The Programmed Cell Death of an Immature Thymocyte Cell Line Transgenic for an αβ TCR and the c-myc Proto-Oncogene

The c-myc proto-oncogene linked to the mouse Thy-1 gene transcriptional unit predisposes mice to development of thymic tumors consisting predominantly of immature CD4+ CD8+ cells. In an attempt to immortalize immature T cells expressing a known T-cell antigen receptor (TCR), Thy-1/c-myc transgenic mice were bred to αβ TCR transgenic mice (F5), and CD4+ CD8+ cell lines were established from thymic tumors in double-transgenic mice. These cells expressed high-level heat-stable antigen (HSA) and were able to undergo programmed cell death upon induction with steroids and CD3 cross-linking, but not with cognate peptide. In addition, one line had rearranged and transcribed endogenous TCR c and genes, in spite of the fact that transgenic α and β genes were also expressed. Furthermore, we show that Thy-1/myc transgenic mice deficient in recombination activating gene-1 (RAG-1) do not develop tumors, in contrast to RAG-1-/- mice, which are also transgenic for both Thy-1/myc and the F5 TCR. This indicates that in order for thymocytes to be transformed by the Thy-myc transgene, they need to proceed to the double-positive stage.


INTRODUCTION
Overexpression of the c-myc proto-oncogene in humans has been implicated in the development of Tand B-cell leukemia, in which chromosomal translocations bring the c-myc in close proximity to either TCRor immunoglobulin-gene regulatory elements (for reviews, see Dalla-Favera et al., 1982;Casares et al., 1993;Gauwerky and Croce, 1993;Stephenson et al., 1993). In an attempt to establish immature thymocyte lines, we previously generated a transgenic mouse model carrying the murine c-myc under the control of regulatory sequences of T-cell-specific Thy-1 gene. Several lines of Thy-1/ c-myc (TM) transgenic mice developed thymic tumors from which cell lines of both nonadherent (thymocytes) and adherent (epithelial cells) phenotypes were established (Spanopoulou et al., 1989). All established thymocyte lines from these mice were CD4 CD8 double-positive and expressed high levels of transgenic c-myc. Analysis of TCR-"Corresponding author. gene rearrangements in the thymic tumors revealed that thymocytes were of oligoclonal origin, suggesting that oncogenic transformation occurred at the CD4 CD8 stage in a stochastic manner. Those data supported the hypothesis that tumorigenesis is a multistep process and cooperation of c-myc with other genetic mutation(s) is required for neoplastic transformation, as had also been shown by others (Adams and Cory, 1992). It was postulated at the time that the secondary event necessary to cooperate with the myc gene and transform the thymocytes occurred at the double-positive (CD4 CD8 /) stage of development. An in vitro system of differentiation from doublepositive to CD4 single-positive stage has been described for an immature thymocyte line (Kaye and Ellenberger, 1992). In order to generate a system that would generate CD8 single-positive cells from immortalized immature thymocytes, we bred TM mice with mice transgenic for an TCR (F5) that recognizes a peptide from influenza virus nucleoprotein in the context of H-2D b (Mamalaki et al., 1992). The F5/TM double-transgenic mice 280 M. MURDJEVA et al. developed thymic tumors from which we were able to derive immature CD4 CD8 / cell lines. Here we show that these double-positive (DP) cell lines are functionally immature and undergo programmed cell death upon induction with steroids and crosslinking of CD3 by antibodies. In contrast, these cell lines were refractory to treatment with antigenic peptides, suggesting either that the levels of the transgenic TCR are insufficient to mediate antigenspecific signals or that the TCP is uncoupled from CD3-1inked downstream signal-transduction machineries.
In order to avoid complications of endogenous T-cell receptors interfering with selection events in the thymus and affecting the transformation processes F5/TM double-transgenic mice were bred into a Recombination Activating Gene 1 (RAG-l) deficient background .
F5/TM/RAG-1 -/mice developed thymic tumors from which cell lines were isolated that expressed transiently the transgenic TCR. In contrast, TM/RAG-1 -/mice do not develop thymic tumors, indicating that for transformation to take place, thymocytes have to enter the double-positive (CD4 / 8 /) stage of development.

Development of Thymic Tumors in F5.TM25
Double-Transgenic Mice It has previously been shown that different TM mouse lines give rise to thymic tumors of various phenotypes with respect to CD4, CD8, and HSA expression (Spanopoulou et al., 1989; Y.T. unpublished observation). Thus, in an attempt to obtain thymic tumors and immortalized cell lines of immature phenotype, the TM25 line that produced HSAhigh cell lines at a high frequency was backcrossed from (CBA x B10) F to B10 background until they were H-2 bb and then to F5 TCR transgenic mice.
Double-transgenic mice (F5/TM25bb) developed thymic tumors after similar latency periods to the parental TM25 mice. Peripheral lymphadenopathy was also observed in some F5/TM25 mice.
In order to determine the phenotype of cells expanded in the thymic tumor, thymic tumor cells were stained for CD4, CD8, and CD3, V11(transgenic TCR), or HSA; lymphnode cells were also stained with CD4 and CD8. Figure 1 shows analysis of lymphoid tissues from a F5.TM25.1 mouse that developed a thymic tumor but did not have peripheral metastasis. The thymic tumor consisted largely of CD4+CD8 doublepositive (DP) cells (76.7%), but also contained CD8 / (5.2%) and CD4 / (12.4%) single-positive (SP) cells (Fig. l a). The number of SP cells in the lymph node was similar to that of normal mice, but there was a skewing toward CD8 / cells due to positive selection of the class MHC-restricted F5 TCR transgenic T cells (Fig. l b). Thymic tumor cells in the F5.TM25.1 mouse expressed CD3 ( Establishment of Immature DP Cell Lines from F5.TM25.1 Mouse In an attempt to establish immature T-cell lines expressing a known TCR, F5.TM25.1 thymic tumor cells were cultured in vitro as described before (Spanopoulou et al., 1989). Several cell lines and clones were derived, all of which expressed high levels of CD4 and CD8. Figure 2a   RNA was extracted and reverse-transcribed to generate complementary DNA. For PCR amplification of transgenic TCR chains, sense primers for V(z4 and V11 were used in combination with antisense primers for a part of the expression cassette (exon 2 of human CD2 gene located downstream of the cDNA inserts) . As shown in Figure 3a, PCR products of predicted sizes were detected for both transgenic (z and chains, providing evidence that 25.1.2 cells transcribe the F5 TCR 0 and transgenes. Next, in order to assess whether endogenous (z and TCR genes were expressed by these cells, panels of sense primers for different V0 and V segments were used with antisense primers for 0 or constant regions. Figure 3b shows the presence of PCR products from V(z4 (transgenic) and V08 segments, demonstrating that an endogenous TCR (z gene is rearranged and transcribed in 25. In order to examine if 25.1.2 retained characteristics of immature DP thymocytes, 25.1.2 cells were treated for 12 hr with 1 x 10-6 M hydrocortisone, a known steroid inducer of programmed cell death in immature thymocytes (Wyllie, 1980). Total DNA was extracted after treatment and was analyzed in an agarose gel. As shown in Fig. 4a, the steroidinduced programmed cell death (apoptosis) in 25.1.2 cells, produced a prominent ladder of fragmented DNA. F5 TCR-bearing immature double-positive thymocytes respond in vivo or in vitro to cognate peptide (influenza nucleoprotein 1968 (c366-374) by undergoing apoptosis. Thus, we studied the effects of cognate peptide on the F5.TM cell line. 25.1.2 cells were cultured with peritoneal macrophages in the presence of various concentrations of the peptide recognized by the F5 TCR (1968 NP). Figure   4b shows viable cell numbers relative to those seen in cultures without peptide. It appeared that 25.1.2 cells were not affected by cognate peptide, whereas F5 thymocytes were deleted at as low as 10-9 M concentration of the peptide as reported before (Tanaka et al., 1993). CD4 or CD8 levels on 25.1.2 cells did not change after peptide treatment (data not shown). These data are in contrast to the previous observation of steroid-induced programmed cell death of 25.1.2 cells. Such a resistance of 25.1.2 cells to peptide treatment could be either due to insufficient level of TCR expression (already lower than that observed on fresh tumor cells) or a defect in TCR-linked signal-transduction pathways. Thus, if 25.1.2 cells expressed endogenous TCR and ( chains instead of transgenic F5 TCR, as discussed before (Fig. 3), 25.1.2 cells might not recognize the 1968 NP. Alternatively, if the low levels of F5 TCR present on the 25.1.2 cells were sufficient to confer antigen-specific response, there could be uncoupling of the TCR and CD3 or its downstream signal-transduction system in the cell.
In order to resolve this issue, 25.1.2 cells were stimulated by direct cross-linking of CD3a with the antibody 145-2Cll (Leo et al., 1987) (Figure 4c). In contrast, addition of 2Cll at 70 ng/ml or 10 g/ml concentration caused reduction in numbers of viable 25.1.2 cells, which was associated with DNA fragmentation, as shown in Fig. 4d. These data demonstrate that the 25.1.2 cell line has an intact CD3-1inked signaltransduction mechanism, and that the most likely reason for their nonresponsiveness to the antigenic peptide is the low level of transgenic T-cell receptor on these cells.

Unable to Rearrange Endogenous TCR Receptors
In order to avoid the complications of endogenous T-cell receptors being expressed during thymic development of F5/Thy-myc double-transgenic mice, use was made of a mouse line that is unable to rearrange endogenous T-cell-receptor genes. These mice are deficient in the Recombination Activating Gene 1 (RAG-l) . In RAG-1 -/mice, thymocyte development is arrested at the double-negative CD4-8stage. However,s if such mice are crossed to transgenic mice carrying an already rearranged -chain transgene, they proceed to the double-positive (CD4 / CD8 /) stage (Mombaerts et al., 1992).
F5/Thy-myc mice in RAG-1 -/background were generated and these developed thymic tumors at 2 to 6 months of age. Thymic tumor cells were stained with CD4-, CD8-, V11-(transgenic), and HSAspecific antibodies, analyzed on a fluorescenceactivated sorter, and the results are shown in Table  1. Interesting, but without explanation at the moment, is the high incidence of metastasis to peripheral lymphoid organs of these mice as opposed to F5/TM/RAG / / / mice. Cell lines were isolated from these tumors and were analyzed for expression of CD4, CD8, V11, and HSA. Such cells were also unresponsive to antigenic stimulus and parallel studies indicated that whereas fresh tumor cells were expressing high levels of F5 TCR, the lines showed low or undetectable levels (data not shown).
RAG-l-/-/Thy-l-myc Mice Do Not Develop Thymic Tumors As mentioned before, in order for Thy-l-myc thymocytes to be transformed and immortalized, a secondary event is necessary. It has been postulated before that the secondary event probably occurs at the double-positive stage. Thy-l-myc/RAG-1 -lmice offer a unique genetic combination that could further resolve this question. Indeed, we examined over 30 Thy-l-myc/RAG-1-/mice at ages from 3 months to 1 year and none of them showed any signs of tumors. In contrast, as described before, most of the F5 //Thy-l-myc //RAG-l-/mice developed thymic tumors. This observation proves that in order for Thy-myc thymocytes to be transformed, they have to proceed to the double-positive stage in development. There could be two reasons why Thy-l-myc/ RAG-l-/mice do not develop tumors. It is possible that the secondary event that cooperates with the myc transgene for transformation occurs at the double-positive stage. Alternatively, it is possible that the myc transgene is not expressed in the double-negative stage at levels that can transform the thymocytes. In order to assess the latter possibility, RAG-1 -/and RAG-1 +// thymocytes were stained with Thy-1 antibodies. Figure 5 shows that in RAG-l-/mice, the double-negative CD4-8thymocytes express Thy-1 at high levels. This indicates that the Thy-1 promoter is active in these cells, excluding the possibility of transcriptional inactivity of the Thy-myc transgene.

DISCUSSION
Immature T-cell lines were established from thymic tumors in mice transgenic for Thy-1/c-myc and F5 TCR. The 25.1.2 cells express CD4 and CD8 at high levels and are immature as judged by high-level  HSA and their ability to undergo programmed cell death upon induction with steroids. In contrast to steroid treatment, however, 25.1.2 cells did not respond to cognate peptide. This could be due to uncoupling of the TCR/CD3 complex to signaling pathways within the cell. To test this possibility, CD3 was cross-linked on those cells; this led to marked apoptosis, indicating that the CD3 complex is linked functionally to signal-transduction molecules. Therefore, the inability to induce apoptosis in these cells with antigenic peptide is probably due to insufficient levels of expression of the F5 TCR on their surface. Alternatively, these cells may repre-sent that population of immature thymocytes that are resistant to negative selection through TCR cross-linking, because not all of their TCR molecules are coupled to the CD3 complex (Finkel et al., 1989). Thus, signaling through CD3 would be efficient in such cells and lead to apoptosis, whereas engaging the T-cell receptor might be less effective. 25.1.2 cells transcribed both transgenic z and chains, but also contained mRNA for rearranged endogenous c and chain genes. It is generally accepted that allelic exclusion of TCR genes is strict and introduction of rearranged TCR [ genes in transgenic mice prohibits thymocytes from rearranging their endogenous TCR genes. The fact that mRNA for endogenous rearranged TCR genes was found in the 25.1.2 cells could be due to allelic inclusion in F5 TCR transgenic mice. In fact, the CD4 cells developing in these mice are selected exclusively on endogenous receptor chains (z and or ) (Corbella et al., 1994). Alternatively, it is possible that immortalized thymocytes continue to rearrange TCR genes (Malissen et al., 1992).
Allelic exclusion of TCR {z genes, on the other hand, is less strict and expression of dual TCR z chains has been demonstrated in normal T cells (Padovan et al., 1993). Our data on expression of rearranged endogenous TCR genes in the F5 TCR transgenic 25.1.2 cell line support the hypothesis that this line is able to express, at the RNA level, two T-cell receptors. CD3 levels on these cells also suggest that possibly both receptors are expressed on the surface.
In order to exclude the complications arising from allelic inclusion or continuous rearrangement, we crossed the F5/Thy-myc mice to mice deficient for the RAG-1 gene. F5/Thy-myc/RAG-1-/mice developed thymic tumors, thus excluding the possibility that endogenous receptor expression (and signaling) was necessary for transformation of thymocytes. The tumors themselves from both F5/ TM/RAG / and F5/TM/RAG-/mice contained cells that expressed the transgenic F5 chain at appreciable levels. However, when placed in culture, these cells would gradually downregulate the expression of the transgenic receptor in a course of 6 months. It is not clear at the moment why this occurs. One possibility is that they lose copies of the transgene. Nevertheless, transgenic mRNA could still be detected making this an unlikely explanation.
An interesting point that arose from breeding the Thy-myc transgene into the RAG-l-/background was that TM/RAG-1 -/mice did not develop IMMATURE THYMOCYTE CELL LINES 287 thymic tumors. Given that RAG-l-/thymocytes do not differentiate past the double-negative stage, this finding supports our earlier hypothesis that transformation of thymocytes in Thy-myc mice is a rare event and needs, in addition to the myc gene, a secondary event that occurs at the double-positive stage.
For analysis of programmed cell death, 25.1.2 cells were treated with 1 x 10 -6 M hydrocortisone (Sigma) for 12 hr. Total DNA was extracted from 25.1.2 cells and was run in 2% agarose gel containing ethidium bromide as described before (Kawabe and Ochi, 1992).
Peritoneal macrophages from B10 mice were prepared according to Morikawa et al., 1993). Briefly, the mouse peritoneal cavity was washed twice with cold PBS, and adherent cells were collected after incubation for 2 hr at 37C. Adherent cells were further treated with 100 IU/ml murine recombinant IFN-, for 12 hr, and were incubated with different concentrations of 1968 NP peptide (366-374) (synthesized at the National Institute for Medical Research (London) for 90 min prior to coculture with 25.1.2 cells. 2 x 105 peritoneal macrophages were cocultured with 5 x 105 25.1.2 cells for 12 hr. Lymphoid cells were recovered by gentle pipetting, viable cell numbers were counted using trypan blue dye exclusion, and cells were stained for CD4, CD8, and V11. As a control, thymocytes from adult F5 transgenic mice were cocultured with macrophages as before.
A murine macrophage cell line J774A.1 (TIB 67, ATCC) was kindly provided by Dr. G. Stockinger (NIMR). 25.1.2 cells were cocultured with J774A.1 in the presence or absence of anti-CD3 antibody (2Cl1.145) at 0.07 or 10 tg/ml. After culture for 12 hr, fragmentation of DNA in 25.1.2 cells was analyzed as described before. In a separate experiment, 25.1.2 cells and F5 thymocytes were treated with various concentrations of anti-CD3 antibody, and cell numbers and expression of CD4, CD8, and V11 were analyzed.

RNA-PCR Analysis
Total RNA was extracted from 25.1.2 cells by the guanidinium thiocyanate method (Sambrook et al., 1989), and was reverse-transcribed (RT) by Moloney murine leukemia virus reverse transcriptase (Promega) using oligo(dT) primers (Promega). Onesixth of the RT products was amplified by polymerase chain reaction (PCR) in 1.5 mM MgC12, 60 mM KC1, 15 mM Tris-HC1 (pH 8.3), 6.75% glycerol, and 2U Taq polymerase (Perkin Elmer) using thermal cycler (Hybaid). Sense primers for TCR Vc and V gene segments and an antisense primer for TCR c constant region have been described before (Casanova et al., 1991). In addition, the following primers specific for transgenic F5 TCR were synthesized in our institute, V(4: ACCAGAcAAGCT-TCACCTGCcAAGATAT; V(4': CAGTATCCCGG AGAAGGTC; VI 1: CAAGCTCcTATAGAT-GATTC; hCD2: TCAAAATCAGAAGGAAGCTGG; and c: CCTTGGGTGGAGTCACATTTC. PCR reactions were also performed without cDNA as mock controls. The PCR products were analyzed in 2% agarose gels containing ethidium bromide. A molecular size marker consisting of multimers of 123 bp DNA fragments (Pharmacia) was used.