Diversity of the CD8⁺ T cell repertoire elicited against an immunodominant epitope does not depend on the context of infection

Abstract

The diversity of the pathogen-specific T cell repertoire is believed to be important in allowing recognition of different pathogen epitopes and their variants and thereby reducing the opportunities for mutation-driven pathogen escape. However, the extent to which the TCR repertoire can be manipulated by different vaccine strategies so as to obtain broad diversity and optimal protection is incompletely understood. We have investigated the influence of the infectious/inflammatory context on the TCR diversity of the CD8⁺ T cell response specific for the immunodominant epitope in C57BL/6 mice, gB-8p, derived from glycoprotein B (gB) of the Herpes Simplex virus type 1 (HSV-1). To that effect, we compared TCR V segment utilization, CDR3 length and sequence diversity of the response to natural HSV-1 infection to those elicited by either Listeria monocytogenes or Vaccinia virus expressing gB-8p. We demonstrate that although the type of infection in which the epitope was encountered can influence the magnitude of the CD8⁺ T cell responses, TCR β-chain repertoires did not significantly differ in among the three infections. These results suggest that widely different live vaccine vectors may have little impact upon the diversity of the induced CTL response, which has important implications for the design of live CTL vaccine strategies against acute and chronic infections.

Introduction

While vaccination provides the most successful means for prevention of infectious disease, the development of protective vaccines against various pathogens still remains largely empirical. CD8⁺ T cells play a central role in combating intracellular infections and tumors, and much attention has been devoted to understanding the immune correlates of protection and efficacious CD8⁺ T cell control of infection. Historically, CD8⁺ T cell responses have been evaluated largely on the basis of their magnitude, with immune responses eliciting a greater number of antigen-specific CD8⁺ T cells generally considered more favorable. However, we have previously shown that the magnitude of the CD8⁺ T cell response does not always correlate with immune defense (1), and this quantitative measurement by itself does not reflect the underlying complexity and quality of the responding CD8⁺ T cells. It is now generally accepted that one key determinant of CD8⁺ T cell mediated immune protection correlates with the diversity of T cell receptor (TCR) usage, since TCR diversity within an epitope-specific response may enrich for higher avidity T cells (1), limit viral escape mutants in chronic infections (2, 3) and promote heterologous immunity (4-6). Collectively, these reports suggest that describing potential vaccine candidates in terms of the diversity of epitope-specific T cell clonotypes may help enable rational vaccine design.

A major challenge for vaccine development against chronic viral infections like HIV and HCV is to induce responses characterized by a diverse CD8⁺ T cell repertoire, which is more likely to recognize viral escape mutants that emerge in the course of infection (1, 7). This is especially difficult in those with physiological (e.g. the elderly, since aging of the immune system is associated with decreased TCR diversity and functionality) or pathophysiological/iatrogenic (immunodeficiency, chemotherapy, bone marrow transplantation) constrictions of the T-cell repertoire. Surprisingly, the extent to which the diversity of the CD8⁺ TCR repertoire can be manipulated by different vaccine strategies has never been thoroughly investigated. Recently, Malherbe et. al demonstrated that vaccine adjuvants can differentially alter the clonal composition of CD4⁺ T cells responding to pigeon cytochrome c (PCC) (8). While both depot (alum, CFA, IFA) and dispersable adjuvants (CpG, monophospahtidyl- lipid A- MPL) promoted clonal expansion of PCC specific CD4⁺ T cells, the vaccine adjuvants with known TLR activity (CpG, MPL) preferentially skewed the TCR repertoire towards clonotypes that exhibited higher TCR avidity. Importantly, these findings addressed the effects on CD4+ T cell repertoire of subunit vaccines and adjuvants. Viral ‘escape’ from immune recognition is, however, most prevalent in the context of CD8+ T cell responses. Moreover, most current vaccines targeting CD8+ T cell recognition use live vectors, as opposed to adjuvants. To address this question with regard to CD8+ T cells and live vaccines, we examined the extent of TCR diversity elicited in response to different infectious vectors by sequencing CD8⁺ T cells specific for the HSV-1 immunodominant epitope (gB_498-505; referred to hereon as the gB-8p epitope) from mice infected with Listeria monocytogenes (Lm-gB) or Vaccinia Virus (VACV-gB) expressing the immunodominant peptide (SSIEFARL). Both of these vectors have previously shown to be protective against a subsequent challenge with HSV(9, 10), however the effects of these different infections on the TCR repertoire has not been studied.

To better understand the molecular complexity of CD8⁺ T cell responses, our present analysis focused on two specific questions: a) Do infections with various recombinant vectors alter the diversity of the epitope-specific CD8⁺ TCR repertoire? (ie: do they modify the number of clonotypes responding, or the clonal dominance structure of the response?); and b) Does the epitope-specific CD8⁺ response to different infections select different TCRs? (ie: do the responding clones have different patterns of CDR3 amino acid sequences). Current estimates indicate that there are ~2 × 10⁶ different TCRs in an uninfected mouse (11, 12), of which 15-900 different TCRs respond to a given epitope (12, 13). The link between the selective forces involved in the preferential recruitment and expansion of different TCR clonotypes and their degree of avidity for cognate antigen (14, 15) suggests likely variation among infections with different antigen availability, kinetics of antigen abundance (microbial growth), and extent of CD8⁺ T cell division. We investigated whether these differences between a gram positive bacterium and two large and different DNA viruses would likely favor the proliferation of CD8⁺ T cells bearing different TCR clonotypes.

Unexpectedly, we found comparable TCRβ repertoire diversity despite major differences in the nature of the pathogen, magnitude of the response, and route of inoculation. Moreover, the characteristics of the gB-8p specific TCRβ repertoires associated with HSV-1 infection were shared by the gB-8p specific TCRβ repertoires from mice infected with Lm-gB and VACV-gB. These results have profound implications for the design of live vaccines against chronic viral infections and suggest that targeting the epitopes eliciting the most diverse CD8+ T cell responses or covaccinating with common escape variations may be a more efficacious strategy for generating long-lived protective immunity against chronic viral infections than altering the nature or type of vaccine vector.

Materials and Methods

Mice and infections

Male C57BL/6 (B6, H-2b) were purchased from NCI (Frederick, MD) and maintained under pathogen-free conditions in the animal facility at the University of Arizona. Naïve mice were used at 8-10 weeks of age and experiments were conducted by guidelines set by the University of Arizona Institutional Animal Care and Use Committee (IACUC). HSV-1 strain 17 was obtained from Dr. D.J. McGeoch (University of Glasgow, Scotland, U.K.), cloned as a syn+ variant and tittered on Vero cells in our laboratory as previously described(16, 17). Recombinant vaccinia virus expressing the MHC class I-restricted CTL epitope HSV gB_498-505 (SSIEFARL, gB-8p in the text), designated VACV-gB, was generously provided by Dr. S.S. Tevethia (Pennsylvania State University of College Medicine, PA). VACV-gB viral stocks were propagated and quantified in 143B cells. Mice were infected intraperitoneally (i.p.) with 5 × 10⁵ plaque-forming units (PFU) of either HSV-1 or VACV-gB. Lm-gB ΔActA expressing gB-8p was constructed as previously described (10) and grown in BHI containing chloramphenicol (20ug/ml). Prior to infection, the bacteria were grown to log phase (OD₆₀₀ 0.1), and mice were injected intravenously (i.v.) with 1 × 10⁶ colony forming units (CFU) in 100 ul of PBS.

Reagents and flow cytofluorometric (FCM) analysis

The gB-8p:K^b tetramer was obtained from the National Institutes of Health Tetramer Core Facility (Emory University, Atlanta, GA). mAbs anti-CD8α (clone 53-6.7), anti-CD4 (RM4-5), anti-CD11a (2D7), anti-Vβ10 (B21.5), anti-Vβ8 (F23.1) were purchased from commercial sources. FCM data was acquired on the custom-made FACS LSRII instrument equipped with four lasers, using Diva software (BD Biosciences), and analysis was performed using FloJo software (Treestar). At least 10⁵ cells were analyzed per sample, with dead cells excluded by selective gating based on orthogonal and side light scatter characteristics. Markers to score positive cells (events) were set to delimit fluorescence higher than the highest fluorescence of unstained or control-stained cells.

Cell sorting

Spleens were harvested from mice at the peak of infection and CD8+ T cells isolated using positive immunomagnetic selection of CD8+ cells (Miltenyi Biotec, Auburn, CA). Highly enriched CD8+ T cells were stained with gB-8p:K^b tetramers conjugated to streptavadin APC, anti-CD8α-PE Texas Red, anti-CD4-FITC, and anti-Vβ10-PE for one hour and washed twice. Cells were then transferred to sorting buffer and CD8+CD4-gB-8p:K^b+ Vβ10+ lymphocytes were isolated as single cells using the FACSAria cell sorter system (BD Biosciences). Control wells without sorted cells were included on every plate to identify any possible contamination.

Single-cell RT-PCR

Our RT-PCR protocol was adapted from of Kedzierska et al. (14) and Hamrouni et al (18). Single cells with appropriate surface markers were sorted directly into 96 well PCR plates contain 5 ul of cDNA reaction mix. The cDNA reaction mixture contained 0.25 ul of Sensiscript revere transcriptase (Qiagen), 1X cDNA buffer (Qiagen), 0.5mM dNTPS (Qiagen), 100 ug/ml tRNA (Invitrogen), 50 ng oligodT_12-18 (Invitrogen), 20 units of RNAse Out (Invitrogen), and 0.1% TritonX-100 (Sigma Aldrich). cDNA synthesis was performed immediately after sorting by incubating plates at 37°C for 90 mins, followed by 5 min at 95°C, and plates were stored at −80°C. The Vβ10 transcripts were amplified by nested PCR and the entire 5 ul cDNA reaction was used for the first PCR reaction in a final 25 ul volume containing 1.5 U Taq polymerase in manufactuers 1X Buffer A with 1.5mM MgCl₂ (Fischer Scientific), 200 uM of each dNTP (Fischer Scientific), and 100 nM of the external sense Vβ10 primer (5’ AAACTCTGGGCCACGATACT-3’) and external antisense Vβ10 primer (5’-CTCAGCTCCACGTGGTCA-3’). The PCR conditions for the first PCR program began with 95°C for 2 mins, followed by 40 cycles of 10s at 95°C, 45s at 59°C and 45s at 72°C, and ending with 5 min at 72°C. A 1.0 ul aliquot of the first-round PCR was used for the second PCR reaction with the internal sense Vβ10 primer (5’-gcaactcattgtaaacgaaaca-3’) and internal antisense Vβ10 primer (5’-CGAGGGTAGCCTTTTGTTTG-3’). The second PCR program began with 95°C for 2 min then 72°C for 5s, followed by 35 cycles of 10s at 95°C, 60s at 61°C, and 30s at 72°C, and then ends with 5 min at 72°C. PCR products were resolved on a 2% agarose gel, purified with the MinElute 96 UF PCR purification kit (Qiagen), and sequenced with 12 pmol of Vβ10 sequencing primer (5’-AGGCGCTTCTCACCTCAGTCTTC-3’) using an Applied Biosystems 3730XL DNA Analyzer at the Genomic Analysis and Technology Core (University of Arizona). The internal sense primer was also used as a sequencing primer for the mice infected with VACV-gB.

TCRβ repertoire analysis

The gB-8p specific CD8⁺ TCRβ repertoires were characterized by sequentially aligning each of the TCRβ sequences with the Vβ10 (TRBV4 in IMGT nomenclature) gene, the best-match Jβ gene, the best-match Dβ gene, and then identifying the CDR3β sequence. The reference alleles for the Mus musculus germline TRB genes from IMGT (19) were used for the analysis of the TCRβ repertoires.

To assess whether the diversities of the TCRβ repertoires specific for the gB-8p-epitope differed between the HSV-1, Lm-gB and VACV-gB infections, both the number of different amino acid sequence clonotypes and Simpson's diversity index (20) were used as measures of clonotypic diversity. Simpson's diversity index accounts for both the variety of amino acid sequence clonotypes and their clone sizes, and ranges in value from 0 (minimal diversity) to 1 (maximal diversity). To account for the differences in the samples sizes obtained between mice, the TCRβ repertoire diversities were estimated as if 39 TCRβ sequences had been obtained for each sample. This estimation involved randomly drawing from each sample a subset of 39 TCRβ sequences and estimating the diversity of this subset. This process was repeated 10,000 times to produce a distribution of TCRβ repertoire diversities, from the median of which, we determined a diversity estimate for each sample (20). The diversity analysis was performed using Matlab (The Mathworks, Natick, MA).

Statistical analysis

The percentage of gB-8p specific CD8⁺ T cells and the diversities of the gB-8p specific CD8⁺ TCRβ repertoires at day 6 post-infection were compared between the groups of mice infected with HSV-1, Lm-gB and VACV-gB using a Kruskal-Wallis test and Dunn's multiple comparison post-test. The dominant Vβ usages of the gB-8p specific CD8⁺ T cells involved in the immune responses in mice in the three different infection groups were compared using a non-parametric two-way ANOVA and Bonferroni post-tests (i.e. based on ranks). All statistical analyses were performed using GraphPad Prism software (GraphPad Software Inc, San Diego, CA).

Results

Kinetics and magnitude of the gB-8p specific CD8⁺ T cell responses in mice infected with HSV-1, Lm-gB and VACV-gB

Over 90% of the acute CD8⁺ T cell response in HSV-1-infected C57BL/6 mice is directed against the single immunodominant gB-8p epitope (21) . To determine if the context of infection alters the kinetics and magnitude of the CD8⁺ T cell response, we first monitored the expansion of gB-8p specific CD8⁺ T cells in mice infected with HSV-1, Lm-gB, and VACV-gB. The dramatic increase of gB-8p specific CD8⁺ T cells was tracked in the spleen. While the overall kinetics of CD8⁺ T cell expansion were similar among infections, the frequency and total number of gB-8p specific CD8⁺ T cells were 2-3 times higher in mice infected with HSV-1 and VACV-gB as compared to those infected with Lm-gB (Fig. 1). This quantitative difference in clonal expansion between different vectors is most likely explained by the attenuated nature of the Listeria vector (Lm-AcA-gB), which is expected to provide lower stimulation of the immune system.

Figure 1. Profiles of gB-8p specific CD8⁺ T cell responses in mice infected with HSV-1, Lm-gB, and VACV-gB.

The kinetics of the gB-8p specific CD8⁺ T cell responses was tracked in the spleen of mice infected with HSV-1, Lm-gB, and VACV-gB (A) and the magnitude of the gB-8p specific CD8⁺ T cell responses were compared for all three infections at day 6 post-infection (B). Significant differences in the percentages of gB-8p specific CD8⁺ T cells at day 6 post-infection between the Lm-gB and both the HSV-1 and VACV-gB infections were determined by a Kruskal-Wallis test. Results are representative of at least 2 separate experiments,with n≥4 mice/group. *, p<0.05, mean±SEM.

The gB-8p specific TCRβ repertoires from mice infected with HSV-1, Lm-gB, and VACV-gB exhibit similar Vβ segment and sequence compositions

To uncover any broad differences in the CD8⁺ TCR repertoire among infections, we next asked whether Vβ usage was skewed in mice infected with HSV-1, Lm-gB, and VACV-gB. Published studies, including these from our laboratory, have documented significant Vβ bias in gB-8p specific CD8⁺ T cells in HSV-1 infected mice, typically consisting of ~55% Vβ10 and ~20% Vβ8 TCR β-chain participation (22-26). When we compared Vβ usage within gB-8p specific CD8⁺ T cells from mice infected with HSV-1, Lm-gB, and VACV-gB we observed a similar Vβ bias in response to all three infections at day 6 post-infection (Fig 2.). While there was no significant differences in Vβ10 usage of gB-8p specific CD8⁺ T cells between HSV-1 infection and either Lm-gB or VACV-gB infections (where 50.0%, 57.8% and 55.4% of gB-8p specific CD8⁺ T cells used Vβ10 for Lm-gB, VACV-gB, and HSV-1, respectively), there was a small but statistically significant difference in Vβ10 usage in mice infected with Lm-gB compared with VACV-gB (p<0.01; non-parametric two-way ANOVA and Bonferroni post-test). We do not consider this difference biologically significant, given that Vβ10 makes up an overwhelming part of this response in each of the three responses. By contrast, Vβ8 usage was extremely similar between all three infections (i.e. 22.0%, 22.3% and 22.3% of gB-8p specific CD8⁺ T cells for the HSV-1, Lm-gB and VACV-gB infections, respectively). Overall Vβ usage profiles among infections were largely comparable. Thus, despite major differences in the magnitude of the CD8⁺ T cell response, different intracellular pathogens (HSV-1, Lm-gB, and VACV-gB) do not appear to shape the CD8⁺ TCR repertoire by selecting widely different Vβ elements.

Figure 2. Vβ gene usage in the gB-8p specific CD8⁺ T cell responses in the HSV-1, Lm-gB, and VACV-gB infections.

A comparison of the percentages of gB-8p specific CD8⁺ T cells using the Vβ10 and Vβ8 genes between all three infections at day 6 post-infection. There were no significant differences in the Vβ10 or Vβ8 gene usage of gB-8p specific CD8⁺ T cells between the HSV-1 infection and either the Lm-gB or VACV-gB infections. However, the Vβ10 usage of gB-8p specific CD8⁺ T cells in the Lm-gB infection was found to be significantly lower than for the VACV-gB infection using a non-parametric two-way ANOVA and Bonferroni post-test. The graphs represent pooled results from 2 separate experiments with n=4 mice/group. *, p<0.01, mean±SEM.

For a more in-depth assessment of the composition of the TCRβ repertoires involved in the gB-8p specific CD8⁺ T cell responses in the HSV-1, Lm-gB, and VACV-gB infections, the CDR3 portions of the TCRβ repertoires using the Vβ10 gene were sequenced. These data are summarized in Table I. Given the enormous amount of inter-individual variability in the TCR repertoire, large-scale analysis was required to resolve these questions and conclusively examine key factors that shape epitope-specific TCR repertoires. A total of 1608 Vβ10⁺ TCRβ sequences were obtained across the HSV-1, Lm-gB, and VACV-gB infection groups. The HSV-1 and LmgB infection groups consisted of ten mice each while the VACV-gB infection was studied in six mice. The number of TCRβ sequences per mouse varied between 39 and 82 TCRβ sequences, with an average of 61.8 TCRβ sequences per mouse. Examples of the gB-8p specific Vβ10⁺ TCRβ amino acid sequence repertoires for one mouse from each of the HSV-1, Lm-gB, and VACV-gB infection groups are shown in Table II.

Table I.

Summary of the data used to compare the Vβ10⁺ TCRβ repertoires for the gB-8p specific CD8⁺ T cell responses in the HSV-1, Lm-gB, and VACV-gB infections.

	HSV-1	Lm-gB	VACV-gB
No. of mice	10	10	6
No. of TCRβ sequences across all mice	645	524	439
Range of no. of TCRβ sequences per mouse	53-73	39-67	60-82
Mean no. of TCRβ sequences per mouse	64.5	52.4	73.2

Table II.

The Vβ10⁺ TCRβ repertoires involved in the gB-8p specific CD8⁺ T cell responses in one mouse each for the HSV-1, Lm-gB, and VACV-gB infections.

HSV-1: Mouse 6				Lm-gB: Mouse 5				VACV-gB: Mouse 6
CDR3βa	Jβ	length	clone size	CDR3βa	Jβ	length	clone size	CDR3βa	Jβ	length	clone size
CASRDWGVAEQFF	2.1	13	1	CASRQSGGSDYTF	1.2	13	3	CASSYWGGTEVFF	1.1	13	1
CASRDWGVYAEQFF	2.1	14	1	CASRQWGAGNTLYF	1.3	14	1	CASRGSGYSDYTF	1.2	13	1
CASRGWGGGAEQFF	2.1	14	1	CASRQGGTERLFF	1.4	13	2	CASRDWGVAEQFF	2.1	13	1
CASRYWGGGAEQFF	2.1	14	3	CASRYWGGNAEQFF	2.1	14	2	CASRDWGVYAEQFF	2.1	14	1
CASSHWGGGAEQFF	2.1	14	3	CASSFWGADAEQFF	2.1	14	1	CASRLWGNYAEQFF	2.1	14	1
CASSLWGNYAEQFF	2.1	14	2	CASSNWGNYAEQFF	2.1	14	2	CASRRWGAIAEQFF	2.1	14	1
CASSSWGGSAEQFF	2.1	14	1	CASSVWGGGAEQFF	2.1	14	1	CASSHWGNYAEQFF	2.1	14	1
CASSSWGGYAEQFF	2.1	14	5	CASSYWGGSAEQFF	2.1	14	2	CASSPWGGGAEQFF	2.1	14	2
CASSYWGGAAEQFF	2.1	14	1	CASRLWGGNYAEQFF	2.1	15	2	CASSYWGGAAEQFF	2.1	14	1
CASRDWGVGYAEQFF	2.1	15	1	CASSFWGYTGQLYF	2.2	14	1	CASRNWGTGQLYF	2.2	13	1
CASRDWGNTGQLYF	2.2	14	1	CASSYWGNTGQLYF	2.2	14	1	CASSFWGNTGQLYF	2.2	14	2
CASSHWGNTGQLYF	2.2	14	1	CASRNWGANTGQLYF	2.2	15	1	CASSLWGYTGQLYF	2.2	14	1
CASSPWGDTGQLYF	2.2	14	1	CASSEGLGNTGQLYF	2.2	15	1	CASSSWGGAGQLYF	2.2	14	1
CASRNWGTNTGQLYF	2.2	15	1	CASSYWGEDTGQLYF	2.2	15	3	CASSYWENTGQLYF	2.2	14	1
CASSHWGAETLYF	2.3	13	1	CASRDWGSSAETLYF	2.3	15	1	CASKDWGTNTGQLYF	2.2	15	1
CASRIWGGAETLYF	2.3	14	5	CASRIWGASAETLYF	2.3	15	1	CASKDWGVGTGQLYF	2.2	15	1
CASSFWGGRETLYF	2.3	14	1	CASRNWGTSAETLYF	2.3	15	3	CASRDWGTNTGQLYF	2.2	15	1
CASRDWGTSAETLYF	2.3	15	1	CASSYWGGSAETLYF	2.3	15	1	CASRNWGEGVTLYF	2.3	14	1
CASSHWGASAETLYF	2.3	15	2	CASRSWGVQNTLYF	2.4	14	1	CASRYWGSAETLYF	2.3	14	6
CASSLWGGRAETLYF	2.3	15	1	CASSHWGSQNTLYF	2.4	14	1	CASSFWGGAPETLYF	2.3	15	1
CASRNWGANTLYF	2.4	13	1	CASSIWGGQNTLYF	2.4	14	7	CASRDWGQNTLYF	2.4	13	1
CASRDWGVQNTLYF	2.4	14	1	CASSNWGVQNTLYF	2.4	14	1	CASRDWGVQNTLYF	2.4	14	2
CASRRWGSQNTLYF	2.4	14	1	CASSQWGAQNTLYF	2.4	14	3	CASRHWGSQNTLYF	2.4	14	1
CASSLWGSQNTLYF	2.4	14	1	CASSRWGGEGTLYF	2.4	14	2	CASRTWGSQNTLYF	2.4	14	1
CASSYWGGQNTLYF	2.4	14	1	CASSSWGGQNTLYF	2.4	14	1	CASSHWGSQNTLYF	2.4	14	1
CASRNWGEVSQNTLYF	2.4	16	1	CASSSWGGRNTLYF	2.4	14	1	CASSLWGGQNTLYF	2.4	14	1
CASRAWGSQDTQYF	2.5	14	2	CASSYWGGQNTLYF	2.4	14	2	CASSYWGGQNTLYF	2.4	14	1
CASRHWGGQDTQYF	2.5	14	5	CASSYWGSQNTLYF	2.4	14	1	CASRNWGTSQNTLYF	2.4	15	1
CASSSWGGADTQYF	2.5	14	2	CASRDWGQDTQYF	2.5	13	2	CASRDWGKDTQYF	2.5	13	1
CASSYWGGRGTQYF	2.5	14	1	CASRPWGGQDTQYF	2.5	14	1	CASRPTGVDTQYF	2.5	13	1
CASRNWGANQDTQYF	2.5	15	1	CASSLWGGKDTQYF	2.5	14	1	CASRYWGEDTQYF	2.5	13	1
CASRTWGGGQDTQYF	2.5	15	4	CASSNWGAQDTQYF	2.5	14	1	CASSRWGGGDTQYF	2.5	14	1
CASRYWGGGQDTQYF	2.5	15	2	CASSSWGGEDTQYF	2.5	14	2	CASSYWGGAGTQYF	2.5	14	6
CASQNTGGGAKTPSYF	2.5	16	1	CASSYWGAQDTQYF	2.5	14	1	CASSYWGGQDTQYF	2.5	14	1
CASRDWATYEQYF	2.7	13	2	CASRPWGANQDTQYF	2.5	15	1	CASRNWGEKEQYF	2.7	13	6
CASRDWGVYEQYF	2.7	13	1	CASRVWGGEEDTQYF	2.5	15	3	CASRNWGGYEQYF	2.7	13	1
CASSYWGGGEQYF	2.7	13	2	CASRGTGSYEQYF	2.7	13	1	CASRNWGGAREQYF	2.7	14	1
CASRNWGSSYEQYF	2.7	14	1	CASRHWGVSYEQYF	2.7	14	2	CASRNWGSSYEQYF	2.7	14	2
CASRSWGGDDEQYF	2.7	14	2	CASRYWGESYEQYF	2.7	14	2	CASRYWGGADEQYF	2.7	14	1
CASSFWGGAGEQYF	2.7	14	1					CASRYWGSSYEQYF	2.7	14	6
CASSYWGGSYEQYF	2.7	14	1					CASSTWGGAGEQYF	2.7	14	1
CASRDWGVISYEQYF	2.7	15	1
			69				66				65

^aWG doublets in the CDR3β are shown in bold.

The compositions of the gB-8p specific Vβ10⁺ TCRβ repertoires for the three infections are summarised by the distributions of CDR3β lengths (Fig. 3A; here the CDR3β sequence is inclusive of the conserved cysteine in the Vβ-region and the conserved phenylalanine in the Jβ-region, as shown in Table II) and Jβ gene usage (Fig. 3B) for each mouse. The variation in CDR3β length and Jβ gene usage between mice in the different infection groups was comparable to that between mice in the same infection group. The gB-8p specific Vβ10⁺ TCRβ repertoires were also pooled across all mice within each infection group and the proportion of TCRβ clonotypes (i.e. ignoring the number of copies of each TCRβ sequence) with a particular CDR3β length (Fig. 3C) and using a particular Jβ gene (Fig. 3D) compared between the HSV-1, Lm-gB, and VACV-gB infections. The dominant length of the CDR3β amino acid sequences was 14, with at least 58% of the pooled gB-8p specific Vβ10⁺ TCRβ clonotypes for each infection having a CDR3β length of 14 amino acids. CDR3β sequences of length 13 and 15 amino acids were also common, comprising greater than 14% of the gB-8p specific Vβ10⁺ TCRβ clonotypes for each of the three infections (Fig. 3C). Preferential usage of the Jβ2 group of genes was found, with at least 95% of the gB-8p specific Vβ10⁺ TCRβ clonotypes for each infection using Jβ2 genes (Fig. 3D). This bias towards use of the Jβ2 group of genes is consistent with observations in previous studies of the HSV-1 gB-8p specific TCRβ repertoire (22, 23, 27).

Figure 3. Comparison of the compositions of the Vβ10⁺ TCRβ repertoires involved in the CD8⁺ T cell responses to the gB-8p epitope in the HSV-1, Lm-gB, and VACV-gB infections.

Shown are the proportion of the gB-8p specific Vβ10⁺ TCRβ repertoires in each mouse, within an infection group, with a particular CDR3β length (A) and using a particular Jβ gene (B). The CDR3β length refers here to the length of the sequence inclusive of the conserved cysteine in the Vβ-region and the conserved phenylalanine in the Jβ-region, as shown in Table II. The mice in each infection group are shown in numerical order of their mouse numbers (i.e. Mouse 1, Mouse 2, ...), from left to right. The gB-8p specific Vβ10⁺ TCRβ clonotypes for all mice in each infection group were pooled to compare the proportion of gB-8p specific Vβ10⁺ TCRβ clonotypes with a particular CDR3β length (C) and using a particular Jβ gene (D) between the HSV-1, Lm-gB, and VACV-gB infections.

This comparison demonstrates that the gB-8p specific TCRβ repertoires involved in the CD8⁺ T cell responses in infections with the live recombinant vectors Lm-gB and VACV-gB have similar Vβ and Jβ gene usage and CDR3β length to the HSV-1 gB-8p specific TCRβ repertoires, suggesting that the type of infection in which the gB-8p epitope was encountered has little influence on the composition of the responding TCRβ repertoire.

Infection with different pathogens does not alter the diversity of the TCR repertoire responding to the gB-8p epitope

The diversity of the TCRs responding to an epitope is thought to be one determinant of the ease of immune escape by viruses. Diversity is often measured as the number of different TCRs seen in an animal. However this doesn't take into account either the clonal dominance among these TCRs (ie: if there are 10 different TCRs, do they each make up 10% of the response, or does one make up 91%, and the other 1% each?), or the number of TCRs sequenced. To directly address whether the type of infection in which an epitope is encountered influences the sequence diversity of the epitope-specific CD8⁺ T cell response, the clonotypic diversities of the TCRβ repertoires responding to the gB-8p epitope were compared between the HSV-1, LmgB, and VACV-gB infections. The diversities of the TCRβ repertoires were estimated for each mouse, as if each sample was of the same size, using both the number of different amino acid sequence clonotypes and Simpson's diversity index as diversity measures. The latter diversity measure accounts for both the variety of different TCRβ clonotypes and the clonal dominance hierarchy and varies in value between 0 and 1, which reflect minimal and maximal diversity, respectively (20). The gB-8p specific TCRβ repertoires were very diverse in all mice in all three infection groups, with at least 19 different clonotypes and Simpson's diversity indices greater than 0.92 (Fig. 4A, B). For the HSV-1, Lm-gB, and VACV-gB infections, the median numbers of clonotypes were 24.5, 28, and 27 and the median Simpson's diversity indices were 0.96, 0.98, and 0.97, respectively. There were no significant differences in the diversities of the gB-8p specific TCRβ repertoires between the HSV-1, Lm-gB, and VACV-gB infections.

Figure 4. Comparison of the diversities of the Vβ10⁺ TCRβ repertoires involved in the CD8⁺ T cell responses to the gB-8p epitope in the HSV-1, Lm-gB, and VACV-gB infections.

TCRβ clonotypic diversity was measured using both the number of different TCRβ clonotypes (A) and Simpson's diversity index (B). Simpson's diversity index accounts for both the number of different TCRβ clonotypes and the number of copies of each clonotype (i.e. the clonal dominance hierarchy) and varies between 0 (minimum diversity) and 1 (maximum diversity). The TCRβ diversities were estimated for all samples having an equal sample size of 39 TCRβ sequences. The diversity of the CDRβ sequences was considered by determining the prevalence of the tryptophan-glycine (WG) doublet in CDR3β positions 6 and 7, as shown in Table II (which correspond to CDR3 positions 3 and 4 using the Chothia definition (29)). Shown are the percentages of the TCRβ clonotypes per mouse (C) and the percentages of the TCRβ repertoires per mouse (i.e. including the number of copies of the clonotypes) (D) that feature a WG doublet in CDR3β positions 6 and 7. A Kruskal-Wallis test was used to determine that there were no significant differences in TCRβ repertoire diversity between the HSV-1, Lm-gB, and VACV-gB infections.

The variety of TCR clonotypes is not the only measure of diversity important in an epitope-specific T cell response, because conserved patterns of amino acid usage in the CDR3 can indicate a restricted TCR repertoire (28). In the case of CD8⁺ T cell responses to the HSV-1 gB epitope, preferential usage of the Dβ2 gene in a single reading frame encodes a tryptophanglycine (WG) doublet in CDR3β positions 3 and 4 (using the Chothia definition (29)), resulting in a relatively conserved junctional sequence (27). The example gB-8p specific TCRβ repertoires shown in Table II suggest that the WG doublet is also prevalent in the Lm-gB and VACV-gB infections (where positions 6 and 7 of the CDR3β sequences in Table II correspond to Chothia CDR3β positions 3 and 4 (29)). Our investigation of the prevalence of the WG doublet in CDR3β positions 6 and 7 (i.e. Chothia CDR3β positions 3 and 4 (29)) in the gB-8p specific TCRβ repertoires found that at least 88% of both the TCRβ repertoire, and TCRβ clonotypes, in all mice exhibited the WG-signature, regardless of the infection (Fig. 4C, D). Moreover, there were no significant differences in the prevalence of the WG doublet between the HSV-1, Lm-gB and VACV-gB infection groups. In all three infections, this tryptophan-glycine doublet was predominantly encoded by the Dβ2 gene, as reported in previous studies of the TCRβ repertoires responding to the HSV-1 gB-epitope (27).

Discussion

The fundamental question we wished to address in this report was whether we can modify the diversity and specificity of the CD8+ T cell repertoire simply by changing the vector used to deliver the epitope? We addressed this question by comparing the TCRβ repertoires involved in the CD8⁺ T cell responses to a specific immunodominant epitope in mice immunized with HSV-1, Lm-gB and VACV-gB. Surprisingly, the gB-8p specific CD8⁺ TCRβ repertoires from mice infected with HSV-1 and live recombinant vectors Lm-gB and VACV-gB all have similar diversities, suggesting that the nature of infection in which the gB-epitope was encountered has no significant influence on CD8 T-cell repertoire diversity. These conclusions were established on the basis of large-scale analyses demonstrating the following results: 1) Vβ and Jβ gene usage and CDR3β lengths were similar among all infections; 2) No significant differences in TCRβ clonotypic diversity were detected among all infections; and 3) No significant differences in the usage of the prominent CDR3 WG signature motifs was found amongst different infections.

It was surprising that the diversity in CD8⁺ TCRβ repertoires responding to a particular epitope was not altered among infections with large DNA viruses (VACV-gB, HSV-1) that generate robust expansion of CD8⁺ T cells and an attenuated gram-positive bacteria (Lm-gB ΔActA) that is incapable of spreading to nearby cells and results in a more diminished response. The context of infection between VACV-gB, HSV-1, and Lm-gB may also differ in their levels of cytopathicity, infected cell types, inflammatory cytokines, and antigen kinetics. Furthermore, there are likely marked differences in protein synthesis, degradation and processing of antigens derived from bacterial versus viral vectors (30-32), yet none of these dissimilarities resulted in a detectable difference in the diversity of responding CD8+ T cells.

Given that in our experimental system diversity of the responding CD8 T-cells was not influenced by the microbial context of the pathogen under widely different conditions, it is important to highlight a few key differences between the report of Malherbe et. al, where differences in CD4+ T cell repertoire were reported (8) and our current study. First, previous reports indicate that different priming requirements may be optimally required to activate CD4⁺ and CD8⁺ T cells. While CD8⁺ T cells require just a brief episode of stimulation (33-35), CD4⁺ T cells require a more prolonged period of antigen stimulation (36, 37) which may result in different levels of sensitivity in the clonal selection process. Secondly, it is likely that more redundancy is present in the immune response initiated with live infectious vectors compared to peptide/protein immunizations. The cytokine millieu can provide important signals involved in CD8⁺ T cell expansion, but some of these signals (IL-12 and type I IFN) have been shown to be redundant in generating CD8⁺ memory T cells (38). It is likely that a greater number of redundant signals are present in bacterial and viral infections (compared to adjuvants), leading to minimization of differences in the inflammatory environments among HSV-1, Lm and VACV infections. Another notable difference between Malherbe et. al's findings (8) and our present study relates to the shape of the TCR repertoire that was examined. Whereas the TCR repertoire specific for the I-E^k-restricted PCC_94-103 epitope is narrow and public, the repertoire specific for the K^b-restricted gB_498-505 epitope is substantially more diverse with less inter-individual sharing of TCR clonotypes observed. It would be interesting to examine all of these issues more closely and determine if the clonal composition of gB-8p specific clonotypes could be altered with different protein vaccination strategies, or whether I-E^k-restricted PCC_94-103 specific T cells could be differentially skewed by various infections (VACV, Lm, etc). Finally, another potential difference between the Malherbe study and the present analysis is the number of animals and sequences analyzed. The former analyzed three mice per group, with <20 sequences per animal, and the results pooled for the purposes of analysis. Our study included larger groups (6-10 mice), with many more sequences per mouse. Our method of analysis compared diversity amongst individual mice, instead of the pooled repertoire, and thus avoided the potential problem of a ‘rogue mouse’ biasing the overall repertoire. It will be important to repeat studies of TCR diversity in the CD8+ and CD4+ T cell receptor repertoire on other epitopes and with larger groups of animals to resolve the differences between these studies.

Regardless of the vaccine strategy, the fundamental unit of T cell responses is the individual clonotype, and T cell immunity must be described in these terms if we are to understand the exact structure of a T cell response that is essential for controlling acute and chronic pathogens. In a recent report, Price et. al demonstrated that certain CD8⁺ TCR repertoire features could predict vaccine efficacy in SIV-challenged rhesus macaques(39). Understanding the vaccine strategies that can generate and maintain these ideal repertoires are of obvious importance, but this requires us to understand whether these features are inherent to specific epitopes or whether they can be altered by different vaccination approaches. Our data indicates that equally diverse CD8⁺ TCR repertoires can be generated by widely different live vaccine vectors. Thus, many attributes of the TCR repertoire may not be easily manipulated by vaccination and finding ideal epitopes may be a more productive goal for efficacious immunizations.

ABBREVIATIONS USED IN THIS PAPER

CDR

complementarity-determining region of the TCR

CpG

unmethylated oligonucleotides composed of cytidine-phospho-guanine repeats

gB

glycoprotein B of the Herpes Simplex virus

gB-8p

immunodominant gB epitope, SSIEFARL

Lm-gB

recombinant Listeria monocytogenes carrying the gB-8p epitope

VACV-gB

recombinant vaccinia virus carrying the gB-8p epitope