Introduction

MHC class I staining reagents

Different types of staining reagents

Fluorescently-labeled soluble MHC I-peptide complexes, often referred to as “tetramers”, are widely used to quantitate, isolate and characterize (e.g. phenotypic or TCR repertoire analysis) antigen-specific CD8+ (and CD4+) T cells (1, 2, 3, 4). They are prepared by enzymatic biotinylation of monomeric MHC I-peptide complexes containing a C-terminal biotinylation sequence peptide (BSP) and subsequent conjugation by reaction with phycoerythrin (PE) [or allophycocyanin (APC)] -streptavidin (1). Due to the large size of PE (and APC), their conjugates with streptavidin are heterogeneous in terms of stoichiometry and accessibility and orientation of the biotin binding sites; therefore such reagents are multimeric (5, 6).

Other types of MHC I-peptide staining reagents have been described, such as:

  1. Streptamers (7) and desthiobiotin (DTB) multimers (8), which have the same structural scaffolding, but allow rapid dissociation into monomer subunits, which permits sorting of antigen-specific cells in the absence of adverse cell activation.
  2. Pentameric complexes in which five monomeric MHC I-peptide complexes spontaneously assemble into pentamers by virtue of added peptidic zipper sequences. Pentamers are a trademark of PROIMMUNE and their production and properties have not been published. All these conjugates contain PE as fluorochrome and their staining performances are comparable (8, 9).
  3. Quantum dots loaded with MHC I-peptide complexes, which allow simultaneous use of multiple MHC I-peptide specificities in polychrome flow cytometry (10).
  4. Well-defined Cy5-labeled dimeric, tetrameric and octameric MHC I-peptide complexes prepared by site-specific alkylation of MHC I-peptide monomers containing a free cysteine at the α3 C-terminus (275C) with maleimide-containing linkers of variable length and flexibility (4, 5, 11, 12, 13). For long rigid linkers, polyproline linkers were used that assume a rigid proline II helix, in which one residue spans 3.12 Å, in aqueous media (12).
  5. Dextramers, a trademark of Immudex, consist of dextran conjugated with fluoresceine and MHC I-peptide complexes. These reagents have been reported to be suitable for staining of antigen-specific CD8+ T cells in sections (14).
  6. Dimeric (MHC I-peptide)-immunoglobulin (Ig) fusion proteins, which have been used for probing T cell membrane organization (15).

Although the number of different MHC I-peptide staining reagents is large, the conventional streptavidin-based tetramers remain those most commonly used (2). However, for specific applications some of these other reagents have their advantages.

Molecular basis of CD8+ T cell staining

Interactions of T cell antigen receptors (TCRs) with MHC I-peptide monomers are characterized by micromolar dissociation constants (KD) and half-lives in the range of seconds (16). However, on living cells, the coordinate binding of CD8 to TCR-associated MHC I-peptide complexes can considerably strengthen their binding, namely by decreasing the dissociation rate (17, 18). CD8 undergoes differentiation- and activation-dependent changes in the glycosylation and sialylation of its β chain, which can profoundly affect cognate and non-cognate MHC I-peptide binding (19, 20). Non-cognate CD8 binding to MHC I-peptide complexes has been reported to increase non-specific multimer binding and therefore multimers should be used that contain the CD8 binding weakening mutation A245V in the MHC α3 domain (6). We found that when working in the usual concentration range of 5 to 30 nM (2.5 to 15 µg/ml), with very few exceptions, non-specific tetramer staining is scant and that instead the A245V mutation can significantly decrease the staining of CD8-dependent T cells (4 and unpublished results).

Conjugation of MHC I-peptide monomers to oligomeric complexes, e.g. in conventional multimers, substantially increases the overall binding avidity and decreases the dissociation rate to half-lives in the order of hours (16, 21). This makes possible the detection of T cells bearing specific TCRs with such reagents by flow cytometry or their isolation by FACS (or MACS). The good staining usually obtained with conventional multimers is explained, at least in part, by their heterogeneity which allows different MHC I-peptide species to preferentially bind to different subsets of cells (4).

Special applications of CD8+ T cell staining

Binding of soluble oligomeric MHC I-peptide complexes at 37°C elicits T cell activation events, such as intracellular calcium mobilization, diverse tyrosine phosphorylation and endocytosis of MHC I-peptide engaged TCR/CD8 (5, 7, 8, 11, 12, 21). Especially for effector CTLs, MHC I-peptide complex driven cell activation can induce their death via FasL-dependent apoptosis or severe mitochondrial damage (11, 22). On one hand, this can cause a serious problem when FACS sorting or cloning antigen-specific cells, which can be circumvented by using staining reagents that are reversible (11, 22) or that fail to activate CD8+ T cells (11, 12, 13). On the other hand, cell death inducing MHC I-peptide complexes can be deliberately used to eradicate antigen-specific CD8+ CTLs (11, 22).

Another special application is to gauge the state of activation or differentiation of CD8+ T cells by means of MHC I-peptide complexes, such as Ig-based dimers, or dimers, tetramers and octamers containing linkers of defined length and flexibility (4, 11, 12, 13, 15). In contrast to heterogeneous multimers, binding studies with such defined complexes can reveal differentiation- and activation-dependent differences of the cells under study (4, 15). This is explained mainly by differentiation- and activation-dependent changes in glycosylation and sialylation of T cell surface molecules involved in antigen recognition, e.g. by affecting CD8 participation in MHC I-peptide binding and/or aggregation of TCR and CD8 (4, 12, 19, 23).

When analyzing populations, especially when these have low cell numbers, it is of interest to combine multimer staining with intracellular cytokine staining. To this end, the cells need to be stimulated with cognate peptide to induce cytokine production; this results in TCR (and CD8) down modulation, which reduces multimer staining. To circumvent this, multimer staining should be performed first and at 37°C, which allows endocytosis of multimers (24).

MHC I-peptide multimers containing mutations in the MHC α3 domain that ablate CD8 binding (D227K, T228A in human and D227K, Q226A in mouse) can be used to quantitate and to sort CD8-independent T cells, which typically express high affinity TCRs (4, 25, 26). CD8+ T cells specific for tumor antigens, namely differentiation antigens (e.g. MELAN-A/Mart-1, gp100, or tyrosinase), tend to express low affinity TCRs. CD8 binding-deficient multimers can be used to sort infrequent CD8+ T cells expressing high affinity TCRs. Such cells have been shown to efficiently kill tumor cells (26). Moreover, CD8 binding-deficient multimers have been reported to selectively induce FasL (CD95L) expression, resulting in apoptosis of antigen-specific CTLs (22).

Practical notes on staining of antigen-specific CD8+ T cells – What staining conditions are best used?

At 37°C MHC I-peptide multimers (and other complexes) can trigger events that affect staining of CD8+ T cells, such as cell death or TCR/CD8 down modulation (8, 11, 21, 22, 24). Conversely, multimer binding in the cold (0-4°C), where membranes are solidified, tends to be slow. Our preferred staining conditions therefore are at ambient temperature in the presence of EDTA (5 mM) and sodium azide (0.02%) to inhibit cell activation. Under these conditions multimer binding is rapid and, after 30 min, steady-state binding is reached. Importantly, as the multimer concentration giving maximal binding can vary considerably, it is crucial to test various concentrations in the range of 5-50 nM (2.5-25 µg/ml). Ideally binding isotherm should be assessed, from which critical binding parameters can be assessed, such as the dissociation constant (KD) and the maximal level of binding (Bmax). In order to assess non-specific background staining, corresponding irrelevant MHC I-peptide complexes must be included in each staining experiment. In addition, it should be noted that anti-CD8 antibodies can have profound and diverse effects on MHC I-peptide complex staining of CD8+ T cells (17, 18, 27). Therefore when counterstaining of CD8 is used, the multimer staining should precede CD8 staining and anti-CD8 antibodies should be used that do not inhibit multimer binding (27).

There are several tricks that can be used to increase MHC I-peptide multimer staining which can be useful, especially when the avidity or the frequency of the T cells is low. For example, staining can be increased by inhibiting TCR down modulation with the protein kinase inhibitor dasatinhib or when anti-CD8 antibodies are used that increase MHC I-peptide staining (21, 27, 28). Moreover, when the frequency of antigen-specific cells is low (<0.1%), scarce antigen-specific cells can be enriched e.g. by MACS using staining with conventional multimers followed by incubation with magnetic beads coated with anti-PE antibody (29).

MHC class II staining reagents

MHC II-peptide tetramers and the ins and outs of their applications for the detection of antigen-specific CD4+ T cells have been reviewed previously (30, 31, 32). In the following we briefly describe observations that we found to be of special importance.

Production and types of MHC II–peptide complexes

While MHC I-peptide complexes are obtained by refolding with peptides of interest, soluble recombinant MHC class II proteins are usually produced by insect expression systems, such as Drosophila S2 cells or baculovirus and sf9 cells (30, 31). With very few exceptions, deletion of the transmembrane (TM) domains of the α and β chains results in the dissociation of the two sub-unit chains. Chain pairing can be re-established by addition of leucine zippers. For “tetramer” formation, a BSP sequence is added after the leucine zipper (after the acidic zipper in our reagents) and enzymatic biotinylation and tetramerization is performed as for MHC I-peptide multimers (29, 30, 31). If the MHC molecule is sufficiently stable without peptide cargo (e.g. HLA DRB1*0101 or DRB1*0401), it can be loaded after purification with the peptides of interest. The efficiency of peptide loading strongly depends on its binding strength to the respective MHC II molecule. If the binding is below a critical threshold, peptide loading is inefficient and the resulting complexes are of limited stability, both physically and conformationally.

If this strategy is not feasible, peptides can be tethered to the N-terminus of the β chain via a flexible linker (29, 30, 31). This strategy works fine for some, but not all, complexes. Also, although the peptide is a part of the molecule, in the case of weak binding peptides it may not bind or may bind incorrectly in the peptide binding groove.

Difficulties in staining CD4+ T cells with MHC II-peptide multimers

Application of the same staining strategy to CD4+ T cells with MHC II-peptide multimers is often less satisfactory compared to the MHC class I system. Frequently the staining obtained is faint or not detectable and the frequency of stained cells ex vivo very low, usually necessitating prior in vitro peptide stimulation to allow conclusive detection. Several factors contribute to this. First, the low frequency of antigen-specific CD4+ T cells observed ex vivo by multimer staining is explained, at least in part, by the fact that CD4 epitopes are generated and presented as different truncates and that CD4+ T cells differentially recognize such truncates (33). Moreover, peptide binding to MHC class II molecules tends to be degenerate, i.e. given peptides can bind to more than one MHC molecule (34). This, together with the vast polymorphism of HLA class II molecules, explains why staining with MHC II multimers containing one given peptide and one given MHC class II molecule detect only a fraction of all CD4+ T cells specific for an antigenic sequence. This also explains divergences between MHC II-peptide multimer staining and functional responses, e.g. intracellular cytokine staining. To circumvent this, we recommend that functional assays, in which the peptide of interest is used on an antigen-presenting cell that expresses only the MHC II molecule of interest, be performed in parallel to staining experiments. Second, while CD8 greatly strengthens MHC-peptide binding to CD8+ T cells, CD4 does not (35). This is a major reason why higher multimers concentrations are typically needed for staining of CD4+ T cells than for CD8+ T cells. Third, while MHC I-peptide complexes obtained by refolding are homogeneous and conformationally uniform, MHC II-peptide complexes obtained by peptide loading of “empty” MHC II proteins or containing tethered-on peptides often are not, which can impact MHC II multimer staining (36).

Optimal MHC II-peptide multimer staining conditions

In most, but not all, cases optimal staining of CD4+ T cells with multimers is obtained upon incubations at 37°C for extended periods of time (30-120 min) (28, 37). Under these conditions, cognate MHC II complexes binding to TCR (and CD4) are internalized and accumulate over time. Moreover, as for the MHC class I system tetramers, certain anti-coreceptor antibodies can increase the staining efficiency (21, 27, 28). Importantly, because the binding avidity of MHC II-peptide multimers on CD4+ T cells is often low, it is crucial to test different (including high) concentrations (e.g. 5-100 nM, i.e. 2.5-50 µg/ml) and to use suitable irrelevant multimers to assess non-specific staining. Moreover, we have often observed a substantial increase in multimer staining upon pre-treatment of the cells with neuraminidase (from Roche Ltd., Basel, Switzerland; 0.03 µ/ml for 30 min at 37°C).

MHC II-peptide multimer based epitope mapping

The availability of “empty” MHC II molecules (e.g. molecules containing no peptide cargo), which can be loaded with peptides of interest, has been exploited for the mapping of epitopes of given antigens (38). It should be noted, however, that epitopes, i.e. peptides that bind poorly to given MHC II molecules, may be missed this way due to inadequate peptide loading.