As a potential means for facilitating studies of NK cell-related molecules, we examined the expression of these molecules on a range of mouse tumor cell lines. Of the lines we initially examined, only EL4 and RMA expressed such molecules, both lines expressing several members of the Ly49 and NKRP1 families. Unexpectedly, several of the NK-related molecules, together with certain other molecules including CD2, CD3, CD4, CD32, and CD44, were often expressed in a mosaic manner, even on freshly derived clones, indicating frequent switching in expression. In each case examined, switching was controlled at the mRNA level, with expression of CD3ζ determining expression of the entire CD3-TCR complex. Each of the variable molecules was expressed independently, with the exception that CD3 was restricted to cells that also expressed CD2. Treatment with drugs that affect DNA methylation and histone acetylation could augment the expression of at least some of the variable molecules. The striking phenotypic similarity between EL4 and RMA led us to examine the state of their TCRβ genes. Both lines had identical rearrangements on both chromosomes, indicating that RMA is in fact a subline of EL4. Overall, these findings suggest that EL4 is an NK-T cell tumor that may have retained a genetic mechanism that permits the variable expression of a restricted group of molecules involved in recognition and signaling.

Natural killer cells are a major population of lymphocytes that are able to recognize and destroy infected and malignant cells. Despite intensive research, the mechanisms by which this recognition takes place are still poorly understood. However, in recent years it has become clear that an important aspect of NK recognition is the ability of NK cells to detect a reduction or loss of class I/peptide complexes at the surface of diseased cells (1) via a series of receptors that upon ligation with class I molecules deliver inhibitory signals. In the mouse these molecules belong to two families of C-type lectin molecules, the CD94/NKG2 family (2, 3) and the Ly49 family (4), both of which are encoded in the NK complex (5). Considerable experimental evidence indicates that individual NK cells express a restricted but largely random selection of the available receptors (6), but the mechanism by which such a diverse repertoire is created is unknown (7). Analysis of this process would be greatly facilitated if it were possible to study it in culture systems. We recently obtained evidence that during the clonal development of fetal NK cells in vitro individual Ly49 molecules were indeed expressed in a stochastic manner, but the frequency with which this occurred was too low to permit mechanistic analysis (8). More recently, we have found that receptors for the nonclassical class I molecule Qa1 are also expressed in a stochastic manner on developing fetal NK cells, but at much higher frequency than Ly49 molecules, permitting the demonstration that their expression is controlled by specific cytokines (9).

In contrast to the growing body of knowledge concerning inhibitory receptors on NK cells, the nature of the receptors that deliver activatory signals upon contact with diseased cells has remained controversial. One view has been that NK cells express specific receptors for target cell molecules. For example, there is convincing evidence that members of another C-type lectin family, the NKRP1 family, can deliver activating signals to NK cells (10, 11). A second possibility, also supported by considerable experimental data, is that NK cells use more generic receptors for this purpose. For example, some NK cells express Fc receptors, principally CD16 (12), but also CD32 (13), that deliver activating signals following interaction with cell-bound Ig, or with as-yet-unknown structures expressed on some target cells (14). NK cells can also receive activating signals via other molecules, such as CD2 (15), CD28 (16), CD40 ligand (17), and LFA-1 (18). Interestingly, during clonal development of fetal NK cells in vitro, several potentially activating or modulatory surface receptors that are expressed on subsets of NK cells in vivo, including members of the NKRP1, Ly6, and CD45 families, are expressed in a heterogeneous manner (8).

Recently, subpopulations of T cells, designated NK-T cells, have been described that appear to utilize the same recognition modalities as NK cells, but which in addition possess an Ag-specific receptor (for a review, see Bendelac et al. (19)). This receptor may be of either αβ or γδ type, and in some subpopulations of NK-T cells is of restricted diversity, one population of αβ NK-T cells having a largely invariant Vα14-Jα281 α-chain paired with a Vβ8, Vβ7, or Vβ2-containing β-chain that endows these cells with the ability to recognize common microbial lipids presented by CD1 molecules. The expression of CD94/NKG2 (20, 21) and Ly49 (22, 23) receptors on NK-T cells has been documented, but the extent to which these cells express and utilize the same activating receptors as NK cells is unknown.

Tumor cells provide a potentially useful means for studying the biochemistry and genetics of molecules that are expressed on small populations of cells that are difficult to grow in large quantities in vitro. With this in mind, we examined a range of tumor cell lines for the expression of molecules that have been implicated in NK cell recognition. We report here the unexpected findings that 1) both EL4 and RMA cells express a number of NK-related C-type lectin molecules, 2) that the expression of these and certain other surface molecules is subject to a high degree of variation, and 3) that EL4 and RMA cells have identical TCRβ gene rearrangements and presumably therefore have a common origin.

RMA, RMA/S, and a TAP2-transfectant of RMA/S, mtp2, were obtained from Drs. R. Glas and K. Karre (Karolinska Institute, Stockholm, Sweden), who also provided the EL4 line used in most of the studies described in this paper. RBL5 cells were kindly provided by Dr. W. Green (Dartmouth Medical School, Hanover, NH). RMA/S lines transfected with Dd and CD1 were kindly provided by Dr. D. Raulet (University of California, Berkeley, CA) and Dr. M. Kronenberg (University of California, Los Angeles, CA), respectively. These cells, and the B cell lymphoma A20, were cultured in DMEM (52100-039; Life Technologies, Paisley, U.K.) supplemented with 2× nonessential amino acids, 5 × 10−5 M 2-ME, and 5% FBS (F-7524; Sigma, Poole, U.K.) in a 10% CO2 atmosphere at 37o. Cells were cloned by adding 100 μl of medium containing 0.5 or 1 cell/ml to flat-bottom 96-well culture plates. Cell lines were regularly tested for mycoplasma as described previously (24) and found to be negative.

Aliquots of ∼1 × 106 EL4 cells were incubated for 24 h with various doses of 5-azacytidine (freshly prepared), trichostatin A (made up from a 10 mM stock in ethanol stored at −20o), or, as control, 5-azaguanine (made up by dissolving at 4 mM in 0.5 M NaOH, then diluting to 1 mM in PBS, and storing at 4o). The cells were then washed and returned to culture for 5–7 days before staining. All chemicals were purchased from Sigma.

The main Abs used in this study were as follows: KT3 anti-CD3ε (25), YTS 191 anti-CD4 (26), 2.4G2 anti-CD16/32 (27), Pgp1 anti-CD44 (28), A1 anti-Ly49A (29), 5E6 anti-Ly49C (30), 4D11 anti-Ly49G (31), PK136 anti-NK1.1 (32), and 10A7 anti-10A7 (33), all kindly provided by the original investigators, with the exception of 2.4G2 which was obtained from the American Type Culture Collection (Manassas, VA). Aliquots of 2 × 105 cells were incubated with either medium or saturating concentrations of these Abs in HBSS supplemented with 2% FBS and 0.2% sodium azide (H2FA) for 20 min at 4o, washed twice with cold H2FA, then incubated with a saturating concentration of FITC sheep anti-mouse Ig (Harlan SeraLab, Loughborough, U.K.) in H2FA for 20 min at 4o. After a further wash in H2FA, cells were resuspended in H2FA, and single viable cells, selected on the basis of forward and side-scatter characteristics, were analyzed on a FACScan (Becton Dickinson, Mountain View, CA). Two-color staining was performed using the above Abs and/or PE- or Cychrome-conjugated Abs of the same specificity purchased from PharMingen (San Diego, CA).

A total of 1–2 × 107 cells were incubated in 500 μl of 10 mM Tris/1 mM EDTA buffer (pH 7.4) containing 2 mg/ml proteinase K and 1% SDS for 4–16 h at 37o with gentle shaking. DNA was then phenol/chloroform extracted, precipitated with ethanol, washed in 75% ethanol, dissolved in Tris/EDTA buffer at pH 8.0, and treated with RNase A at 100 μg/ml for 30 min at 37o. Analysis of rearrangements at the TCRβ locus was performed essentially as described by van Meerwijk et al. (34) using the primers shown in Table I.

Table I.

Primers for PCR analysis

PrimerSequenceTmaRef.
CD2f cgt atg agg tct tag caa acg 63  
CD2r caa gag cac caa gag gag tcc 63  
CD44f gcc atg gac aag ttt tgg tgg 62  
CD44r atc ctg atc tcc agt agg ctg 62  
FcγRII/IIIf ttc cac cac tga caa ttc tgc 60  
FcγRII/IIIf ggt gcc ata gct gga gga ac 60 35 
FcγRIIr gca gct tct tcc aga tca gg 60 35 
FcγRIIIr gga ggc aca tca cta ggg ag 60 35 
Dβ1 ggt aga cct atg gga ggg tc 64 34 
Jβ1 aaa ccc aga gaa gag caa gc 64 34 
Dβ2 gta ggc acc tgt ggg gaa gaa act 63 80 
Jβ2 tga gag ctg tct cct act atc gat t 63 80 
Vβ12 ggc ttt caa gga tcg att ca 60 36 
Cβ2 ggt agc ctt ttg ttt gtt tgc 60 36 
Vβ12(M13) [M13f]-gaa gat ggt ggg gct ttc aag gat c 55 37 
Cβ2(M13) [M13r] ctt ggg tgg agt cac att tct c 55 37 
PrimerSequenceTmaRef.
CD2f cgt atg agg tct tag caa acg 63  
CD2r caa gag cac caa gag gag tcc 63  
CD44f gcc atg gac aag ttt tgg tgg 62  
CD44r atc ctg atc tcc agt agg ctg 62  
FcγRII/IIIf ttc cac cac tga caa ttc tgc 60  
FcγRII/IIIf ggt gcc ata gct gga gga ac 60 35 
FcγRIIr gca gct tct tcc aga tca gg 60 35 
FcγRIIIr gga ggc aca tca cta ggg ag 60 35 
Dβ1 ggt aga cct atg gga ggg tc 64 34 
Jβ1 aaa ccc aga gaa gag caa gc 64 34 
Dβ2 gta ggc acc tgt ggg gaa gaa act 63 80 
Jβ2 tga gag ctg tct cct act atc gat t 63 80 
Vβ12 ggc ttt caa gga tcg att ca 60 36 
Cβ2 ggt agc ctt ttg ttt gtt tgc 60 36 
Vβ12(M13) [M13f]-gaa gat ggt ggg gct ttc aag gat c 55 37 
Cβ2(M13) [M13r] ctt ggg tgg agt cac att tct c 55 37 
a

Annealing temperature.

RNA was prepared using RNAzol B (Biogenesis, Poole, U.K.) at the rate of 1 ml per 107 cells according to the manufacturer’s instructions. cDNA was prepared from ∼1 μg of RNA using an oligo(dT) primer and a Promega (Madison, WI) Reverse Transcription kit. Aliquots (1 μl) of cDNA solution were added to 49 μl of PCR mixture containing 200 μM dNTPs, 1–2 mM Mg2+, 0.75 U Taq polymerase, and 1 μM forward and reverse primers. The Ly49 and CD3 primers have been described previously (38); most of the other primers used, together with their annealing temperature (Tm), are shown in Table I. The specificity of the Ly49 primers has been confirmed by sequencing (38). The specificity of the other primers is based on 1) the work of others (see references in Table I), 2) the presence of a single product band of the correct molecular size on gels, and 3) PCR product being detected only in cells that expressed the relevant molecule at the cell surface. To establish which Vβ region was present in the TCR in EL4 cells, we used the set of primers described by Casanova et al. (37), kindly provided by Dr. A. Mellor (Clinical Sciences Centre, London, U.K.) and Dr. A.G. Diamond (University of Newcastle, Newcastle, U.K.). For sequencing, additional Vβ12 and Cβ primers bearing M13 tags were constructed (Table I). Generally, 35 cycles were run at 95°C for 1 min, Tm for 1 min, 72°C for 30 s, with a final extension time of 10 min. Aliquots of the reaction were examined on agarose gels containing ethidium bromide.

In a search for cells that express NK-related molecules, we examined a panel of tumor cell lines available in our laboratory. Of the lines initially examined, only two, EL4 and RMA, expressed molecules encoded in the NK complex. Surprisingly, in both cases the expression of NK-related molecules was found to be highly variable. Thus, despite the fact that EL4 has previously been reported to express Ly49A (23, 29, 39), one line of EL4, EL4(S), that had been maintained in our laboratory for many years, showed no detectable expression of Ly49A, whereas another line of EL4, obtained from Dr. K. Karre and used in all of the studies described below, did express Ly49A, but only on a proportion of the cells (Fig. 1). Neither EL4 line showed any detectable surface expression of Ly49C, -I, or -G. By contrast, the RMA line stained strongly and uniformly for Ly49A, was negative for Ly49C and -I, but expressed Ly49G in a heterogeneous manner (Fig. 1). Expression of Ly49 molecules was confirmed and extended by RT-PCR analysis: EL4 was found to express mRNA for Ly49B as well as Ly49A, but lacked mRNA for Ly49C, -D, -E, -G1, -G2, -G3, and -H; RMA expressed mRNA for Ly49A, -B, -G1, -G2, and -G3, but not for Ly49C, -D, -E, or -H (data not shown).

FIGURE 1.

EL4 and RMA cells express certain surface molecules in a mosaic manner.

FIGURE 1.

EL4 and RMA cells express certain surface molecules in a mosaic manner.

Close modal

The EL4 and RMA lines also showed significant but low staining with the PK136 anti-NK1.1 mAb that recognizes NKRP1C in C57 strains (40), and expressed at least one other member of the NKRP1 family as shown by staining with the 10A7 mAb which recognizes NKRP1A and B (V. Kumar, personal communication). However, for both cells the molecule(s) detected by 10A7 were expressed in a heterogeneous manner. RMA and EL4 also stained heterogeneously with the 2.4G2 mAb that recognizes CD16 (FcγRIII) and CD32 (FcγRII). PCR analysis, performed as described by Koyasu (35), revealed that it was CD32 and not CD16 that was expressed by EL4 and RMA cells (data not shown), in agreement with previous reports concerning EL4 (41, 42).

A number of other molecules were also found to be expressed heterogeneously on one or both of these tumor cell lines, including CD2, CD3, CD4, and CD44 (Fig. 1). Most molecules examined, however, were consistently expressed in a uniform manner on both the parent EL4 and RMA lines and on all sublines and clones examined; for example, Thy-1, CD5, CD45, Kb, and Db being expressed on all cells, and 2B4, CD8, CD25, and various CD45R isoforms being absent from all cells. Furthermore, the phenomenon of variable Ag expression appeared to be largely confined to EL4 and RMA, as staining of several other tumor cell lines (including YAC-1, R1.1, SL8, LBRM, L1210, P815, A20, C1498, and J774) with a large panel of Abs showed, with very few exceptions, entirely uniform patterns of expression or nonexpression (data not shown).

The heterogeneity in expression of surface molecules on EL4 and RMA could have a number of trivial explanations, including spontaneous mutation during the many years of growth in vitro and in vivo, contamination with other cell lines, or cell cycle-related expression. To investigate these possibilities, the lines were cloned and the clones examined at an early stage of development (10–20 days after cloning) for expression of variable molecules. In a typical experiment in which EL4 was cloned at 0.05 cell/well, where the probability of clones having arisen from a single progenitor was >98% (as judged from Poisson analysis of the frequency of wells having colonies), 15/30 clones were found to be uniformly positive for CD32, 13/30 were uniformly negative, and 2/30 showed heterogeneous staining. When the same clones were stained for Ly49A, 3 were found to be uniformly positive, 1 was uniformly negative, and 26 were heterogeneous. Examples of the staining patterns are shown in Fig. 2. When heterogeneous clones were recloned, heterogeneous Ag expression was again seen on many of the granddaughter clones.

FIGURE 2.

Typical patterns of expression of Ly49A and CD32 on clones derived from EL4.

FIGURE 2.

Typical patterns of expression of Ly49A and CD32 on clones derived from EL4.

Close modal

Because all of the clones grew at a similar rapid rate, the finding of clones that expressed CD32 or Ly49A in either a uniformly positive or uniformly negative manner ruled out the possibility that heterogeneous expression of these molecules was due to their being expressed only at certain stages of the cell cycle. Likewise, the finding of clones that showed heterogeneous expression of CD32 or Ly49A eliminated the possibility that the heterogeneity of the parent lines was due to classical mutagenic processes or to contamination with other cell lines. Similar results and conclusions regarding each of the molecules that showed variable expression on parental EL4 and RMA lines were reached following further cloning experiments. It therefore appears that the expression of these molecules in EL4 and RMA cells is controlled by a stochastic epigenetic process. As judged from the proportions of young clones that showed heterogeneous expression, three molecules displayed a particularly high rate of switching, namely Ly49A, Ly49G, and 10A7. For each of the variable molecules, switching appeared to be an all-or-none phenomenon as even in the youngest heterogeneous clones the level of expression of the variable molecule on positive cells was similar to that seen in the parent line (see for example Fig. 2).

Over a period of time the proportions of cells expressing particular surface molecules fluctuated considerably, and clones that were uniformly positive or negative for a given Ag often became heterogeneous. Examples of two clones that were followed for >1 year are shown in Fig. 3. The finding that clones that initially appeared to be uniformly negative or positive for a given Ag often became heterogeneous for the expression of that Ag, as illustrated for 10A7 in clone 1 and for CD2, CD3, and Ly49A in clone 3D, raised the possibility that clones that appeared to be uniformly negative or positive for a particular Ag were in fact heterogeneous at a level below the sensitivity of detection (about 1%). That this was indeed the case was confirmed in several cases by showing that Ag-positive cells could be obtained from apparently negative populations by positive selection with magnetic beads (data not shown).

FIGURE 3.

Typical examples of time-dependent changes in the proportion of cells expressing different surface molecules in EL4 clones.

FIGURE 3.

Typical examples of time-dependent changes in the proportion of cells expressing different surface molecules in EL4 clones.

Close modal

That at least some of the Ags displaying variable expression on EL4 and RMA cells can vary independently from each other is shown by the data in Fig. 3 in which Ly49A, 10A7, CD32, and CD44 showed heterogeneous expression on clone 1, whereas CD2 and CD3 were expressed uniformly. By contrast, in clone 3D, CD32 and CD44 were expressed uniformly, whereas CD2 and CD3 were expressed heterogeneously. On the other hand, some of the data in Fig. 3 suggest the possibility of linked variation, such as between Ly49A and CD44 in clone 1 and among Ly49A, CD2, and CD3 in clone 3D. To address this issue more directly, cloned populations of EL4 and RMA cells were examined by two-color immunofluorescence. Fig. 4 shows some examples of the results obtained. The most common finding, illustrated by the expression of CD44 and CD32 on EL4 clone 1A, and of CD4 and 10A7 on RMA clone D3, was that the expression of two variable Ags was completely unlinked, the frequency of double positive and double negative cells being close to that expected for random assortment (Table II). In some cases, illustrated by the expression of CD2 and Ly49A on EL4 clone 3E, a partial but significant (p = 2.7 × 10−4) inverse correlation was found between the expression of two variable molecules. Because this inverse correlation was only partial (a significant proportion of double positive cells being present), and was not consistently observed for the expression of these Ags in other clones, it most likely arose either as a consequence of the dominant growth of subclones of cells within the clonal population or as a reflection of a random switching event that occurred at a formative stage in the development of the clone. A similar partial correlation was found between the expression of Ly49G and 10A7 on RMA clone D3 (Fig. 4); this partial coordinate expression of 10A7 and Ly49G was also seen on each of several randomly chosen subclones of D3 (data not shown). However, by far the most striking case of linked expression, was that between CD2 and CD3: among the dozens of clones examined in this study we found not a single instance in which the proportion of CD3+ cells significantly exceeded the proportion of CD2+ cells. The linkage between CD2 and CD3 expression was most strikingly illustrated by EL4 clone 3D in which nearly all cells were either double positive or double negative, and few if any cells expressed CD3 in the absence of CD2 (Fig. 4, Table II).

FIGURE 4.

Typical results obtained in two-color fluorescence staining of EL4 and RMA clones that showed mosaic expression of more than one surface molecule.

FIGURE 4.

Typical results obtained in two-color fluorescence staining of EL4 and RMA clones that showed mosaic expression of more than one surface molecule.

Close modal
Table II.

Analysis of Ag expression on clones of EL4 and RMA cells

CloneExpt.aSubpopulationsRatiosbBias,c AD/BCp Valued
ABCDB/AD/C
EL4 clone 1A  CD44+CD32 CD44+CD32+ CD44CD32 CD44CD32+     
 35.7 26.4 25.8 11.6 0.7 0.4 1.6  
 33.1 35.7 18.9 12.2 1.1 0.6 1.7  
 47.6 33.3 13.3 5.9 0.7 0.4 1.6  
 57.6 29.9 9.6 2.9 0.5 0.3 1.7 7.1 × 10−2 
          
RMA clone D3  CD4+10A7 CD4+10A7+ CD410A7 CD410A7+     
 24.2 28.8 12.5 34.4 1.2 2.8 0.4  
 11.2 25.4 12.5 50.9 2.3 4.1 0.6  
 35.0 35.0 9.3 20.8 1.0 2.2 0.4  
 35.6 48.0 5.5 10.9 1.3 2.0 0.7  
 65.3 26.9 4.4 3.4 0.4 0.8 0.5 1.1 × 10−1 
          
EL4 clone 3E  CD2+Ly49A CD2+Ly49A+ CD2Ly49A CD2Ly49A+     
 12.2 4.5 73.0 10.3 0.4 0.1 2.6  
 11.4 4.9 71.8 11.9 0.4 0.2 2.6  
 10.5 4.0 74.4 11.1 0.4 0.1 2.6 2.7 × 10−4 
          
EL4 clone 3D  CD2+CD3 CD2+CD3+ CD2CD3 CD2CD3+     
 2.8 21.5 74.8 0.9 7.7 0.013 611  
 2.3 11.4 86.3 0.1 4.9 0.001 8425 ND 
CloneExpt.aSubpopulationsRatiosbBias,c AD/BCp Valued
ABCDB/AD/C
EL4 clone 1A  CD44+CD32 CD44+CD32+ CD44CD32 CD44CD32+     
 35.7 26.4 25.8 11.6 0.7 0.4 1.6  
 33.1 35.7 18.9 12.2 1.1 0.6 1.7  
 47.6 33.3 13.3 5.9 0.7 0.4 1.6  
 57.6 29.9 9.6 2.9 0.5 0.3 1.7 7.1 × 10−2 
          
RMA clone D3  CD4+10A7 CD4+10A7+ CD410A7 CD410A7+     
 24.2 28.8 12.5 34.4 1.2 2.8 0.4  
 11.2 25.4 12.5 50.9 2.3 4.1 0.6  
 35.0 35.0 9.3 20.8 1.0 2.2 0.4  
 35.6 48.0 5.5 10.9 1.3 2.0 0.7  
 65.3 26.9 4.4 3.4 0.4 0.8 0.5 1.1 × 10−1 
          
EL4 clone 3E  CD2+Ly49A CD2+Ly49A+ CD2Ly49A CD2Ly49A+     
 12.2 4.5 73.0 10.3 0.4 0.1 2.6  
 11.4 4.9 71.8 11.9 0.4 0.2 2.6  
 10.5 4.0 74.4 11.1 0.4 0.1 2.6 2.7 × 10−4 
          
EL4 clone 3D  CD2+CD3 CD2+CD3+ CD2CD3 CD2CD3+     
 2.8 21.5 74.8 0.9 7.7 0.013 611  
 2.3 11.4 86.3 0.1 4.9 0.001 8425 ND 
a

Cells were stained on separate occasions over a period of several months.

b

The B/A ratio is the ratio of cells that express/do not express Ag 2 among cells that express Ag 1, and the D/C ratio is the corresponding ratio for cells that do not express Ag 1. If the two Ags were expressed randomly within the cell population, the two ratios would be identical.

c

The ratio of the B/A and D/C ratios provides a simple measure of the bias (nonrandomness) of Ag expression in the population. For randomly expressed antigens the bias value would be 1.0.

d

The probability that the distribution of the two Ags is nonrandom, determined by comparing the B/A and D/C ratios by Student’s t test.

As noted above, surface expression of Ly49 molecules on parental EL4 and RMA lines correlated with the presence of the corresponding mRNA. As shown in Fig. 5, this was also true for other molecules expressed in a variable manner. Interestingly, clones that lacked expression of CD3 contained mRNA for CD3γ, CD3δ, and CD3ε, but were deficient in expression of mRNA for CD3ζ.

FIGURE 5.

RT-PCR analysis of the expression of specific RNA sequences in the B cell lymphoma A20 and in four EL4 clones. The clones whose PCR analysis is shown had the following phenotypes: clone A Ly49A80, CD2, CD3, CD32, CD44+; clone B Ly49A1, CD23, CD3, CD32, CD44+; clone C Ly49A+, CD2+, CD3+, CD32+, CD44; clone D Ly49A3, CD2+, CD3+, CD32+, CD44, where superscript numbers are the percentages of positive cells; +, no detectable negative cells present; and −, no detectable positive cells present.

FIGURE 5.

RT-PCR analysis of the expression of specific RNA sequences in the B cell lymphoma A20 and in four EL4 clones. The clones whose PCR analysis is shown had the following phenotypes: clone A Ly49A80, CD2, CD3, CD32, CD44+; clone B Ly49A1, CD23, CD3, CD32, CD44+; clone C Ly49A+, CD2+, CD3+, CD32+, CD44; clone D Ly49A3, CD2+, CD3+, CD32+, CD44, where superscript numbers are the percentages of positive cells; +, no detectable negative cells present; and −, no detectable positive cells present.

Close modal

In many systems gene expression has been found to be linked to the demethylation of critical cytidine residues (43, 44, 45) and to the acetylation of nucleosomal histones in gene regulatory regions (46, 47). In at least some cases, these two events seem to be linked (48). To determine whether these processes might be involved in the mosaic expression of surface molecules on EL4 cells, we examined whether transient exposure of cloned EL4 cells to 5-azacytidine, an inhibitor of DNA methylation, and/or trichostatin A, an inhibitor of histone deacetylase, would affect the expression of variable Ags. Cells were pretreated for 24 h with drugs, then washed and cultured for a further 4–7 days in drug-free medium before analysis. Some clones showed no change in expression of any of the Ags examined. However, several clones showed a dramatic increase in the proportion of cells expressing 10A7 or CD32 when treated for 24 h with 10–50 μM 5-azacytidine. Trichostatin A had no effect on Ag expression when used alone. However, when combined with low doses (1–10 μM) of 5-azacytidine, synergistic effects were often observed. A typical result involving the induction of 10A7 is shown in Fig. 6.

FIGURE 6.

The effect of pretreatment with 5-azacytidine and trichostatin A on the expression of 10A7 on an EL4 clone. Cells were treated for 24 h with medium, 5-azacytidine (5-AC), trichostatin A (TSA), or a combination of both drugs, then cultured for 7 days in the absence of drugs.

FIGURE 6.

The effect of pretreatment with 5-azacytidine and trichostatin A on the expression of 10A7 on an EL4 clone. Cells were treated for 24 h with medium, 5-azacytidine (5-AC), trichostatin A (TSA), or a combination of both drugs, then cultured for 7 days in the absence of drugs.

Close modal

The finding that EL4 and RMA shared not only an unusual NK1.1+, 10A7+, Ly49+, CD3+, CD32+ phenotype, but also the unusual property of mosaic expression of these and other surface molecules, led us to consider the possibility that EL4 and RMA might be directly related. Using a PCR-based method we found that both tumor lines had similar rearrangements at their TCRβ gene loci. Thus, at the Jβ1 locus they both possessed a partial Dβ1-Jβ1.5 rearrangement and lacked a germline Dβ1-Jβ1 band (Fig. 7,A), whereas at the Jβ2 locus they both lacked partial Dβ1-Jβ2 (data not shown) and Dβ2-Jβ2 rearrangements (Fig. 7,B) and displayed a germline Dβ2-Jβ2 band (Fig. 7 B). Sequencing of the Dβ1-Jβ1.5 partial rearrangement showed it to be identical in both EL4 and RMA, having the junctional sequence Dβ1(gggacaggggg)-N(gca)-Jβ1.5(taacaacca… . ). These results agree with an earlier study that demonstrated that EL4 possessed a partial Dβ1-Jβ1 rearrangement and expressed a full-length Cβ2-containing mRNA transcript (49).

FIGURE 7.

Analysis of TCR gene rearrangement in EL4 and RMA cells. DNA prepared from EL4, RMA, A20 (negative control), and fresh thymus (positive control) cells was analyzed for the presence of rearrangements at the Jβ1 (A) and the Jβ2 (B) loci as described in Materials and Methods. GL, germline band. C, RT-PCR analysis of the presence of Vβ12-Cβ sequences in RNA prepared from EL4, RMA, A20 (negative control), and fresh spleen (positive control) cells.

FIGURE 7.

Analysis of TCR gene rearrangement in EL4 and RMA cells. DNA prepared from EL4, RMA, A20 (negative control), and fresh thymus (positive control) cells was analyzed for the presence of rearrangements at the Jβ1 (A) and the Jβ2 (B) loci as described in Materials and Methods. GL, germline band. C, RT-PCR analysis of the presence of Vβ12-Cβ sequences in RNA prepared from EL4, RMA, A20 (negative control), and fresh spleen (positive control) cells.

Close modal

RT-PCR analysis revealed that both EL4 and RMA expressed Vβ12-containing transcripts (Fig. 7 C) but not transcripts for eight other Vβ regions that were examined using the primer set described by Casanova et al. (37). Sequencing of the Vβ12 PCR products from both lines revealed identical junctional sequences, i.e., Vβ12 (… cagcagt)-D/N(accggga)-Jβ2.3(cagaaacgc… ). This sequence matched a single entry in the databases, that reported by Shi et al.(Ref. (50); GenBank accession no. AF020206) for the TIB39 (American Type Culture Collection) line of EL4. Fluorescence staining confirmed the expected expression of Vβ12 at the surface of EL4 and RMA cells (data not shown).

To exclude the possibility that the RMA line had been mixed up with EL4 in our laboratory, we examined several RMA-related lines that were distinguishable from EL4: 1) the RMA/S line derived by Karre et al. (51), which we obtained directly from the originating laboratory, and which we demonstrated to have the expected characteristics of RMA/S (lack of expression of Kb and Db under normal culture conditions, but expression of these at 26oC, and inducibility by Kb and Db specific peptides); 2) mtp2, a TAP-transfected revertant of RMA/S, generated by Powis et al. (52), which we obtained from Drs. R Glas and K. Karre, and which we confirmed to have the expected characterisics (partially but not completely restored expression of Kb and Db, resistance to G418 (in addition, this line is morphologically distinct from all other EL4/RMA-related lines in our laboratory); 3) an RMA line transfected with Dd (53) obtained directly from Dr. D. Raulet and which we confirmed to have the unique characteristics of coexpression of Kb, Db, and Dd, and resistance to G418; and 4) an RMA/S line transfected with CD1 (54) obtained directly from Dr. M. Kronenberg and confirmed to have the unique phenotype of lack of expression of Kb and Db under normal culture condtions, expression of the C57 isoform of β2-microglobulin, extremely high expression of CD1 (at least 10-fold higher than on any other cell line in our laboratory), and resistance to G418. RT-PCR performed on RNA extracted from these lines, but not on RNA extracted in parallel from unrelated tumor cell lines, revealed a Vβ12 product. In each case its sequence was identical to the EL4 Vβ12 sequence. In addition, we found that a line of RBL5 that we obtained, RBL5 being the parent line from which both RMA and RMA/S were derived (51), also expressed a Vβ12 transcript with the same junctional sequence as that found in EL4. Finally, following communication of our findings to another group, they were able to establish that their lines of EL4 and RMA expressed full-length rearranged Vα10-containing transcripts of identical sequence, and that their line of RBL5 possessed both Vβ12 and Vα10 transcripts (T. van Hall, M. Kraakman, C. Melief, F. Ossendorp, and R. Offringa, manuscript in preparation).

The studies described above have revealed that EL4 and RMA tumor cell lines 1) express several members of the Ly49 and NKRP1 families of C-type lectin molecules, 2) frequently express these and certain other cell surface molecules in a mosaic manner, and 3) have a common origin. In the discussion that follows we shall consider EL4 and RMA collectively as EL4 lines.

That EL4 expresses the prototypic member of the Ly49 family, Ly49A, has been known for some time (29, 39). What had not been generally appreciated is that EL4 expresses at least two other members of Ly49 family, Ly49G (confirming another recent report (23)), and Ly49B (at least at the RNA level), and that the expression of Ly49A and Ly49G varies considerably both between and within sublines. The latter observation, coupled with the finding of freshly derived clones showing mosaic expression of these molecules, demonstrates that the genes encoding Ly49A and -G molecules exist in a heritable metastable state, switching on and off in a frequent and apparently random manner. By contrast, we did not detect expression of Ly49C, -D, -E, -F, -H, or -I in EL4 cells. However, EL4 did express certain other NK-related molecules. In particular, expression of NK1.1 (NKRP1C) was found, albeit at a low level, together with expression of the NKRP1 molecules recognized by the mAb 10A7, which are known to include NKRP1A and NKRP1B (V. Kumar, personal communication). The 10A7-defined molecule(s) were almost always expressed in a mosaic manner, even on newly derived clones, and by this criterion were the most variable of all of the molecules we investigated.

Several other molecules, namely CD2, CD3, CD4, CD32, and CD44, were also found to vary on EL4 cells, although, as judged from the frequency of clones that displayed mosaic expression, the rate of switching was much lower than for Ly49A, Ly49G, and 10A7. In every case, expression was controlled at the mRNA level, and apparently independently for each molecule. The striking exception was the strong and consistent linkage of CD3 expression to that of CD2. Because expression of the entire CD3-TCR complex appeared to be controlled by expression of CD3ζ, this finding suggests that, at least in EL4 cells, transcription of the CD3ζ gene rarely occurs in the absence of transcription of the CD2 gene. Linkage between these gene products has previously been reported at the functional level, signaling through CD2 being dependent on the expression of CD3ζ (55, 56). Lack of expression of CD3ζ would also be expected to prevent signaling via GPI-linked surface molecules such as Thy-1 and Ly6 (57) and the recently described NK activating molecule p46 (58). In addition, silencing of the CD3ζ genetic region may block expression of the transcription factor Oct1 that is encoded on the opposite DNA strand to CD3ζ (59), leading to an absence of all those proteins whose expression is dependent on this factor.

At present we have little information on the nature of the regulatory events responsible for the mosaic expression of certain molecules in EL4 cells. One possibility would be that it is caused by the limiting availability of key transcription factors. However, in view of the independent variation of expression of several different molecules, this would require a large number of relatively gene-specific transcription factors to be close to threshold levels. An alternative possibility would be that the binding of readily available transcription factors was affected by changes in local chromatin structure or histone modification. In support of this, our studies showed that the expression of some of the variable Ags could be altered by drugs that affect the methylation levels of DNA or the acetylation levels of histones, in line with a previous report that 5-azacytidine treatment of a Thy-1 variant of EL4 restored expression of Thy-1, and that this was associated with demethylation of the Thy-1 gene promoter (60). Indeed, the mosaic expression of surface molecules on EL4 cells bears a striking resemblance to position effect variegation in which genes located close to the borders of euchromatin and heterochromatin are switched on and off by saltatory shifts in the positions of these borders, perhaps triggered by changes in methylation (61). Increased methylation of genes in cultured cells could lead to the development of a functionally hemizygous state (62) or even the complete loss of expression of nonessential proteins (63, 64). However, we suspect that at least some of the phenotypic variability we have observed in EL4 lines is of a nontrivial nature in that 1) in contrast to the occasional and largely irreversible loss of gene expression that occurs through epigenetic mutation in long-lived cell lines, the changes in gene expression in EL4 are much more frequent and occur in both directions, 2) many, perhaps all, of the molecules that show variable expression on EL4 cells are expressed in a heterogeneous manner on NK and NK-T (see below) cell populations in vivo, and 3) freshly established lines of fetal mouse NK cells derived from single progenitor cells show mosaic expression of several of these same molecules, most importantly those encoded in the NK complex, including the NKRP1 molecules recognized by the mAb 10A7 (8), Ly49 molecules (8), and Qa1 receptors (9). It is interesting to note that in EL4 lines, there is clear evidence of both coordinated and independent regulation of molecules encoded within the NK complex. Thus, whereas EL4 cells can express Ly49 molecules, they frequently express Ly49A in the absence of Ly49G, and never express Ly49C or -I. Similarly, although the NKRP1-encoded molecules recognized by 10A7 tend to be expressed on cells that also express Ly49G, the linkage is only partial, and the expression of these NKRP1 molecules is unlinked to those recognized by the PK136 anti-NK1.1 mAb. Analysis of the factors controlling the expression of these molecules in EL4 cells may therefore provide insights into the mechanism responsible for the stochastic expression of these molecules in vivo.

Although EL4 lacks the canonical Vα14-Jα281 chain (M. Unnikrishnan, unpublished observations) that is expressed by a large proportion of NK-T cells (65), and expresses a Vβ12-containing rather than a Vβ8-, Vβ7-, or Vβ2-containing receptor (66), the findings reported here suggest that EL4 may be a transformed NK-T cell line. First, not only do NK-T cells and EL4 cells express Ly49 and NKRP1 molecules, but they both do so in an apparently stochastic manner. The heterogeneous expression of Ly49 molecules on NK-T cells in vivo has been well documented (23, 67). With regard to NKRP1 molecules, Chen et al. (68) have reported that NK1.1 was frequently lost upon activation of splenic NK-T cells in vitro. Furthermore, two recent studies have provided compelling evidence for the existence of populations of NK-T cells in vivo that lack NK1.1 (69, 70), and when human NK-T cells were cloned by Davodeau et al. (71), four of five clones showed heterogeneous expression of NKRP1A. Second, like NK-T cells in vivo, EL4 exists in alternate forms that are either CD4+CD8 or CD4CD8. Although the relationship between the CD4+ and CD4 subsets of NK-T cells in vivo is unclear, Chen et al. (68) found that the expression of CD4 on CD4+ splenic NK-T cells is unstable, being lost from some but not all cells following activation in vitro. Third, like NK-T cells (72, 73), EL4 secretes IL4 upon stimulation (74).

The most unexpected finding in our study was that the EL4 and RMA lines that we investigated, which included the RMA and RMA/S lines from the originating laboratory, shared 1) the same Vβ gene rearrangements on both chromosomes, 2) expressed Vβ12 containing receptors on the surface, and 3) contained in-frame Vβ12 RNA transcripts of identical sequence. This sequence was in turn identical to that previously reported for TCRβ transcripts in the TIB39 subline of EL4 obtained from the American Type Culture Collection (50). The finding that RMA and RMA/S are almost certainly variants of EL4 does not affect the validity of any of the important findings relating to Ag processing and NK recognition that have been made using these lines. Indeed, it may lead to important insights and rationalization of existing data. For example, the striking resistance of EL4 and RMA to lysis by activated NK cells, and their apparent sharing of a novel protective class I molecule (9), is now revealed not to be a coincidence. However, our finding of extensive epigenetic instability in the expression of various molecules in EL4/RMA lines provides a strong and salutory warning against the assumption that variant lines of these cells, e.g., RMA and RMA/S, are identical in all respects other than those controlled by the mutated or transfected genes. At the same time, our results have provided rigorous confirmation that RMA, RMA/S, and various transfectant lines derived from them are indeed all variants of the same line, albeit EL4. In addition, we have established that the EL4 Vβ12 sequence is expressed in three widely used variants of EL4 namely the “Salk” variant, EL4(S), which displays different retroviral Ags to those found on other EL4 lines (75), a β2-microglobulin mutant of EL4, C4.4 (76), and its β2-microglobulin transfected revertant, E50 (76), confirming that these are indeed of EL4 origin.

Our finding that a line of the RBL5 tumor, from which both RMA and RMA/S were derived (51), also expresses the EL4 Vβ transcript, strongly suggests that RBL5 had inadvertently been contaminated with EL4 cells before its distribution to the Karolinska Institute. Remarkably, during the preparation of this paper, we found that two other tumor cell lines in our laboratory express the canonical EL4 Vβ12 transcript. One is a line of the E♀K1 tumor (77), and the other a line of MBL2, a tumor that has previously been reported to express Ly49 molecules (39), and which was derived at the same time and in the same laboratory as RBL5 (78). (A third line, E♂ G2 (81), was found to express a Vβ12-containing receptor but this was associated with a Vβ12-Jβ2.1 transcript that had a distinct junctional sequence from that found in EL4 lines.) Although these results require confirmation in other laboratories, it is clear that contamination of cell lines with EL4 has been a widespread but previously undocumented occurrence, with profound implications for the validity of a wide range of studies, especially those on tumor specific Ags. We suggest that all C57 T cell tumor lines, especially those reported to express Ly49 molecules such as TIMI4 (79) and C6VL-B (29), be suspected of being EL4 until proven otherwise.

We are grateful to Prof. Vinay Kumar for providing the 10A7 mAb, and to Drs. Rickard Glas, William Green, Klas Karre, Mitch Kronenberg, and David Raulet for providing us with cell lines. We are grateful to Drs. William Green, Klas Karre, and Hans-Gustaf Ljunggren for helpful discussions, and to Ferry Ossendorp and colleagues in Leiden for informing us of their unpublished results.

1

This work was supported by grants from the University of Newcastle, the Medical Research Council, the Biotechnology and Biological Sciences Research Council, and the Cancer Research Campaign (U.K.). K.P.F. is the recipient of a William Ross Scholarship from the Cancer Research Campaign (U.K.). Z.M.A.C.-L. held a Sir Henry Wellcome Commemorative Award for Innovative Research at the time this work was performed.

1
Ljunggren, H. G., K. Karre.
1990
. In search of the “missing self”: MHC molecules and NK cell recognition.
Immunol. Today
11
:
237
2
Lopez-Botet, M., M. Carretero, T. Bellon, J. J. Perez-Villar, M. Llano, F. Navarro.
1998
. The CD94/NKG2 C-type lectin receptor complex.
Curr. Top. Microbiol. Immunol.
230
:
41
3
Vance, R. E., J. R. Kraft, J. D. Altman, P. E. Jensen, D. H. Raulet.
1998
. Mouse CD94/NKG2A is a natural killer cell receptor for the nonclassical major histocompatibility complex (MHC) class I molecule Qa-1(b).
J. Exp. Med.
188
:
1841
4
Takei, F., J. Brennan, D. L. Mager.
1997
. The Ly-49 family: genes, proteins and recognition of class I MHC.
Immunol. Rev.
155
:
67
5
Brown, M. G., A. A. Scalzo, K. Matsumoto, W. M. Yokoyama.
1997
. The natural killer gene complex: a genetic basis for understanding natural killer cell function and innate immunity.
Immunol. Rev.
155
:
53
6
Raulet, D. H., W. Held, I. Correa, J. R. Dorfman, M. F. Wu, L. Corral.
1997
. Specificity, tolerance and developmental regulation of natural killer cells defined by expression of class I-specific Ly49 receptors.
Immunol. Rev.
155
:
41
7
Vance, R. E., D. H. Raulet.
1998
. Toward a quantitative analysis of the repertoire of class I MHC-specific inhibitory receptors on natural killer cells.
Curr. Top. Microbiol. Immunol.
230
:
135
8
Manoussaka, M. S., R. J. Smith, V. Conlin, J. A. Toomey, C. G. Brooks.
1998
. Fetal mouse NK cell clones are deficient in Ly49 expression, share a common broad lytic specificity, and undergo continuous and extensive diversification in vitro.
J. Immunol.
160
:
2197
9
Toomey, J. A., M. Salcedo, L. A. Cotterill, M. M. Millrain, Z. Chrzanowska-Lightowlers, J. Lawry, K. Fraser, F. Gays, J. H. Robinson, S. Shrestha P., et al
1999
. Stochastic acquisition of Qa1 receptors during the development of fetal NK cells in vitro accounts in part but not in whole for the ability of these cells to distinguish between class I-sufficient and class I-deficient targets.
J. Immunol.
163
:
3176
10
Chambers, W. H., N. L. Vujanovic, A. B. DeLeo, M. W. Olszowy, R. B. Herberman, J. C. Hiserodt.
1989
. Monoclonal antibody to a triggering structure expressed on rat natural killer cells and adherent lymphokine-activated killer cells.
J. Exp. Med.
169
:
1373
11
Karlhofer, F. M., W. M. Yokoyama.
1991
. Stimulation of murine natural killer (NK) cells by a monoclonal antibody specific for the NK1.1 antigen: IL-2-activated NK cells possess additional specific stimulation pathways.
J. Immunol.
146
:
3662
12
Perussia, B..
1998
. Fc receptors on natural killer cells.
Curr. Top. Microbiol. Immunol.
230
:
63
13
Metes, D., L. K. Ernst, W. H. Chambers, A. Sulica, R. B. Herberman, P. A. Morel.
1998
. Expression of functional CD32 molecules on human NK cells is determined by an allelic polymorphism of the FcγRIIC gene.
Blood
91
:
2369
14
Mandelboim, O., P. Malik, D. M. Davis, C. H. Jo, J. E. Boyson, J. L. Strominger.
1999
. Human CD16 as a lysis receptor mediating direct natural killer cell cytotoxicity.
Proc. Natl. Acad. Sci. USA
96
:
5640
15
Siliciano, R. F., J. C. Pratt, R. E. Schmidt, J. Ritz, E. L. Reinherz.
1985
. Activation of cytolytic T lymphocyte and natural killer cell function through the T11 sheep erythrocyte binding protein.
Nature
317
:
428
16
Nandi, D., J. A. Gross, J. P. Allison.
1994
. CD28-mediated costimulation is necessary for optimal proliferation of murine NK cells.
J. Immunol.
152
:
3361
17
Carbone, E., G. Ruggiero, G. Terrazzano, C. Palomba, C. Manzo, S. Fontana, H. Spits, K. Karre, S. Zappacosta.
1997
. A new mechanism of NK cell cytotoxicity activation: the CD40-CD40 ligand interaction.
J. Exp. Med.
185
:
2053
18
Helander, T. S., O. Carpen, O. Turunen, P. E. Kovanen, A. Vaheri, T. Timonen.
1996
. ICAM-2 redistributed by ezrin as a target for killer cells.
Nature
382
:
265
19
Bendelac, A., M. N. Rivera, S. H. Park, J. H. Roark.
1997
. Mouse CD1-specific NK1 T cells: development, specificity, and function.
Annu. Rev. Immunol.
15
:
535
20
Battistini, L., G. Borsellino, G. Sawicki, F. Poccia, M. Salvetti, G. Ristori, C. F. Brosnan.
1997
. Phenotypic and cytokine analysis of human peripheral blood γδ T cells expressing NK cell receptors.
J. Immunol.
159
:
3723
21
Carena, I., A. Shamshiev, A. Donda, M. Colonna, G. D. Libero.
1997
. Major histocompatibility complex class I molecules modulate activation threshold and early signaling of T cell antigen receptor-γ/δ stimulated by nonpeptidic ligands.
J. Exp. Med.
186
:
1769
22
Brennan, J., D. Mager, W. Jefferies, F. Takei.
1994
. Expression of different members of the Ly-49 gene family defines distinct natural killer cell subsets and cell adhesion properties.
J. Exp. Med.
180
:
2287
23
Ortaldo, J. R., R. Winkler-Pickett, A. T. Mason, L. H. Mason.
1998
. The Ly-49 family: regulation of cytotoxicity and cytokine production in murine CD3+ cells.
J. Immunol.
160
:
1158
24
Wayner, E. A., C. G. Brooks.
1984
. Induction of NKCF-like activity in mixed lymphocyte-tumor cell culture: direct involvement of mycoplasma infection of tumor cells.
J. Immunol.
132
:
2135
25
Tomonari, K..
1988
. A rat antibody against a structure functionally related to the mouse T-cell receptor/T3 complex.
Immunogenetics
28
:
455
26
Cobbold, S. P., A. Jayasuriya, A. Nash, T. D. Prospero, H. Waldmann.
1984
. Therapy with monoclonal antibodies by elimination of T-cell subsets in vivo.
Nature
312
:
548
27
Unkeless, J. C..
1979
. Characterization of a monoclonal antibody directed against mouse macrophage and lymphocyte Fc receptors.
J. Exp. Med.
150
:
580
28
Trowbridge, I. S., J. Lesley, R. Schulte, R. Hyman, J. Trotter.
1982
. Biochemical characterization and cellular distribution of a polymorphic, murine cell-surface glycoprotein expressed on lymphoid tissues.
Immunogenetics
15
:
299
29
Nagasawa, R., J. Gross, O. Kanagawa, K. Townsend, L. L. Lanier, J. Chiller, J. P. Allison.
1987
. Identification of a novel T cell surface disulfide-bonded dimer distinct from the α/β antigen receptor.
J. Immunol.
138
:
815
30
Sentman, C. L., J. Hackett, Jr, V. Kumar, M. Bennett.
1989
. Identification of a subset of murine natural killer cells that mediates rejection of Hh-1d but not Hh-1b bone marrow grafts.
J. Exp. Med.
170
:
191
31
Mason, L., S. L. Giardina, T. Hecht, J. Ortaldo, B. J. Mathieson.
1988
. LGL-1: a non-polymorphic antigen expressed on a major population of mouse natural killer cells.
J. Immunol.
140
:
4403
32
Koo, G. C., J. R. Peppard.
1984
. Establishment of monoclonal anti-NK-1.1 antibody.
Hybridoma
3
:
301
33
Sentman, C. L., J. Hackett, Jr, T. A. Moore, M. M. Tutt, M. Bennett, V. Kumar.
1989
. Pan natural killer cell monoclonal antibodies and their relationship to the NK1.1 antigen.
Hybridoma
8
:
605
34
van Meerwijk, J. P., H. Bluthmann, M. Steinmetz.
1990
. T-cell specific rearrangement of T-cell receptor β transgenes in mice.
EMBO J.
9
:
1057
35
Koyasu, S..
1994
. CD3+CD16+NK1.1+B220+ large granular lymphocytes arise from both α-β TCR+CD4CD8 and γ-δ TCR+CD4CD8 cells.
J. Exp. Med.
179
:
1957
36
Masuda, K., Y. Makino, J. Cui, T. Ito, T. Tokuhisa, Y. Takahama, H. Koseki, K. Tsuchida, T. Koike, H. Moriya, et al
1997
. Phenotypes and invariant αβ TCR expression of peripheral Vα14+ NK T cells.
J. Immunol.
158
:
2076
37
Casanova, J. L., P. Romero, C. Widmann, P. Kourilsky, J. L. Maryanski.
1991
. T cell receptor genes in a series of class I major histocompatibility complex-restricted cytotoxic T lymphocyte clones specific for a Plasmodiumberghei nonapeptide: implications for T cell allelic exclusion and antigen-specific repertoire.
J. Exp. Med.
174
:
1371
38
Toomey, J. A., S. Shrestha, S. A. de la Rue, F. Gays, J. H. Robinson, Z. M. Chrzanowska-Lightowlers, C. G. Brooks.
1998
. MHC class I expression protects target cells from lysis by Ly49-deficient fetal NK cells.
Eur. J. Immunol.
28
:
47
39
Chan, P. Y., F. Takei.
1986
. Expression of a T cell receptor-like molecule on normal and malignant murine T cells detected by rat monoclonal antibodies to nonclonotypic determinants.
J. Immunol.
136
:
1346
40
Ryan, J. C., J. Turck, E. C. Niemi, W. M. Yokoyama, W. E. Seaman.
1992
. Molecular cloning of the NK1.1 antigen, a member of the NKR-P1 family of natural killer cell activation molecules.
J. Immunol.
149
:
1631
41
Ravetch, J. V., A. D. Luster, R. Weinshank, J. Kochan, A. Pavlovec, D. A. Portnoy, J. Hulmes, Y. C. Pan, J. C. Unkeless.
1986
. Structural heterogeneity and functional domains of murine immunoglobulin G Fc receptors.
Science
234
:
718
42
Perussia, B., M. M. Tutt, W. Q. Qiu, W. A. Kuziel, P. W. Tucker, G. Trinchieri, M. Bennett, J. V. Ravetch, V. Kumar.
1989
. Murine natural killer cells express functional Fcγ receptor II encoded by the FcγRα gene.
J. Exp. Med.
170
:
73
43
Bird, A. P..
1986
. CpG-rich islands and the function of DNA methylation.
Nature
321
:
209
44
Jaenisch, R..
1997
. DNA methylation and imprinting: why bother?.
Trends Genet.
13
:
323
45
Jones, P. A..
1999
. The DNA methylation paradox.
Trends Genet.
15
:
34
46
Kuo, M. H., C. D. Allis.
1998
. Roles of histone acetyltransferases and deacetylases in gene regulation.
BioEssays
20
:
615
47
Kouzarides, T..
1999
. Histone acetylases and deacetylases in cell proliferation.
Curr. Opin. Genet. Dev.
9
:
40
48
Nan, X., H. H. Ng, C. A. Johnson, C. D. Laherty, B. M. Turner, R. N. Eisenman, A. Bird.
1998
. Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex.
Nature
393
:
386
49
Kronenberg, M., J. Goverman, R. Haars, M. Malissen, E. Kraig, L. Phillips, T. Delovitch, N. Suciu-Foca, L. Hood.
1985
. Rearrangement and transcription of the β-chain genes of the T-cell antigen receptor in different types of murine lymphocytes.
Nature
313
:
647
50
Shi, Y., A. Kaliyaperumal, L. Lu, S. Southwood, A. Sette, M. A. Michaels, S. K. Datta.
1998
. Promiscuous presentation and recognition of nucleosomal autoepitopes in lupus: role of autoimmune T cell receptor α chain.
J. Exp. Med.
187
:
367
51
Karre, K., H. G. Ljunggren, G. Piontek, R. Kiessling.
1986
. Selective rejection of H-2-deficient lymphoma variants suggests alternative immune defence strategy.
Nature
319
:
675
52
Powis, S. J., A. R. Townsend, E. V. Deverson, J. Bastin, G. W. Butcher, J. C. Howard.
1991
. Restoration of antigen presentation to the mutant cell line RMA-S by an MHC-linked transporter.
Nature
354
:
528
53
Correa, I., L. Corral, D. H. Raulet.
1994
. Multiple natural killer cell-activating signals are inhibited by major histocompatibility complex class I expression in target cells.
Eur. J. Immunol.
24
:
1323
54
Brossay, L., D. Jullien, S. Cardell, B. C. Sydora, N. Burdin, R. L. Modlin, M. Kronenberg.
1997
. Mouse CD1 is mainly expressed on hemopoietic-derived cells.
J. Immunol.
159
:
1216
55
Moingeon, P., J. L. Lucich, D. J. McConkey, F. Letourneur, B. Malissen, J. Kochan, H. C. Chang, H. R. Rodewald, E. L. Reinherz.
1992
. CD3ζ dependence of the CD2 pathway of activation in T lymphocytes and natural killer cells.
Proc. Natl. Acad. Sci. USA
89
:
1492
56
Howard, F. D., P. Moingeon, U. Moebius, D. J. McConkey, B. Yandava, T. E. Gennert, E. L. Reinherz.
1992
. The CD3ζ cytoplasmic domain mediates CD2-induced T cell activation.
J. Exp. Med.
176
:
139
57
Wegener, A. M., F. Letourneur, A. Hoeveler, T. Brocker, F. Luton, B. Malissen.
1992
. The T cell receptor/CD3 complex is composed of at least two autonomous transduction modules.
Cell
68
:
83
58
Vitale, M., C. Bottino, S. Sivori, L. Sanseverino, R. Castriconi, E. Marcenaro, R. Augugliaro, L. Moretta, A. Moretta.
1998
. NKp44, a novel triggering surface molecule specifically expressed by activated natural killer cells, is involved in non-major histocompatibility complex-restricted tumor cell lysis.
J. Exp. Med.
187
:
2065
59
Lerner, A., L. D’Adamio, A. C. Diener, L. K. Clayton, E. L. Reinherz.
1993
. CD3ζ/η/θ locus is colinear with and transcribed antisense to the gene encoding the transcription factor Oct-1.
J. Immunol.
151
:
3152
60
Sneller, M. C., K. C. Gunter.
1987
. DNA methylation alters chromatin structure and regulates Thy-1 expression in EL-4 T cells.
J. Immunol.
138
:
3505
61
Kioussis, D., R. Festenstein.
1997
. Locus control regions: overcoming heterochromatin-induced gene inactivation in mammals.
Curr. Opin. Genet. Dev.
7
:
614
62
Siminovitch, L..
1976
. On the nature of hereditable variation in cultured somatic cells.
Cell
7
:
1
63
Holliday, R..
1987
. The inheritance of epigenetic defects.
Science
238
:
163
64
Antequera, F., J. Boyes, A. Bird.
1990
. High levels of de novo methylation and altered chromatin structure at CpG islands in cell lines.
Cell
62
:
503
65
Makino, Y., R. Kanno, T. Ito, K. Higashino, M. Taniguchi.
1995
. Predominant expression of invariant Vα14+ TCR α chain in NK1.1+ T cell populations.
Int. Immunol.
7
:
1157
66
Arase, H., N. Arase, K. Ogasawara, R. A. Good, K. Onoe.
1992
. An NK1.1+CD4+8 single-positive thymocyte subpopulation that expresses a highly skewed T-cell antigen receptor Vβ family.
Proc. Natl. Acad. Sci. USA
89
:
6506
67
MacDonald, H. R., R. K. Lees, W. Held.
1998
. Developmentally regulated extinction of Ly-49 receptor expression permits maturation and selection of NK1.1+ T cells.
J. Exp. Med.
187
:
2109
68
Chen, H., H. Huang, W. E. Paul.
1997
. NK1.1+CD4+ T cells lose NK1.1 expression upon in vitro activation.
J. Immunol.
158
:
5112
69
Chen, H., W. E. Paul.
1998
. A population of CD62Llow NK1.1CD4+ T cells that resembles NK1.1+CD4+ T cells.
Eur. J. Immunol.
28
:
3172
70
Chiu, Y. H., J. Jayawardena, A. Weiss, D. Lee, S. H. Park, A. Dautry-Varsat, A. Bendelac.
1999
. Distinct subsets of CD1d-restricted T cells recognize self-antigens loaded in different cellular compartments.
J. Exp. Med.
189
:
103
71
Davodeau, F., M. A. Peyrat, A. Necker, R. Dominici, F. Blanchard, C. Leget, J. Gaschet, P. Costa, Y. Jacques, A. Godard, et al
1997
. Close phenotypic and functional similarities between human and murine αβ T cells expressing invariant TCR α-chains.
J. Immunol.
158
:
5603
72
Arase, H., N. Arase, K. Nakagawa, R. A. Good, K. Onoe.
1993
. NK1.1+CD4+CD8 thymocytes with specific lymphokine secretion.
Eur. J. Immunol.
23
:
307
73
Yoshimoto, T., A. Bendelac, J. Hu-Li, W. E. Paul.
1995
. Defective IgE production by SJL mice is linked to the absence of CD4+, NK1.1+ T cells that promptly produce interleukin 4.
Proc. Natl. Acad. Sci. USA
92
:
11931
74
Dornand, J., M. Bouaboula, A. D. d’Angeac, J. Favero, D. Shire, P. Casellas.
1992
. Contrasting effects of the protein kinase C inhibitor staurosporine on the interleukin-1 and phorbol ester activation pathways in the EL4-6.1 thymoma cell line.
J. Cell Physiol.
151
:
71
75
Green, W. R..
1980
. H-2-restricted cytolytic T lymphocytes specific for a subclass of AKR endogenous leukemia virus-induced tumors: correlation of tumor cell susceptibility with expression of the gross cell surface antigen.
J. Immunol.
125
:
2584
76
Glas, R., K. Sturmhofel, G. J. Hammerling, K. Karre, H. G. Ljunggren.
1992
. Restoration of a tumorigenic phenotype by β2-microglobulin transfection to EL-4 mutant cells.
J. Exp. Med.
175
:
843
77
Gomard, E., J. P. Levy, F. Plata, Y. Henin, V. Duprez, A. Bismuth, T. Reme.
1978
. Studies on the nature of the cell surface antigen reacting with cytolytic T lymphocytes in murine oncornavirus-induced tumors.
Eur. J. Immunol.
8
:
228
78
Glynn, J. P., J. L. McCoy, A. Fefer.
1968
. Cross-resistance to the transplantation of syngeneic Friend, Moloney, and Rauscher virus-induced tumors.
Cancer Res.
28
:
434
79
Chan, P. Y., F. Takei.
1989
. Molecular cloning and characterization of a novel murine T cell surface antigen, YE1/48.
J. Immunol.
142
:
1727
80
Anderson, S. J., K. M. Abraham, T. Nakayama, A. Singer, R. M. Perlmutter.
1992
. Inhibition of T-cell receptor β-chain gene rearrangement by overexpression of the non-receptor protein tyrosine kinase p56lck.
EMBOJ.
11
:
4877
81
Old, L. J., E. A. Boyse, E. Stockert.
1965
. The G (Gross) leukemia antigen.
Cancer Res.
25
:
813