Cross-Reactivity of HLA-B^*1801-Restricted T-Lymphocyte Clones with Target Cells Expressing Variants of the Human Cytomegalovirus 72kDa-IE1 Protein

Abstract

The impact of natural polymorphism in a cytomegalovirus-dominant HLA-B^*1801-restricted epitope, IE1_199-206, on the specific responses of T-cell clones was assessed by measuring their cytolytic activity against target cells expressing mutated recombinant IE1 proteins. Our results suggest an in vivo selection of T lymphocytes that cross-react with multiple IE1 variants.

CD8⁺ cytotoxic T lymphocytes (CTLs) play a major role in protection against cytomegaloviruses and in maintenance of their latency (14, 18, 21, 24, 25). The memory response appears to be focused on a small set of viral proteins (for reviews, see references 16 and 17). Two immunodominant proteins from human cytomegalovirus (HCMV) antigens, pp65 (9, 12, 27) and 72kDa-IE1 (1, 4, 8, 9), have been identified.

To date, characterization of HCMV antigenic epitopes almost always relied on stimulation assays using target cells infected by virions of the laboratory strain AD169, expressing AD169 recombinant proteins or loaded with synthetic peptides designed from AD169 sequence (1, 4, 8, 12, 27). Such experimental procedures that do not take in account genetic variation in natural HCMV strains may introduce biases in the present assessment of HCMV gene product immunogenicity.

In this study, we analyzed the impact of natural variation in the IE1_199-206 epitope on the reactivity of an HLA-B^*1801-restricted subpopulation of four distinct memory CTL clones (C5, C6, C7, and C8) with target cells expressing recombinant variants of IE1 or loaded with mutant synthetic peptides. These T-cell clones were derived from peripheral blood mononuclear cells of healthy HCMV⁺ donor C with an autologous B-lymphoblastoid cell line expressing a natural variant of IE1, IE1-1 (20). Clones C5 (Vβ22S1Jβ2.7) and C6 (Vβ22S1Jβ2.3) were found 10 and 7 times, clone C8 (Vβ6S5Jβ2.7) was found twice, and clone C7 (Vβ9S2Jβ2.1) was found once (20). All clones produced tumor necrosis factor when stimulated by the IE1_199-207 (ELKRKMIYM) epitope (20) but killed autologous target cells pulsed with the IE1_199-206 octapeptide more efficiently than those pulsed with the IE1_199-207 nonapeptide in a standard ⁵¹Cr release assay (data not shown).

A compilation of 70 published sequences, obtained from clinical HCMV isolates of unrelated individuals (19, 28), led to evaluation of natural variation in IE1_199-206 (Table 1). Mutations were targeted to residues 201 (K, R, N, or T) and 205 (I or M), which corresponded to positions 3 and 7 within the octapeptide. Amino acid sequences ELKRKMIY and ELRRKMMY corresponded to the sequences of recombinant proteins encoded respectively by our original clinical strain, IE1-1(KI), and by the AD169 strain [recombinant protein IE1-2(RM); this paper] (target residues for natural polymorphism are indicated in boldface). All together, they represented 80% of the compiled sequences. Four less frequent variants were found that carried amino acid combinations NI, KM, RI, and TI at positions 201 and 205.

TABLE 1.

Published data on natural polymorphism in a IE1_199-206 epitope

IE1199-206 sequencea	No. of variant strains		Total no. of variant strains	% of total no. of strains investigated
	B. Zweygberg Wirgart et al.b	C. Retière et al.c
ELKRKMIYd	13	13	26	37.1
ELRRKMMYe	20	10	30	42.9
ELNRKMIY	5	0	5	7.1
ELKRKMMY	5	0	5	7.1
ELRRKMIY	2	1	3	4.3
ELTRKMIY	1	0	1	1.4

^aTarget residues for natural polymorphism are indicated in boldface.

^bReference 28; accession no. AF099552 to 099597.

^cReference 19.

^dOriginal clinical strain sequence which corresponds to IE1-1(KI) sequence in this paper.

^eAD169 strain sequence which corresponds to IE1-2(RM) sequence in this paper.

We performed site-directed mutagenesis with IEI-1(KI) cDNA, using a QuickChange Site-Directed Mutagenesis kit (Stratagene) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis-purified oligonucleotides (Genset, Paris, France), to generate vectors encoding IE1-1(NI), IE1-1(KM), IE1-1(RI), and IE1-1(TI) recombinant proteins that carried the amino acid combinations NI, KM, RI, and TI, respectively, at positions 201 and 205. U373 MG cells, which express HLA-B^*1801 molecules, were transfected with pcDNA3-IRESneo vectors (15) carrying the mutant cDNAs. After G418 selection, all six recombinant cell lines were labeled with E13-fluorescein isothiocyanate monoclonal antibody to check homogeneous IE1 expression by flow cytometry analysis (data not shown). A DNA region including the IE1_199-206 coding sequence was amplified from all six recombinant cell lines and sequenced to check that each population of selected cells expressed the appropriate mutant protein.

Using clones C5 to C8 as effector cells, the immunogenicity of recombinant proteins bearing the six IE1_199-206 mutant epitopes was then assessed by cytotoxicity assays (Fig. 1). All four CTL clones killed efficiently U373 MG/IE1-1(KI) cells that expressed the original IE1 variant. They did not recognize the U373 MG negative control. The two immunodominant clones, C5 and C6, were able to kill U373 MG cells expressing 5/6 and 3/6 IE1 variants, respectively. The two minor clones, C7 and C8, cross-reacted with 3/6 and 2/6 IE1 variants, respectively. No clone was found to be cytotoxic for IE1-1(TI), which carried an infrequent IE1₂₀₁IE1₂₀₅ TI combination. This lack of antigenicity could be interpreted as a defect in processing and/or presentation or recognition of the mutant epitope bearing a T at IE1₂₀₁.

FIG. 1.

Cytotoxic activity of HLA-B^*1801-restricted T-cell clones C5 to C8 against target cells U373 MG/IE1-1(KI), U373 MG/IE1-2(RM), U373 MG/IE1-1(NI), U373 MG/IE1-1(KM), U373 MG/IE1-1(RI), and U373 MG/IE1-1(TI), which stably express IE1 proteins that carry natural mutations listed in Table 1 in epitope IE1_199-206. Relative antigenic activities of the recombinant target cells were measured by a standard ⁵¹Cr release assay at E/T ratios of 5:1, 10:1, 20:1 and 40:1. U373 MG cells were used as a negative control. Results were reproducible in three independent experiments.

Homologous virus variants (3, 5, 11), as well as heterologous viruses carrying cross-reactive epitopes (2, 10, 22, 23, 26), may stimulate the memory CD8⁺ T-cell repertoire previously generated against one virus. Consistent with these data, T-cell repertoire directed at a first virus may confer a protective immunity against viral variants (3, 11) and unrelated cross-reactive pathogens (2, 23), and selective amplification of cross-reactive T-lymphocyte subpopulations for variants of homologous (5) and heterologous viruses (2) influences immunodominance hierarchies. HCMV healthy carriers are frequently infected with multiple HCMV strains (13), and consequently, their immune systems may ensure protection against homologous viruses carrying antigenic polymorphism. Our results (showing that HLA-B^*1801-restricted memory CTL clones focused on the IE1_199-206 epitope are able to cross-react with most of the IE1 recombinant proteins bearing mutations encountered within clinical strains) are consistent with these data and with results obtained in the HLA-A^*0201 context (15). They suggest a selection process in which T cells that recognize most of the IE1 variant epitopes are amplified.

Clone C5, which corresponded to 10/21 clones in the HLA-B^*1801 subpopulation sorted from donor C and recognizing the IE1_199-206 (ELKRKMIY) epitope (20), was the only clone that reacted with IE1-2(RM) or IE1-2(NI) variants (Fig. 1). Therefore, specificity analyses based on the sequence of a single viral strain may considerably underestimate the magnitude of responses to a polymorphic epitope of an HCMV antigenic protein.

Cytotoxic responses of T-cell clones C5 to C8 were measured for U373 MG cells pulsed with various concentrations of synthetic IE1_199-206 octapeptides P(KI), P(RM), P(NI), P(KM), P(RI), and P(TI) (Genosys, Pampisford, United Kingdom) and carried amino acid combinations KI, RM, NI, KM, RI, and TI, respectively, at positions 3 and 7 (Fig. 2). Consistent with experiments performed with target cells expressing mutant IE1 proteins, P(KI) was the only one mutant peptide that was efficiently recognized by all four T-cell clones at 10⁻⁶ M and the peptide P(TI) did not stimulate any of the four T-cell clones. Unexpectedly, clones C5 to C8 did not respond or poorly responded to synthetic peptides P(RM), P(NI), P(KM), and P(RI), which corresponded to IE1 variants efficiently recognized when naturally processed in U373 MG cells. Similar reactivity patterns were obtained when cells of the autologous B-lymphoblastoid cell line pulsed with the octapeptides were used as target cells (data not shown). Although we cannot completely exclude the possibility that peptides different in length or modified in their structure are presented by HLA-B^*1801 molecules and recognized by CTL clones when IE1 proteins are intracellularly processed, our data strongly suggest that except for P(KI), all of the 8-mer IE1_199-206 peptides had a very low affinity for HLA-B^*1801 molecules and were not able to displace efficiently endogenous peptides complexed with HLA molecules. Whatever the explanation may be, such an observation raises questions about the interpretation and functional relevance of data exclusively drawn from screening of CTL responses to synthetic peptides (6-8).

FIG. 2.

Cytotoxic responses of HLA-B^*1801-restricted T-cell clones C5 to C8 against U373 MG target cells pulsed with the 8-mer synthetic peptide P(KI), P(RM), P(NI), P(KM), P(RI), or P(TI) that carried the natural mutations listed in Table 1. U373 MG cells were loaded with increasing peptide concentrations at an E/T ratio of 20:1. Their relative antigenic activities were measured by a standard ⁵¹Cr release assay. Results were reproducible in three independent experiments.

Our data might be interpreted as a special feature of donor C, and it would be necessary in further studies to determine whether they are broadly applicable to the HCMV⁺ HLA-B^*1801 subpopulation. Nevertheless, they strongly suggest that cross-reactive HLA-B^*1801-restricted CTL responses targeted to the IE1_199-206 immunodominant epitope are sufficient to confer cellular immune protection against most natural HCMV variants and should have implications in vaccinal strategies.