Characterizing the Impact of CD8 Antibodies on Class I MHC Multimer Binding
http://www.100md.com
免疫学杂志 2005年第7期
Abstract
Many studies have suggested that CD8 Abs affect the binding of class I MHC tetramers/multimers to CD8+ T cells, which has led to the interpretation that CD8 participates directly in multimer binding. In contrast, a recent publication has argued that CD8 Abs instead cause reorganization of TCR distribution and hence have an indirect effect on multimer binding to the TCR alone. We address these issues by testing the role of CD8 and the impact of CD8 Abs on the binding of normal and mutant multimers to Ag-specific mouse T cells. Our data suggest that, in this system, CD8 Abs act directly on CD8 and only mediate their effects on multimer binding when CD8 is capable of binding to the multimer. These data reinforce the paradigm that CD8 plays an active and direct role in binding of class I MHC multimers.
Introduction
Peptide/MHC multimers have revolutionized the capacity to identify Ag-specific T cells. Because of their high valency, multimers can compensate for the inherent low affinity of the TCR-peptide/MHC interactions and allow detection of specific T cells by flow cytometry and other techniques.
Initial studies on CD4 T cells indicated that peptide/MHC class II tetramer binding occurred independently of the coreceptor. Although CD4 was found to be critical for responses induced by tetramers, the tetramer binding per se was equally efficient whether or not CD4 was available (1, 2). In contrast, we and others (3, 4, 5, 6, 7, 8, 9) argued that binding of class I MHC multimers was influenced strongly by the participation of CD8, consistent with earlier data using monomeric class I MHC ligands (10). The requirement for CD8 in multimer binding varied with the particular TCR-peptide/MHC combination from mild to absolute, in keeping with previous functional studies, which have suggested CD8-dependent and -independent T cell interactions. However, in nearly all cases, some role for CD8 in multimer binding could be observed.
For the most part, the role for CD8 in multimer binding was tested using Abs to CD8. Intriguingly, some Abs were found to inhibit or enhance multimer binding (3, 4, 5, 6, 7, 8, 9). Inhibitory Abs might be expected to occlude class I MHC binding sites, although it has been proposed that enhancing Abs mediate their effects by stabilizing higher affinity conformations of CD8 (11).
However, these interpretations have been challenged by recent data from Wooldridge et al. (12). These authors presented intriguing data that argued that CD8 Abs have dramatic effects on peptide/MHC multimer binding even when the multimer was not expected to engage CD8. Thus, using class I MHC multimers bearing crippling mutations in key CD8 binding sites, they were still able to observe enhancement and inhibition of tetramer binding by anti-CD8 Abs when tested on human and mouse CTL lines. These authors suggested that, rather than directly influencing CD8-class I multimer binding, CD8 Ab binding was influencing the ability of the TCR to engage the multimer. They went on to propose that, although some of these effects could be mediated by steric hindrance, a more likely model was that anti-CD8 Abs were causing changes in TCR distribution on the cell surface, leading to altered multimer binding (12). These data raised serious doubts about the proposed role for direct CD8 participation in multimer binding and also the validity of using anti-CD8 Abs to explore such a role.
In this study, we revisited the role of CD8 in multimer binding, using the 2C TCR-transgenic system, in which both CD8-dependent and -independent multimer binding can be analyzed. We studied CD8+ and CD8– 2C T cells and the effects of various CD8 Abs on binding of specific class I MHC tetramers possessing normal vs mutant CD8 binding sites. These data indicate that CD8 actively participates in tetramer binding and that CD8 Abs only impact tetramer binding when CD8 is capable of engaging the class I ligand. These findings were consistent over various multimer staining conditions and held for both naive T cells and CTL lines. Hence, these data support the model that CD8 binding directly contributes to class I multimer binding.
Materials and Methods
Mice
2C TCR-transgenic mice were maintained on a B6 or a Thy1.1,Rag-1–/– background under specific pathogen-free conditions at the University of Minnesota and were used in accordance with Institutional Animal Care and Use Committee guidelines.
2C cell line
2C splenocytes were stimulated with irradiated (1500 cGy) BALB/c splenocytes in 24-well tissue culture plates. Recombinant human IL-2 (Tecin, supplied by the Biological Resources Facility, National Cancer Institute/National Institutes of Health) was added every 3–4 days to a final concentration of 500 U/ml. The line was restimulated weekly with fresh irradiated BALB/c splenocytes and IL-2.
Peptides and multimers
The peptides SIY (SIYRYYGL) and A6 (SIYRYAGL) were obtained from Research Genetics and Invitrogen Life Technologies. The Kb 227K mutant was generated by site directed mutagenesis of the Kb-bsp plasmid (a kind gift from J. Altman, Emory University (Atlanta, GA)) using the Quick change kit according to the manufacturer’s protocol (Stratagene). Primers used were 5'-AATGGGGAGGAGCTGATCCAGAAGATGGAGCTTGTGGAGACC-3' and 5'-GGTCTCCACAAGCTCCATCTTCTGGATCAGCTCCTCCCC-3' (nucleotides changed are italicized). Monomers containing either SIY or A6 were generated and biotinylated as described previously (3). Refolded MHC monomers were purified by size exclusion, concentrated to 1 mg/ml, dialyzed against water, and stored in 25 μl of aliquot at –70°C. Multimers were made fresh before use by mixing an equal concentration of purified monomers and streptavidin:phycoerythrin (PE)3 or streptavidin:allophycocyanin (Molecular Probes), followed by incubation at room temperature for at least 1 h. Multimers were used individually at a final concentration of 10 μg/ml or as indicated in the figure legend.
Flow cytometry
Naive lymph node (LN) cells (0.15–1 x 106) and a 2C line (0.15 x 106) were preincubated on ice with an anti-CD16/32 Ab mixture (Fc block) (eBioscience) for 10 min in FACS buffer (PBS, 1% FCS, 0.02% sodium azide) before costaining with multimers and Abs for 1 h. The FITC-conjugated anti-CD8 Ab 53-6.7 (BD Biosciences or BioLegend) and CT-CD8 (Caltag Laboratories) were used at 5 μg/ml for enhancing and blocking multimer staining, respectively. In some experiments, allophycocyanin-conjugated anti-CD8 Abs were used for enhancing (53-6.7; eBioscience) and blocking (CT-CD8; Caltag Laboratories) multimer staining. Allophycocyanin-conjugated Thy1.1 or Thy1.2, FITC-conjugated Thy1.2 (eBioscience), and PerCP-CD4 (BD Biosciences) were used for identifying T cells. Data was collected on a BD Biosciences FACSCalibur or LSRII instrument and analyzed using FlowJo (Tree Star).
Results
The 2C TCR system exhibits CD8-dependent and -independent multimer binding
To test the role of CD8 in class I MHC multimer binding, we selected the 2C TCR-transgenic system. These mice generate both a CD4–CD8+ and a CD4–CD8– population of mature T cells bearing the transgenic TCR, allowing for analysis of the role of CD8 in the absence of anti-CD8 Abs. Previous reports (3, 6) have suggested 2C T cells can bind some ligands, including SIY/Kb multimers, even in the absence of CD8. Accordingly, we find SIY/Kb multimer binds CD8+ 2C cells only slightly (but reproducibly) more intensely than CD8– 2C cells (Fig. 1, A–C). SIY/Kb binding to CD8+ 2C cells can be enhanced or diminished by anti-CD8 Abs (Fig. 1, A–C), but again, the effects are moderate compared with other CD8 T cells, including the OT-I TCR-transgenic model (3, 6). In contrast, binding of SIY/Kb monomers to CD8+ vs CD8– 2C cell lines was shown to be strongly CD8 dependent (13), indicating that our use of SIY/Kb multimers may partially overcome the requirement for CD8. Furthermore, we found that 2C T cell binding to different peptide/MHC multimers was highly influenced by the presence of CD8. An alanine substitution at position 6 of the SIY peptide (here called "A6") does not influence peptide binding to Kb (14) but does impact 2C recognition, the A6/Kb ligand acting as a TCR antagonist for 2C cells (15). Multimers prepared with this ligand bound to 2C cells in a strictly CD8-dependent way (Fig. 1, D–F). First, this multimer bound only detectably to CD8+ 2C cells (Fig. 1, D and E). Although staining in the absence of CD8 Abs (gray histogram, Fig. 1D) precludes identification of the CD8+/CD8– pools, we note that the frequency of A6/Kb multimer-binding cells matches the frequency of CD8+ 2C cells (Fig. 1 legend). Second, A6/Kb binding to 2C cells was influenced strongly by anti-CD8 Abs, which were enhanced by 53-6.7 and inhibited by CT-CD8 (Fig. 1, E and F), in keeping with previously reported effects of these Abs (3, 6). Thus, these data suggested considerable variation in the contribution of CD8 to multimer staining of 2C T cells depending on the peptide/MHC ligands involved, offering an opportunity to additionally define the role of CD8 in multimer binding.
FIGURE 1. CD8-dependent and -independent multimer binding on naive 2C T cells. The impact of anti-CD8 Abs on binding of SIY/Kb (A–C) or A6/Kb (D–F) multimers to CD8+ and CD8– 2C T cells was determined. 2C LN cells were incubated with multimers alone or together with the indicated anti-CD8 Abs. A and D, Multimer staining on bulk 2C T cells (Thy1.2+, CD4–) in the presence of 53-6.7 (open histogram), CT-CD8 (filled histogram), or absence of CD8 Abs (gray histogram) is shown. Multimer staining on non-T cells (Thy1.2–, CD4–) is shown as a dotted histogram. For the cells stained with A6/Kb multimer in the absence of anti-CD8 Abs (gray line, D), a marker (arrow) was introduced to determine the frequency of A6/Kb multimer low and high populations. A6/Kb multimer-positive cells represented 66% of the bulk 2C (Thy1.2+, CD4–) pool, closely matching the frequency of CD8+ 2C cells in this sample (70%, data not shown). Multimer binding is also shown for 2C cells stained in the presence of the anti-CD8 Abs 53-6.7 (B and E) or CT-CD8 (C and F), gating on the CD8+ (open histogram) and CD8– (filled histogram) populations. B and C, The mean fluorescent intensity for the CD8+ and CD8– populations is given for comparison.
Staining with class I multimers carrying a mutant CD8 binding site is not influenced by anti-CD8 Abs
A central conclusion by Wooldridge et al. (12) was that CD8 Abs might influence multimer binding beyond their direct effects on CD8-class I interactions. This model was based in part on the finding that binding of class I ligands that cannot engage CD8 was still influenced by anti-CD8 Abs. To test this in our model, we constructed multimers in which the Kb molecule was mutated from DK at residue 227. This mutation influences a key binding region between CD8 and class I alleles, and such mutants have been shown to cripple CD8-class I interactions (10, 16, 17). The mutant SIY/Kb 227K multimers could stain 2C T cells efficiently, but changes in binding were observed (Fig. 2). First, binding of SIY/Kb 227K multimers was virtually identical on both CD8+ and CD8– 2C cells. Second, and more importantly for this discussion, binding of SIY/Kb 227K multimers to CD8+ 2C cells was not influenced by anti-CD8 Abs (Fig. 2). So, although binding of wild-type SIY/Kb multimers to 2C cells was reduced (to a similar extent) by anti-CD8 blockade and by lack of CD8 expression, binding to SIY/Kb 227K multimers was not influenced by either expression or blockade of CD8.
FIGURE 2. Binding of CD8-null multimer on 2C T cells is not affected by anti-CD8 Abs. The SIY/Kb 227K multimer used to stain 2C LN cells binds equally well in the presence or absence of CD8 molecules. A, Multimer staining on bulk 2C T cells (Thy1.2+, CD4–) in the presence of 53-6.7 (open histogram), CT-CD8 (filled histogram), or absence of CD8 Abs (gray histogram). Multimer staining on non-T cells (Thy1.2–, CD4–) is shown as a dotted histogram. Multimer staining in the presence of 53-6.7 (B) or CT-CD8 (C) is shown for the CD8+ (open histogram) and CD8– (filled histogram) populations.
It was possible that the CD8 dependence of multimer binding was influenced by the dose of multimer used. Use of saturating doses of multimers might minimize the contribution of CD8 and mask an effect of the anti-CD8 Abs. To address this, we titrated both wild-type and 227K mutant SIY/Kb multimers on 2C cells in the presence or absence of anti-CD8 Abs. As expected, titration of both ligands leads to decreased multimer staining, yet the qualitative impact of anti-CD8 Abs was preserved (Fig. 3). Hence, for SIY/Kb staining 53-6.7 enhanced multimer binding to CD8+ 2C cells, whereas multimer staining in the presence of the blocking CT-CD8 Ab was similar on CD8+ and CD8– 2C cells (Fig. 3, A–C, and data not shown). In contrast, SIY/Kb 227K multimer binding to 2C cells was similar regardless of CD8 expression or the presence of either CD8 Ab (Fig. 3, D–F, and data not shown). Thus, these data argue that the dose of multimer does not influence the overall pattern of CD8 requirement in binding to 2C cells.
FIGURE 3. Multimer titration does not change the effects of CD8 Abs in SIY/Kb and SIY/Kb 227K multimer binding. 2C cells were stained with SIY/Kb (A–C) or SIY/Kb 227K (D–F) multimers in the presence of 53-6.7 (open histogram) or CT-CD8 (shaded histogram). Multimer staining on non-T cells (Thy1.2–, CD4–) is shown as a dotted histogram. Multimers were diluted 1/4 (A and D), 1/32 (B and E), or 1/512 (C and F) relative to the usual concentration (10 μg/ml).
In contrast to these effects, neither CD8+ nor CD8– 2C cells were capable of binding to the A6/Kb 227 mutant multimers (data not shown). Furthermore, anti-CD8 Abs had no effect on binding of this multimer, consistent with the idea that its engagement was completely dependent on an intact CD8 binding site.
Thus, in aggregate, these data are contrary to the model of Wooldridge et al. (12), which predicts that anti-CD8 Abs would influence binding of both CD8-dependent and CD8-independent multimers.
Assessing direct interactions between CD8 and class I MHC multimers
The experiments described above show that class I multimer binding is influenced by CD8 expression and anti-CD8 Abs, consistent with a direct role for CD8 in multimer binding. However, additional experiments were needed to define whether CD8 itself engages the multimers. This was a particular concern for the experiments using SIY-containing multimers because SIY/Kb binds only slightly better to CD8+ vs CD8– 2C cells and because SIY/Kb 227 mutant multimers stain 2C cells almost as efficiently as wild-type multimers. Such data raised the question of whether CD8 actually engages SIY/Kb multimers when they are bound to the 2C TCR.
Initially, we determined peptide/MHC multimer colocalization with CD8 using fluorescence resonance energy transfer (FRET) (6, 7). The emission spectrum of the fluorochrome PE overlaps the excitation wavelengths for allophycocyanin. Hence, if molecules of PE and allophycocyanin are in close proximity and suitable orientation, excitation of PE will produce FRET in the allophycocyanin molecule, which can be detected by flow cytometry as a FL-3 signal, and this approach has been used previously to study peptide/MHC binding to TCR and CD8 (6, 7, 18). As expected, based on previous reports (6, 7), we could detect a FRET signal when PE-labeled SIY/Kb multimers were allowed to bind to 2C cells in the presence of allophycocyanin-conjugated 53-6.7 but not in the presence of allophycocyanin-conjugated CT-CD8 (Fig. 4, top panels), despite the fact that multimer staining of 2C cells was evident in the presence of either CD8 Ab. The FRET signal was specific to the fluorochrome combination, in that using FITC-conjugated anti-CD8 Abs with PE-multimers did not yield a FL-3 (FRET) signal (data not shown).
FIGURE 4. FRET is observed only between multimers that can associate with CD8. LN cells from a 2C Rag–/– mouse were stained with PE-labeled multimers as indicated in the presence of allophycocyanin-labeled anti-CD8 Abs 53-6.7 or CT-CD8 as indicated. The y-axis shows the FRET between PE and allophycocyanin, as detected by a signal in FL3.
Using this system, we could further explore the role of CD8 and anti-CD8 Abs on multimer binding. Notably, although SIY/Kb 227K multimers bind 2C T cells efficiently (Fig. 2), no FRET signals were induced by this interaction, even in the presence of 53-6.7 (Fig. 4, middle panels). It is most likely that the absence of a FRET signal in this case indicates that the multimer and CD8 (bound to Ab) are not in close proximity, although the same result could potentially occur if the orientation of the two molecules was changed significantly. These data then suggest that the FRET signal requires direct interaction between CD8 and the multimer, rather than reflecting a constitutive bystander association of CD8 with the TCR. A weak but reproducible FRET signal was also produced when A6/Kb multimers were bound to 2C cells in the presence of 53-6.7, again reflecting the key involvement of CD8 in binding to this ligand (Fig. 4, bottom panels, and data not shown).
Next, we used another approach to test the impact on 2C binding of mutating the CD8 binding site on SIY/Kb multimers. Although 2C T cells can bind both SIY/Kb and SIY/Kb 227K multimers, if CD8 bound directly to the multimer, we would expect that the presence of the CD8 binding site on the wild-type Kb multimers would offer these multimers a competitive advantage in binding to CD8+ 2C cells. Hence, we developed a competition assay, in which PE-SIY/Kb and allophycocyanin-SIY/Kb 227K were mixed and used to stain 2C T cells. In the presence of the enhancing CD8 Ab (53-6.7), two populations of multimer-binding cells were observed: one pool (CD8– 2C cells) bound both multimers, whereas the other population (CD8+ 2C cells) bound preferentially to the SIY/Kb multimers (Fig. 5, A and B). A similar result was seen when no anti-CD8 Ab was used in the staining; although we cannot be certain which cells are CD8+ vs CD8– in this case, the frequency of cells bound to one vs both multimers indicate a similar pattern of staining in the presence or absence of the 53-6.7 Ab. These data suggest that, for CD8– 2C cells, both wild-type and 227K mutant multimers had equal opportunity to stably bind, whereas for CD8+ 2C cells, the presence of a CD8 binding site on wild-type multimers gave them a selective competitive advantage. When blocking CD8 Abs were used in the staining, both CD8+ and CD8– 2C cells bound both multimers equally well (Fig. 5C). These data suggest that the CD8 interaction directly participates in stabilizing the SIY/Kb multimer binding to the 2C TCR and that CD8– 2C cells behave similarly to CD8+ T cells treated with blocking CD8 Abs.
FIGURE 5. Competition between SIY/Kb and SIY/Kb 227K multimers for binding to 2C T cells. 2C LN cells were incubated simultaneously with PE-conjugated SIY/Kb 227K and allophycocyanin-conjugated SIY/Kb multimers in the absence of anti-CD8 Abs (A) or presence of FITC-conjugated 53-6.7 (B) or FITC-conjugated CT-CD8 (C). The dot plots show staining for both multimers. Cells binding either one or both multimers were gated, and the histograms in B and C indicate multimer staining on CD8+ cells (open histogram) or CD8– cells (shaded histogram). Staining of multimer-negative cells is shown as a dotted histogram.
Participation of CD8 in multimer binding to CTL lines
The experiments described above and in our previous studies involved naive T cells, whereas those of Wooldridge et al. (12) used long-term CTL lines. It was possible that the role of CD8 in multimer binding was altered in naive vs effector CTL, which might lead to altered sensitivity to anti-CD8 Abs.
To test this in our model, we developed long-term 2C T cell lines by regular Ag stimulation in vitro. These cells maintained expression of CD8 and the 2C TCR, although CD8 expression levels broadened compared with naive 2C T cells (data not shown). Importantly, multimer binding to the 2C line showed the same characteristics as we observed for naive 2C cells; binding of SIY/Kb and A6/Kb multimers was enhanced or blocked by anti-CD8 Abs as expected, whereas binding of SIY/Kb 227K multimers was not influenced by anti-CD8 Abs (Fig. 6). Similar results were found when 2C cells were analyzed after one primary in vitro stimulation compared with five rounds of restimulation, suggesting the longevity of the line did not influence the pattern of multimer staining (data not shown). Hence, similar to naive 2C T cells, we conclude that multimer binding on the 2C CTL lines was influenced only by anti-CD8 Abs when the multimer possessed an intact CD8 binding site.
FIGURE 6. Impact of anti-CD8 Abs on multimer binding to a 2C T cell line. A 2C CTL line was generated and maintained by restimulation for 6 wk in culture. The cells were stained with SIY/Kb (A), SIY/Kb 227K (B), and A6/Kb (C) multimers in the absence of anti-CD8 Abs (gray histogram) or in the presence of 53-6.7 (open histogram) or CT-CD8 (filled histogram).
Discussion
This study was designed to test the role of CD8 in binding to class I MHC multimers. Our previous reports and those of others (3, 4, 5, 6, 10, 19) were interpreted to mean that enhancement or blockade of multimer binding by CD8 Abs implied participation of CD8 in the binding. In contrast, a recent report (12) reached the surprising conclusion that CD8 Abs may be indirectly influencing the ability of TCR to engage peptide/MHC multimers and that CD8-class I interactions per se were not required for multimer binding.
Our findings support the former conclusions, i.e., that CD8 Ab effects reflect participation of CD8 in binding to the multimer. We show this in several ways. First, in our hands, binding of specific peptide/MHC multimers to CD8+ 2C cells in the presence of blocking CD8 Abs mimics the binding of these multimers to CD8– T cells. Second, binding of multimers carrying the D227K mutation in the CD8 binding site is equivalent on both CD8+ and CD8– T cells and is not affected by anti-CD8 Abs. Third, we observed a FRET signal between anti-CD8 and peptide/MHC multimers only when the latter was capable of binding to CD8. Fourth, a role for CD8 was manifested by the ability of SIY/Kb multimers to outcompete SIY/Kb 227K multimers for binding to CD8+ 2C T cells, but this competitive advantage was negated by the absence or blockade of CD8 on the 2C T cells.
Wooldridge et al. (12) discuss our previous data in the 2C system, making the valid point that multimer binding to CD8– 2C cells may be difficult to interpret because of the possible impact of CD8 on TCR distribution. However, we observed striking parallels of binding for D227K mutant multimers binding to 2C cells compared with wild-type multimers binding to either CD8– cells or CD8+ T cells treated with blocking anti-CD8 Abs. Likewise, neither CD8– 2C T cells nor CD8+ 2C cells cultured in the presence of the blocking CT-CD8 Ab could bind A6/Kb multimers. Hence, in our studies, analysis of CD8– 2C T cells was fully consistent with other methods used to test the role of CD8-class I interactions.
In additional experiments, we explored the role of CD8 in binding to another TCR-transgenic model system, OT-I, reactive to OVA/Kb. As in previous studies, binding of specific multimers to these CD8+ T cells was highly influenced by anti-CD8 Abs (3, 6). The correlation for a direct role of CD8 binding to class I was supported by the inability of OVA/Kb227K to bind to OT-I T cells (our unpublished observations). Likewise, Schott and Ploegh (20) reported that OVA/Kb multimers carrying other CD8 binding site mutants (E223K or a Q226L, D227N double mutation) were completely unable to bind OT-I T cells. Although the overarching dependence of CD8 availability for multimer binding in the OT-I system ironically limits its usefulness for the current studies, these observations are fully consistent with a direct role for CD8-class I binding in regulating multimer binding to OT-I T cells.
That anti-CD8 Abs influence multimer binding to CD8 directly, rather than mediating an indirect effect on TCR engagement, is also suggested by reports from our group and others describing noncognate binding of CD8 to class I multimers (i.e., interactions in which CD8 but not the TCR bind the peptide/MHC ligand). Such noncognate interactions are strongly influenced by anti-CD8 Abs, and the patterns of enhancement or inhibition by specific anti-CD8 Ab clones are identical for both cognate and noncognate binding (Refs. 3 , 11 , 21 , 22 and our unpublished observations). For noncognate multimer binding, it has been shown that the presence or absence of the TCR is irrelevant (11, 21, 22). Hence, these data additionally support the direct interpretation of how anti-CD8 Abs mediate their effects.
In aggregate, our studies were unable to confirm the model proposed by Wooldridge et al. (12), suggesting additional effects of anti-CD8 Abs (beyond the direct effect on CD8-class I MHC interactions) on multimer binding. The basis for the discrepancy between our results and those of Wooldridge et al. (12) is currently unclear. The most obvious concern is that the CD8-null multimers used in the earlier study were, in fact, still able to bind CD8 to some extent. Functional and biophysical studies argue that the mutants used in that (and in this) study show compromised CD8 binding, but the magnitude of the loss in binding is difficult to determine because the affinity of CD8 binding to wild-type class I MHC is itself very low. However, such a model is hard to sustain with the range of CD8-null multimers used by Wooldridge et al. (12) and is difficult to reconcile with the greater resistance of the CD8-null multimers to anti-CD8 Ab blockade reported by that group.
The majority of studies reported by Wooldridge et al. (12) involved CTL lines or clones, whereas we focused primarily on naive T cells. As the association between CD8 and TCR could alter with activation, this might be a source of the discrepancy. However, we also analyzed long-term 2C CTL lines and concluded that, as for naive cells, CD8-class I interactions contributed directly to class I multimer binding (Fig. 6). Another potential experimental difference between our studies is the sequential order and incubation temperatures used for staining with anti-CD8 Abs and multimers. However, using the staining protocol described by Wooldridge et al. (12), we still did not reveal additional effects of anti-CD8 Abs on multimer binding (data not shown).
Importantly, in the report by Wooldridge et al. (12), the most clear-cut experiments demonstrating a disconnect between CD8-class I binding and anti-CD8 Ab effects on multimer binding involved analysis of human CTL lines. Although mouse CD8 T cells were also studied, leading those authors to reach similar conclusions about the indirect effects of anti-CD8 Abs on multimer binding, Wooldridge pointed out that there are well-defined differences in the affinity and dependency for CD8 in human vs mouse systems. Therefore, it is certainly possible that studies in the two species cannot be fully reconciled, either due to inherent differences in the nature of CD8-class I binding or of the specific properties of the panels of anti-CD8 Abs available. Hence, our data should not be taken to contradict the conclusions of Wooldridge et al. (12) with respect to the human system. However, our findings do reinforce the model that, in the mouse, the effect of anti-CD8 Abs can be interpreted as a direct role of CD8-class I MHC engagement. Furthermore, our data do not imply that anti-CD8 Abs can have no effect other than to alter class I MHC binding. As with the CD4 coreceptor, several studies have demonstrated signaling events induced by Ab-mediated CD8 cross-linking (12, 23, 24), and recent data suggest such interactions can lead to premature death of immature thymocytes (25). Our current studies were designed to test what effect CD8 plays in multimer binding. However, experimental approaches that permit signaling from anti-CD8 Ab engagement may well alter multimer binding by additional means, and this should be considered in design of experiments using multimer staining approaches.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank the members of the Jameson/Hogquist labs for their valuable input and Larry Pease (Mayo Clinic, Rochester, MN) for a timely gift of 2C mice.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grant AI52163 (to S.C.J.).
2 Address correspondence and reprint requests to Dr. Stephen C. Jameson, MMC 334, 420 Delaware Street SE, Minneapolis, MN 55455. E-mail address: james024@umn.edu
3 Abbreviations used in this paper: PE, phycoerythrin; LN, lymph node; FRET, fluorescence resonance energy transfer.
Received for publication October 19, 2004. Accepted for publication January 7, 2005.
References
Boniface, J. J., J. D. Rabinowitz, C. Wulfing, J. Hampl, Z. Reich, J. D. Altman, R. M. Kantor, C. Beeson, H. M. McConnell, M. M. Davis. 1998. Initiation of signal transduction through the T cell receptor requires the multivalent engagement of peptide/MHC ligands. Immunity 9:459.
Crawford, F., H. Kozono, J. White, P. Marrack, J. Kappler. 1998. Detection of antigen-specific T cells with multivalent soluble class II MHC covalent peptide complexes. Immunity 8:675
Daniels, M. A., S. C. Jameson. 2000. Critical role for CD8 in T cell receptor binding and activation by peptide/major histocompatibility complex multimers. J. Exp. Med. 191:335.
Denkberg, G., C. J. Cohen, Y. Reiter. 2001. Critical role for CD8 in binding of MHC tetramers to TCR: CD8 antibodies block specific binding of human tumor-specific MHC-peptide tetramers to TCR. J. Immunol. 167:270
Campanelli, R., B. Palermo, S. Garbelli, S. Mantovani, P. Lucchi, A. Necker, E. Lantelme, C. Giachino. 2002. Human CD8 co-receptor is strictly involved in MHC-peptide tetramer-TCR binding and T cell activation. Int. Immunol. 14:39.
Block, M. S., A. J. Johnson, Y. Mendez-Fernandez, L. R. Pease. 2001. Monomeric class I molecules mediate TCR/CD3/CD8 interaction on the surface of T cells. J. Immunol. 167:821.
Lee, P. U., D. M. Kranz. 2003. Allogeneic and syngeneic class I MHC complexes drive the association of CD8 and TCR on 2C T cells. Mol. Immunol. 39:687.
Dutoit, V., P. Guillaume, M. Ayyoub, C. S. Hesdorffer, I. F. Luescher, D. Valmori. 2003. Decreased binding of peptides-MHC class I (pMHC) multimeric complexes to CD8 affects their binding avidity for the TCR but does not significantly impact on pMHC/TCR dissociation rate. J. Immunol. 170:5110.
Buslepp, J., S. E. Kerry, D. Loftus, J. A. Frelinger, E. Appella, E. J. Collins. 2003. High affinity xenoreactive TCR:MHC interaction recruits CD8 in absence of binding to MHC. J. Immunol. 170:373
Luescher, I. F., E. Vivier, A. Layer, J. Mahiou, F. Godeau, B. Malissen, P. Romero. 1995. CD8 modulation of T-cell antigen receptor-ligand interactions on living cytotoxic T lymphocytes. Nature 373:353.
Devine, L., M. E. Hodsdon, M. A. Daniels, S. C. Jameson, P. B. Kavathas. 2004. Location of the epitope for an anti-CD8 antibody 53.6.7 which enhances CD8-MHC class I interaction indicates antibody stabilization of a higher affinity CD8 conformation. Immunol. Lett. 93:123.
Wooldridge, L., S. L. Hutchinson, E. M. Choi, A. Lissina, E. Jones, F. Mirza, P. R. Dunbar, D. A. Price, V. Cerundolo, A. K. Sewell. 2003. Anti-CD8 antibodies can inhibit or enhance peptide-MHC class I (pMHCI) multimer binding: this is paralleled by their effects on CTL activation and occurs in the absence of an interaction between pMHCI and CD8 on the cell surface. J. Immunol. 171:6650
Cho, B. K., K. C. Lian, P. Lee, A. Brunmark, C. McKinley, J. Chen, D. M. Kranz, H. N. Eisen. 2001. Differences in antigen recognition and cytolytic activity of CD8+ and CD8– T cells that express the same antigen-specific receptor. Proc. Natl. Acad. Sci. USA 98:1723.
Brock, R., K.-H. Weismuller, G. Jung, P. Walden. 1996. Molecular basis for the recognition of two structurally different major histocompatibility complex/peptide complexes by a single T-cell receptor. Proc. Natl. Acad. Sci. USA 93:13108.
Daniels, M. A., S. L. Schober, K. A. Hogquist, S. C. Jameson. 1999. Cutting edge: a test of the dominant negative signal model for TCR antagonism. J. Immunol. 162:3761.
Shen, L., T. A. Potter, K. P. Kane. 1996. Glu227Lys substitution in the acidic loop of major histocompatibility complex class I 3 domain distinguishes low avidity CD8 coreceptor and avidity enhanced CD8 accessory functions. J. Exp. Med. 184:1671.
Potter, T. A., T. V. Rajan, R. F. Dick, 2nd, J. A. Bluestone. 1989. Substitution at residue 227 of H-2 class I molecules abrogates recognition by CD8-dependent, but not CD8-independent, cytotoxic T lymphocytes. Nature 337:73.
Batard, P., J. Szollosi, I. Luescher, J. C. Cerottini, R. MacDonald, P. Romero. 2002. Use of phycoerythrin and allophycocyanin for fluorescence resonance energy transfer analyzed by flow cytometry: advantages and limitations. Cytometry 48:97
Choi, E. M., J. L. Chen, L. Wooldridge, M. Salio, A. Lissina, N. Lissin, I. F. Hermans, J. D. Silk, F. Mirza, M. J. Palmowski, et al 2003. High avidity antigen-specific CTL identified by CD8-independent tetramer staining. J. Immunol. 171:5116.
Schott, E., H. L. Ploegh. 2002. Mouse MHC class I tetramers that are unable to bind to CD8 reveal the need for CD8 engagement in order to activate naive CD8 T cells. Eur. J. Immunol. 32:3425.
Moody, A. M., Y. Xiong, H. C. Chang, E. L. Reinherz. 2001. The CD8 co-receptor on double-positive thymocytes binds with differing affinities to the products of distinct class I MHC loci. Eur. J. Immunol. 31:2791
Daniels, M. A., L. Devine, J. D. Miller, J. M. Moser, A. E. Lukacher, J. D. Altman, P. Kavathas, K. A. Hogquist, S. C. Jameson. 2001. CD8 binding to MHC class I molecules is influenced by T cell maturation and glycosylation. Immunity 15:1051.
Tomonari, K., S. Spencer. 1990. Epitope-specific binding of CD8 regulates activation of T cells and induction of cytotoxicity. Int. Immunol. 2:1189.
Ravichandran, K. S., S. J. Burakoff. 1994. Evidence for differential intracellular signaling via CD4 and CD8 molecules. J. Exp. Med. 179:727.
Grebe, K. M., R. L. Clarke, T. A. Potter. 2004. Ligation of CD8 leads to apoptosis of thymocytes that have not undergone positive selection. Proc. Natl. Acad. Sci. USA 101:10410.(Philmore O. Holman, Eliza)
Many studies have suggested that CD8 Abs affect the binding of class I MHC tetramers/multimers to CD8+ T cells, which has led to the interpretation that CD8 participates directly in multimer binding. In contrast, a recent publication has argued that CD8 Abs instead cause reorganization of TCR distribution and hence have an indirect effect on multimer binding to the TCR alone. We address these issues by testing the role of CD8 and the impact of CD8 Abs on the binding of normal and mutant multimers to Ag-specific mouse T cells. Our data suggest that, in this system, CD8 Abs act directly on CD8 and only mediate their effects on multimer binding when CD8 is capable of binding to the multimer. These data reinforce the paradigm that CD8 plays an active and direct role in binding of class I MHC multimers.
Introduction
Peptide/MHC multimers have revolutionized the capacity to identify Ag-specific T cells. Because of their high valency, multimers can compensate for the inherent low affinity of the TCR-peptide/MHC interactions and allow detection of specific T cells by flow cytometry and other techniques.
Initial studies on CD4 T cells indicated that peptide/MHC class II tetramer binding occurred independently of the coreceptor. Although CD4 was found to be critical for responses induced by tetramers, the tetramer binding per se was equally efficient whether or not CD4 was available (1, 2). In contrast, we and others (3, 4, 5, 6, 7, 8, 9) argued that binding of class I MHC multimers was influenced strongly by the participation of CD8, consistent with earlier data using monomeric class I MHC ligands (10). The requirement for CD8 in multimer binding varied with the particular TCR-peptide/MHC combination from mild to absolute, in keeping with previous functional studies, which have suggested CD8-dependent and -independent T cell interactions. However, in nearly all cases, some role for CD8 in multimer binding could be observed.
For the most part, the role for CD8 in multimer binding was tested using Abs to CD8. Intriguingly, some Abs were found to inhibit or enhance multimer binding (3, 4, 5, 6, 7, 8, 9). Inhibitory Abs might be expected to occlude class I MHC binding sites, although it has been proposed that enhancing Abs mediate their effects by stabilizing higher affinity conformations of CD8 (11).
However, these interpretations have been challenged by recent data from Wooldridge et al. (12). These authors presented intriguing data that argued that CD8 Abs have dramatic effects on peptide/MHC multimer binding even when the multimer was not expected to engage CD8. Thus, using class I MHC multimers bearing crippling mutations in key CD8 binding sites, they were still able to observe enhancement and inhibition of tetramer binding by anti-CD8 Abs when tested on human and mouse CTL lines. These authors suggested that, rather than directly influencing CD8-class I multimer binding, CD8 Ab binding was influencing the ability of the TCR to engage the multimer. They went on to propose that, although some of these effects could be mediated by steric hindrance, a more likely model was that anti-CD8 Abs were causing changes in TCR distribution on the cell surface, leading to altered multimer binding (12). These data raised serious doubts about the proposed role for direct CD8 participation in multimer binding and also the validity of using anti-CD8 Abs to explore such a role.
In this study, we revisited the role of CD8 in multimer binding, using the 2C TCR-transgenic system, in which both CD8-dependent and -independent multimer binding can be analyzed. We studied CD8+ and CD8– 2C T cells and the effects of various CD8 Abs on binding of specific class I MHC tetramers possessing normal vs mutant CD8 binding sites. These data indicate that CD8 actively participates in tetramer binding and that CD8 Abs only impact tetramer binding when CD8 is capable of engaging the class I ligand. These findings were consistent over various multimer staining conditions and held for both naive T cells and CTL lines. Hence, these data support the model that CD8 binding directly contributes to class I multimer binding.
Materials and Methods
Mice
2C TCR-transgenic mice were maintained on a B6 or a Thy1.1,Rag-1–/– background under specific pathogen-free conditions at the University of Minnesota and were used in accordance with Institutional Animal Care and Use Committee guidelines.
2C cell line
2C splenocytes were stimulated with irradiated (1500 cGy) BALB/c splenocytes in 24-well tissue culture plates. Recombinant human IL-2 (Tecin, supplied by the Biological Resources Facility, National Cancer Institute/National Institutes of Health) was added every 3–4 days to a final concentration of 500 U/ml. The line was restimulated weekly with fresh irradiated BALB/c splenocytes and IL-2.
Peptides and multimers
The peptides SIY (SIYRYYGL) and A6 (SIYRYAGL) were obtained from Research Genetics and Invitrogen Life Technologies. The Kb 227K mutant was generated by site directed mutagenesis of the Kb-bsp plasmid (a kind gift from J. Altman, Emory University (Atlanta, GA)) using the Quick change kit according to the manufacturer’s protocol (Stratagene). Primers used were 5'-AATGGGGAGGAGCTGATCCAGAAGATGGAGCTTGTGGAGACC-3' and 5'-GGTCTCCACAAGCTCCATCTTCTGGATCAGCTCCTCCCC-3' (nucleotides changed are italicized). Monomers containing either SIY or A6 were generated and biotinylated as described previously (3). Refolded MHC monomers were purified by size exclusion, concentrated to 1 mg/ml, dialyzed against water, and stored in 25 μl of aliquot at –70°C. Multimers were made fresh before use by mixing an equal concentration of purified monomers and streptavidin:phycoerythrin (PE)3 or streptavidin:allophycocyanin (Molecular Probes), followed by incubation at room temperature for at least 1 h. Multimers were used individually at a final concentration of 10 μg/ml or as indicated in the figure legend.
Flow cytometry
Naive lymph node (LN) cells (0.15–1 x 106) and a 2C line (0.15 x 106) were preincubated on ice with an anti-CD16/32 Ab mixture (Fc block) (eBioscience) for 10 min in FACS buffer (PBS, 1% FCS, 0.02% sodium azide) before costaining with multimers and Abs for 1 h. The FITC-conjugated anti-CD8 Ab 53-6.7 (BD Biosciences or BioLegend) and CT-CD8 (Caltag Laboratories) were used at 5 μg/ml for enhancing and blocking multimer staining, respectively. In some experiments, allophycocyanin-conjugated anti-CD8 Abs were used for enhancing (53-6.7; eBioscience) and blocking (CT-CD8; Caltag Laboratories) multimer staining. Allophycocyanin-conjugated Thy1.1 or Thy1.2, FITC-conjugated Thy1.2 (eBioscience), and PerCP-CD4 (BD Biosciences) were used for identifying T cells. Data was collected on a BD Biosciences FACSCalibur or LSRII instrument and analyzed using FlowJo (Tree Star).
Results
The 2C TCR system exhibits CD8-dependent and -independent multimer binding
To test the role of CD8 in class I MHC multimer binding, we selected the 2C TCR-transgenic system. These mice generate both a CD4–CD8+ and a CD4–CD8– population of mature T cells bearing the transgenic TCR, allowing for analysis of the role of CD8 in the absence of anti-CD8 Abs. Previous reports (3, 6) have suggested 2C T cells can bind some ligands, including SIY/Kb multimers, even in the absence of CD8. Accordingly, we find SIY/Kb multimer binds CD8+ 2C cells only slightly (but reproducibly) more intensely than CD8– 2C cells (Fig. 1, A–C). SIY/Kb binding to CD8+ 2C cells can be enhanced or diminished by anti-CD8 Abs (Fig. 1, A–C), but again, the effects are moderate compared with other CD8 T cells, including the OT-I TCR-transgenic model (3, 6). In contrast, binding of SIY/Kb monomers to CD8+ vs CD8– 2C cell lines was shown to be strongly CD8 dependent (13), indicating that our use of SIY/Kb multimers may partially overcome the requirement for CD8. Furthermore, we found that 2C T cell binding to different peptide/MHC multimers was highly influenced by the presence of CD8. An alanine substitution at position 6 of the SIY peptide (here called "A6") does not influence peptide binding to Kb (14) but does impact 2C recognition, the A6/Kb ligand acting as a TCR antagonist for 2C cells (15). Multimers prepared with this ligand bound to 2C cells in a strictly CD8-dependent way (Fig. 1, D–F). First, this multimer bound only detectably to CD8+ 2C cells (Fig. 1, D and E). Although staining in the absence of CD8 Abs (gray histogram, Fig. 1D) precludes identification of the CD8+/CD8– pools, we note that the frequency of A6/Kb multimer-binding cells matches the frequency of CD8+ 2C cells (Fig. 1 legend). Second, A6/Kb binding to 2C cells was influenced strongly by anti-CD8 Abs, which were enhanced by 53-6.7 and inhibited by CT-CD8 (Fig. 1, E and F), in keeping with previously reported effects of these Abs (3, 6). Thus, these data suggested considerable variation in the contribution of CD8 to multimer staining of 2C T cells depending on the peptide/MHC ligands involved, offering an opportunity to additionally define the role of CD8 in multimer binding.
FIGURE 1. CD8-dependent and -independent multimer binding on naive 2C T cells. The impact of anti-CD8 Abs on binding of SIY/Kb (A–C) or A6/Kb (D–F) multimers to CD8+ and CD8– 2C T cells was determined. 2C LN cells were incubated with multimers alone or together with the indicated anti-CD8 Abs. A and D, Multimer staining on bulk 2C T cells (Thy1.2+, CD4–) in the presence of 53-6.7 (open histogram), CT-CD8 (filled histogram), or absence of CD8 Abs (gray histogram) is shown. Multimer staining on non-T cells (Thy1.2–, CD4–) is shown as a dotted histogram. For the cells stained with A6/Kb multimer in the absence of anti-CD8 Abs (gray line, D), a marker (arrow) was introduced to determine the frequency of A6/Kb multimer low and high populations. A6/Kb multimer-positive cells represented 66% of the bulk 2C (Thy1.2+, CD4–) pool, closely matching the frequency of CD8+ 2C cells in this sample (70%, data not shown). Multimer binding is also shown for 2C cells stained in the presence of the anti-CD8 Abs 53-6.7 (B and E) or CT-CD8 (C and F), gating on the CD8+ (open histogram) and CD8– (filled histogram) populations. B and C, The mean fluorescent intensity for the CD8+ and CD8– populations is given for comparison.
Staining with class I multimers carrying a mutant CD8 binding site is not influenced by anti-CD8 Abs
A central conclusion by Wooldridge et al. (12) was that CD8 Abs might influence multimer binding beyond their direct effects on CD8-class I interactions. This model was based in part on the finding that binding of class I ligands that cannot engage CD8 was still influenced by anti-CD8 Abs. To test this in our model, we constructed multimers in which the Kb molecule was mutated from DK at residue 227. This mutation influences a key binding region between CD8 and class I alleles, and such mutants have been shown to cripple CD8-class I interactions (10, 16, 17). The mutant SIY/Kb 227K multimers could stain 2C T cells efficiently, but changes in binding were observed (Fig. 2). First, binding of SIY/Kb 227K multimers was virtually identical on both CD8+ and CD8– 2C cells. Second, and more importantly for this discussion, binding of SIY/Kb 227K multimers to CD8+ 2C cells was not influenced by anti-CD8 Abs (Fig. 2). So, although binding of wild-type SIY/Kb multimers to 2C cells was reduced (to a similar extent) by anti-CD8 blockade and by lack of CD8 expression, binding to SIY/Kb 227K multimers was not influenced by either expression or blockade of CD8.
FIGURE 2. Binding of CD8-null multimer on 2C T cells is not affected by anti-CD8 Abs. The SIY/Kb 227K multimer used to stain 2C LN cells binds equally well in the presence or absence of CD8 molecules. A, Multimer staining on bulk 2C T cells (Thy1.2+, CD4–) in the presence of 53-6.7 (open histogram), CT-CD8 (filled histogram), or absence of CD8 Abs (gray histogram). Multimer staining on non-T cells (Thy1.2–, CD4–) is shown as a dotted histogram. Multimer staining in the presence of 53-6.7 (B) or CT-CD8 (C) is shown for the CD8+ (open histogram) and CD8– (filled histogram) populations.
It was possible that the CD8 dependence of multimer binding was influenced by the dose of multimer used. Use of saturating doses of multimers might minimize the contribution of CD8 and mask an effect of the anti-CD8 Abs. To address this, we titrated both wild-type and 227K mutant SIY/Kb multimers on 2C cells in the presence or absence of anti-CD8 Abs. As expected, titration of both ligands leads to decreased multimer staining, yet the qualitative impact of anti-CD8 Abs was preserved (Fig. 3). Hence, for SIY/Kb staining 53-6.7 enhanced multimer binding to CD8+ 2C cells, whereas multimer staining in the presence of the blocking CT-CD8 Ab was similar on CD8+ and CD8– 2C cells (Fig. 3, A–C, and data not shown). In contrast, SIY/Kb 227K multimer binding to 2C cells was similar regardless of CD8 expression or the presence of either CD8 Ab (Fig. 3, D–F, and data not shown). Thus, these data argue that the dose of multimer does not influence the overall pattern of CD8 requirement in binding to 2C cells.
FIGURE 3. Multimer titration does not change the effects of CD8 Abs in SIY/Kb and SIY/Kb 227K multimer binding. 2C cells were stained with SIY/Kb (A–C) or SIY/Kb 227K (D–F) multimers in the presence of 53-6.7 (open histogram) or CT-CD8 (shaded histogram). Multimer staining on non-T cells (Thy1.2–, CD4–) is shown as a dotted histogram. Multimers were diluted 1/4 (A and D), 1/32 (B and E), or 1/512 (C and F) relative to the usual concentration (10 μg/ml).
In contrast to these effects, neither CD8+ nor CD8– 2C cells were capable of binding to the A6/Kb 227 mutant multimers (data not shown). Furthermore, anti-CD8 Abs had no effect on binding of this multimer, consistent with the idea that its engagement was completely dependent on an intact CD8 binding site.
Thus, in aggregate, these data are contrary to the model of Wooldridge et al. (12), which predicts that anti-CD8 Abs would influence binding of both CD8-dependent and CD8-independent multimers.
Assessing direct interactions between CD8 and class I MHC multimers
The experiments described above show that class I multimer binding is influenced by CD8 expression and anti-CD8 Abs, consistent with a direct role for CD8 in multimer binding. However, additional experiments were needed to define whether CD8 itself engages the multimers. This was a particular concern for the experiments using SIY-containing multimers because SIY/Kb binds only slightly better to CD8+ vs CD8– 2C cells and because SIY/Kb 227 mutant multimers stain 2C cells almost as efficiently as wild-type multimers. Such data raised the question of whether CD8 actually engages SIY/Kb multimers when they are bound to the 2C TCR.
Initially, we determined peptide/MHC multimer colocalization with CD8 using fluorescence resonance energy transfer (FRET) (6, 7). The emission spectrum of the fluorochrome PE overlaps the excitation wavelengths for allophycocyanin. Hence, if molecules of PE and allophycocyanin are in close proximity and suitable orientation, excitation of PE will produce FRET in the allophycocyanin molecule, which can be detected by flow cytometry as a FL-3 signal, and this approach has been used previously to study peptide/MHC binding to TCR and CD8 (6, 7, 18). As expected, based on previous reports (6, 7), we could detect a FRET signal when PE-labeled SIY/Kb multimers were allowed to bind to 2C cells in the presence of allophycocyanin-conjugated 53-6.7 but not in the presence of allophycocyanin-conjugated CT-CD8 (Fig. 4, top panels), despite the fact that multimer staining of 2C cells was evident in the presence of either CD8 Ab. The FRET signal was specific to the fluorochrome combination, in that using FITC-conjugated anti-CD8 Abs with PE-multimers did not yield a FL-3 (FRET) signal (data not shown).
FIGURE 4. FRET is observed only between multimers that can associate with CD8. LN cells from a 2C Rag–/– mouse were stained with PE-labeled multimers as indicated in the presence of allophycocyanin-labeled anti-CD8 Abs 53-6.7 or CT-CD8 as indicated. The y-axis shows the FRET between PE and allophycocyanin, as detected by a signal in FL3.
Using this system, we could further explore the role of CD8 and anti-CD8 Abs on multimer binding. Notably, although SIY/Kb 227K multimers bind 2C T cells efficiently (Fig. 2), no FRET signals were induced by this interaction, even in the presence of 53-6.7 (Fig. 4, middle panels). It is most likely that the absence of a FRET signal in this case indicates that the multimer and CD8 (bound to Ab) are not in close proximity, although the same result could potentially occur if the orientation of the two molecules was changed significantly. These data then suggest that the FRET signal requires direct interaction between CD8 and the multimer, rather than reflecting a constitutive bystander association of CD8 with the TCR. A weak but reproducible FRET signal was also produced when A6/Kb multimers were bound to 2C cells in the presence of 53-6.7, again reflecting the key involvement of CD8 in binding to this ligand (Fig. 4, bottom panels, and data not shown).
Next, we used another approach to test the impact on 2C binding of mutating the CD8 binding site on SIY/Kb multimers. Although 2C T cells can bind both SIY/Kb and SIY/Kb 227K multimers, if CD8 bound directly to the multimer, we would expect that the presence of the CD8 binding site on the wild-type Kb multimers would offer these multimers a competitive advantage in binding to CD8+ 2C cells. Hence, we developed a competition assay, in which PE-SIY/Kb and allophycocyanin-SIY/Kb 227K were mixed and used to stain 2C T cells. In the presence of the enhancing CD8 Ab (53-6.7), two populations of multimer-binding cells were observed: one pool (CD8– 2C cells) bound both multimers, whereas the other population (CD8+ 2C cells) bound preferentially to the SIY/Kb multimers (Fig. 5, A and B). A similar result was seen when no anti-CD8 Ab was used in the staining; although we cannot be certain which cells are CD8+ vs CD8– in this case, the frequency of cells bound to one vs both multimers indicate a similar pattern of staining in the presence or absence of the 53-6.7 Ab. These data suggest that, for CD8– 2C cells, both wild-type and 227K mutant multimers had equal opportunity to stably bind, whereas for CD8+ 2C cells, the presence of a CD8 binding site on wild-type multimers gave them a selective competitive advantage. When blocking CD8 Abs were used in the staining, both CD8+ and CD8– 2C cells bound both multimers equally well (Fig. 5C). These data suggest that the CD8 interaction directly participates in stabilizing the SIY/Kb multimer binding to the 2C TCR and that CD8– 2C cells behave similarly to CD8+ T cells treated with blocking CD8 Abs.
FIGURE 5. Competition between SIY/Kb and SIY/Kb 227K multimers for binding to 2C T cells. 2C LN cells were incubated simultaneously with PE-conjugated SIY/Kb 227K and allophycocyanin-conjugated SIY/Kb multimers in the absence of anti-CD8 Abs (A) or presence of FITC-conjugated 53-6.7 (B) or FITC-conjugated CT-CD8 (C). The dot plots show staining for both multimers. Cells binding either one or both multimers were gated, and the histograms in B and C indicate multimer staining on CD8+ cells (open histogram) or CD8– cells (shaded histogram). Staining of multimer-negative cells is shown as a dotted histogram.
Participation of CD8 in multimer binding to CTL lines
The experiments described above and in our previous studies involved naive T cells, whereas those of Wooldridge et al. (12) used long-term CTL lines. It was possible that the role of CD8 in multimer binding was altered in naive vs effector CTL, which might lead to altered sensitivity to anti-CD8 Abs.
To test this in our model, we developed long-term 2C T cell lines by regular Ag stimulation in vitro. These cells maintained expression of CD8 and the 2C TCR, although CD8 expression levels broadened compared with naive 2C T cells (data not shown). Importantly, multimer binding to the 2C line showed the same characteristics as we observed for naive 2C cells; binding of SIY/Kb and A6/Kb multimers was enhanced or blocked by anti-CD8 Abs as expected, whereas binding of SIY/Kb 227K multimers was not influenced by anti-CD8 Abs (Fig. 6). Similar results were found when 2C cells were analyzed after one primary in vitro stimulation compared with five rounds of restimulation, suggesting the longevity of the line did not influence the pattern of multimer staining (data not shown). Hence, similar to naive 2C T cells, we conclude that multimer binding on the 2C CTL lines was influenced only by anti-CD8 Abs when the multimer possessed an intact CD8 binding site.
FIGURE 6. Impact of anti-CD8 Abs on multimer binding to a 2C T cell line. A 2C CTL line was generated and maintained by restimulation for 6 wk in culture. The cells were stained with SIY/Kb (A), SIY/Kb 227K (B), and A6/Kb (C) multimers in the absence of anti-CD8 Abs (gray histogram) or in the presence of 53-6.7 (open histogram) or CT-CD8 (filled histogram).
Discussion
This study was designed to test the role of CD8 in binding to class I MHC multimers. Our previous reports and those of others (3, 4, 5, 6, 10, 19) were interpreted to mean that enhancement or blockade of multimer binding by CD8 Abs implied participation of CD8 in the binding. In contrast, a recent report (12) reached the surprising conclusion that CD8 Abs may be indirectly influencing the ability of TCR to engage peptide/MHC multimers and that CD8-class I interactions per se were not required for multimer binding.
Our findings support the former conclusions, i.e., that CD8 Ab effects reflect participation of CD8 in binding to the multimer. We show this in several ways. First, in our hands, binding of specific peptide/MHC multimers to CD8+ 2C cells in the presence of blocking CD8 Abs mimics the binding of these multimers to CD8– T cells. Second, binding of multimers carrying the D227K mutation in the CD8 binding site is equivalent on both CD8+ and CD8– T cells and is not affected by anti-CD8 Abs. Third, we observed a FRET signal between anti-CD8 and peptide/MHC multimers only when the latter was capable of binding to CD8. Fourth, a role for CD8 was manifested by the ability of SIY/Kb multimers to outcompete SIY/Kb 227K multimers for binding to CD8+ 2C T cells, but this competitive advantage was negated by the absence or blockade of CD8 on the 2C T cells.
Wooldridge et al. (12) discuss our previous data in the 2C system, making the valid point that multimer binding to CD8– 2C cells may be difficult to interpret because of the possible impact of CD8 on TCR distribution. However, we observed striking parallels of binding for D227K mutant multimers binding to 2C cells compared with wild-type multimers binding to either CD8– cells or CD8+ T cells treated with blocking anti-CD8 Abs. Likewise, neither CD8– 2C T cells nor CD8+ 2C cells cultured in the presence of the blocking CT-CD8 Ab could bind A6/Kb multimers. Hence, in our studies, analysis of CD8– 2C T cells was fully consistent with other methods used to test the role of CD8-class I interactions.
In additional experiments, we explored the role of CD8 in binding to another TCR-transgenic model system, OT-I, reactive to OVA/Kb. As in previous studies, binding of specific multimers to these CD8+ T cells was highly influenced by anti-CD8 Abs (3, 6). The correlation for a direct role of CD8 binding to class I was supported by the inability of OVA/Kb227K to bind to OT-I T cells (our unpublished observations). Likewise, Schott and Ploegh (20) reported that OVA/Kb multimers carrying other CD8 binding site mutants (E223K or a Q226L, D227N double mutation) were completely unable to bind OT-I T cells. Although the overarching dependence of CD8 availability for multimer binding in the OT-I system ironically limits its usefulness for the current studies, these observations are fully consistent with a direct role for CD8-class I binding in regulating multimer binding to OT-I T cells.
That anti-CD8 Abs influence multimer binding to CD8 directly, rather than mediating an indirect effect on TCR engagement, is also suggested by reports from our group and others describing noncognate binding of CD8 to class I multimers (i.e., interactions in which CD8 but not the TCR bind the peptide/MHC ligand). Such noncognate interactions are strongly influenced by anti-CD8 Abs, and the patterns of enhancement or inhibition by specific anti-CD8 Ab clones are identical for both cognate and noncognate binding (Refs. 3 , 11 , 21 , 22 and our unpublished observations). For noncognate multimer binding, it has been shown that the presence or absence of the TCR is irrelevant (11, 21, 22). Hence, these data additionally support the direct interpretation of how anti-CD8 Abs mediate their effects.
In aggregate, our studies were unable to confirm the model proposed by Wooldridge et al. (12), suggesting additional effects of anti-CD8 Abs (beyond the direct effect on CD8-class I MHC interactions) on multimer binding. The basis for the discrepancy between our results and those of Wooldridge et al. (12) is currently unclear. The most obvious concern is that the CD8-null multimers used in the earlier study were, in fact, still able to bind CD8 to some extent. Functional and biophysical studies argue that the mutants used in that (and in this) study show compromised CD8 binding, but the magnitude of the loss in binding is difficult to determine because the affinity of CD8 binding to wild-type class I MHC is itself very low. However, such a model is hard to sustain with the range of CD8-null multimers used by Wooldridge et al. (12) and is difficult to reconcile with the greater resistance of the CD8-null multimers to anti-CD8 Ab blockade reported by that group.
The majority of studies reported by Wooldridge et al. (12) involved CTL lines or clones, whereas we focused primarily on naive T cells. As the association between CD8 and TCR could alter with activation, this might be a source of the discrepancy. However, we also analyzed long-term 2C CTL lines and concluded that, as for naive cells, CD8-class I interactions contributed directly to class I multimer binding (Fig. 6). Another potential experimental difference between our studies is the sequential order and incubation temperatures used for staining with anti-CD8 Abs and multimers. However, using the staining protocol described by Wooldridge et al. (12), we still did not reveal additional effects of anti-CD8 Abs on multimer binding (data not shown).
Importantly, in the report by Wooldridge et al. (12), the most clear-cut experiments demonstrating a disconnect between CD8-class I binding and anti-CD8 Ab effects on multimer binding involved analysis of human CTL lines. Although mouse CD8 T cells were also studied, leading those authors to reach similar conclusions about the indirect effects of anti-CD8 Abs on multimer binding, Wooldridge pointed out that there are well-defined differences in the affinity and dependency for CD8 in human vs mouse systems. Therefore, it is certainly possible that studies in the two species cannot be fully reconciled, either due to inherent differences in the nature of CD8-class I binding or of the specific properties of the panels of anti-CD8 Abs available. Hence, our data should not be taken to contradict the conclusions of Wooldridge et al. (12) with respect to the human system. However, our findings do reinforce the model that, in the mouse, the effect of anti-CD8 Abs can be interpreted as a direct role of CD8-class I MHC engagement. Furthermore, our data do not imply that anti-CD8 Abs can have no effect other than to alter class I MHC binding. As with the CD4 coreceptor, several studies have demonstrated signaling events induced by Ab-mediated CD8 cross-linking (12, 23, 24), and recent data suggest such interactions can lead to premature death of immature thymocytes (25). Our current studies were designed to test what effect CD8 plays in multimer binding. However, experimental approaches that permit signaling from anti-CD8 Ab engagement may well alter multimer binding by additional means, and this should be considered in design of experiments using multimer staining approaches.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank the members of the Jameson/Hogquist labs for their valuable input and Larry Pease (Mayo Clinic, Rochester, MN) for a timely gift of 2C mice.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grant AI52163 (to S.C.J.).
2 Address correspondence and reprint requests to Dr. Stephen C. Jameson, MMC 334, 420 Delaware Street SE, Minneapolis, MN 55455. E-mail address: james024@umn.edu
3 Abbreviations used in this paper: PE, phycoerythrin; LN, lymph node; FRET, fluorescence resonance energy transfer.
Received for publication October 19, 2004. Accepted for publication January 7, 2005.
References
Boniface, J. J., J. D. Rabinowitz, C. Wulfing, J. Hampl, Z. Reich, J. D. Altman, R. M. Kantor, C. Beeson, H. M. McConnell, M. M. Davis. 1998. Initiation of signal transduction through the T cell receptor requires the multivalent engagement of peptide/MHC ligands. Immunity 9:459.
Crawford, F., H. Kozono, J. White, P. Marrack, J. Kappler. 1998. Detection of antigen-specific T cells with multivalent soluble class II MHC covalent peptide complexes. Immunity 8:675
Daniels, M. A., S. C. Jameson. 2000. Critical role for CD8 in T cell receptor binding and activation by peptide/major histocompatibility complex multimers. J. Exp. Med. 191:335.
Denkberg, G., C. J. Cohen, Y. Reiter. 2001. Critical role for CD8 in binding of MHC tetramers to TCR: CD8 antibodies block specific binding of human tumor-specific MHC-peptide tetramers to TCR. J. Immunol. 167:270
Campanelli, R., B. Palermo, S. Garbelli, S. Mantovani, P. Lucchi, A. Necker, E. Lantelme, C. Giachino. 2002. Human CD8 co-receptor is strictly involved in MHC-peptide tetramer-TCR binding and T cell activation. Int. Immunol. 14:39.
Block, M. S., A. J. Johnson, Y. Mendez-Fernandez, L. R. Pease. 2001. Monomeric class I molecules mediate TCR/CD3/CD8 interaction on the surface of T cells. J. Immunol. 167:821.
Lee, P. U., D. M. Kranz. 2003. Allogeneic and syngeneic class I MHC complexes drive the association of CD8 and TCR on 2C T cells. Mol. Immunol. 39:687.
Dutoit, V., P. Guillaume, M. Ayyoub, C. S. Hesdorffer, I. F. Luescher, D. Valmori. 2003. Decreased binding of peptides-MHC class I (pMHC) multimeric complexes to CD8 affects their binding avidity for the TCR but does not significantly impact on pMHC/TCR dissociation rate. J. Immunol. 170:5110.
Buslepp, J., S. E. Kerry, D. Loftus, J. A. Frelinger, E. Appella, E. J. Collins. 2003. High affinity xenoreactive TCR:MHC interaction recruits CD8 in absence of binding to MHC. J. Immunol. 170:373
Luescher, I. F., E. Vivier, A. Layer, J. Mahiou, F. Godeau, B. Malissen, P. Romero. 1995. CD8 modulation of T-cell antigen receptor-ligand interactions on living cytotoxic T lymphocytes. Nature 373:353.
Devine, L., M. E. Hodsdon, M. A. Daniels, S. C. Jameson, P. B. Kavathas. 2004. Location of the epitope for an anti-CD8 antibody 53.6.7 which enhances CD8-MHC class I interaction indicates antibody stabilization of a higher affinity CD8 conformation. Immunol. Lett. 93:123.
Wooldridge, L., S. L. Hutchinson, E. M. Choi, A. Lissina, E. Jones, F. Mirza, P. R. Dunbar, D. A. Price, V. Cerundolo, A. K. Sewell. 2003. Anti-CD8 antibodies can inhibit or enhance peptide-MHC class I (pMHCI) multimer binding: this is paralleled by their effects on CTL activation and occurs in the absence of an interaction between pMHCI and CD8 on the cell surface. J. Immunol. 171:6650
Cho, B. K., K. C. Lian, P. Lee, A. Brunmark, C. McKinley, J. Chen, D. M. Kranz, H. N. Eisen. 2001. Differences in antigen recognition and cytolytic activity of CD8+ and CD8– T cells that express the same antigen-specific receptor. Proc. Natl. Acad. Sci. USA 98:1723.
Brock, R., K.-H. Weismuller, G. Jung, P. Walden. 1996. Molecular basis for the recognition of two structurally different major histocompatibility complex/peptide complexes by a single T-cell receptor. Proc. Natl. Acad. Sci. USA 93:13108.
Daniels, M. A., S. L. Schober, K. A. Hogquist, S. C. Jameson. 1999. Cutting edge: a test of the dominant negative signal model for TCR antagonism. J. Immunol. 162:3761.
Shen, L., T. A. Potter, K. P. Kane. 1996. Glu227Lys substitution in the acidic loop of major histocompatibility complex class I 3 domain distinguishes low avidity CD8 coreceptor and avidity enhanced CD8 accessory functions. J. Exp. Med. 184:1671.
Potter, T. A., T. V. Rajan, R. F. Dick, 2nd, J. A. Bluestone. 1989. Substitution at residue 227 of H-2 class I molecules abrogates recognition by CD8-dependent, but not CD8-independent, cytotoxic T lymphocytes. Nature 337:73.
Batard, P., J. Szollosi, I. Luescher, J. C. Cerottini, R. MacDonald, P. Romero. 2002. Use of phycoerythrin and allophycocyanin for fluorescence resonance energy transfer analyzed by flow cytometry: advantages and limitations. Cytometry 48:97
Choi, E. M., J. L. Chen, L. Wooldridge, M. Salio, A. Lissina, N. Lissin, I. F. Hermans, J. D. Silk, F. Mirza, M. J. Palmowski, et al 2003. High avidity antigen-specific CTL identified by CD8-independent tetramer staining. J. Immunol. 171:5116.
Schott, E., H. L. Ploegh. 2002. Mouse MHC class I tetramers that are unable to bind to CD8 reveal the need for CD8 engagement in order to activate naive CD8 T cells. Eur. J. Immunol. 32:3425.
Moody, A. M., Y. Xiong, H. C. Chang, E. L. Reinherz. 2001. The CD8 co-receptor on double-positive thymocytes binds with differing affinities to the products of distinct class I MHC loci. Eur. J. Immunol. 31:2791
Daniels, M. A., L. Devine, J. D. Miller, J. M. Moser, A. E. Lukacher, J. D. Altman, P. Kavathas, K. A. Hogquist, S. C. Jameson. 2001. CD8 binding to MHC class I molecules is influenced by T cell maturation and glycosylation. Immunity 15:1051.
Tomonari, K., S. Spencer. 1990. Epitope-specific binding of CD8 regulates activation of T cells and induction of cytotoxicity. Int. Immunol. 2:1189.
Ravichandran, K. S., S. J. Burakoff. 1994. Evidence for differential intracellular signaling via CD4 and CD8 molecules. J. Exp. Med. 179:727.
Grebe, K. M., R. L. Clarke, T. A. Potter. 2004. Ligation of CD8 leads to apoptosis of thymocytes that have not undergone positive selection. Proc. Natl. Acad. Sci. USA 101:10410.(Philmore O. Holman, Eliza)