) , and femaleTcR dp mice (c, 9. Sections in Fig. a-c show CD8 expression; sections in Fig. d-f show CD4 expression. C = cortex, M = medulla ( X 100).
CD8 (as indicated in Fig. 2c).This figure shows in particular that, in contrast to B6 control mice, now virtually all medullary thymocytes strongly express CD8. Hence, CD8+ cells accumulate in the medulla of female Tg mice.
are even bright CD4+. Expression of CD4 on cortical thymocytes in the female thymus (Fig. 2f) is comparable to CD4 expression in B6 control mice. The medulla of the thymus of femaleTg mice also contains CD4+,though fewer than the B6 thymus, some of which stain brightly.
3.1.5 Expression of CD4 (Fig. 26, e, f) 3.1.6 Expression of Mel-14 (Fig. 3a, b, c )
A typical CD4 staining pattern is observed in the thymus of B6 control mice (Fig. 2d). Again, CD4 is expressed on virtually all cortical thymocytes, and also the majority of medullary thymocytes is CD4+. In the male thymus, CD4 expression is generally low in the cortex and virtually absent on cells at the cortico-medullary junction (Fig. 2e). Some medullary thymocytes clearly express CD4 and some
Mel-14 staining was included in the present study since a high expression of this cell surface marker is present on a subpopulation of immature CD4-8- thymocytes [41, 421. In B6 mice some individual cortical thymocytes express Mel-14 (Fig. 3a). By contrast, both in the male and female Tg mice, a considerable proportion of cortical cells stains
Eur. J. lmmunol. 1990. 20: 129-137
Immunohistology of TcR ./p transgenic mice
133
Figure 3. Frozen sectionsof thymi of C57BL (a), maleTcR a/p Tg (b), male female TcR a l p Tg mice (c). All sections are incubated with mAb Mel-14. C = cortex, M = medulla ( X 100).
brightly with the Mel-14 antibody (Fig. 3b, resp. Fig. 3c). Moreover, taken into account the centripetally oriented migration of cortical thymocytes [43], we argue that these cells accumulate at the cortico-medullary junction. In the medulla of both Tg mice dull Mel-14+ cells are observed. Such cells occur also in the control B6 thymus, albeit at a lower frequency.
3.2.3 Medullary epithelial cells, stained with ER-TR5 (Fig. 4, e, f) The differences in ER-TR5 staining between medullary reticular elements in the B6 control, male, and female Tg mice are minor (c.f. Fig. 4d, e, f). In the thymus of femaleTg mice medullary epithelial cells show a pattern comparable t o control mice; male mice show slightly more and smaller ER-TRS+ cells in the thymus.
3.2 Thymic microenvironments 3.2.1 Introductory remarks In the second part of the results section we focus upon the architecture of the thymic stroma in normal B6 and a l p Tg mice. For this purpose, sections were stained with antibodies directed against cortical or medullary epithelial cells (ER-TR4, resp. ER-TRS), antibodies directed to fibrocytes (ER-TR7) and antibodies directed to class I and class I1 MHC molecules.
3.2.4 Fibrocytes, revealed with ER-TR7 (Fig. 5a, b, c) Comparison of the distribution of ER-TR7+ fibrocytes in B6, male and female Tg mice reveals a different staining pattern in the male Tg thymus. Normally, ER-TR7 reacts with fibrocytes in the capsula, trabecula, and with reticular cells around blood vessels (perivascular spaces) and with some individual fibrocytes in the medullary stroma (Fig. 5a). Such a pattern is also observed in the female Tg mouse thymus (Fig. 5c). In the male thymus, however, an extensive net of fibroblasts is observed, both in the medulla and in the cortex (Fig. 5b).
3.2.2 Cortical epithelial cells stained by ER-TR4 (Fig. 4a, b, c) 3.2.5 Class I MHC expression The ER.TR4 antibody is directed against cortical epithelial stromal cells [36]. Frozen sections stained by this antibody reveal a fine open lacework in the thymic cortex of B6 mice (Fig. 4a, b, c). MaleTg mice show a marked difference in the distribution of ER-TR4+ cells. As shown in Fig. 4b, the staining intensity of cortical stromal cells appears higher and the stromal cells are more closely packed, especially in the deeper part of the cortex. The staining intensity of ER-TR4 in the female cortex is comparable to that in B6 control mice (Fig. 4c), but the cortical epithelial cells are less dispersed compared to the distribution of these cells in the control thymus.
Since the original Tcell clone was H-2Db restricted, it was of obvious interest to study H-2Db expression on stromal cells in the Tg mice. H-2Db molecules are clearly present on cortical epithelial cells in control B6 mice, as shown in Fig. 6a. This distribution pattern is in clear contrast to that of H-2Kb molecules (data not shown). The latter molecules are mostly expressed in the thymic medulla, as previously reported [44, 451. There is also expression of H-2Db on medullary cells, but the “confluent” staining pattern usually observed with anti-H-2K antibodies was not detected. Both cortical and medullary thymocytes show a low, but specific
134
Eur. J. Immunol. 1990. 20: 129-137
W.Van Ewijk, F! Kisielow and H.Von Boehmer
Figure 4. Frozen sections of thymi of C57BL (a, d), maleTcR a / p Tg (b, e), and femaleTcR a/ p mice (c, f). Fig. a-c showsTR4 expression (cortical epithelial-reticular cells), Fig. d-f shows TR5 expression (medullary epithelial-reticular cells). C = cortex, M = medulla ( x loo).
H-2Db expression. Thymic expression of H-2Db in the male mice is grossly aberrant. Here, in the outer cortex, H-2Db staining is confined to a few stubbed reticular cells. As indicated in Fig. 6b, the fine reticular lacework as observed in the B6 control thymus is absent. H-2Db expression in the thymic cortex of female Tg mice is comparable with the H-2Db expression in control mice. Some cortical areas in the female thymus appear free from reticular elements (Fig. 6c, asterisk). Such areas are not specific for the female transgenic thymus since they also occur in the thymi of control B6 mice (data not shown).
staining pattern is “confluent”. Previous studies have shown that in this compartment both thymocytes, epithelial cells and interdigitating cells are MHC+ [45]. Comparison of the staining patterns in B6 and Tg mice shows a marked increase in the intensity of I-Abstaining in both transgenic mice. The I-Abstaining pattern observed in the medulla of Tg mice differs from B6 control mice; now a more flocculent staining pattern is observed indicating that not all medullary stromal cells in Tg mice express MHC class I1 antigens.
3.2.6 Class I1 MHC expression (Fig. 6d, e,
4 Discussion
f)
Class I1 MHC molecules (I-Ab) are clearly expressed both on cortical and medullary stromal cells. In the medulla this
The data described in this report are in good agreement with data obtained by surface staining of thymocytes in
Eur. J. Immunol. 1990. 20: 129-137
Immunohistology of TcR a/p transgenic mice
135
Figure 5. Frozen sections of thymi of C57BL (a), male TcR cdp Q (b), and femaleTcR a/@Tg mice (c). All sections are incubated with mAb ER.TR7 detecting fibrocytes. C = Cortex, M = medulla, c = thymic capsule, p = perivascular lining ( x 100).
suspension.They extend previous results by describing the anatomical location of the various thymocyte subpopulations as well as by analyzing the stromal elements in the various thymuses. The female thymus of the Tg mice resembles in many aspects the thymus of normal B6 mice.There are, however, some important differences which are relevant in considering positive and negative selection of thymocytes as studied in TcR Tg mice. First, it is noted that practically all cells in the thymus express the transgenicTcR p chain. This includes CD4-8- cells in the outer subcapsularcortex. Such cells are among the earliest intrathymic precursors of Tcells and are TcR- in normal mice [7, 401. Second, all of the cortical CD4+8+ express the transgenic receptor, albeit at lower density than their CD4-8- precursors [15]. As can be observed with the light microscope, there is no apparent clustering of the Tg TcR on thymocytes in the outer cortex. Redistribution of TcR was once thought to signify an interaction of lymphoid cells with epithelial cells, a process implicated in positive selection [40]. It appears that a massive redistribution of receptors is not essential for positive selection. Third, in the medulla of female Tg animals we clearly see an over-representation of CD4-8+ thymocytes, which indicates that positive selection occurs in these animals [26-281. We know from previous experiments that most CD4-8+ cells express the transgenic a / p heterodimer, while the medullary population of CD4+8cells expresses the transgenic p but not the a chain. Apparently, the deeper cortex and more specifically the cortico-medullary junction plays an important role in Tcell differentiation: here in normal Tcell development, thymocytes switch off either of the two CD4 and CD8 accessory molecules (as illustrated in Fig. 2a), whereas the level of TcR expression increases (Fig. la, c). Furthermore, in the Tg mice, Mel-14+ cells accumulate in the cortico-medullary microenvironment which may signify a high extent of positive selection in these mice. Overall, it appears that in
the thymus of femaleTg mice most of the outer cortical cells are not affected, and that phenotypic changes related to positive selection are only apparent in the cortico-medullary microenvironment. However, our observations do not exclude the possibility that thymocyte-epithelial cell interaction required for positive selection [26] occurs initially in the outer cortex. In this context we have recently observed that a / p Tg SCZD mice which lack H-2Db molecules do not contain thymocytes with medullary phenotypes at all [28]. This latter observation indicates that the signal for positive selection is not delivered in the microenvironment of the medulla itself. The most obvious change in the thymus from male TcR Tg mice is the virtual absence of CD8 molecules in cortical and medullary cells. Because this staining would also detect intracellular CD8 molecules, the data argue that CD8 expressing cells are absent, rather than having endocytosed surface CD8 molecules. This correlates well with the reduced number of thymocytes in the male thymus and indicates that CD8+ cells expressing the Tg receptor are deleted at an early stage of Tcell maturation. This leaves behind a population of CD4-8- cells expressing relatively high levels of the Tg receptor which are now found all over the cortex of the male mice. Some CD4+8- bright cells are detected in the medulla. From other studies we know that such cells express the p but not the a Tg TcR chain [27]. Finally we have observed rare phenotypes in the periphery of male a/p Tg mice, like cells which have deleted both Tg and express endogenous TcR genes [31]. Cells with this phenotype may account for the Vg8.2 dull cells in the medulla. The strongly reduced number of lymphocytes in the cortex of male mice goes together with a densely packed network of stromal elements. Most likely, the accumulation of cortical epithelial cells represents condensation of epithelial cells and not necessarily an outgrowth. The abnormal
136
W. Van Ewijk, F! Kisielow and H. Von Boehmer
Eur. J. Immunol. 1990.20: 129-137
Figure 6. Frozen sections of thymi of C57BL (a, d), maleTcR a / p Tg (b, e), and femaleTcR a/fi Tg mice (c, f). Sections in panels a-c show MHC class I (H-2Db) expression; sections in panels d-f show MHC class I1 (I-Ab) expression. C = cortex, M = medulla ( x 100).
architecture of the cortex in male a / p Tg mice is also reflected in the abnormal staining with Db MHC antigenspecific antibodies: the fine lacework of reticular cells observed in control C57BL mice is absent and the Db staining is much more intense in the outer compared to the deeper cortex.These data may indicate that the structure of the cortical microenvironment itself is influenced by the presence of developing thymocytes. Similar observations have been described before in the post-irradiation regenerating murine thymus [46], and very recently we have shown that the thymic microenvironment of SCID mice can be altered by repopulation of SCZD mice with BM cells from normal littermates (Van Ewijk, Shores and Singer, unpublished observation).
Earlier experiments on tolerance of class I1 MHC-restricted Tcells in non-Tg mice [12, 141 did not provide any data indicating that CD4+8+ cortical thymocytes could be the target of deletion. On the contrary, the data were interpreted to indicate that immature C D 4 W thymocytes were not affected. We have discussed several reasons for this apparent discrepancy, like different antigen representation in the cortex, different densities of TcR on cortical cells in transgenic and nontransgenic animals and different sensitivity of assays detecting the partial deletion of cortical cells [15]. In the meantime, functional antibody blocking experiments [47-501 and more careful staining analysis [51] indicated that also in those systems the target of deletion are CD4+8+ cortical thymocytes. The reason that not all
Eur. J. Immunol. 1990.20: 129-137
CD4+8+ thymocytes expressing the relevant TcR are deleted in these mice [45] may indicate either that the deletion requires a certain density of TcR expression or that the deleting antigens are differently distributed. Because of an early expression of theTgTcR in our mice, most CD4+8+ thymocytes express the Tg receptor at densities within the top range of that of normal CD4+8+cells. It may be also for this reason that the deletion of CD4+8+is most obvious in our male TcR a/@Tg mice. The authors appreciate secretarial assistance from Ms. Cary Meijerink and Nicole Schoepflin, and photographical artwork from Mr. Tar van 0s. Received August 15, 1989.
5 References 1 Ezine, S., Weissman, I. L. and Rouse, R.V., Nature 1984.309: 629. 2 Kingston, R., Jenkinson, E. J. and Owen, J. J.T., Nature 1985. 317: 811. 3 Von Boehmer, H., Annu. Rev. Immunol. 1988. 6: 309. 4 Kisielow, P.,Leiserson, W. and Von Boehmer, H . , J. Immunol. 1984. 133: 1117. 5 Raulet, D. H., Garman, R. D., Saito, H. and Tonegawa, S., Nature 1985. 314: 103. 6 Jenkinson, E. J., Kingston, R. and Owen, J. J.T., Nature 1987. 329: 160. 7 Crisanti, A., Colantoni, A., Snodgrass, R. and Von Boehmer, H., EMBO J. 1986. 5: 2837. 8 Havran, W. L. and Allison, J. P., Nature 1988. 335: 443. 9 Pardoll, D. M., Fowlkes, B. J., Bluestone, J. A., Kruisbeek, A., Maloy,W. L., Coligan, J. E. and Schwartz, R. H., Nature 1987. 326: 79. 10 Williams, G . T., Kingston, R., Owen, M. J., Jenkinson, E . J., Owen, J. J. T., Nature 1986. 324: 63. 11 Fowlkes, B. J. and Pardoll, D. H., Adv. Immunol. 1989. 44: 207. 12 Kappler, J. W., Roehm, N. and Marrack, P., Cell 1987. 49: 207. 13 Kappler, J. W., Staerz, U., White, J. and Marrack, P., Nature 1988. 332: 35. 14 McDonald, H. R., Schneider, R., Lees, R. K., Howe, R. C., Acha-Orbea, H., Festenstein, H., Zinkernagel, R. M. and Hengartner, H., Nature 1988. 332: 40. 15 Kisielow, F!, Bliithmann, H., Staerz, U. D., Steinmetz, M. and Von Boehmer, H., Nature 1988. 333: 742. 16 Wortis, H. H., Nehlson, S. and Owen, J. J. T., J. Exp. Med. 1971. 134: 681. 17 Jenkinson, E. J. ,Van Ewijk,W. and Owen, J. J.T., J. Exp. Med. 1981. 153: 280. 18 Van Ewijk, W., Lab. Invest. 1988. 59: 579. 19 Bevan, M. J., Nature 1977. 269: 417. 20 Zinkernagel, R. M., Callahan, G. N., Althage, A., Cooper, S., Klein, P. A. and Klein, J., J. Exp. Med. 1978. 147: 882. 21 Von Boehmer, H., Haas,W. and Jerne, N. K., Proc. Natl. Acad. Sci. USA 1978. 75: 2439.
Immunohistology of TcR a/P transgenic mice
137
22 Von Boehmer, H. and Schubiger, K.. Eur. J. Immunol. 1983. 14: 1048. 23 Jenkinson, E. J., Ihittay, P. and Owen, J. J. T., Transplantation 1985. 39: 331. 24 Lo, D. and Sprent, J., Nature 1986. 319: 7672. 25 Van Ewijk,W., Ron,Y., Monaco, J., Kappler, J., Marrack, P.,Le Mew, H., Gerlinger, P., Durrand, B., Benoist, C. and Mathis, D., Cell 1988. 53: 357. 26 Kisielow, P.,Teh, H. S., Bliithmann, H. and Von Boehmer, H., Nature 1988. 335: 730. 27 Teh, H. S., Kisielow, P., Scoll, B., Kishi, H., Uematsu, Y.. Bliithmann, H. and Von Boehmer, H., Nature 1988. 335: 229. 28 Scott, B., Bliithmann, H., Teh, H. S. and Von Boehmer. H.. Nature 1989. 338: 591. 29 Sha, W. C., Nelson, C. A., Newberry, R. D., Kranz, D. M., Russel, J. H. and Loh, D.Y., Nature 1988. 335: 271. 30 Sha, W. C., Nelson, C. A., Newberry, R. D., Kranz. D. M., Russel, J. H. and Loh, D.Y., Nature 1988. 336: 73. 31 Bliithmann, H., Kisielow, P., Kematsu, Y., Malissen, M., Krirnpenfort, P., Berns, A.,Von Boehmer, H. and Steinmctz. M., Nature 1988. 334: 156. 32 Staerz, U., Rammensee, H., Benedetta, J. and Bevan, M. J., J. Immunol. 1985. 134: 3994. 33 Behlke, M. A., Henkel, T. J., Anderson, S. J., Lan, N. C., Hood, L., Braciale,V.. Braciale, T. J. and Loh, D. Y., J. Exp. Med. 1987. 165: 257. 34 Ledbetter, J. A. and Herzenberg, L. A., Irnmunol. Reb: 1979. 47: 63. 35 Pierres, A., Naquet, P.,Van Agthoven. A., Bekkhoucha, F., Benizot, F., Mishal, Z., Schmitt-Verhulst, A . M. and Pierres. M., J. Immunol. 1984. 132: 2775. 36 Van Vliet, E., Melis, M. and Van Ewijk, W., Eur. J. Immunol. 1984. 14: 524. 37 Gallatin, W. M., Weissrnann, I. L. and Butcher, E. C.. Nafure 1983. 304: 30. 38 Ozato, K. and Sachs, D. H., J. Immunol. 1981. 126: 317. 39 Bhattacharaya, A., Dort, M. E. and Springer, T. A., J. Immunol. 1981. 127: 2488. 40 Farr, A. G., Anderson, K. S., Marrack. P.and Kappler. J., Cell 1985. 43: 543. 41 Reichert, R. A., Weissmann, I. L. and Butcher, E. C.. J. Imrnunol. 1986. 136: 3521. 42 Shortman, K., Wilson, A., Van Ewijk, W. and Scollay, R., J. Immunol. 1987. 138: 342. 43 Weissman, I. L., J. Exp. Med. 1973. 137: 504. 44 Rouse, R.V.,Van Ewijk,W., Jones, P. D. and Weissman, I . L..L Immunol. 1979. 122: 2508. 45 VanEwijk,W., Rouse, R.V, Jones, I? D. and Weissman, 1. L.. J. Histochem. Cytochem. 1980. 28: 1089. 46 Huiskamp, R.,Van Vliet, E. and Van Ewijk, W.. J. Immunol. 1985. 143: 2170. 47 Fowlkes, B. J., Schwartz, R. H. and Pardoll, D. M., Nuture 1988. 334: 620. 48 McDonald, H. R., Hengartner, H. and Pedrazzini. P., Nature 1988. 335: 174. 49 McCarthy, S. A., Kruisbeek, A. M., Lippenkamp, J. K.. Sharrow. S. 0. and Singer, A., Nature 1988. 336: 76. 50 Kyewski, B. A., Schirrmacher, V. and Allison, J.. Eur. J. Immunol. 1989. 19: 857. 51 White, J., Herman, A., Pullen, A. M.. Kubo, R., Kappler. J. and Marrack, P., Cell 1989. 56: 27.
