Aug 21, 1989 - Julius,M.H., Mossmann,H., Carsetti,R. and Kohler,G. (1989) Eur. J. Immunol., 19, 459-468. Ledbetter,J.A. and Herzenberg,L.A. (1979) Immunol.
The EMBO Journal vol.8 no. 1 2 pp.3719 - 3726, 1989
Elimination of CD8 + thymocytes in transgenic mice expressing an anti-Lyt2.2 immunoglobulin heavy chain gene
Frank Brombacher, Marinus CLamers, Georges Kohier and Hermann Eibel Max-Planck-Institut fiir Immunbiologie. Stubeweg 51. 7800 Freiburg i.B., FRG Communicated by Georges Kohler
Individual T cell populations are characterized by specific surface proteins, namely by the T cell receptor complex (TCR) and by two accessory molecules, CD8 (Lyt2) and CD4 (L3T4). CD8 and CD4 are required for T cell interactions with class I or class II major histocompatibility complex molecules. In the thymus, immature CD8-4- TCR- cells differentiate, possibly via a short stage of CD8+4- thymocytes, into CD8+4+ TCR+ T cells and mature further into the main T cell populations, the CD8+4- TCR+ cytotoxic T lymphocytes and the CD4+8- TCR+ T helper cells. In order to analyse the differentiation steps involving CD8, we generated transgenic mice expressing ,u heavy chain genes from an anti-Lyt2.2 hybridoma. Transgenic lines expressing either the complete (s'm) or only the secreted A protein (uS) suffer from a severe depletion of their CD8+4+ thymocytes affecting also the mature CD8+4- and CD4+8populations. The depletion is correlated to the expression of transgenic t-chain proteins within thymocytes. This intrathymocyte expression of the A chain prevents CD8-4- thymocytes from further differentiation, most probably via intracellular interactions between A heavy chain and CD8 proteins. These results show that CD8 plays an important role during thymocyte maturation. Key words: CD8 expression/T cell differentiation/transgenic mice
Introduction In the thymus, T lymphocyte differentiation is characterized by positive as well as negative selection mechanisms (Marrack et al., 1988; Sha et al., 1988a). Positive selection allows antigen recognition in the context of self major histocompatibility complex molecules (MHC; Kisielow et al., 1988a; McDonald et al., 1988a). Negative selection serves to eliminate self-reactive T cells from the repertoire (Kappler, 1987; Kisielow et al., 1988b; MacDonald et al., 1988b; Teh et al., 1989). Mature T cell populations are characterized by the surface expression of two different glycoproteins, the accessory molecules CD4 (L3T4,) or CD8 (Lyt2) (Kisielow et al., 1975; Ledbetter and Herzenberg, 1979; Dialnyas et al., 1983). Both accessory molecules are thought to increase the avidity of the T cells for their targets (Dembic et al., 1987; Gebert et al., 1987). CD4+CD8- T cells are mainly T helper cells and restricted to MHC class II proteins; cytotoxic T lymphocytes are mostly CD8+CD4- and recognize antigen in the context of MHC ©IRL Press
class I molecules. During development, immature CD8-4thymocytes differentiate via a stage of CD8+4+ cells into mature CD8+4- or CD4+8- T cells (reviewed in Adkins et al., 1987; von Boehmer, 1988). Both types of accessory molecules are supposed to play an important role in this developmental process: selection of the appropriate T cell receptor (TCR) CD8/CD4 combination at the CD8+4+ stage is thought to direct differentiating thymocytes towards each of the major single positive, mature T cell populations (Sha et al., 1988b; Teh et al., 1988). The generation of transgenic mice expressing TCR c and / chains helped in the analysis of these developmental processes (Kisielow et al., 1988a,b; Sha et al., 1988a,b; Pircher et al., 1989). In vivo studies carried out with injections of anti-CD8 and antiCD4 antibodies, however, lead to conflicting results (Smith, 1987; MacDonald et al., 1988c). In order to dissect the differentiation steps involving interactions of CD8, we therefore decided to generate transgenic mice expressing an anti-CD8 (Lyt2.2) IgM (A) heavy (H) chain gene. We expected that the transgenic I H-chain molecules would associate with endogenous light (L) chains and form, at a certain frequency, anti-CD8 antibodies reactive against CD8+ T lymphocytes. Antibodies with CD8 specificity, however, were not found in the transgenic mouse lines. To our surprise, the analysis of the transgenic lines revealed that the intracellular expression of the transgenic /i gene within thymocytes severely interferes with their differentiation, most probably by preventing CD8 surface expression. This impairment inhibits CD4-8- thymocytes from further maturation. As a consequence, CD8+4+ cells are severely depleted from the thymus of transgenic mice. The onset of CD8 expression appears therefore to be an important signal for CD8-4- cells to proceed in further differentiation. In addition, the CD8+4- and CD4+ 8- populations in the thymus as well as in the periphery were significantly reduced. This result shows that CD8+4+ cells can serve as precursors for both of the mature, single-positive T cell
populations.
Results Production of transgenic mice expressing either the
/u1 or the us form
The source for the anti-Lyt2.2 immunoglobulin (Ig) H-chain gene was the hybridoma line 19/178 producing a mouse monoclonal anti-Lyt2.2 IgG2Alkappa antibody (Haimmerling, 1978). A genomic 4.8 kb EcoRI fragment containing the productively rearranged Ig H-chain variable (V) region and the Ig H-chain enhancer was cloned and used to generate plasmid pCit19 (described in the Material and methods). Transfection experiments, flow cytometric analyses and complement-mediated cytolysis showed that the genetically engineered A chain antibody had retained the original antiLyt2.2 specificity encoded by the V-region (data not shown). In order to generate transgenic mice expressing both the
3719
F.Brombacher et al.
membrane-bound and the secreted form of the transgenic we microinjected an AatII-SphI restriction fragment of pCAQ19 into one of the pronuclei of fertilized mouse eggs (Figure 1, tl9sm). Nine transgenic lines were obtained. All of them showed expression of the transgene in the periphery as judged from serum levels of the transgenic IgMa allotype at 1-10 Ag/ml. The results obtained from six of these lines (M4, M8A, M8Y, Mll, M12, M13) are reported here. In line M8Y the integration site of the transgene was mapped to the Y chromosome, the other lines had the transgene integrated on autosomal loci. Four different transgenic founder mice (M41, M42, M85, M86), expressing the secreted form of the it H-chain (,S), were obtained by microinjecting the AatH-KpnI fragment of pC,ul9 (1d19s), which lacks exons 5 and 6 of the constant part encoding the transmembrane part of the ,u chain (Figure 1,
Iu H-chain (s'm),
I PC).J19sm _
o
E E E c IIo m co
p
m co
LU
VDJ4
Ig enh
p.A
CpICP2CIP3 CU-
14.2 kb 72 kb
1pCJJl9SI =ccr "
0c
O0
..
L.1Iu
..1T3m VDJ4
el 1g eo,h
mm mm p.A _
_
CP IC P2 CP3 CP4
12.2 kb
Fig. 1. Restriction map of the microinjected DNA fragments. Mouse genomic DNA is represented by thick, pUC19 DNA by thin lines. the closed boxes symbolize the J4 element and the exons of the it constant part (c LA -c t6), open boxes the V and D segments (V,D), a rhombus the Ig H-chain enhancer (Ig enh), dots polyadenylation sites (p.A.). The 7.2 kb BamHI fragment was indicative for the transgenes in Southern blots (data not shown). Offspring of eggs injected with /l9Sm express both the secreted (uS) and the membrane form (us') of the transgene, offspring of eggs injected with yl9S only express is.
jd95). M42 had the transgene integrated on the X chromosome, the other lines on autosomal loci (data not shown). All of these lines expressed the transgene in their serum (data not shown) and were used in further analyses. The expression of the transgenic it H-chain on the surface of splenic B-cells was analysed by flow cytometry (Table I). Transgenic A was present on most of the B cells in transgenic lines carrying t19sm (M4, M8A, M8Y, Ml1, M13) encoding both the secreted (/As) and the membrane (u') form of the tt H-chain. Here, its expression was accompanied by suppression of the endogenous immunoglobulin H-chain gene synthesis due to a mechanism analogous to allelic exclusion (Table I; Rusconi and Kohler, 1985; Weaver et al., 1986; Iglesias et al., 1987; Storb, 1987). Transgenic lines expressing A 19s (M41, M42, M85, M86) produce only the secreted (us) and not the membrane form(am) of the it chain. Offspring from M41 and M42 did not express transgenic A (IgMa) on the surface of their B lymphocytes; mice of the M85 and M86 lines showed surface expression within some ofthe B lymphocytes (Table I). This surface expression of transgenic A chains was always accompanied by the expression of endogenous A chains and could be due to Fc receptor binding of transgenic A chains. Alternatively, B lymphocytes could carry mixed IgM molecules on their surface. They could be composed from transgenic and endogenous It chains and linked to the cell surface via the transmembrane domains of the endogenous / proteins. Depletion of CD8+4+ thymocytes and impairment of thymocyte development in transgenic mice During the analysis of our transgenic mouse lines we observed that their thymus was consistently smaller compared to the controls. This suggested to us that the expression of the transgenic anti-Lyt2 /A H-chain gene might have disturbed thymocyte development.
Table I. Expression of cell surface determinants on spleen cells as analysed by flow cytometry % of positive spleen cells
Mouse line
IgMa+
IgMb+
36.2 i 12.9 28.0 ± 3.0 0.25 + 0.25 10.8f ± 5.0 0
4.9 29.0 46.5 55.6 61.3
M4 M8A Ml M13a
M8Yb M41 M42C M85 M86d
Controle
Cells/spleen (x M4 M8A Mll M13 M8Y M41 M42 M85 M86 Control
107)
2.0 5.8 8.0 5.1 10.0
i ± + ± A
i
CD4+8-
9.2 13.0 5.5 9.8 8.8
± 2.7 + 2.0 ± 1.5
12.0 ± 4.0 20.0 + 1.0 11.0 + 5.0
± 2.8 + 2.6
11.8 ± 3.8 14.4 ± 3.2
IgMb+
CD8+4-
CD4+8-
1.2 29.0 46.5 28.3 61.3
1.9 14.0 5.5 4.8 8.8
4.5 8.0 13.5 5.4 6.0
Positive cells/spleen (x 106)
IgMa+ 0.7 0.1 1.0 1.8 3.2
+ + X ±
CD8+4-
7.3 28.0 0.2 5.1 0
+ + + +
3.8 3.0 0.1 2.7
+ 1.0 8.0 ± 13.5 + 10.3 ± 6.0
± 0.9 ± 1.0 + 1.5 f 1.4 + 2.6
2.5 19.5 11.0 5.7 14.4
+ + + + ±
Data represent the average of individually analysed spleen cells. analysed: M4.10 (17 weeks old), M4.17 (4 weeks), M4.18 (7 weeks), M8A.48 (5 weeks), M11.332 (11 weeks), M13.34 (7
aSix mice weeks).
(/19sm)
bTwo mice (jl9Sm) analysed: M8Y.231, M8Y.381, both were 10 weeks old. M41.176, M42.184 (4-5 weeks). dFive mice (jd9S) analysed: M85.346 (7 weeks), M85.347 (4 weeks), M85.348
cTwo mice (jl9S) analysed:
eFourteen control mice were analysed (4-17 weeks of age). fTransgenic IgMa was co-expressed with endogenous IgMb.
3720
(4 weeks), M86.284 (7 weeks), M86.289 (7 weeks).
1.2 0.5 5.0 2.0 3.0
.
Elimination of CD8+ thymocytes in transgenic mice
cells/thymus). Double fluorescence analysis with anti-CD8 and anti-CD4 antibodies showed that in the A19s" mice the CD8+4+ cell population was strongly reduced to an average of 40% of all thymocytes (Figures 2 and 3A). The fraction of CD8-4- T cells was greatly increased from 4% to an average of 48% of total thymocytes (Figure 3A). Comparison of the actual numbers of thymocytes in 19sm mice revealed that the double negative T cells remained constant if compared to controls (8 x 106 cells/thymus; Figure 3B). The number of CD8+4+ cells, however, was
In order to exclude position effects due to the integration site of the transgene, we analysed offspring of all 13 lines. The results of 10 representative lines are reported here. For the analysis of l9sm transgenic mice we used offspring derived from six lines (M4, M8A, M8Y, MI1, M12, M13). For the A19' mice, offspring of all four lines (M41, M42, M85, M86) were analysed. The differences were striking: thymi of Atl9Sm mice (with the exception of M8Y, see below) contained 10 times less thymocytes (2 x 107 cells/thymus) compared to the controls (2 x 108 -
-
TRANSG EN IC PC 9sm M11,332
.I2@,
CD8
CD8 4 I
4% a21o -1 2
iia
-CD84
1 -7
CONTROL LITTERMATE 149P, I39,2
2, 130 9!9. II4 19 CD8 4
CD8 4
CD8 4
26%
3.5%
84%
.
, 12l
1
++ I, I
9992 1012
a0
-'3141
C-----=
.64ONI
+4 , _il t4
CD84
4
6%
... I'll . . ._
I B;
IO-
CD8 4
I
T 1111 II 1
,S3
I>2
,B
1a
8.9%
3.8% 1 19( I2
T]
CD
I
T Ur
4 9I9
CD4
THYMOCYTE POPULATIONS
Fig. 2. Flow cytometric analysis of thymocytes. Thymocytes were stained first with biotinylated anti-CD8 antibodies and counterstained with streptavidin-phycoerythrin complex and FITC-labelled anti-CD4 antibodies and subjected to double-fluorescence analysis. The four-quadrant analysis of thymocytes isolated from transgenic (MI1.332) and control mice represents a typical example for the depletion of the CD8+4+ subpopulation in pC/l9Sm mnice.
A w
90
>-
80-
o
70.
I
605040.
X U.
0 X
B
10 0
2 U1)
::: ::: :::
x
I
i
,
1E
U1) -j
30 20-
-j
LU
10-
h..--;. ;
A B C D E
A B C D E
A B C D E
CD8 4 CD8 4 CD8+4 THYMOCYTE POPULATIONS
Al
v.1
I
.X-
A B C D E
A B C D E
CD4 8
CD8 4 CD8 4 CD8 4 THYMOCYTE POPULATIONS
A B C D E
A B C D E
A B C D E
CD4 8
Fig. 3. Flow cytometric analysis of thymocytes derived from transgenic and control mice. Thymocytes were stained as described in Figure 2, the plotted data represent the average.of individual analyses. The individual thymocyte populations of the transgenic lines are shown in their relative size (A) and in their absolute size (B). (A) represents the average of 11 analysed mice from lines M4, M8A, Ml 1, M12 and M13 (transgenic for sl9s19, high expression in thymocytes); (B) the average of four mice of the M8Y line (u19S', very low intrathymocyte expression); (c) the average of three mice from lines M41 and M42 (ju9S, very low intrathymocyte expression); (D) the average of 10 mice from lines M85 and M86 (19s, high expression in thymocytes). (E) represents the average of 16 control littermates analysed in parallel to the transgenic lines. The standard deviation (thin line) is given in (A).
3721
F.Brombacher et al.
drastically changed and reduced to an average of 7.5 x 106 cell/thymus (controls: 1.6 x 108/thymus). The number of mature CD8+4- cells was reduced by a factor of 7.5, the number of CD4+8- thymocytes by a factor of 15. As the immature CD8+4+ thymocytes are located primarily
in the cortex, their reduction in population size leads also to visible morphological differences between normal and transgenic thymi. In a normal thymus, the cortex is densely populated by CD8+4+ cells. The thymus of an anti-CD8 vsm transgenic mouse, however, is not only much smaller
4.
E* ."~k~ .
;
,; z
.'
)(
.
C,
WI~~~~I C
Fig. 4. Histological analysis of thymi derived from transgenic and control mice. Frozen 6 pm sections were stained with hematoxylin/eosin (A and B, lOOX), or with FITC-labelled anti-p antibodies (C and D, 430x). As shown by the H/E staining, the transgenic thymus (M12.102, A) is much less populated with T cells and lacks a clear distinction between cortical (C) and medullary (M) areas if compared to a control littermate (B). Staining with FITC-labelled anti-p antibodies visualizes the expression of the pl9sm transgene within thymocytes of M12.102 (C). In the control, only intrathymic B cells are stained (D).
A
S
Fig. 5. Immunofluorescence with thymocytes. (A) Transgenic and (B) control cells. Thymocytes from M4.460 were stained in suspension with FITC-labelled anti-CD8 antibodies (green fluorescence), fixed with 2.5% acetic acid, 97.5% methanol at 0°C for 20 min and stained for intracellular pt expression with Texas Red labelled anti-p antibodies (red fluorescence). Transgenic p is expressed in CD8+ as well as in CD8- thymocytes. The strongly fluorescing cell in (A) is an intrathymic B cell.
3722
Elimination of CD8+ thymocytes in transgenic mice
in size but also lacks a clear distinction between cortical and medullary areas (Figure 4A and B). Flow cytometric analysis with thymocytes from Al9s mice revealed a similar reduction of CD8+4+ cells for the offspring of M85 and M86 lines. Mice of the M41 and M42 lines did not suffer from this depletion and were similar to the M8Y line. The depletion of CD8+4+ thymocytes depends on intracellular expression of the u transgene There are several explanations for the elimination of the double-positive thymocytes. The integration of the transgene could have directly perturbed a gene involved in T cell development. This is extremely unlikely since we analysed offspring of all founder mice. The reduction of CD8+4+ thymocytes could also be caused by interactions between soluble anti-CD8 IgM antibodies and CD8 accessory molecules on thymocytes or by direct contacts between double-positive thymocytes and intrathymic B cells expressing anti-CD8 IgM antibodies. Such interactions could potentially inhibit thymocyte differentiation and therefore result in the depletion of CD8+4+ T cells. We therefore searched in the B cell repertoire of /1l9sd mice for the presence of anti-CD8 antibodies formed by combination of an appropriate endogenous L-chain with the transgenic tl9sm H-chain. As shown by Zhu et al. (1984), L-chains could complement H-chains to their original specificities and antibodies reactive against the hapten TNP or red blood cells were regenerated in hybridomas at frequencies ranging from 1/10 to 1/200. We generated hybridomas from LPS-stimulated spleen cells of A19sm mice and analysed them for the expression of the transgenic H-chain (IgMa allotype) and their potential antiCD8 specificity. Among > 5000 IgMa-positive hybridomas we were unable to detect the original anti-Lyt2.2 specificity or binding to surface CD8. In a control experiment we transfected pCjt19 into the myeloma line X63.Ag8.0 and used a it-positive transfectant for fusions with LPS-stimulated C57BL/6 spleen cells. Here, anti-Lyt2.2-specific hybridomas were found in 6/2400 IgMa-positive clones. This result shows that in Itl9sm transgenic mice the appropriate combination with an endogenous L-chain occurs at an at least 10 times lower frequency. As interpretation we favour the deletion of anti-Lyt2.2-specific B cells from the repertoire of transgenic B lymphocytes (manuscript in preparation). In addition, complement-dependent cytolysis and immunofluorescence staining analysis did not reveal anti-CD8 specificities in the serum of our transgenic mice. In contrast to our mice, other transgenic lines expressing It genes contained serum antibodies of the orginal specificity associated with the it chains (Grosschedl et al., 1984; Storb, 1987; our own unpublished results). Transgenic anti-CD8 antibodies might also have caused a specific depletion of CD8+ peripheral T cells. This, however, was also not observed. Therefore, the elimination of thymocytes in the transgenic mice is most likely not caused by the action of anti-CD8 antibodies formed by transgenic H- and endogenous L-chain molecules. Alternatively, the mechanism of T cell depletion could reside within the T cell compartment itself: immunoglobulin i transgenes are frequently expressed in thymocytes as intracellular it proteins (Storb, 1987; Lamers et al., 1989; Figure 4C). These H-chain proteins are not part of functional immunoglobulin molecules, since L-chain genes are neither
rearranged nor expressed in the thymus. This, however, does not exclude intracellular interactions between the transgenic i H-chain and CD8 proteins within thymocytes. They might potentially prevent CD8 surface expression and, as a consequence, lead to the observed disturbances in T cell differentiation. In order to test this hypothesis, we analysed thymocytes derived from the transgenic mice for intracellular A expression by immunofluorescence (Figure 5) and by flow cytometry of permeabilized cells (Table II). Both tests revealed intracellular expression of the transgenic i protein in 50% of the cells within all populations of thymocytes. The flow cytometric analysis showed that in the remaining CD8+4+ and CD8+4- thymocytes subsets, i.e. cells which escaped their elimination, the levels of CD8 surface -
Table II. Intracellular expression of transgenic A in thymocytes
Thymocyte population
% of total thymocytes
% of j+'I+/total in % thymocytes
CD8+4CD8+4+
10.4 50.8 28.2 9.7
7.4 21.2 11.6 5.2
CD8-4-
CD4+8-
71 42 41 52
Thymocytes of the 19Asm transgenic mouse M13.420 were stained first for surface CD8 and CD4 expression, permeabilized and then stained for intracellular i expresson (described in detail in Materials and methods). The individual populations were analysed on the flow cytometer. All four major thymocyte populations show intracellular expression of the transgenic A chain.
E)
E
F G H
Fig. 6. Immunoprecipitation of transgenic 1t chains from thymocytes. jt chains from thymocytes of transgenic mice were labelled in vivo
with [35S]methionine and separated by PAGE. In all lanes the precipitates of 107 lysed cells were loaded except for lane B where 5 x 106 cells were applied to the gel. In parallel, total thymocyte populations were analysed by flow cytometry for CD8 and CD4 surface expression. Lane A shows the 1t chain precipitates of X63Ag8.0 myeloma cells transfected with pCj19; it serves as control for the two different 1t forms expressed from the vector DNA: the 73 kd precursor protein of the membrane form of y and the 70 kd precursor of the secretory form. Lane B shows the precipitate of M4.542 (tu9sm), lane C of M8A.520 (ul9sm), lane D of M86.576 (ul9s), lane E of M42.585 (u19s), lane F of M8Y.566 (ul9Sm). Lanes G (H19.152) and H (C57BL/6) serve as controls: H19.152 is a transgenic mouse line that carries a construct similar to Al9sm containing a V-segment of different specificity (a-TCR V,B8).
3723
F.Brombacher et al.
expression were unchanged and comparable to those found in control littermates. By immunoprecipitation of intrathymocyte It protein combined with flow cytometric analysis carried out in parallel we were able to correlate the differences is the it expression levels within the transgenic lines to the degree of thymocyte depletion: mice with a high level of intracellular Itl9S' or /Al9s expression (M4.542, M8A.520, M86.576; Figure 6, lanes B-D) suffered also from a strong depletion of CD8+4+ cells (3.5 x 106, 1.9 X 107, 6.7 x 106 CD8+4+ cells/thymus respectively). Mouse M42.585 (Figure 6, lane E), which expressed ji19s at a lower level, had a normal population of CD8+4+ (1.6 x 108 cells/thymus) if compared to the C57BL/6 control (1.6 x 108 CD4+8+ cells/thymus; Figure 6, lane H). In the M8Y line (M8Y.566; Figure 6, lane F) intrathymocyte expression of the ,ul9sm chain was very low too. Again, the thymocyte compartment remained unchanged (1.9 x 108 CD8+4+ cells/thymus). As an additional control we used thymocytes from another transgenic mouse line containing a constuct very similar to ,u19sm (H19.152; Figure 6, lane G). It differs by an exchange of its V-region which was derived from the hybridoma line F23.1 producing monoclonal antibodies specific for TCR V38 chains (unpublished results). These mice also express their transgenic It H-chain in thymocytes at high levels but do not suffer from an elimination of double-positive thymocytes (1.2 x 108
CD8+4+ cells/thymus).
Discussion Our experiments demonstrate that the T cell depletion found in the /t 19Sm and /t 19s mice is correlated to the specificity and to the expression level of the transgene within thymocytes. An effect mediated by high-affinity, anti-CD8 antibodies is very unlikely since we were unable to detect them in our transgenic mice. Several lines of evidence support the hypothesis that transgenic it chain expression in thymocytes leads to the elimination of CD8+4+ cells and of their descendants. Mice of the M8Y line, which hardly show intrathymocyte expression of the ,Lsm transgene, do not suffer from the T cell depletion although they do express the transgene in their B lymphocytes (Table I and Figure 3). A contribution of low-affinity, anti-CD8 antibodies to the CD8+4+ cell elimination also seems improbable. B cells producing these antibodies might have remained undetected in our analysis and might have escaped their elimination from the repertoire. But they should be present in the M8Y mice as well as in the other It 19S' lines and their effects should therefore be observed in both types of transgenic mice. This, however, was not the case. In addition, flow cytometric analysis of thymocytes obtained from thymic lobe cultures of transgenic (M4, It 19Sm) and control lobes revealed also a strong depletion of transgenic CD8+4+ cells (K.Eichmann, unpublished results). Here, any influence from peripheral anti-CD8-specific B lymphocytes can be excluded. Although we do not yet have direct evidence for binding of the transgenic It chain to CD8 proteins, we favour-in line with our experimental datathe hypothesis that the depletion of CD8+ cells from the thymus is caused by intracellular interactions between the transgenic it chain and Lyt2 molecules. Binding of either H-chain or L-chain molecules to antigen has been observed although at lower affinities if compared to the complete
3724
antibody (Day, 1972; Nisonoff et al., 1975). Insertion of the Asm form into internal membranes appears to facilitate the depletion. It is more effective in some of the A'sm mice although these lines express less intrathymocyte A protein if compared to corresponding As mice (Figure 6). Membrane insertion of yi and CD8 would allow high local concentrations of both proteins favouring protein-protein interactions. The As protein which lacks the transmembrane part of It present in the Asm form would not remain inserted in membranes and therefore be less favoured to interact with CD8. These interactions depend on the presence of the specific V-region of At. Other transgenic Asm mice with different V-regions show also H-chain expression in thymocytes (Figure 6, lane G; Storb, 1987; Lamers et al., 1989). An effect on their thymocyte populations, however, was not detected. Our experiments therefore suggest specific intracellular interactions between transgenic i and CD8 proteins as the most probable cause for the depletion of CD8+4+ thymocytes and their descendants. In particular, this is implied by the fact that the CD8-4- precursor cells remain unaffected, whereas CD8+4+ cells and their differentiation products are efficiently eliminated. Such interactions could, for example, prevent the transport of CD8 to the cell surface. Retention of /L and concomitantly of CD8 within thymocytes could be enhanced by BiP, a ubiquitously expressed binding molecule of the endoplasmic reticulum (ER) (Haas and Wabl, 1983; Munro and Pelham, 1986). BiP prevents H-chain transport to the cell surface by associating post-translationally with nascent H-chains in the ER until they assemble with L-chain molecules (Hendershot et al., 1987; Hendershot and Kearney, 1988). During their maturation pathway from CD8-4- to CD8+4+ cells, thymocytes proliferate extensively. CD8, which was found to be associated with a tyrosine kinase activity, is thought to exert growth regulatory activities required for the activation and proliferation of T cells (Eichmann et al., 1987; Barber et al., 1989). The inhibition of CD8 surface expression might result in a differentiation block in the transition from the CD8-4precursors to the immature CD8+4+ thymocyte population at the stage of early CD8+4- cells which are thought to develop into double positives (Schwartz, 1989). This would, as a consequence, not allow the proliferation required to build up the large population of CD8+4+ cells. Both mature thymocyte populations, the CD8+4- and the CD4+8- cells, are also highly reduced. This suggests that CD4+8- cells mature from the population of immature CD8+4+ thymocytes. A precisely timed CD8 surface expression at the transition step from CD8-4- to CD8+4+ cells seems therefore to be a prerequisite for all further events in thymocyte development. Our experiments also demonstrate that specific immunoglobulin chains may be used to interfere with normal development. This opens new applications using targeted Ig gene expression in transgenic mice to study complex differentiation pathways.
Materials and methods Cell lines The anti-CD8 (Lyt2.2) hybridoma cell line 19/178 was obtained from G.Hammerling (Hammerling, 1978). A 19/178 H-chain loss variant was isolated by analysing individual clones obtained from serial dilutions for IgG2A synthesis in a specific ELISA. A gamma2A-/kappa+ subclone (19/178.5-) was used for further transfection experiments.
Elimination of CD8+ thymocytes in transgenic mice DNA constructions and transfection experiments Plasmid pCA 19 was constructed as follows: the productively rearranged Hchain allele of 19/178 was identified on a 4.8 kb EcoRI fragment by comparing Southern hybridization patterns of 19/178 and its H-chain-negative variant 19/178.5- using a H-chain enhancer probe (pJl 1). Size-selected, EcoRI-digested genomic 19/178 DNA was cloned into Xgt/wes. Phage libraries were screened with pJ I 1, EcoRI fragments of positive phage clones were further subcloned into pUC 19. Plasmid pC,u was generated by ligating a 9.1 kb EcoRl -XoI fragment carrying the complete constant part of the it gene (allotype a) and a 1.7 kb BamHI-Sall fragment of pXX529 containing a G418 selection marker equipped with a V-segment promoter and the HSV-tk gene poly A site into EcoRI-BamHI cleaved pUC19 vector DNA. The genomic 4.8 kb EcoRI fragment encoding the anti-Lyt2.2 V-segment of 19/178, promoter and Ig H-chain enhancer sequences was then introduced into the unique EcoRI site of pCy in order to generate plasmid
pC,ul9. DNA transfections into 19/178.5 - hybridoma cells were carried out either by electroporation (Chu et al., 1987) using 10-50 Atg of AatIIlinearized pCy19 DNA or by lipofection using DOTMA (Felgner et al., 1987), a generous gift of Synthex Research, Palo Alto, with supercoiled or linearized DNA. G418-resistant tranfectants were selected with 1 mg/ml G418. The transfection frequencies in the lipofection experiments were 10-100 times higher than for electroporation.
Production of transgenic mice Transgenic mice were produced by microinjecting purified fragments of pCA19 into one of the pronuclei of fertilized eggs obtained from superovulated C57BL/6 females mated to Fl (C57BL/6 x SJL) males. Viable eggs were reimplanted into the oviduct of pseudopregnant NMRI females. Tail DNA was analysed for the presence of the transgene by Southern blotting. Transgenic i expression in serum was tested by an IgMa allotype-specific ELISA. Founder mice were back-crossed to C57BL/6 mice.
Immunological techniques and flow cytometry Cytotoxicity tests with culture supernatants of transfectants or with sera from transgenic mice were carried out by incubating 5 x 106 C57BL/6 or 2,B16 (congenic Lyt2. 1 mice) derived thymocytes with 50 y1 supernatant containing the relevant antibody and 1/8 vol of baby rabbit complement (Gibco) at 37°C for 20 min. Dead cells were identified by inclusion staining using Trypan blue. IgMa was tested by a specific ELISA: 96-well microtiter plates were coated with the IgMa-specific antibody RS3.1 (Schuppel et al., 1987), incubated with serial dilutions of culture supernatant or serum and developed with a phosphatase or peroxidase conjugated ut-specific antibody (Southern Biotechnology). Immunoprecipitations were carried out using lysates from [35S]methionine in vivo labelled thymocytes. Transgenic 1t chain was precipitated with 5 yg rabbit anti-ai serum and protein A-Sepharose. Precipitates corresponding to 107 thymocytes were separated by PAGE. Labelled proteins were visualized by autoradigraphy. Flow cytometric experiments were performed on a Becton Dickinson FACScan using the following FITC- or biotin-conjugated antibodies: M41 for ti (Leptin and Melchers, 1983), RS3. 1 for IgMa (Schuppel et al., 1987), MB86 for IgMb, RA33.A1 for B220 (Coffmann and Weismann, 1981) F23. 1 for V,B8 (Staerz et al., 1985), GK1.5 for CD4 (Dialnyas et al., 1983), 19/178 for Lyt2.2 (Hammerling et al., 1978), anti-CD8, anti-thy 1.2, antiA and anti-thy 1.2 were from Becton Dickinson or Southern Biotechnology. Spleen or thymus cells (106) were stained for 20 min at 0°C with biotinand/or FITC-conjugated antibodies and, if required, with phycoerythrinstreptavidin complex in PBS buffer containing 3 % BSA and 0.1 % sodium azide. For each sample, 7000 or 10 000 cells were analysed by the flow cytometer. Intracellular 1z expression in thymocytes was analysed by staining cells first for surface proteins and fixing them thereafter for 4 h by adding 3 vol of a 1:1 0.2 M Na-cacodylate, pH 7.4, 8% paraformaldehyde solution to cells which were resuspended in a PBS, 10% FCS, 0.5% DMEM buffer. Fixed cells were washed three times in PBS and incubated for 30 min in PBS, 30% FCS at 0°C. Then, 106 cells were stained with 50 A1 FITClabelled anti-it antibodies for 1 h, washed in PBS and analysed in the flow cytometer as described above. (J.Hartsuiker and D.Opstelten, personal communication.) Histology and immunofluorescence Thymus cryosections were performed on a Reichert cryostat. Sections (6 itm) were fixed in acetone for 2 s and stored at -200C. The tissue sections were stained with FITC-labelled anti-t antibodies (Southern Biotechnology) diluted 1:10 in PBS containing 0.5% horse serum. Cell suspensions were stained for CD8 with FITC-labelled 19/178 antibodies as described for flow
cytometric analysis. Subsequently, the cells were fixed in 95% methanol, 5 % acetic acid for 10 min at 0°C and intracellular it protein was visualized with Texas Red labelled anti-ti antibodies (Southern Biotechnology). For the hematoxylin/eosin staining, sections were fixed first at 0°C for 3 h in a buffer containing 2% paraformaldehyde, 0.01 M Na-perjodate, 17 mM Na2HPO4 and 33 mM lysine, pH 7.4, and then washed in the same buffer lacking paraformaldehyde (McLean and Nakane, 1974).
Acknowledgements We wish to thank Sabine Kleinhans for her excellent technical assistance and for her help in preparing the manuscript. We want to thank Pascale Renard for her skilful technical assistance and for her help in generating and maintaining transgenic mouse lines. We also want to thank Klaus Eichmann and Michael Reth for critically reading the manuscript. The hybridoma line 19/718 was provided by G.Hammerling.
References Adkins,B., Mueller,C., Okada,C.Y., Reichert,R.A., Weissman,I. and Spangrude,G.J. (1987) Annu. Rev. Immunol., 5, 325-365. Barber,E.K., Dasgupta,J.D., Schlossman,S.F., Trevillyan,J.M. and Rudd,C.E. (1989) Proc. Natl. Acad. Sci. USA, 86, 3277-3281. Benjamin,R.J. and Waldmann,H. (1986) Nature, 320, 449-451. Chu,G., Hayakawa,H. and Berg,P. (1987) Nucleic Acids Res., 15, 1311-1326. Coffmann,R.L. and Weismann,I.L. (1981) Nature, 289, 681-683. Day,E.D. (1972) Advanced Immunochemistry. Williams and Wilkins, Baltimore, pp. 211-213. Dembic,Z., Haas,W., Zamoyska,R., Parnes,J., Steinmetz,M. and von Boehmer,H. (1987) Nature, 326, 510-511. Dialnyas,D.P., Quan,Z.S., Wall,K.A., Pierres,A., Quintans,J., Loken,M.R. and Fitch,F.W. (1983) J. Immunol., 131, 2445-2451. Eichmann,K., Johnson,J.-I., Falk,I. and Emmerich,F. (1987) Eur. J. Immunol., 17, 643-650. Felgner,P.L., Gadek,T.R., Holm,M., Roman,R., Chan,H.W., Wenz,M., Northrop,J.P., Ringold,G.M. and Danielsen,M. (1987) Proc. Nati. Acad. Sci. USA, 84, 7413-7417. Gebert,J., Couglet,C., Zamoyska,R., Parnes,J.R., Schmitt-Verhulst,A. and Malissen,B. (1987) Cell, 50, 545-554. Grosschedl,R., Weaver,D., Baltimore,D. and Constantini,F. (1984) Cell, 38, 647-658. Haas,I. and Wabl,M. (1983) Nature, 306, 387-389. Hammerling,G.J., Lemke,H., Hammerling,U., Hohmann,C., Wallich,R. and Rajewsky,K. (1978) Curr. Top. Microbiol. Immunol., 81, 100-106. Hendershot,L.M. and Kearney,J.F. (1988) Mol. Immunol., 25, 585-595. Hendershot,L.M., Bole,D., Kohler,G. and Kearney,J.F. (1987) J. Cell. Biol., 104, 761 -767. Iglesias,A., Lamers,M. and Kohler,G. (1987) Nature, 330, 482-484. Kappler,J.W., Roehm,N. and Marrack,P. (1987) Cell, 49, 273-280. Kisielow,P., Bluthmann,H., Staerz,U.D., Steinmetz,M. and von Boehmer,H. (1988b) Nature, 333, 742-746. Kisielow,P., Hirst,J., Shiku,H., Beverly,P., Hoffman,N.K., Boyse,E. and Oettgen,H. (1975) Nature, 253, 219-200. Kisielow,P., Teh,H.S., Blithmann,H. and von Boehmer,H. (1988a) Nature, 335, 730-733. Lamers,M.C., Vakil,M., Kearney,J.F., Langhorne,J., Paige,C.J., Julius,M.H., Mossmann,H., Carsetti,R. and Kohler,G. (1989) Eur. J. Immunol., 19, 459 -468. Ledbetter,J.A. and Herzenberg,L.A. (1979) Immunol. Rev., 47, 63-96. Leptin,M. and Melchers,F. (1983) J. Immunol. Methods, 59, 53-61. MacDonald,H.R., Lees,R.K., Schneider,R., Zinkernagel,R.M. and Hengartner,H. (1988a) Nature, 336, 471-473. MacDonald,H.R., Schneider,R., Lees,R., Howe,R.C., Acha-Orbea,H., Festenstein,H., Zinkernagel,R.M. and Hengartner,H. (1988b) Nature, 332 40-45. MacDonald,H.P., Hengartner,H. and Pedrozzini,T. (1988c) Nature, 335, 174-176. MacDonald,H.R, Howe,R.C., Pedrazzini,T., Lees,R.K., Budd,R.C., Schneider,R., Liao,N.S., Zinkernagel,R.M., Louis,J.A., Raulet,D.H., Hengartner,H. and Miescher,G. (1988d) Immunol. Rev., 104, 157-182. Marrack,P., Lo,D., Brinster,R., Palmiter,R., Burkly,L., Flavell,R.H. and Kappler,J. (1988) Cell, 53, 627-634. McLean,I.W. and Nakane,P.K. (1974) J. Histochem. Cytochem., 22,
1077-1083.
Munro,S. and Pelham,H.R.B. (1986) Cell, 46, 291-300.
3725
F.Brombacher et al. Nisonoff,A., Hopper,J.E. and Spring,S.B. (1975) The Antibody Molecule. Academic Press, London, pp. 254-259. Pircher,H., Mak,T.W., Lang,R., Ballhausen,W., Ruedi,E., Hengartner,H., Zinkernagel,R.M. and Burki,K. (1989) EMBO J., 8, 719-727. Rusconi,S. and Kohler,G. (1985) Nature, 314, 330-334. Schuppel,R., Wilke,J. and Weiler,E. (1987) Eur. J. Immunol., 17, 739-741. Schwartz,R.H. (1989) Cell, 57, 1073-1081. Sha,W.C., Nelson,C.A., Newberry,R.D., Kranz,D.M., Russel,J.H. and Loh,D.Y. (1988a) Nature, 336, 73-76. Sha,W.C., Nelson,C.A., Newberry,R.D., Kranz,D.M. Russel,J.H. and Loh,D.Y. (1988b) Nature, 335, 271-274. Smith,L. (1987) Nature, 326, 798-800. Staerz,U.D., Rammensee,H., Benedetto,J. and Bevan,M. (1985) J. Inmmunol., 134, 3994-4000. Storb,U. (1987) Annu. Rev. Immunol., 5, 151-174. Teh,H.S., Kisielow,P., Scott,B., Kishi,H., Uematsu,Y., Bluthmann,H. and von Boehmer,H. (1988) Nature, 335, 229-233. Teh,H.S., Kishi,H., Scott,B. and von Boehmer,H. (1989) J. Exp. Med., 169, 795-806. von Boehmer,H. (1988) Annu. Rev. Immnunol., 6, 309-326. Weaver,D., Reis,M.H., Albanese,C., Constantini,F., Baltimore,D. and Imanishi-Kari,T. (1986) Cell, 45, 274-259. Zhu,D., Lefkovits,I. and Kohler,G. (1984) J. Exp. Med., 160, 971-986.
Received on July 5, 1989; revised on August 21, 1989
3726
