1992 Harwood Academic Publishers GmbH Printed in the United Kingdom Analysis of Immature (CD4-CD8-) Thymic Subsets in T-Cell Receptor Transgenic Mice

Introduction of a transgenic αß TCR (Vα2, Vß8.1) specific for lymphocytic choriomeningitis virus (LCMV), in the context of H-2Db into the genome of C57BL/6 mice, has many effects on the development and selection of T cells in both the thymus and the periphery. These mice produce increased numbers of CD4–8+ mature T cells, all of which express the transgenic TCR, and small numbers of CD4–8+cells using endogenous TCRs are also produced. This study follows the intrathymic development of T cells in these TCR αß transgenic mice, in particular the earliest CD4–8– stages. As expected, the transgenic TCR is expressed on the cell surface at an earlier developmental stage than endogenous TCRs in nontransgenic littermate controls. Of the three major subsets expressing the heat–stable antigen (HSA), only the most mature, the CD25–CD44– expresses the transgenic TCR, and the earlier CD25–CD44+ and CD25+CD44– do not. Furthermore, in contrast to other TCRαßtransgenic lines, TCR γδ lineage cells appear to develop normally.


INTRODUCTION
The development of mature T lymphocytes is a complex process, involving proliferation and differentiation of immature precursor cells through a series of quite well-defined stages. CD4-CD8prothymocytes (originating in the fetal liver or bone marrow) colonize the thymus and subsequently proliferate and differentiate to give rise to CD4/8 / intermediates and ultimately functional cells expressing either CD4 or CD8 (Fowlkes et al., 1985;MacDonald et al., 1988b;Scollay et al., 1988). During this process, the T-cell receptor (TCR) c/ and ,3 genes are rearranged and expressed, and various other phenotypic markers are acquired or lost (Fowlkes and Pardoll, 1989;Pearse et al., 1989). Positive and negative selection events limit the repertoire of expressed TCR c]/genes to avoid autoreactivity and account for MHC-restricted antigen recognition (Kappler and Marrack, 1987;von Boehmer, 1990). This latter selection phenomenon has come under closer scrutiny due to the *Corresponding author. recent availability of transgenic mice expressing TCR c]/with defined antigen and MHC specificity (Sha et al., 1988;Teh et al., 1988;Berg et al., 1989;Kaye et al., 1989;Pircher et al., 1989a). In the present study, we have analyzed immature CD4-CD8-T cell subsets that would normally not express surface TCR from TCR o]/ transgenic mice in order to assess the effects of premature TCR expression on early T-cell development.
'j:."'.'. i.'.:",", "":"' ":"''" log fluorescence ture CD4-CD8-subset is increased both in percentage and in absolute numbers in the transgenics. On the H=2 d background, where there is no positive selection of the transgene-bearing cells, the thymi are of a similar size to their littermate controls, but still have increased numbers and percentage of the CD4-CD8-subset (Table 1).

Expression of Transgenic TCR on CD4-CD8-Thymocytes
The transgenic TCR expressed in these mice can be detected by mAbs directed against V]/8.1 (KJ16) or Ve2 (B20.1). As shown in Fig. 2, a large proportion of CD4-8-thymocytes expressed both the transgenic TCR c and chains. Furthermore, staining with mAbs directed against CD3 or TCR c]/gave similar patterns as V8.1 or V2 staining, indicating that few (if any) nontransgenic TCR c fl were expressed. The same staining patterns were observed on CD4-8-thymocytes in the presence (H-2b), or absence (H-2d) of the selecting element, indicating that this has no effect on the expression of the transgenic TCR.
Subsets of CD4-8-Thymocytes in TCR Transgenic Mice CD4-8-thymocytes can be subdivided into subsets on the basis of expression of various surface markers. Among the most useful markers are heat-stable antigens (HSA), IL-2 receptor cz chain (CD25), and Pgp-1 (CD44), which together define four major subsets (MacDonald et al., 1988b;Scollay et al., 1988). According to current views (Pearse et al., 1989;Petrie et al., 1990), HSA+CD25-CD44 / immature precursors give rise sequentially to HSA/CD25+CD44 and subsequently HSA+CD25-CD44 subsets, the latter being immediate precursors of CD4/8 / cells Petrie et al., 1990). The fourth subset (HSA-CD25-CD44/) is believed not to belong to the normal generative lineage and may result from an aberrant selection process Scollay et al., 1988). Negative controls (second stages alone) have been omitted from this plot to simplify it, but were the same as the negative peaks for the littermate control.

Subsets
The CD4-8-subsets described before were analyzed for expression of the transgenic TCR. Transgenic TCR (Vfl8.1 or Vow2) was found on the relatively mature HSA/CD25-CD44 subpopulation and was not detected on the less mature + + + HSA CD25-CD44 and HSA CD25 CD44-subsets (Fig. 4, and data not shown). The HSAsubset could not be analyzed f-or TCR expression in the transgenics due to the small number of cells present; however, this subset was readily detectable in the control littermates, where many cells expressed V]/8.2 (data not shown; Budd et al., 1987;Fowlkes et al., 1987).

Expression of TCR ?'c in TCR o]/Transgenics
Several studies have suggested that expression of TCR ]/ transgenes leads to allelic exclusion of endogenous TCR ' genes as well as TCR ]/ (Fenton et al., 1988;Uematsu et al., 1988;von Boehmer et al., 1988;Pircher et al., 1990). As shown in Fig. 5 and Table 3, TCR ,c + cells were readily detectable in the CD4-8-subset of the H-2 b or H-2 d TCR o]/transgenics described here.
Although the relative percentage of TCR ,+ cells was lower in transgenics than in littermate controls, the absolute number was not significantly different between the two groups (Table  CD4-CDSthymocytes. Purified CD4-CDSthymocytes were stained in three colors as follows: culture supernatants for the anti-TCR antibodies, second-stage fluorescein conjugates, and rat Ig for blocking were as described in Figs. 2 and 3; this was followed by anti-CD25-PE (direct conjugate), anti-HSA-biotin and finally TR-PE-avidin. Cells were analyzed and live gating performed as described in Fig. 1, except that 100,000 events were collected. The HSA+CD25 and HSA/CD25 were selected by a second gating and the TCR expression on these subpopulations obtained. Solid lines and shading correspond to transgenic and dotted lines to littermate controls. CD25 + thymocytes are lacking in the TCR transgenic thymus. However, CD4-8 / cells with a lower level of CD8 are found in the transgenic thymi in numbers comparable o that of the littermate controls (1.5+0.4% and 1.6+0.6%, respectively; see Fig. 1). As noted elsewhere, this phenotype may correspond to relative immaturity of these cells . The recently described CD4/8-3 "immature" thymocyte showing a similar phenotype and morphology Matsumoto et al., 1989;Hugo et al., 1990) was also virtually undetectable in these transgenic mice.

DISCUSSION
It is well established that the precursors of mature T lymphocytes are contained within the CD4-8-thymocyte subset (Fowlkes et al., 1985;Crispe et al., 1987;Shimonkevitz et al., 1987;Scollay et al., 1988). The d.evelopmental sequence occurring within that population is now more clearly understood (Pearse .et al., 1989;Petrie et al., 1990), although the precise events that direct this differentiation are not. In this communication, we have analyzed CD4-8-thymocytes from mice expressing a transgenic TCR c]/ in order to gain further insights into early T-cell development. In normal mice, TCR c]/ is first expressed at the CD4/8 + stage of thymus development and hence TCR expression in transgenic mice allows cross-correlations with other phenotypic markers believed to define early developmental stages.
The current consensus view of lineage relationships among major CD4-8-subsets (Fowlkes and Pardoll, 1989;Petrie et al., 1990) Teh et al. (1990) were large genomic fragments (Teh et al., 1988;Uematsu et al., 1988), whereas the TCR constructs used here were TCR c and ]/ cDNA clones coupled to a H-2K promoter and Ig heavy-chain enhancer (Pircher et al., 1989b). It is thus possible that these differences in regulatory elements lead. to differential ontogeny of expression of the two TCR transgenes in vivo. Failure to express the transgene on the early CD4-8-subsets is not due to lack of function of the H-2K promoter because H-2K b is expressed at high levels on all of these subsets (data not shown).
In contrast to the three major HSA / subsets of CD4-8thymocytes, HSA-CD4-8-cells appear not to belong to the generative lineage Fowlkes and Pardoll, 1989). Furthermore, this HSAsubset is peculiar in that it appears late in development and most of the cells express TCR o]/ with a strong preference for the V]/8. variable domain (Budd et al., 1987;Fowlkes et al., 1987). Interestingly, HSA-CD4-8-thymocytes were dramatically reduced in our TCR o]/transgenic mice. In contrast, Teh et al. (1990) reported large numbers of HSAthymocytes (~20% of  show the position of the negative controls, and analyses were performed as described in Fig. 1 (von Boehmer, 1990)  HSA log fluorescence FIGURE 7. The "immature" CD4-CD8 thymocyte. CD4thymocytes were isolated from H-2 transgenics or their littermate controls and stained in three colors with, in the following order, anti-CD3 (17A2 culture supernatant), goat antirat Ig PE, rat Ig for blocking, anti-HSA FITC, anti-CD8 biotin, and finally TR-PE-Av. Live gating to exclude dead cells and debris was performed as in Fig. 1, then the CD8 cells were selected by a second live gate on the FL3 (TR-PE-Av/) cells. Single-parameter profiles are shown for the FLS (reflecting relative cell size), CD3, and HSA expression of the CD4-CD8 thymocytes. Solid lines correspond to transgenic and dotted lines to littermate controls.
Transition of thymocytes from the CD4-8to CD4/8 + compartment may involve an intermediate CD4-8 / (MacDonald et al., 1988a; or CD4/8 Matsumoto et al., 1989) cell expressing little or no detectable surface TCR. In mouse, the "immature CD8 population is rapidly cycling and exhibits high levels of HSA. By these criteria, no immature CD4-8 thymocytes could be detected in TCR ofl transgenic mice. The absence of this intermediate subset may result from accelerated differentiation of the immediate CD4-8-precursors of these cells. Indeed, HSA/CD25-CD44 thymocytes were present in elevated numbers in TCR c]/transgenic mice of both the H-2 b (selecting) and H-2 d (nonselecting) backgrounds. In addition, this subset expressed enhanced levels of several phenotypic markers (CD5, Mel-14) and a decreased proportion of cycling cells, both traits that are usually associated with T-cell maturation. Therefore, the premature expression of surface transgenic TCR, even if it cannot be positively selected (as on the H-2 d background), may be associated with accelerated maturation of immature thymocytes. Furthermore, these data provide evidence that the "immature," CD4/8-3 or CD4-8/3 subsets are not obligatory stages in the differentiation from CD4-8to CD4/8 + cells Petrie et al., 1990).
Considerable controversy surrounds thedevelopmental relationship between TCR o]/and TCR ' lineages (Fowlkes and Pardoll, 1989). In this context, several studies have indicated that introduction of a rearranged TCR ]/transgene effectively inhibits further gene rearrangement at the TCR 'locus, presumably due to some transacting allelic exclusion mechanism (Fenton et al., 1988;von Boehmer et al., 1988), and other reports have suggested that a partial inhibitory effect of transgenic TCR ]/on TCR/,.rearrangement occurs, but that the TCR y population remains unchanged . In other studies using and/or ' transgenic mice, complete inhibition (Bonneville et al., 1989), no effect (Dent et al., 1990;Ishida et al., 1990), or enhancement (Ferrick et al., 1989) of the TCR ?'c lineage have been reported. In the TCR ofl transgenic mice reported here, TCR?' + cells were readily detected in the CD4-8-thymocyte subset and absolute numbers of these cells were identical to the values obtained in control littermates. In addition, the TCR 'expressing dendritic epidermal T cells (DETC) isolated from the skin of these TCR transgenic mice were shown to be indistinguishable (both in absolute numbers and V'3 usage) from those found in their littermate controls (P. Ohashi et al., unpublished data In conclusion, the premature expression of a defined TCR c]/ transgene provides a useful model system in which to test the predicted developmental lineage of immature (CD4-8-) thymocyte subsets. Furthermore, the comparison of CD4-8thymocytes expressing different transgenic Vfl domains may ultimately lead to new insights regarding the developmental origin of the HSA-CD4-8-subset.

MATERIALS AND METHODS
bit Company, Buxted, UK) and 10 units/mL of DNase I (Boehringer Mannheim GmbH, Mannheim, West Germany) to prevent excessive clumping of dead cells and debris. Dead cells were removed by centrifugation over Isopaque-Ficoll (Pharmacia, Uppsala, Sweden). To ensure purity of the final preparation, this entire procedure was repeated with the cells at 107/mL. CD4-thymocytes were prepared in the same manner with the omission of the anti-CD8 antibody. In both cases, the rsultant cells were greater than 98% pure.

Mice
All the mice used in this study were bred and maintained at the Swiss Institute for Experimental Cancer Research (Epalinges, Switzerland). The transgenic TCR (VOa,JOfTA31/V8.1,D,J2.4, Cfl2) was derived from a LCMV/H-2Db-specific cytotoxic T-cell clone P14 (Pircher et al., 1987). Transgenic males on either a C57BL/6 (H-2b) or Balb/c (H-2d) genetic background (Pircher et al., 1989a) were heterozygously bred with normal MHC compatible 6-8-week-old females. In this way, a colony of TCR o]/transgenic mice on the selective (H-2b) or nonselective (H-2d) background and their nontransgene bearing littermates were established. Litters were screened for transgene-bearing mice at approximately 4 weeks of age by staining peripheral blood lymphocytes with the monoclonal antibody KJ16, which binds the V/8.1 and V/8. TCRs, followed by goat antirat Ig-FITC, and then counterstaining with a mixture of anti-CD8 biotin and either anti-CD4-PE or anti-CD4 biotin and finally PE-streptavidin. The percent of V]/8-bearing T cells was 15-20% for the nontransgene-bearing littermates and 85-90% for the transgenics.
Analysis of these stained cells was done with a FACS II for the single staining and a FACScan for the two-and three-color fluorescence (both Flow Cytometers were supplied by Becton Dickinson, Mountain View, CA).

ACKNOWLEDGMENTS
We thank Christian Knabenhans and Pierre Zaech for their patience and skill with the flow cytometric analysis, Bernard Malissen for his kind gift of the B20.1 monoclonal antibody, and Anna Zoppi for her help in preparation of the manuscript. (Received May 14, 1991) (Accepted September 25,1991)