By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Due to essential maintenance the subscribe/renew pages will be unavailable on Wednesday 26 October between 02:00- 08:00 BST/ 09:00 – 15:00 SGT/ 21:00- 03:00 EDT. Apologies for the inconvenience.
J. L. Nayak, Department of Pediatrics, University of Rochester Medical Center, 601 Elmwood Avenue, Box 690, Rochester, NY 14642, USA. Email: email@example.com Senior author: Jennifer L. Nayak
An understanding of factors controlling CD4 T-cell immunodominance is needed to pursue CD4 T-cell epitope-driven vaccine design, yet our understanding of this in humans is limited by the complexity of potential MHC class II molecule expression. In the studies described here, we took advantage of genetically restricted, well-defined mouse strains to better understand the effect of increasing MHC class II molecule diversity on the CD4 T-cell repertoire and the resulting anti-influenza immunodominance hierarchy. Interferon-γ ELISPOT assays were implemented to directly quantify CD4 T-cell responses to I-Ab and I-As restricted peptide epitopes following primary influenza virus infection in parental and F1 hybrid strains. We found striking and asymmetric declines in the magnitude of many peptide-specific responses in F1 animals. These declines could not be accounted for by the lower surface density of MHC class II on the cell or by antigen-presenting cells failing to stimulate T cells with lower avidity T-cell receptors. Given the large diversity of MHC class II expressed in humans, these findings have important implications for the rational design of peptide-based vaccines that are based on the premise that CD4 T-cell epitope specificity can be predicted by a simple cataloguing of an individual’s MHC class II genotype.
Endosomal processing of bacterial and viral pathogens results in the production of a large number of peptides, of which only a subset are able to bind to MHC class II molecules expressed in the host and initiate a CD4 T-cell-mediated immune response. This phenomenon, known as immunodominance, has been an important area of research focus as many seek to understand the factors underlying preferences in T-cell specificity. The human MHC genotype is complex, with co-expression of the HLA-DR, DQ and DP isotypes, potential expression of up to two different DR β-chains simultaneously, and heterozygosity at most loci leading to the possible simultaneous expression of up to 12 distinct MHC class II molecules on the cell surface.1,2 For this reason, most published work addressing immunodominance has used mouse models, both because of the availability of inbred strains that allow MHC restriction to be easily defined and the relative ease of performing comprehensive experiments using a reproducible system.3–6 One system that can be used to better understand the effect of increasing MHC complexity on the immunodominance hierarchy is the F1 mouse model. These mice, generated through the crossing of two different homozygous inbred strains, double MHC class II expression when compared with the parental strains while still avoiding the unmanageable complexity present in the human population.
There are several factors that could alter the immunodominance hierarchy when MHC class II diversity is increased. First, gene dosage effects in heterozygous individuals could lead to a lower surface density of individual MHC class II molecules expressed on the cell surface. Second, there could be alterations in positive and negative selection because of the presence of additional MHC class II molecules. Third, there might be increased competition among T cells because of a larger number of epitopes available for recognition. Finally, there might be preferences in MHC assembly or interactions with machinery involved in peptide acquisition by MHC class II, such as invariant chain or HLA-DM, that result in preferences for one MHC molecule over another. These factors could have many effects on the CD4 T-cell response and could lead to alterations in immunodominance, either individually or in combination.
Although there have been previous attempts to understand the effect of increasing MHC genotype complexity on T-cell immunodominance hierarchies, most of these studies have focused on CD8 T cells. Some of these studies described a broadening of the CD8 T-cell repertoire to include representation of epitopes from both parental strains in diminished abundance,5,7,8 whereas others found more dramatic changes in the CD8 T-cell immunodominance hierarchy following diversification of MHC class I expression.9–12 Few studies have examined the effect of increasing MHC class II diversity on the CD4 T-cell response, with much of the existing data based on the study of T-cell proliferative responses following immunization with protein antigens where the CD4 T-cell repertoire is heavily compromised by homology with self.13–15 There are several aspects of MHC class II antigen presentation that may further complicate the effects of added genetic complexity on the specificity of the CD4 T-cell response. Unlike MHC class I molecules, MHC class II molecules are composed of two polymorphic chains. Hence, in F1 animals there may be expression of novel MHC class II molecules in addition to the parental molecules as the result of cross-allelic α-chain and β-chain pairing. Although this cross-allelic pairing is known to occur in both mice and humans,16–21 the extent to which these molecules participate in CD4 T-cell responses has been debated.22,23 Additionally, the effect of increasing MHC class II diversity could vary depending on the type of antigen being presented and the route by which it is being given. CD4 T-cell responses initiated by immunization with protein antigens tend to be focused on a small number of epitopes that bind to MHC class II with high affinity because of antigen-presenting cells (APC) only taking up a limited amount of antigen and the editing effects of the HLA-DM molecule.3,4,24–30 When there is more abundant antigen, such as with active infection of APC, the number of specificities elicited could be much greater.3
In the studies described here, we examined the consequences of expanded MHC class II protein diversity on the primary CD4 T-cell repertoire and immunodominance hierarchy following infection with influenza virus, a complex pathogen responsible for yearly epidemics and occasional pandemics of disease. Our previous studies have shown that the primary immune response to influenza virus can include from 20 to > 80 different CD4 T-cell specificities, depending on the strain of mouse and the alleles of MHC class II expressed.31–33 In all cases this diversity far exceeds that of any of the model systems used to examine the effects of increasing MHC class II diversity to date and is much more relevant to our understanding of CD4 T-cell immunity to complex antigens and vaccines in humans. Overall, our results demonstrate that CD4 T-cell recognition of antigens depends not only upon the expression of a given restriction element, but also on the array of class II molecules expressed in the host. Current efforts in vaccine design, particularly peptide-based strategies, combine advances in epitope discovery with sophisticated bioinformatic and computational strategies to predict peptides that will be likely to elicit an immune response in large portions of a population.34–38 Our results suggest that not only the presence or absence of a given restriction element, but also an individual’s complete MHC class II genotype, may need to be considered as we move towards the rational design of peptide-based vaccines against pathogens such as influenza virus.
Materials and methods
Influenza virus production
Influenza virus was produced in the allantoic cavity of embryonated eggs as described previously.32 Embryonated eggs from SPAFAS Inc. (North Franklin, CT) were incubated at 21° and 100% humidity for 9 days and then infected with 100 μl of the A/New Caledonia/20/99 human influenza virus (H1N1) at 103 50% egg infective doses (EID50) per millilitre. The eggs were incubated at 37° for 48 hr and for 24 hr at 4° then the allantoic fluid was harvested under sterile conditions. Following centrifugation at 2000 g for 20 min at 4°, aliquots were frozen at −70°. Virus titre was determined by infecting embryonated eggs with serial dilutions of the virus and harvesting and immediately titrating the resulting allantoic fluid using a haemagglutination assay of chicken red blood cells according to the procedure recommended by the 2005–2006 World Health Organization bulletin for the identification of influenza virus isolates from subjects with influenza.
C57BL/10J mice (referred to as B10; H-2b) were purchased from Jackson Laboratory (Bar Harbor, ME), B10.S-H2S/SqMcdJ mice (referred to as B10.S; H-2s) were originally purchased from Jackson Laboratory and bred at the University of Rochester, and (B10 × B10.S) F1 mice (H-2b×s) were bred at the University of Rochester. All animals were housed in specific pathogen-free facilities and maintained according to institutional guidelines. All animal protocols used in this study adhere to the AAALAC, International, the Animal Welfare Act and the PHS Guide. All protocols were approved by the University of Rochester Committee on Animal Resources; Animal Welfare Assurance Number A3291-01. The protocols under which these studies were conducted were approved on 4 March 2006 (protocol no. 2006–030) and 10 April 2008 (protocol no. 2008–023).
Seventeen-mer peptides overlapping by 11 amino acids to encompass the entire sequences of the haemagglutinin (HA) and neuraminidase (NA) proteins from the A/New Caledonia/20/99 influenza virus (H1N1), the non-structural protein 1 (NS1) sequence from the A/New York/444/2001 influenza virus (H1N1; > 99% conserved compared with A/New Caledonia/20/99), and the nucleoprotein (NP), matrix protein 1 (M1), polymerase acidic protein (PA), and polymerase basic 1 protein (PB1) sequences from A/New York/348/2003 influenza virus (H1N1; > 99% conserved compared to A/New Caledonia/20/99) were used. The following reagents were obtained through the NIH Biodefense and Emerging Infections Research Resources Repository, National Institute of Allergy and Infectious Diseases, National Institutes of Health: Peptide Arrays, Influenza Virus A/New Caledonia/20/1999 (H1N1) HA protein, NR-2602; and NA protein, NR-2606; Peptide Array, Influenza Virus A/New York/444/2001 (H1N1) NS1, NR-2612; and Peptide Array, Influenza Virus A/New York/348/2003 (H1N1) NP, NR-2611; M1, NR-2613; PB1, NR-2617; and PA, NR-2618. The peptides were reconstituted to 10 mm in PBS with added dimethyl sulphoxide for hydrophobic peptides and 1 mm dithiothreitol for cysteine-containing peptides. Stocks were stored at −20°, and working stocks at concentrations of 1 mm were prepared in Dulbecco’s modified Eagle’s medium (Invitrogen Corp., Carlsbad, CA), filter sterilized, and also stored at −20°. The final concentrations of individual peptides used in the ELISPOT assays were either 2 μm (pools containing all peptides for a given protein) or 10 μm (single peptide assays).
Mice that were 2–6 months of age were anaesthetized by intraperitoneal injection of 200–300 μl per mouse of a 20 mg/ml solution of tribromoethanol (Avertin; provided by Dr. D. Topham at the University of Rochester). They were then infected intranasally with A/New Caledonia/20/99 influenza (H1N1) at a dose of 50 000 EID50 per mouse in 30 μl PBS, as previously described.32 Ten to 15 days after infection the mice were euthanized and spleens and mediastinal (draining) lymph nodes were excised and used as a source of CD4 T cells for ELISPOT analysis. Syngeneic splenocytes from uninfected mice were used as a source of APC unless otherwise noted.
Cells were depleted of B cells, CD8-positive cells and class II-positive cells either by antibody binding and complement-mediated lysis as previously described32,33 or by negative selection using MACS no-touch CD4 purification (Miltenyi Biotec, Gladbach, Germany) as per the manufacturer’s instructions. Cell lines that produced the antibody used for complement-mediated lysis were obtained from the American Type Culture Collection and included 3.155 (anti-CD8), RA3/3A1/6.1 (anti-B220), M5/114 (anti-I-Ab-expressing cells), and 10.2.16 (anti-I-As-expressing cells). Syngeneic splenocytes from uninfected mice were depleted of T cells using antibody-mediated complement depletion with anti-Thy-1.2 Ab (J1j.10). The purity of the resulting cell populations was assessed by analytical flow cytometry by staining for expression of the CD4 and CD8 cell surface markers.
ELISPOT assays were performed as previously described.32,33 Ninety-six-well ELISPOT plates (Millipore, Billerica, MA) were coated with 50 μl purified rat anti-mouse interferon-γ (IFN-γ; clone AN-18; BD Biosciences, San Jose, CA) in PBS at room temperature for 2 hr or overnight at 4°. The antibody was removed and plates were washed and incubated with cell culture media to block non-specific binding. A total of 500 000 syngeneic T-cell-depleted APC were co-cultured with 100 000–400 000 CD4-enriched T cells and peptide at a concentration of either 2 or 10 μm, and incubated at 37° in 5% CO2 for 16–18 hr. The cells were then removed and the plates were washed with PBS containing 0·1% Tween-20 (ELISPOT wash buffer) and processed as previously described.32,33 Plates were analysed on an Immunospot reader series 2A using Immunospot software, version 3.2 (Cellular Technology Ltd., Cleveland, OH). Data were calculated and presented as cytokine ELISPOTs per million CD4 T cells, with background values subtracted and the number of spots adjusted to account for the purity of the CD4 T-cell population as determined by analytical flow cytometry. The range in negative values obtained was similar to that seen with irrelevant peptides and was typically < 20 spots per well.
Purity after CD4 T-cell enrichment was assessed by flow cytometry by staining with CD4-FITC (clone RM4-4; BD Biosciences) and CD8a-FITC (Ly-2 clone 53–6.7; eBiosciences, San Diego, CA). Data were analysed using Cell Quest software (Becton Dickinson, San Jose, CA). No contaminating CD8 T cells were detectable, but the purity of CD4 T cells varied depending on both the mouse strain and the purification method. To account for this, samples were normalized by dividing the number of cytokine ELISPOTs obtained per million cells by the percentage of CD4 T cells present in each sample after purification. The expression of MHC class II present on each strain was assessed by staining with -FITC (clone 25-9-17; BD Biosciences), -FITC (clone 10-3.6; BD Biosciences; cross-reacts with I-As), CD19-PE (clone 1D3; BD Biosciences), CD3ε-PE-Cy7 (clone 145-2C11; eBiosciences), and CD11c-APC (clone HL3; BD Biosciences). Data were analysed using flowjo software (Tree Star, Ashland, OR).
The major goal of our studies was to examine the consequences of doubling MHC class II genetic diversity on the CD4 T-cell repertoire and immunodominance hierarchy following primary infection with influenza virus, a complex viral pathogen. One of the simplest possibilities was that the response would expand to include specificities from both parental strains. In this situation, the overall pattern of I-Ab and I-As restricted responses in F1 animals would be preserved when compared with the parental strains. If resources and space were limiting, the broadening of responses to include both parental types might be accompanied by a decrease in the magnitude of all specificities as the result of competition for resources limiting the number of effector cells able to be supported during clonal expansion.39,40
To examine the above possibility, B10, B10.S and (B10 × B10.S) F1 mice were infected in parallel with A/New Caledonia/20/99 influenza virus, a virus that was originally isolated from humans but has been used extensively by our laboratory in the study of the immune response to influenza using a murine model.31–33 Ten to 15 days following infection (when immunodominance hierarchies were stable), CD4 T cells were isolated from the spleen or mediastinal lymph node by either antibody binding and complement-mediated lysis or by negative selection using MACS no touch CD4 T-cell purification. The resulting CD4 T-cell-enriched populations were used in IFN-γ ELISPOT assays with added splenocytes from naive mice as APC. CD4 T-cell specificity was assessed by restimulating CD4 T cells with synthetic peptides representing known I-Ab and I-As restricted influenza peptide epitopes derived from the HA, NA, NS1, NP and M1 proteins.33 Whereas most of the epitopes studied were restricted to only one of the two MHC class II molecules present on the surface of F1 cells, two of the epitopes (NP 45 and NP 47) were restricted to both parental MHC class II molecules.
Strikingly, rather than seeing the response broadened by a new distribution across both parental types, we instead saw dramatic and unexpected changes in CD4 T-cell response magnitude when individual peptide epitopes were examined in F1 mice. Figure 1 shows the average IFN-γ response per 106 CD4 T cells to all epitopes tested in the B10 (Fig. 1a), B10.S (Fig. 1b), and F1 (Fig. 1c) mouse strains, with each peptide epitope tested in parallel in multiple experiments. Decreases in the magnitude of the response to specific peptide epitopes affected both dominant and subdominant responses, although not all specificities were affected equally. The decrease in the F1 CD4 T-cell response to individual peptide epitopes is more easily visualized in Fig. 2, where the F1 response is shown as a fraction of the parental response for each epitope (calculated by the formula: F1 response/parental response × 100). There were remarkable disparities in the preservation of responses in F1 mice. Interestingly, responses to most of the HA-specific peptide epitopes, including major epitopes such as HA 22, were decreased to < 20% of that seen in the parental strain. Responses to epitopes derived from other influenza proteins were also significantly diminished, including the responses to NP 16 and 17 (30% and 22% of the parental responses, respectively), NP 74 and 75 (around 30% of the parental responses), NA 30 and 31 (decreased to around 15% and 25% of the parental response, respectively), NA 47 (decreased to around 15% of the parental response), and NA 77 (decreased to around 22% of the parental response). Although the majority of responses were significantly diminished, a small number of both dominant and subdominant responses were well preserved in the F1 animal, including the dominant responses to NP 44 and NP 45 and the subdominant responses to NS1 13, NS1 14 and NA 20. On average, F1 CD4 T-cell responses were decreased to about 35% of the respective parental response.
If one MHC class II molecule was highly favoured during thymic selection or peripheral CD4 T-cell activation, CD4 T-cell responses restricted to that MHC molecule might dominate the F1 response. Although responses restricted to both the I-Ab and I-As molecules could be identified (Fig. 1), this did not rule out the possibility of a more subtle allelic preference. To further evaluate the presence of such a preference, responses to peptides presented by the two different class II molecules were quantified in both the parental and F1 strains. These responses were then summed and the parental and F1 strains were compared (Fig. 3). When the number of responding CD4 T cells was compared in the parental strains, between 2 and 2·5 times more influenza-specific CD4 T cells were elicited in B10.S mice than in B10 mice (Fig. 3a). When the distribution of the anti-influenza CD4 T-cell repertoire restricted to the I-As and I-Ab MHC class II molecules was examined in F1 animals this ratio was relatively preserved, indicating that a strong allelic preference was not present (Fig. 3b).
To quantify the overall diversity of the response in F1 and parental animals, the total number of epitopes producing a response of on average > 40 IFN-γ spots per 106 CD4 T cells was summed in each strain. This cut-off was chosen because it produced results that were reliably above background. Using this method, there were 17 epitopes in the B10 strain, 24 in the B10.S strain, and 23 in the F1 strain. Hence, despite the greater number of potential epitopes in F1 animals, the overall response diversity was similar (Fig. 4).
Given the possible existence of hybrid MHC class II molecules ( or ), we wanted to test for the presence of new peptide epitopes unique to the F1 animal. All peptides not recognized by either of the parental strains were pooled into groups that included between eight and 12 peptides, and these pools of peptide were then used to stimulate CD4 T cells in IFN-γ ELISPOT assays with cells isolated from B10, B10.S and (B10 × B10.S) F1 mice. When tested in this manner, one NP pool consistently produced a strong response in F1 mice with a variable response in the B10 parental strain and a response at the level of background in the B10.S parental strain. Individual peptides in this pool were then examined to see if they were unique to the F1 strain. Figure 5 shows the number of IFN-γ ELISPOTs elicited per 106 CD4 T cells using cells derived from either the B10 parental strain or the F1 strain when restimulated with each of the individual peptides that comprised this NP pool. In this enhanced search, two overlapping epitopes that may be preferentially presented by a hybrid MHC molecule were identified.
As published data suggested that there was increased heterogeneity among individual F1 animals,13 we next looked at the CD4 T-cell responses in individual B10, B10.S and F1 mice to determine if there was a greater degree of mouse to mouse variability present in the F1 strain. CD4 T cells were purified from individual B10, B10.S and F1 spleens and used in IFN-γ ELISPOT assays as previously described, with the response recruited by each peptide epitope presented as a percentage of the total measured response by dividing by the sum of all peptides tested in that individual mouse. As shown in Fig. 6, there was a similar amount of variability between all the strains when the pattern among individual animals was assessed. Further, a statistically significant difference in the magnitude of the CD4 T-cell response to most of the epitopes tested was present when comparing the F1 and parental animals.
As discussed previously, HA-specific responses appeared to be more compromised in F1 animals than responses devoted to the other proteins tested. To further explore this, CD4 T-cell reactivity in all three strains of mice was tested using pools of peptides encompassing the entire translated sequences of the PA, PB1, NP, NS1, M1, NA and HA viral proteins in IFN-γ ELISPOT assays using cells from pooled B10, B10.S and F1 animals as before. Although peptides from the polymerase basic 2 protein were not included in these studies, this protein accounted for < 4% of the parental I-Ab or I-As restricted influenza-specific T-cell response in preliminary epitope mapping studies, making it unlikely to account for a significant portion of the F1 response. The bar graphs in Fig. 7(a) show the absolute reactivity to the peptide pools in each strain and the pie graphs depict the reactivity to the pools in B10 (Fig. 7b), B10.S (Fig. 7c) and F1 (Fig. 7d) mice as a percentage of the total response. In agreement with our studies using individual peptides, a greater loss of reactivity to peptide epitopes within the HA protein was seen when F1 mice were compared to B10.S mice. Responses devoted to the NA, PA and PB1 proteins were also relatively compromised in the F1 strain, while the responses devoted to NP and NS1 remained relatively preserved. These findings suggest that there could be a characteristic of certain influenza proteins that results in a more impaired response to epitopes within these proteins in the F1 animal. Interestingly, the total influenza-specific response in F1 mice was consistently found to be between the responses seen in the two parental strains. This argues for the presence of strain-specific ceilings in the CD4 T-cell response generated following intranasal influenza infection.
There were several potential mechanisms that could account for the diminished responses to many peptide epitopes in F1 mice. One possible explanation was that the lower density of each MHC class II molecule on the surface of the F1 APC resulted in activation of fewer CD4 T cells in ELISPOT assays. To evaluate the surface expression of MHC class II, parental and F1 splenocytes were stained for and chain surface expression and the mean fluorescent intensity was compared. This staining revealed that both MHC class II molecules were decreased on CD11c-positive and CD19-positive F1 APC, as expected, with a greater decrease in the expression of the chain seen (data not shown). To examine whether this lower density of MHC class II resulted in the activation of fewer CD4 T cells, we next took CD4 T cells obtained from influenza-infected F1 mice and tested them in a peptide-specific IFN-γ ELISPOT assay with either parental or F1 APC. Figure 8(a) shows representative data using a fixed concentration of peptide, whereas Fig. 8(b) shows dose–response curves comparing the ability of the F1 or parental APC to restimulate F1 CD4 T cells. These experiments revealed little apparent difference in the ability of F1 and parental APC to activate F1 CD4 T cells ex vivo in an ELISPOT assay, indicating that MHC class II gene dosage effects during the restimulation culture did not account for the diminished F1 CD4 T-cell responses seen.
Another possibility was that the lower surface density of each MHC class II molecule in the host influenced the efficiency of priming in F1 mice, resulting in only CD4 T cells of higher avidity being successfully recruited into the CD4 T-cell response. To examine this possibility, the functional avidity of the T cells from F1 and parental mice was tested using a peptide dose–response assay. The results were plotted as a percentage of the maximal response obtained for each peptide in each strain and the approximate concentration of peptide needed to activate 50% of the responding CD4 T cells (EC50) was estimated. As shown in Fig. 9, there was no detectable increase in the avidity of F1 CD4 T cells when compared with parental CD4 T cells for any of the epitopes tested across three individual experiments. Hence, there was no evidence that the diminished responses to some epitopes in the F1 strain were the result of changes in the functional avidity of T cells, either during priming or during restimulation.
Based on the experiments performed, we concluded that primary anti-influenza CD4 T-cell responses to many peptide epitopes in (B10 × B10.S) F1 mice were asymmetrically decreased when compared with the homozygous parental responses, resulting in shifts in the F1 immunodominance hierarchy. The overall response diversity in the F1 strain remained similar to that seen in the parental strains despite the greater number of potential epitopes due to many responses decreasing to the point of becoming undetectable in F1 animals. The smaller number of responding F1 CD4 T cells could not be accounted for solely by the lower MHC class II surface density on the F1 APC or by a failure to activate CD4 T cells with a lower avidity T-cell receptor (TCR) in vivo. Instead, we postulate that our observations may be the result of alterations in CD4 T-cell precursor frequencies because of differences in thymic selection in F1 animals.
Although decreased cell surface density of MHC class II has been shown to affect the magnitude of the CD4 T-cell response, this has been seen mainly in the setting of more severely reduced MHC class II cell surface expression. The effect of lowering the cell surface density of MHC class II was examined using H-2s/s mice carrying different copy numbers of an transgene, such that surface expression of the endogenous molecule was reduced by varying amounts. In this system, decreasing the expression of MHC molecules was shown to diminish antigen presentation and have a modest impact on thymic positive selection, but both of these changes were only present when MHC class II surface expression was severely reduced.41 One study has shown that competition between different MHC alleles can affect the cell surface expression of MHC class I molecules, with allele-specific expression of different molecules varying from the 50% predicted by gene dosage,42 but the decrease in MHC class II present in (B10 × B10.S) F1 mice may not have been to a level sufficient to effect these changes. Although competition for peptides could also theoretically influence the repertoire of peptides presented in F1 mice, relatively few peptide epitopes were shared in the experimental system that we used, and even among closely related MHC class II molecules little difference has been seen when comparing peptides isolated from F1 animals with those present in the parental strain.43
There is some support for the idea that an additional expressed MHC class I molecule can have an effect on positive and negative selection, leading to changes in precursor frequency and corresponding changes in the immunodominance pattern of CD8 T cells. When CD8 T-cell receptor clonotypes were examined in parental and F1 mice following infection with respiratory syncytial virus, there was a loss of several of the parental Vβ subtypes in the subdominant DbM187–195 response in an F1 hybrid that may have contributed to alterations in that epitope’s precursor frequency.44 Alterations in precursor frequency have been documented using tetramer enrichment in studies examining the CD8 T-cell immunodominance pattern in F1 mice following vaccinia virus or influenza virus infection, with several studies postulating that this mechanism may be at least partially responsible for the immunodominance alterations seen.9–11 Whereas in some situations this could be accounted for by deletion of a specific TCR because of cross-reactivity with self, in F1 mice T cells with a greater range of MHC restriction and peptide specificity must be accommodated without large changes in overall T-cell numbers. One way in which this could be accomplished is if the lower levels of MHC cell surface expression in F1 mice result in reduced positive selection of some TCRs,45 possibly through a quantitative reduction in TCR signalling.46 Our results did not show the overall increase in TCR affinity in F1 animals that would be expected if this was the case, but it is possible that the peptide dose–response curves generated were not sufficiently sensitive to detect subtle changes. The relative contribution of these two possibilities in (B10 × B10.S) F1 mice could be explored in more depth using the technique of tetramer enrichment described by Moon et al.47 to determine the precursor frequency of specific CD4 T-cell epitopes, with detailed characterization of TCRVβ usage and single cell CDR3β sequence analysis to further characterize the potential role of changes in positive and negative selection in the immunodominance alterations described.
Given the complexity of the human MHC class II genotype, MHC class II related effects could play a very important role in epitope-specific immune responses in the human population. In the F1 mouse system that we used, only a single restriction element was added. This can be contrasted with the human population, where up to 12 different MHC class II molecules can be expressed on the cell surface simultaneously, with the large amount of allelic diversity making it rare for two people to express the exact same MHC class II phenotype.1 An influence of HLA type on human CD8 T-cell responses has been documented. Boon et al.48 demonstrated a lower frequency of cytotoxic T-lymphocyte responses specific for the HLA-B8-restricted epitope NP380–388 in HLA-B27-positive donors and a higher frequency of HLA-A1-restricted epitope NP44–52 responses in HLA-A1-, -A2-, -B8-, and -B35-positive donors than in other donors. These observations suggest that the HLA phenotype of an individual can affect the specificity and magnitude of the cytotoxic T-lymphocyte response generated. However, in certain situations, the TCR repertoire may have sufficient versatility to ensure that deletion of prominent TCR clonotypes does not severely compromise the magnitude of a given T-cell response.49
If the presence of other MHC class II molecules influences the magnitude and specificity of the CD4 T-cell response, as our study suggests, a CD4 T-cell epitope present in one individual could have greatly decreased expression, or even no expression, despite preservation of that epitope’s restriction element in individuals with a different HLA genotype. Current bioinformatic approaches to vaccine design attempt to predict population-based CD4 T-cell responses to an identified peptide epitope in silico based on the binding of that epitope to given MHC class II molecules.35–38 Our results suggest that this method may have limitations, as CD4 T-cell reactivity may not always be accurately predicted by simply determining the MHC class II molecules expressed in a given individual. Hence, when attempting to rationally design peptide-based vaccines, CD4 T-cell reactivity to a potential epitope should first be confirmed in individuals expressing a diverse array of MHC class II molecules that includes the restriction element in question. It is likely that the epitopes that will be most useful for peptide-based vaccines will be those whose immunogenicity is stable in the presence of other restriction elements, making responses relatively preserved in the majority of individuals sharing expression of a particular MHC class II molecule.
Tetramer studies have the potential to further our understanding of the effects of MHC class II diversity in the human immune response. Although precursor frequencies to common infections such as seasonal influenza virus could be determined using this method, these results would be heavily affected by an individual’s infection and vaccination history. To better understand the influence of HLA phenotype on precursor frequency independent of exposure history, epitopes within pathogens to which the majority of a given population is expected to be naive (such as malaria or avian influenza in non-endemic regions) could be used. Tetramers could also be exploited to determine whether there are preferences in MHC class II restriction either within alleles or across different isotopes and help to determine which epitopes are stable in the setting of diverse MHC genotypes. The pursuit of such studies will help to elucidate predictable rules governing immunodominance in the setting of MHC diversification. If such rules exist, it will greatly increase our understanding of human immunology and help to make the rational design of peptide-based vaccines in outbred human populations a reality.
The authors would like to thank Dr Frank Gigliotti, Scott Leddon and Katherine Richards for editorial suggestions during preparation of this manuscript. This work was supported by grant HHSN266200700008C and grants T32 AI007464 and 1K12HD068373-01 from the National Institutes of Health. The authors would also like to thank the Department of Pediatrics for financial support.
The authors have no financial conflicts of interest to disclose.