We have quantitated the major families of peptides from hen egg lysozyme (HEL) presented by MHC class II I-Ak molecules. One striking feature is that the four epitopes are presented at levels that differ by as much as 200- to 300-fold. In these studies, we describe the CD4+ T cell response to each epitope after immunization with several doses of hen egg lysozyme protein. Although fewer T cells were generated at lower doses, the surprising finding was the responses to all four peptides were maintained. The relative number of T cell clones to each of the four epitopes was influenced to a very limited degree by their levels of presentation at the lowest dose. In conclusion, under strong stimulatory conditions, there is not a direct relationship between levels of peptide presentation and the T cell responses.
We examine the relationship between levels of peptide presentation by the MHC class II I-Ak molecules and the magnitude and clonal specificity of the CD4+ T cell response after immunization. To examine whether the levels of peptide-MHC presentation correlate with the extent of the CD4+ T cell response, three important pieces of information are required: the first is information on the chemistry of the naturally presented epitope after processing of protein, so as to include all potential T cell epitopes; the second is on the direct biochemical quantitation of each peptide presented by APCs, and the third is on the direct measurement of the number of T cells responding to each epitope.
Our laboratory has examined the peptides presented by APCs expressing MHC class II I-Ak molecules after processing and presenting the model Ag hen egg lysozyme (HEL).2 These peptide families have been chemically isolated, sequenced by mass spectometry, and their levels of presentation have been established using a quantitative ELISA (1, 2). The importance of identifying the naturally presented sequence is illustrated by the 48–63 family of peptides, where 50% of the responding T cells require the tryptophans at positions 62 and 63 (3). These T cells would be missed if studies used the minimal epitope required for binding (epitope 52–61), or the tryptic fragment (fragment 46–61) (4). The 48–63 family of peptides is chemically dominant, in that it is presented at levels that are 60-fold higher than the 31–47 family, and greater than 200-fold higher than the 18–33 and 115–129 families (2, 5, 6). It is very important to stress that these ratios of presentation among the four peptide families are similar in all APCs tested to date, which includes B cells, macrophages, dendritic cells, and splenocytes and thymic APCs from HEL-transgenic mice. Although the ratios of presentation remain constant, the absolute amount of presentation varies between APCs, with dendritic cells presenting the highest amount of each. Finally, we have developed a sensitive limiting dilution analysis (LDA) that allows for the determination of the number and specificity of individual T cell clones responding to each epitope after immunization (7, 8).
A previous study revealed a surprising lack of correlation between epitope abundance and the T cell response upon immunization with high doses of HEL in both normal and HEL-transgenic mice (7). The expectation was that the T cell response to the chemically dominant 48–63 epitope would be immunodominant. The term immunodominant has been used to describe a strong T cell response to one epitope compared with others (9, 10), and there is both evidence for (11, 12, 13) and against (14, 15, 16, 17) the correlation between epitope abundance and immunodominance in the CD8+ T cell response. Examination of this relationship in the CD4+ response has proven to be a more difficult task, in part due to a lack of identification and quantitation of MHC class II epitopes and the difficulty in quantitating CD4+ T cell responses. We extend our previous studies by examining the CD4+ T cell response to the four well-characterized epitopes of HEL after immunization with several doses of HEL in CFA.
Materials and Methods
Mice and immunizations
B10.BR mice (ages 6–10 wk) were obtained from The Jackson Laboratory (Bar Harbor, ME) and bred in our facility. CBA/J mice (ages 6–8 wk) were also obtained from The Jackson Laboratory. Mice of both sexes were immunized in the hind footpads with 10 nmol (144 μg), 1 nmol, or 0.1 nmol of HEL emulsified in CFA containing Mycobacterium tuberculosis H37Ra (Difco, Detroit, MI).
Serum Ab assays
Serum was collected 7 days after mice were immunized with HEL in CFA, and the presence of Abs to HEL was determined by ELISA. Nunc Maxisorp ELISA plates (Nunc, Roskilde, Denmark) were precoated with 10 μg/ml HEL (Sigma-Aldrich, St. Louis, MO) protein in 0.1 M sodium bicarbonate buffer (pH 8.8) overnight at 4°C. Wells were washed with PBS-0.05% Tween 20, blocked with PBS-1% BSA for 1 h at room temperature, and washed three times. Serum was then added in serial dilutions and incubated at room temperature for 2 h and washed three times. After washing, 100 μl of secondary Ab was added to each well (goat anti-mouse IgG-peroxidase 1/1500) in PBS-1% BSA and incubated at room temperature for 1 h. After washing, the ELISAs were developed with 1 mM 2-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) in citrate buffer with 0.03% H2O2 (Roche, Basel, Switzerland), and the absorbency was measured at 415 nm.
Limiting dilution assays
Draining lymph nodes were harvested from mice immunized 7 days previously and single-cell suspensions were used to set up the LDA described previously (7). Briefly, freshly isolated lymph node cells were plated immediately after harvesting in limiting numbers ranging from 20,000 to 1,000 cells/well in 96-well U-bottom plates in the presence of 50 U/ml IL-2, and 5 × 10−5 irradiated splenocytes from a transgenic mouse expressing a membrane form of HEL under the MHC class II promoter (7). Cells were then visually scored for growth on day 7 to determine the frequency of growth-positive cells. Growth-positive wells were expanded 7–10 days later under the same conditions and then tested in a proliferation assay in the absence of exogenous IL-2 to determine the percentage of growth-positive clones that were HEL specific. The frequency of HEL growth-positive cells was determined by Poisson distribution (7). HEL-specific T cell clones were further expanded under the same conditions and tested 7–10 days later in a proliferation assay using 5 μM of each HEL peptide (18–33, 31–47, 48–63, and 115–129). Proliferation was determined by adding 0.5 μCi of [3H]thymidine for the last 24 h of a 72-h culture.
T cell proliferation assays
T cell clones generated during the LDA after immunizing B10.BR mice with either 10 or 0.1 nmol of HEL in CFA and determined to be specific for the 48–63 peptide, were expanded under the conditions detailed above. Individual T cell clones were then cocultured with 5 × 10−5 irradiated B10.BR splenocytes per well with the indicated amount of HEL protein, and proliferation was measured by adding 0.5 μCi of [3H]thymidine the last 24 h of a 72-h culture.
The Ab response to HEL is proportional to the amount of HEL used to immunize
Mice immunized with 10 nmol (144 μg) of HEL generated a strong Ab response, with activity in the serum that titered out to a 1/3000 dilution (Fig. 1). Mice immunized with 1 nmol (14.4 μg) made significantly less, while mice immunized with 0.1 nmol (1.4 μg) had undetectable levels of Abs to HEL protein. These data indicated that the doses of HEL tested in these studies, from 10 to 0.1 nmol, encompassed a range of responses ranging from strong to poor.
Decreased frequency of T cells responding to lower doses of HEL
The frequency of HEL-reactive T cells responding at each dose of HEL was determined by LDA. Our assay was previously standardized to detect most of the HEL-reactive T cells. T cells were isolated from lymph nodes 7 days after immunizing mice, which represents the peak time of the response. The T cells were placed in culture with optimal amounts of IL-2 and stimulated by APCs that were previously determined to be presenting all four HEL epitopes. The frequency of HEL-reactive T cells was 1 in 5,000, 1 in 9,500, and 1 in 50,000 lymph node cells in mice immunized with 10, 1, and 0.1 nmol of HEL, respectively (Fig. 2 A). Thus, while a modest decrease was seen after decreasing the dose from 10 to 1 nmol, the number of responding T cells was decreased by a factor of 10 when the dose was decreased to 0.1 nmol of HEL.
Relative distributions of T cells responding to high and low abundance peptides
We expanded individual T cell clones isolated from the LDA, and determined their specificity for the four HEL epitopes presented by MHC class II I-Ak molecules. The specificity analysis of 162 individual T cell clones after immunizing with 10 nmol of HEL in CFA is shown in Fig. 2 B. In agreement with our previously published results, at this high dose of Ag, the numbers of T cells responding to the 48–63, 18–33, and 115–129 families of peptides are similar. Thus, despite levels of presentation that differed by >200-fold, the numbers of T cells reacting against the chemically dominant epitope (48–63), 30%, was very similar to the chemically subdominant epitope (115–129), 25%. The other chemically subdominant epitope, 20–35, which was presented at levels 200-fold less than 48–63 generated 17% of the HEL response.
The response to the 31–47 epitope was consistently represented by a lower proportion of T cells, 9%. It is of note that although fewer T cells responded to 31–47, this family is second in terms of abundance, represented at levels 60-fold less than the 48–63 family, but higher than the 20–35 family (2, 6). The binding affinity of 20–35 and 31–47 for I-Ak is about the same. We do not believe that the reasons for the weak clonal response to 31–47 can be attributed to any influence of the response to murine lysozyme: the murine lysozyme peptides of 31–47 are never presented because of hindering residues (18). We concluded that there was no strong correlation between the density of various peptide-MHC complexes and the relative proportion of clones generated to each.
Our initial concern was that the amount of HEL used in these and the previous studies, 10 nmol, was a saturating amount of Ag that could be masking a relationship between epitope density and T cell responses. We therefore lowered the dose to determine the effect on the T cell frequency and specificity. If a relationship existed between epitope density and T cell responses, we expected that at lower doses the T cell response would focus on the chemically dominant 48–63 epitope. The relative proportion of T cells responding to each of the four epitopes changed, albeit to a variable extent among the four epitopes, when immunizing with the lesser amounts of HEL of 1 and 0.1 nmol. Changes took place particularly at the 0.1-nmol dose compared with the 10 nmol, but these varied depending on the peptide. As noted, the 0.1 was a limiting dose which led to an undetectable B cell response.
The response to the chemically dominant 48–63 epitope increased slightly, from 30% of HEL-reactive T cells at the highest dose to 38 and 37% at 1 and 0.1 nmol, respectively. Thus, although T cells specific for 48–63 constituted the major single response, that of the other clones together still represented the majority, i.e., 63%.
There was no major change in the relative proportion of cells specific for the chemically subdominant 20–35 family, which constituted between 15 and 17% of the HEL response at each dose (Fig. 2 B). Thus, the ratio of T cells specific for 48–63 and 20–35 varied from 1.8 to 2.5 to 2.2 at each of the immunizing doses, 10, 1, and 0.1 nmol, respectively. Since the 0.1 dose was limiting and the difference in amounts presented between both epitopes was ∼200-fold in favor of 48–63, our expectation was that the 20–35 response would be markedly affected, which was not the case.
The proportion of T cells responding to the least abundant 115–129 epitope was decreased by ∼2-fold at the 0.1-nmol dose. The 115–129 peptides are presented in very low amounts that we have had trouble quantitating, but clearly >200-fold less than 48–63. The relative number of cells decreased as the dose was lowered by about half, i.e., the ratios of T cells specific for 48–63 and 115–129 went from 1.2 to 2.2 to 3.4, respectively.
The proportion of cells responding to the 31–47 epitope, which consistently generated fewer T cells, was reduced further at the lower dose. Other HEL epitopes comprised ∼20% of the response and are represented among several peptides, e.g., the 84–96 epitope presented by the MHC class II molecule I-Ek. The overall proportion increased slightly as the dose became limiting from 20 to 32%. Thus, despite limiting the HEL response by lowering the Ag dose, there was little direct correlation observed between epitope abundance and relative numbers of T cells generated to each epitope. Similar proportions of T cells responding to each epitope were also observed 30 days after immunization (data not shown).
We also examined the response in a different strain of mice to examine a possible role of the preimmune repertoire. CBA/J mice responded similarly to B10.BR despite having differences in the expression of TCR genes (Table I). Thus, 31–47 maintained a low response, while 115–129 was the closest to 48–63, the major responding peptide.
|Mouse .||Dose (nmol) .||Frequency .||% Specificity .||.||.||.|
|.||.||.||48–63 .||31–47 .||18–33 .||115–129 .|
|Mouse .||Dose (nmol) .||Frequency .||% Specificity .||.||.||.|
|.||.||.||48–63 .||31–47 .||18–33 .||115–129 .|
Summary of the frequencies of the responses to immunization with HEL in CFA and the distribution of the responses to four individual epitopes. The 48–63 is chemically dominant in that it is presented at levels that are 60-fold higher than 31–47 and greater than 200-fold higher than 18–33 and 115–129.
Finally, to test whether T cells primed at lower doses had higher affinity TCRs, the amount of Ag required to stimulate 50% proliferation (functional avidity) of individual clones from high- and low-dose immunizations was determined. Six individual T cell clones isolated after immunization with 10 and 0.1 nmol of HEL, and specific for 48–63, were tested for their sensitivity to HEL protein. The average of the six clones at each dose is shown in Fig. 2 C. We found no evidence of T cells having higher affinities at the lower Ag dose, as similar functional avidities were observed in clones generated after immunizing with 10 and 0.1 nmol. The selection for higher affinity T cells at lower levels of Ag may well occur in the secondary response rather than the primary response (19, 20, 21).
The first issue raised here is a technical one having to do with the nuances of quantitating CD4+ T cells by LDA. We have standardized the system by using carefully evaluated amounts of IL-2 and, very importantly, testing with APCs from HEL-transgenic mice that present high levels of all four of the naturally presented epitopes. This assay has been demonstrated to identify between one and two or three transgenic T cells specific for the 48–63 epitope. This finding along with the reproducibility of our results gives us confidence that the LDA is giving an accurate measurement of the representation of clones with the potential to divide in response to Ag. There is the valid issue that the LDA readout is an average of growth/death programs which may vary among different clones, but the fact that similar results were obtained 30 days after the initial immunization argues against a selective differential in the kinetics of responses among clones to different epitopes. Whether this assay misses cytokine-producing cells that are not entering cell division is another issue that will be addressed in future experiments.
A second issue concerns the levels of peptide-MHC complexes required to stimulate various T cell clones. This level has been examined in vitro, albeit in a limited number of experiments. The number required to activate naive CD4+ T cells has been estimated in two independent studies to require on the order of 300 complexes per APC (22, 23). The threshold for T cell activation can be lowered in previously activated T cells (24, 25). Another way to enhance T cell activation and promote expansion is through activation of APCs by administering Ag in strong adjuvants (26, 27, 28). This adjuvant effect can alter the biology of APCs, leading to higher expression of costimulatory molecules, cytokine secretion, and other auxiliary molecules that can lower the threshold for stimulation of T cell responses.
In our situation, surprisingly, despite an overall drop in the absolute number of HEL-reactive T cells, T cell clones were generated to all epitopes; from those displayed at very high levels (like 48–63) to those displayed in very low abundance (like 20–35 and 115–129). Thus, the level of presentation of each epitope was above the minimal threshold required to activate naive T cells in the environment brought about by immunization in adjuvant. We are currently measuring the peptide presentation by different APCs in the draining lymph nodes after immunization. Perhaps the amounts of HEL are highly concentrated in very few APCs, allowing for presentations of epitopes above the minimal threshold. It is of note that a similar lack of correlation between epitope abundance and T cell responses has been observed in CD8 T cell responses. For example, the response to a Listeria monocytogenes epitope required achieving a certain level of presentation and enhanced presentation did not increase the response (29), and responses to several different epitopes correlated poorly with their presentation levels (17, 30).
Finally, related to this last issue is that of the relative distribution of the T cell response toward several epitopes. Despite the large differential in peptide display and the use of different amounts of immunogen, all families of peptides generated T cell responses, albeit the relative proportion tended to change at limiting doses (summarized in Table I). Thus, whether the epitope is presented at hundreds of copies per APC, such as the 115–129 epitope, or several thousand, such as the 48–63 epitope, each stimulated T cells, very strikingly, differed by only 3-fold at most. Presumably, this was due to the use of CFA, which provides an environment to optimally stimulate T cell expansion, making the T cell response practically a “go or no go” response. Factors that may influence the extent of the go response will be the minimal density of peptide-MHC complexes, the preimmune repertoire, the expression of costimulatory molecules, and/or signaling through regulatory molecules. We conclude that under inflammatory conditions, such as protein administered in CFA, mechanisms which are yet undefined equalize large differences in the levels of peptide presentation, allowing for the generation of T cell response to high- and low-density epitopes and superseding any direct relationship between the level of peptide presentation, or chemical dominance, and the number of responding T cells.
We thank Kathy Frederick for her assistance with mice and Dr. Osami Kanagawa and Dr. Paul Allen for their helpful discussions.
Abbreviations used in this paper: HEL, hen egg lysozyme; LDA, limiting dilution assay.