Coreceptor-Independent T Cell Activation in Mice Expressing MHC Class II Molecules Mutated in the CD4 Binding Domain¹

Generation of the mature T cell repertoire is driven by the stringent processes of positive and negative selection. The outcome of selection is thought to be determined by the avidity of interaction between the developing thymocyte and the selecting epithelium (1). Thymocytes possessing a requisite reactivity to self-MHC are positively selected, ensuring effective Ag presentation in the periphery. Cells recognizing self too efficiently undergo negative selection, thereby eliminating a potentially autoreactive population. Those unable to effectively engage their selecting ligands die by neglect. While the avidity of interaction is critically dependent on the intrinsic affinity of the TCR for the peptide-MHC complex, factors such as receptor/ligand density and coreceptor engagement are also integral contributors to the overall avidity (for review, see Refs. 1–3).

The CD4 and CD8 coreceptors play a critical role in the outcome of the TCR-peptide/MHC interaction. Binding of the coreceptor to the appropriate MHC molecule stabilizes the TCR recognition complex, primarily by decreasing the rate of dissociation (4), and enhances signaling through the engaged TCR via recruitment of the protein tyrosine kinase, p56^lck (5). The trimer of TCR, coreceptor, and MHC molecule is thought to provide the optimal signal to the engaged T cell during thymic selection and peripheral activation (6, 7).

Alterations in coreceptor expression levels have been shown to significantly disrupt the selection of transgenic TCRs. Decreased coreceptor expression renders certain TCRs unable to meet the avidity threshold for positive selection, while forced overexpression raises the avidity of interaction above the threshold where negative selection occurs (8, 9, 10, 11, 12, 13). Coreceptor-mediated enhancement of TCR signaling has also been shown to play a critical role in thymic selection (14). Transgenic mice expressing tail-less coreceptors that are unable to participate in lck recruitment have been generated (15, 16, 17). These mice are dramatically impaired in their ability to select the corresponding T cell lineage. Interestingly, CD4 T cell development in the tailless CD4 mice is restored as the tailless transgene is increasingly overexpressed (15), arguing that a high degree of surface adhesive interactions can compensate for the lack of signaling through the coreceptor.

While coreceptor-mediated enhancement of TCR adhesion and signaling clearly modulates selection of the mature TCR repertoire, coreceptor engagement does not appear to be absolutely required for development of T helper cell function. Helper cell activity is apparent in CD4 knockout mice, as evidenced by the generation of an effective Th1 response upon challenge with Leishmania (18); this activity is attributed to an expanded population of double negative TCR-αβ⁺ T cells (19, 20, 21). However, CD4 null mice were subsequently shown to be unable to mount an effective Th2 response to infection with Nippostrongylus brasiliensis (22). Further studies have confirmed a generalized defect in Th2-mediated responses in CD4 knockout mice, suggesting that CD4 may actually be quite important in the development of effector function (22, 23). Alternatively, these findings might be explained if the absence of CD4 results in the selection of a restricted repertoire of Th cell specificities.

Our studies were undertaken to address the role of the CD4 coreceptor and its interaction with the MHC class II molecule during thymic selection and T cell development. The absence of a CD4⁺ T cell compartment in class II-deficient (C2D) mice provides evidence that class II molecules are necessary for selection of CD4⁺ T cells (24, 25). However, the effect of the class II-TCR interaction vs the class II-CD4 coreceptor interaction cannot be specifically addressed in these animals. To specifically examine the role of the class II-CD4 interaction, we have generated mice expressing class II molecules mutated in the CD4 binding domain (26). Class II knockout (I-A^b β-chain-deficient) mice were reconstituted with either a wild-type I-A^b β-chain transgene (Aβ WT)⁴, or an I-A^b β-chain transgene encoding the EA137/VA142 double mutation in the CD4 binding site in the β2 domain (Aβ MUT). Functional abrogation of the class II-CD4 interaction was demonstrated by increased TCR and CD4 levels on double positive thymocytes from the class II MUT mice (26). Importantly, whereas binding of the class II molecule to the CD4 coreceptor is disrupted by the mutation, peptide presentation to the TCR remains intact. Another critical feature of this model is that selection of all class II-restricted T cells occurs on class II molecules that are unable to functionally engage CD4. This is in contrast with other transgenic systems expressing MHC class I molecules mutated in the CD8 binding domain, which examine the selection only of a single TCR specificity (27, 28).

Abrogation of the CD4-class II interaction is predicted to inhibit selection of CD4⁺ T cells by decreasing the avidity of interaction between the immature thymocyte and the selecting ligand. As predicted, in vivo disruption of the class II-CD4 interaction significantly impaired selection into the CD4 compartment (26). The number of CD4⁺ T cells was reduced by one-half in the thymus and by two-thirds in the periphery of the class II MUT mice. Furthermore, selection of CD4⁺ T cells expressing the class II-restricted AND TCR transgene (29) was completely eliminated on the class II MUT background, arguing that the CD4-class II interaction is essential for the selection of at least certain TCR specificities.

In this report, we describe the phenotype of the CD4⁺ T cells that are successfully selected in the class II MUT mice. CD4⁺ T cells that develop under these conditions are predicted to be a CD4-independent subset of the preselection repertoire, bearing TCRs of sufficient affinity for the class II ligand to undergo selection in the absence of a productive class II-CD4 interaction. We show that CD4⁺ T cells from class II MUT mice, unlike the residual population of CD4⁺ T cells in C2D mice, can respond to T-dependent Ags and support Ig isotype switching. Furthermore, we show that CD4⁺ T cells from the class II MUT mice are indeed CD4 independent with respect to their peripheral activation requirements. Surprisingly, we find that CD4⁺ T cells from the class II MUT mice respond equally well when stimulated with either wild-type or mutant class II molecules. This unexpected result suggests that CD4⁺ T cells from the class II MUT mice, having been selected in the absence of a productive class II-CD4 interaction, fail to functionally engage CD4 even when subsequently provided with a wild-type class II ligand.

Wild-type Aβ transgenic and mutant Aβ (EA137/VA142) transgenic mice were generated as described (26). Mice were housed in a specific pathogen-free barrier facility (Duke University, Durham, NC).

Single cell suspensions of lymphocytes (1–2 × 10⁶ cells) were incubated for 30 min on ice in a final volume of 50 μl with Fc block (50 μg/ml; PharMingen, San Diego, CA), and the indicated combinations of Abs directly conjugated to FITC or phycoerythrin (PE). Anti-CD4 (H129.19), CD8α (53-6.7), and CD3ε (29B) were from Life Technologies (Grand Island, NY). Analyses were performed on a FACScan flow cytometer (Becton Dickinson, Mountainview, CA) using CELLQUEST software.

For the proliferation assays, 8- to 10-wk-old male mice were injected i.p. with 100 μg of keyhole limpet hemocyanin (KLH; Sigma, St. Louis, MO) or at the base of the tail with 100 μg (50 μg per side) OVA (Sigma). Antigens were solubilized in PBS (1 μg/μl) and emulsified 1:1 with CFA (Sigma) before injection. For the isotype-switching experiments, 8-wk-old mice were injected i.p. with 100 μg of 2,4 dinitrophenyl-conjugated KLH (DNP-KLH; Calbiochem-Novabiochem, La Jolla, CA) emulsified in CFA and boosted 21 days later.

Mesenteric lymph nodes from KLH-immunized mice were harvested 5–8 days postimmunization. Single cell suspensions were prepared by mincing of lymph nodes with forceps. Cell suspensions were washed with PBS, resuspended in RPMI-10 media (RPMI 1640 without sodium bicarbonate, 10% newborn calf serum, 20 mM HEPES) at 40 × 10⁶ cells/ml and incubated on anti-IgM-coated plates for 30 min at room temperature. Anti-IgM-coated plates were prepared by incubating tissue culture dishes with 100 μg/ml anti-IgM (μ-chain) Ab (Cappel, ICN Pharmaceuticals, Aurora, OH) overnight at 4°C. After passage over anti-IgM-coated plates, cells were collected, washed, and resuspended at 10⁶ cells/ml in complete media (RPMI 1640, 10% FCS, 2 mM glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 10⁻⁵ M 2-ME). In some experiments anti-IgM-treated cells were depleted of either CD4⁺ or CD8⁺ T cells by incubation with anti-CD4 or anti-CD8 Ab-coated magnetic beads (Dynabeads M-450; Dynal, Oslo, Norway). Cell suspensions were adjusted to 10–40 × 10⁶ cells/ml in RPMI-10 media, and rocked for 30 min at 4°C at a bead-to-target ratio of 3:1 before placement against a magnet for 2 min. The supernatant was collected, and the T-depleted populations were washed, resuspended at 10⁶ cells/ml in complete media, and used as responders in proliferation assays. FACS analysis of lymphocytes after T cell depletion revealed 99% elimination of the depleted CD4⁺ or CD8⁺ subset. OVA-specific lymphocytes were harvested from inguinal and periaortic lymph nodes 8–9 days post base-of-tail immunization. Single cell suspensions were generated, and cells were washed and resuspended at 10⁶ cells/ml in complete media for use in proliferation assays.

GK1.5 ascites (obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health: L3T4 monoclonal Ab (GK1.5), operated by Ogden BioSevices Corporation, Rockville, MD) was passed over a T-gel purification column (Pierce, Rockford, IL). The purified Ig was concentrated in a Centricon 100 filter (Amicon, Beverly MA) and dialyzed overnight at 4°C against dialysis buffer (10 mM EDTA, 20 mM phosphate, pH 7.0) in a 3-ml Slide-A-Lyzer pouch (Pierce). Fab fragments were generated by incubation of the dialyzed Ig with immobilized papain (Pierce) as per manufacturer’s instructions. The papain-treated Ig was then incubated with protein G (Life Technologies) for 45 min at 37°C to remove any undigested Ab. The remaining Fab fragments were dialyzed against media, quantified by determination of OD, and filter sterilized before use in proliferation assays.

Proliferation assays were performed in 96-well U-bottom plates with 10⁵ responders and 10⁵ stimulators per well in a total volume of 200 μl. To prepare APCs, splenocytes from nonimmunized mice were cultured overnight in 10 ml of complete media at 4 × 10⁶ cells/ml in the presence of the indicated concentration (0–100 μg/ml) of KLH or OVA_258–276 peptide (IINFEKLTEWTSSNVMEER, Peptide Synthesis Facility, Department of Microbiology, University of North Carolina, Chapel Hill, NC). The pulsed stimulators were then washed, irradiated (3000 rad), and resuspended in complete media at 10⁶ cells/ml for plating. Anti-CD4 Fabs (GK1.5; 6 μg/ml), anti-CD8α Abs (53-6.7; 30 μg/ml, PharMingen), and anti-MHC II Abs (25-9-17; 30 μg/ml, PharMingen) were included as indicated. Plates were pulsed at 78 h with 1.0 μCi [³H]thymidine in 25 μl media/well and harvested at 96 h with a PhD cell harvester. Filters were suspended in Betafluor (National Diagnostics, Atlanta, GA), and [³H]thymidine incorporation was measured on a Packard 1900CA Tri-Carb Liquid Scintillation Analyzer.

Sera from DNP-KLH-immunized mice were collected by retroorbital bleed at days 0, 7, 14, 21, and 28 and stored at −20°C until Ab titers were measured. Relative Ab levels from individual serum samples were determined using ELISA plates (Costar, Cambridge, MA) coated overnight at 4°C with DNP-conjugated BSA (5 μg/ml; Calbiochem-Novabiochem). Plates were washed three times in Tris-buffered saline and blocked for one h at 37°C in a solution containing 2% BSA and 1% gelatin. Following three washes, test sera (diluted 1/1,000) were added in duplicate and incubated at room temperature for 2 h. Plates were washed three times and incubated for 1 h at room temperature with alkaline phosphatase-conjugated goat anti-mouse isotype-specific Abs (2 μg/ml; Southern Biotechnology Associates, Birmingham, AL). The plates were washed five times, and alkaline phosphatase activity was detected following addition of p-nitrophenyl phosphate (Sigma). The OD was measured at a wavelength of 405 nm, and the mean OD from duplicate wells was compared between groups of animals.

C2D mice have virtually no conventional class II-restricted CD4⁺ T cells (24, 25). The residual CD4⁺ T cells are largely NK1.1⁺ T cells restricted by the nonclassical CD1 molecule (30). These mice do not mount responses to T-dependent Ags and cannot support Ig isotype switching (24, 25). In the EA137/VA142 Aβ MUT mice, the number of CD4⁺ T cells is reduced by one-half in the thymus and by two-thirds in the periphery (Fig. 1 and Table I). Moreover, nearly one-third of the peripheral CD4⁺ T cells are NK1.1⁺ (26). Of note, this does not represent an increase in the absolute number of CD4⁺ NK1.1⁺ T cells in the Aβ MUT mice; rather, their proportion is high because the number of conventional class II-restricted CD4⁺ T cells is so dramatically reduced. However, given that the Aβ MUT mice develop so few CD4⁺ T cells and that a significant proportion belong to the NK1.1⁺ subset, it was important to determine whether Aβ MUT mice were capable of mounting class II-restricted proliferative and helper responses.

We first tested the ability of lymphocytes from the Aβ MUT mice to respond to the T-dependent Ag, keyhole limpet hemocyanin (KLH). Aβ WT and EA137/VA142 Aβ MUT mice were immunized by i.p. injection with KLH emulsified in CFA. Draining lymph nodes were harvested 5–8 days following immunization, and lymphocytes were restimulated in vitro with Ag-pulsed splenocytes from either class II WT or MUT transgenic mice. In all experiments, mice transgenic for the WT Aβ-chain were used as controls for mice bearing the MUT Aβ-chain transgene. Lymphocytes from Aβ WT mice proliferated vigorously in response to WT APCs but responded poorly to APCs derived from class II MUT mice (Fig. 2,A), emphasizing that the EA137/VA142 mutation disrupts functional class II-CD4 interactions. In striking contrast, lymphocytes from the class II MUT mice proliferated robustly when provided with either the WT or MUT APCs (Fig. 2 B), arguing that CD4⁺ T cells selected in the Aβ MUT mice can respond to the mutated class II molecule. Unexpectedly, lymphocytes from the Aβ MUT mice did not proliferate more vigorously when stimulated with WT APCs than when stimulated with MUT APCs. Identical results were obtained whether the readout was proliferation or IL-2 secretion, or when splenocytes from nontransgenic wild-type C57BL/6 animals were used as APCs (data not shown).

We also tested the ability of the Aβ MUT mice to respond to the protein Ag OVA, for which the immunodominant class II-restricted peptide has been identified (OVA_258–276) (31, 32). OVA-immunized class II MUT mice clearly mounted an Ag-specific response upon in vitro restimulation with the OVA peptide (Fig. 2, C and D). Furthermore, for both Aβ WT and Aβ MUT mice, the profile of the OVA-specific response was nearly identical to that of the KLH-specific response.

To determine whether the KLH-specific T cells from the Aβ MUT mice were responding independently of CD4 engagement, the proliferation assays were repeated in the presence of anti-CD4-blocking Fab fragments. Fab fragments were used since whole Ab to CD4 has been shown to induce negative signaling and would thus inhibit a response (33). As expected, inclusion of anti-CD4 Fabs significantly inhibited (>66%) the proliferation of the KLH-specific lymphocytes from the class II WT mice. However, the response of lymphocytes from the class II MUT mice was not susceptible to inhibition by the anti-CD4 Ab when stimulated by either WT or MUT APCs (Fig. 2 E).

Thus, in contrast to the residual CD4⁺ T cells found in C2D mice, CD4⁺ T cells from the EA137/VA142 Aβ MUT mice are functional as measured by proliferation to the T-dependent Ags, KLH and OVA. Significantly, the KLH-specific CD4⁺ T cells derived from the class II MUT mice are CD4 independent in their peripheral activation requirements; they respond to Aβ MUT APCs and are not susceptible to inhibition by anti-CD4 Fabs. These results suggest that CD4⁺ T cells from the class II MUT mice, having been selected in the absence of a productive class II-CD4 interaction, are subsequently able to undergo peripheral activation in a coreceptor-independent manner.

Of note, the time course of the KLH-specific response (but not of the OVA-specific response) was reproducibly delayed by several days in the Aβ MUT vs Aβ WT mice. In the class II MUT mice, the maximal KLH-specific response was detected in lymph nodes harvested 7–8 days postimmunization, whereas in the class II WT mice, the maximal response occurred at days 4–5 postimmunization and was virtually undetectable by day 8 (data not shown). This may reflect a decrease in the precursor frequency of KLH-specific CD4⁺ T cells in the class II MUT mice and is not surprising, inasmuch as a restriction in the peripheral repertoire would be a predicted consequence of thymic selection in the absence of a functional class II-CD4 interaction.

To confirm that the KLH-specific response observed in the class II MUT mice was mounted by the CD4⁺ T cells, proliferation assays were repeated with purified T cell populations. Mesenteric lymph node cells were depleted of B cells by passage over anti-IgM coated plates and further purified by incubation with either anti-CD4- or anti-CD8-coated magnetic beads. FACS analysis revealed that the purified populations were less than 1% positive for the depleted T cell subset (data not shown). The Ag-specific response was clearly present in the purified CD4⁺ T cell compartment of both the class II WT and MUT responders (Fig. 3). Of note, CD4⁺ T cells from Aβ WT mice required 10-fold more Ag to proliferate when stimulated with MUT APCs than when stimulated with WT APCs (Fig. 3,A). However, as observed in the experiments with whole T cell populations, purified CD4⁺ T cells from the Aβ MUT mice proliferated equally well in response to either WT or MUT APCs (Fig. 3,B). The inclusion of anti-I-A^b mAb inhibited the response of purified CD4⁺ T cells from both the class II WT and MUT mice, confirming that the responding cells were indeed MHC class II-restricted CD4⁺ T cells (Fig. 3 C).

Interestingly, purified CD8⁺ T cells from the class II MUT mice also gave rise to a detectable KLH-specific response (Fig. 3,B). Although minor compared with the CD4⁺ T cell response, this finding was reproducible in subsequent experiments (Fig. 4,A). The inclusion of anti-CD8 Abs inhibited the response observed in the CD8⁺ compartment (Fig. 4,B). This response was also inhibited by anti-class II Abs, demonstrating the presence of class II-restricted Ag-specific CD8⁺ T cells in the class II MUT mice (Fig. 4 C).

Inasmuch as CD4⁺ T cells present in Aβ MUT mice are capable of mounting an Ag-specific response to the T-dependent Ag KLH, we tested their ability to support a fully differentiated Ig response. Mice were immunized with DNP-KLH at day 0, boosted at day 21, and bled at days 0, 7, 14, and 28. Serum was stored at −20°C for Ab isotype titering by anti-DNP ELISA. The response of Aβ MUT mice was compared with C2D mice as negative controls and Aβ WT and nontransgenic C57BL/6 mice as positive controls. DNP-specific Ab titers were determined for IgM, IgG1, IgG2a, IgG2b, IgG3, and IgA (Fig. 5). As previously shown, no Ig isotypes were detected following immunization of C2D mice, apart from a moderate IgM response (24, 25). In contrast, immunization of Aβ MUT, Aβ WT, and C57BL/6 mice resulted in the detectable presence of each Ab isotype examined throughout the time course of the experiment. Differences in Ab titers between the Aβ MUT and Aβ WT controls were not statistically significant.

We have previously shown that in vivo disruption of the MHC class II-CD4 interaction significantly impairs selection of CD4⁺ T cells (26). However, a functional class II-CD4 interaction is clearly not required for selection of all CD4⁺ T cells, since these cells still comprise 7–12% of all peripheral lymphocytes in mice expressing the Aβ MUT class II protein. Here we describe the functional characteristics of the CD4⁺ T cells selected in the Aβ MUT background. KLH-immunized class II MUT mice clearly mounted an Ag-specific proliferative response. Importantly, the majority of the response could be attributed specifically to the activity of CD4⁺ T cells. No expansion in the number of double negative T cells was observed in the Aβ MUT mice (data not shown), nor was any proliferative activity detected in this population, since the small response mounted by the CD4-depleted T cells was eliminated by the inclusion of anti-CD8 Abs. This is significant inasmuch as an expanded population of double negative T cells is thought to be responsible for the helper activity observed in CD4-deficient mice (20, 21, 34). The Aβ MUT mice were also capable of supporting a fully differentiated Ig response, similar in magnitude and kinetics to that observed in Aβ WT mice.

Interestingly, whereas the time course of the maximal KLH response was delayed in the Aβ MUT vs the Aβ WT mice, the time course of the OVA response was very similar (days 9–11 for both). This finding is consistent with the idea that the class II MUT mice possess T cells with the relevant OVA specificity, such that the time course of the response is not delayed, but that the delay in the KLH response reflects a narrowing in the repertoire of KLH-specific T cells. This is not unlikely, given that KLH is a far more complex Ag than OVA, so that the KLH response of the Aβ MUT mice as compared with the response of the Aβ WT mice could consist of different TCR specificities or be directed at different epitopes of the KLH protein.

An important predicted consequence of having been selected in an environment devoid of functional class II-CD4 interactions is that CD4⁺ T cells in the class II MUT mice will be CD4 independent in their peripheral activation requirements. The experiments presented in this report support this hypothesis. KLH-specific T cells from Aβ MUT mice proliferated vigorously when stimulated with MUT APCs. Furthermore, the response was not inhibited by the inclusion of anti-CD4 Fab fragments, demonstrating that the T cells from the class II MUT mice are responding in a coreceptor-independent manner. Similarly, TCR transgenic CD8⁺ T cells, in mice expressing class I molecules unable to bind CD8, displayed CD8-independent lysis of target cells (35).

Intriguingly, the response of T cells from the Aβ MUT mice was not augmented when stimulated with APCs expressing WT class II proteins. These results are in striking contrast to those obtained with CTLs derived from mice expressing class I molecules with suboptimal affinity for CD8 (35). In these studies, the magnitude of the lytic response was significantly increased when the CD8⁺ T cells were provided with target cells bearing class I molecules that could productively engage CD8. The results in the Aβ MUT mice were also surprising inasmuch as the introduction of wild-type CD4 into a CD4⁻ variant of an Ag-specific hybridoma has been shown to enhance IL-2 production (36). Similarly, Ag-specific CD4⁻ (or CD8⁻) hybridomas were shown to respond at lower peptide doses when CD4 (or CD8) was provided (17, 36, 37, 38, 39).

Our observations in the Aβ MUT mice might not be surprising if the T cell were already being maximally stimulated, despite the absence of coreceptor engagement. However, if this were the case, the presence of effective coreceptor engagement would be expected to have an effect at lower Ag doses, such that a shift in the dose-response curve would be observed in the presence of WT APCs. The absence of such a shift suggests that there may be an intrinsic defect in CD4 recruitment or signaling pathways in the CD4⁺ T cells from the class II MUT mice. Viola et al. have shown that triggered TCRs and coreceptors are coordinately down-modulated, even when the MHC-coreceptor interaction does not occur (40). This coordinated down-regulation is thought to be due to the recruitment of CD4-associated lck by CD3ζ/ZAP-70. It will be critical to determine whether this recruitment of CD4 still occurs in lymphocytes from the Aβ MUT mice or whether CD4 is sequestered from the TCR complex such that it can no longer be recruited, even in the presence of a WT class II ligand.

Alternatively, our observations may be more simply explained if the TCRs expressed on CD4⁺ T cells in the Aβ MUT mice are of higher intrinsic affinity for MHC/peptide complexes. A model for how higher affinity TCR-MHC/peptide interactions may translate mechanistically into CD4-independent T cells was proposed by Davis and colleagues (39), based on a “sequential engagement” model wherein CD4 acts to stabilize preformed clusters of TCR-MHC/peptide complexes (41, 42). In this model, CD4 recruitment occurs following TCR-MHC/peptide engagement and only for those complexes that can engage the TCR for at least 1–2 s at 25°C, which is in the range of measured affinities seen with typical CD4-dependent T cells interacting with an agonist peptide/MHC complex (43). Recruited CD4 may then serve to stabilize the complex and allow the generation of a full activation signal. TCR-MHC/peptide interactions, which are intrinsically more stable will last longer, and, although they too may recruit CD4, the recruited CD4 may be of little consequence for the generation of a complete activation signal. Thus, in the case of higher affinity or “CD4-independent” T cells, CD4 recruitment may not appear to contribute to the response, qualitatively or quantitatively. The phenotype of the CD4⁺ T cells selected in the Aβ MUT mice is consistent with this hypothesis.

The presence of a class II-restricted Ag-specific response in the CD8⁺ T cell compartment is another notable outcome of selection in the Aβ MUT background. Selection of these class II-restricted CD8⁺ T cells can be potentially accounted for by several mechanisms. In the Aβ MUT mice, selection of class II-restricted TCRs must be occurring independently of CD4 engagement. Therefore, any T cell bearing a TCR with high affinity for MHC II could potentially undergo positive selection, regardless of which coreceptor it maintains. The increased TCR levels on double-positive thymocytes in the Aβ MUT mice (26) might also facilitate selection of TCRs that would normally require coreceptor-mediated interactions. Alternatively, TCR-independent MHC class I-CD8 interactions might facilitate selection of class II-restricted CD8⁺ T cells in the Aβ MUT mice. Kirbirg et al. have demonstrated that, in mice transgenic for a class II-restricted TCR, both CD4⁺CD8⁻ and CD4⁻CD8⁺ T cells expressing the transgenic TCR develop, but that full maturation of the CD8⁺ transgenic T cells requires expression of class I as well as class II MHC molecules (44). Likewise, Matachek et al. have demonstrated that, in CD4 knockout mice, selection of the class II-restricted AND TCR transgene into the CD8 compartment requires both class I and class II molecules (45). As we have previously reported, selection of the AND TCR transgene into the CD4 compartment is completely eliminated in the Aβ MUT mice (26). However, a small population of AND TCR⁺ CD8⁺ T cells are present in the periphery of AND/Aβ MUT mice (16% vs 5% in AND/Aβ WT mice, data not shown). Furthermore, this population is reduced to background (5%) if Aβ MUT mice expressing the AND TCR transgene are crossed onto a β₂-microglobulin^−/− background (manuscript in preparation). These data argue that selection of the class II-restricted CD8⁺ T cells in the Aβ MUT mice may indeed utilize TCR-independent class I-CD8 interactions.

While we favor the hypothesis that the CD4⁺ T cells present in the Aβ MUT mice have been selected based on expression of high affinity TCRs, TCR-independent class I-CD8 interactions could be facilitating selection of both CD4⁺ and CD8⁺ T cells. To determine whether TCR-independent class I-CD8 interactions are required for selection of CD4⁺ T cells in the Aβ MUT mice, we have crossed the Aβ MUT transgenic line onto an MHC knockout background, such that the only MHC molecules available for mediating selection are the EA137/VA142 MUT class II molecules. As expected, selection of CD8⁺ T cells was completely eliminated in the absence of class I molecules (data not shown). However, no decrease in the efficiency of CD4⁺ T cell selection was observed, arguing that TCR-independent class I-CD8 interactions are not required for selection of CD4⁺ T cells in the Aβ MUT mice.

We have demonstrated that the CD4⁺ T cells present in the Aβ MUT mice are functional T helper cells and are coreceptor-independent in their peripheral activation requirements. Ongoing experiments with Ag-specific hybridomas derived from the Aβ MUT mice are designed to more specifically examine the issues of TCR affinity and alterations in peripheral T cell repertoire. Recent evidence suggests that class II-CD4 interactions are important in Th2 vs Th1 subset differentiation (22, 23, 46, 47). Future experiments will utilize the Aβ MUT mice to examine Th development in the absence of productive class II-CD4 engagement.

We thank Dr. Michael S. Krangel for critical review of the manuscript, Mike Cook and Lynn Martinek for flow cytometry, Paula Farless and Timothy Floreth for technical assistance, and Rommel Shipman for animal care.