Search tips
Search criteria 


Logo of jcinvestThe Journal of Clinical InvestigationCurrent IssueArchiveSubscriptionAbout the Journal
J Clin Invest. 2005 December 1; 115(12): 3602–3612.
Published online 2005 November 23. doi:  10.1172/JCI25078
PMCID: PMC1288832

Cross-reactive influenza virus–specific CD8+ T cells contribute to lymphoproliferation in Epstein-Barr virus–associated infectious mononucleosis


The marked proliferation of activated CD8+ T cells is pathognomonic of EBV-associated infectious mononucleosis (IM), common in young adults. Since the diversity and size of the memory CD8+ T cell population increase with age, we questioned whether IM was mediated by the reactivation of memory CD8+ T cells specific to previously encountered pathogens but cross-reactive with EBV. Of 8 HLA-A2+ IM patients, 5 had activated T cells specific to another common virus, as evidenced by a significantly higher number of peripheral blood influenza A virus M158–66–specific T cells compared with healthy immune donors. Two patients with an augmented M1 response had tetramer-defined cross-reactive cells recognizing influenza M1 and EBV-BMLF1280–288, which accounted for up to one-third of their BMLF1-specific population and likely contributed to a skewed M1-specific T cell receptor repertoire. These epitopes, with only 33% sequence similarity, mediated differential effects on the function of the cross-reactive T cells, which may contribute to alterations in disease outcome. EBV could potentially encode an extensive pool of T cell epitopes that activate other cross-reactive memory T cells. Our results support the concept that cross-reactive memory CD8+ T cells activated by EBV contribute to the characteristic lymphoproliferation of IM.


There is a high degree of individual variation in disease severity associated with human virus infections, and age is one of many factors that can contribute to such variation. Childhood infection with EBV is often subclinical while the same infection is frequently symptomatic in adolescents and adults and presents as infectious mononucleosis (IM). The 3 classic criteria for IM diagnosis are the following: (a) lymphocytosis, the marked expansion of lymphocyte numbers in the peripheral blood caused by the proliferation of EBV-specific CD8+ T cells; (b) clinical symptoms that include fever, pharyngitis, and lymphadenopathy; and (c) a positive serologic test (1, 2). IM can vary in duration from a few weeks to 6 months, and the symptoms can vary in severity (3). Complications, such as pneumonia and fulminant hepatitis, are more common in older adults and have been linked to the infiltration of activated T cells and EBV-infected B cells into these tissues (46). When comparing IM and asymptomatic cases of acute EBV infection, Silins et al. found that the magnitude of the CD8+ T cell response, not viral load, correlated with the presence of disease (7). Furthermore, treatment of IM patients with antiviral drugs, although decreasing viral load, did not have any effect on the disease course (8, 9). These data suggest that a massive CD8+ T cell response can be counterproductive and mediate the disease pathology. It is still unclear why this massive CD8+ T cell proliferation occurs more frequently in older individuals.

Based on animal models of heterologous immunity that showed that T cells specific to a previously encountered virus may enhance immunopathology during a second, unrelated virus infection and based on the increasing number of reports documenting virus-specific CD8+ T cell cross-reactivity, we hypothesized that cross-reactive memory T cells specific to previously encountered pathogens contribute to the lymphoproliferation characteristic of EBV-associated IM (1015). In support of this, there is well-documented evidence that at least a proportion of the CD8+ T cells activated by EBV can have alternative specificities for allogeneic MHC molecules, self peptides, and bacterial antigens (1619). Testing this hypothesis is challenging due to individual variation in HLA allele expression, the history of infections, and the private specificity of the responding T cell repertoire (2022).

In order to optimize our chances of detecting whether cross-reactive T cells were contributing to the EBV-induced CD8+ T cell response during IM, we focused our studies on individuals with a common HLA allele, A*0201, and their responses to a commonly encountered virus, influenza A. The majority of the world’s population, starting at a young age, are repeatedly infected with influenza virus A. In almost all HLA-A2+ individuals, the CD8+ T cell response to influenza A virus is focused on an immunodominant epitope, M158–66, which is derived from the matrix protein (23). This protein is well conserved among different virus strains, ensuring the maintenance of M1-specific T cells in an individual’s memory pool. While there is significant individual variation regarding which EBV antigen(s) drive a dominant CD8+ T cell response, we focused on the response to BMLF1280–288, an immunodominant epitope within an early lytic protein that is also consistently recognized by all HLA-A2+ individuals during the acute response (2426). In this study, we detected a cross-reactive response with specificity for these 2 dissimilar epitopes, influenza M1 and EBV-BMLF1, in bulk T cell cultures and at the clonal level, and we demonstrate that these cross-reactive cells participate in the characteristic CD8+ T cell–mediated pathology observed during acute IM.


Maintenance of cross-reactive CD8+ T cells in the memory pools of healthy donors.

Using ex vivo tetramer staining, we found a small subset of T cells in a healthy immune donor that recognized both influenza M1 (GILGFVFTL) and EBV-BMLF1 (GLCTLVAML) epitopes (Figure (Figure1A).1A). In order to further verify this observation and enhance the detection of cross-reactive T cells, we cultured CD8+ T cell lines in the presence of either M1 or BMLF1 peptide, and their specificity was then tested with an intracellular stain for the production of IFN-γ. As expected, a large proportion (33%) of an M1-specific T cell line derived from donor D-002 was able to produce IFN-γ following M1 stimulation (Figure (Figure1B).1B). However, a subset (3%) of this same M1-specific line were able to produce a high level of IFN-γ following BMLF1 stimulation (Figure (Figure1B).1B). The IFN-γ production in response to BMLF1 was considered antigen specific because stimulation with HLA-A2–presented peptides derived from HIV-gag or tyrosinase, a self antigen, resulted in very little IFN-γ production. In the case of donor D-002, a putatively cross-reactive subset of cells was also detected within the BMLF1-specific CD8+ T cell line (Figure (Figure1C).1C). Not only did the majority (59%) of the line produce IFN-γ following BMLF1 stimulation, but at least 1% of the line produced a low level of IFN-γ following M1 stimulation. Table Table11 summarizes the results of several intracellular IFN-γ stains on both M1- and BMLF1-specific T cell lines derived from 8 healthy donors with previous exposure to EBV and, presumably, influenza A virus. Using this technique, we were able to detect cytokine production to the heterologous peptide over background levels in cell lines from 3 out of 8 healthy donors.

Figure 1
T cell lines grown in the presence of 1 peptide can respond to stimulation with a second unrelated peptide. (A) CD8+ T cells were isolated ex vivo from healthy donor D-002 and costained with M1- and BMLF1-loaded tetramers; 106 events were collected. ...
Table 1
CD8+ T cell lines from multiple healthy donors responding to M1 and BMLF1 stimulation

The observation that a greater proportion of the M1-specific population compared with the BMLF1-specific population were cross-reactive T cells in the blood (5% of M1-tetramer+ cells versus 0.3% of BMLF1-tetramer+ cells, Figure Figure1A)1A) and in culture (3% of an M1-specific line versus 1% of a BMLF1-specific line, Figure Figure1,1, B and C) prompted us to compare the relative avidities of the cross-reactive interaction with M1 versus BMLF1. To estimate TCR avidity, we performed an intracellular IFN-γ assay, using a peptide titration, on an M1-specific T cell line derived from donor D-002. A concentration of 5 × 10–8 M of M1 peptide compared with 10–7 M of BMLF1 peptide resulted in about half of the maximum amount of IFN-γ produced by this M1-specific T cell line (effective concentration eliciting 50% of the maximum response [EC50]) (Figure (Figure1D).1D). Thus, the avidity for these 2 epitopes was slightly different, but both were within the avidity range previously reported for M1-specific and BMLF1-specific T cell clones using IFN-γ production or cytotoxicity as the readout (27, 28). Although the stronger M1 stimulus may have been more efficient at promoting the growth of cross-reactive cells, it is likely that the cross-reactive population growing in the M1-specific T cell line is composed of a distinctly different subset of T cell clones than that growing in a BMLF1-specific T cell line. In support of this, the same peptide titration assay was performed using the BMLF1-specific T cell line, and this resulted in a cross-reactive population with a much greater discrepancy between the avidities for these 2 epitopes. Interestingly, the EC50 in response to the cross-reactive M1 stimulation was 10–9 M compared with 5 × 10–7 M in response to BMLF1 stimulation (Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/JCI25078DS1). Thus, the cross-reactive cells grown in the presence of only BMLF1 peptide had an even higher avidity for the M1 epitope than did the cross-reactive cells grown in the presence of M1 peptide. The stimulating antigen used in culture appeared to drive differences in the clonal composition of M1- and BMLF1-specific T cell lines derived from the same donor, and this likely explains the lack of reciprocity in the frequency of the cross-reactive cells found within them.

Cross-reactive stimulation also resulted in the production of macrophage inflammatory protein-1β (MIP-1β), an antiviral C-C (β) chemokine. Nearly every cell of an M1-specific line derived from donor D-002 was able to produce some level of MIP-1β following stimulation with BMLF1 but, importantly, not in response to stimulation with HIV-gag or tyrosinase (Figure (Figure1B).1B). The production of MIP-1β, therefore, appeared to be more sensitive than the production of IFN-γ for the measurement of cross-reactivity in lines that had IFN-γ–producing cross-reactive T cells. In cell lines that did not have IFN-γ–producing cross-reactive T cells, no MIP-1β was produced in response to cross-reactive stimulation, suggesting that the protocol used here did not support the outgrowth of cross-reactive cells from the CD8+ T cell populations of donors D-046 and D-048 and that the production of MIP-1β was specifically induced by TCR engagement with antigen (Table (Table1).1). Thus, MIP-1β production was a more sensitive measurement of cross-reactivity than IFN-γ production, but this required cross-reactive cells to be present. There was no detectable cross-reactive MIP-1β production in the BMLF1-specific line derived from donor D-002, which again likely reflects differences in the clonotypic composition of this BMLF1-driven T cell line compared with the M1-driven T cell line (Figure (Figure11C).

Breadth and quality of cross-reactivity revealed through alternative techniques.

The growth of cross-reactive cells specific to both M1 and BMLF1 improved when we cultured CD8+ T cells with both peptides simultaneously. The frequency of cells that costained with both M1- and BMLF1-loaded tetramers increased to a range of 0.3–1.1%. The cross-reactive cells that bound both tetramers were also able to respond functionally to both epitopes. They produced MIP-1β, IFN-γ, and TNF-α specifically following either M1 or BMLF1 stimulation, but BMLF1 stimulation appeared to result in a more robust production of all 3 cytokines (Figure (Figure2A).2A). However, a peptide titration assay revealed that these M1+ BMLF1+ cross-reactive T cells actually had a slightly higher avidity for the M1 epitope. The EC50 of the cross-reactive response to the M1 peptide was 10–8 M compared with 10–7 M in response to the BMLF1 peptide (Figure (Figure3A).3A). The more robust functional response to BMLF1 initially observed using a 5 μM concentration of peptide appeared to be an effect of significant TCR downregulation, which decreased the sensitivity of M1-tetramer binding.

Figure 2
Culturing with M1 and BMLF1 peptides simultaneously promotes the growth of cross-reactive cells. A CD8+ T cell line derived from healthy donor D-002 was grown for 4 weeks in the presence of both M1 and BMLF1 peptide–pulsed T2 cells. Cells ...
Figure 3
Tetramer-defined subsets of cross-reactive T cells differ in their avidity for the 2 epitopes. A similar intracellular IFN-γ assay was performed on the same T cell line described in Figure Figure2,2, which had been grown in the presence ...

The tetramer-based frequency of cross-reactive cells within this T cell line was lower than the frequency based on function. The subset of cells only able to bind the M1-loaded tetramer produced MIP-1β and IFN-γ but very little TNF-α in response to BMLF1 stimulation (Figure (Figure2B).2B). The cells that bound only BMLF1-loaded tetramer also showed some degree of functional cross-reactivity. At least 35% of the BMLF1+ cells produced a low level of MIP-1β in response to M1 stimulation although this cross-reactive stimulation was not as efficient at inducing IFN-γ or TNF-α production (Figure (Figure2C).2C). A separate peptide titration experiment on this single BMLF1 tetramer–positive subset revealed that the cross-reactive cells present in this subpopulation had a similar avidity for the M1 and BMLF1 epitopes, where the EC50 was 5 × 10–7 M of either peptide (Figure (Figure3).3). Overall, it would appear that how a cross-reactive T cell interacts with its alternative ligand is highly variable and that cross-reactive T cell populations are indeed heterogeneous. Hence, multiple techniques are required to detect T cell cross-reactivity, including tetramer staining and different functional assays. As shown here, TCR avidity is an important factor to consider when detecting cross-reactive T cell responses. An interaction between a cross-reactive T cell and its alternative ligand may be too weak to stably bind tetramer but may still be sufficient to induce a distinct hierarchy of cytokine production. This is analogous to the observation that certain non–cross-reactive influenza M1–specific clones were unable to bind M1-loaded tetramers but produced IFN-γ following M1 peptide stimulation (29).

Cross-reactivity at a clonal level.

We next sought to clone these cross-reactive cells from a polyclonal T cell line, using the experimental design outlined in Figure Figure4A.4A. Briefly, we allocated single T cells that costained with both M1- and BMLF1-loaded tetramers into microwells. The single cells were propagated for 2 weeks (referred to as clones from here on) and then assessed for functional specificity. As expected from the T cell line data, there was tremendous variability in the functional characteristics of each clone in response to either antigen. Of all the clones that grew, 8% produced IFN-γ following stimulation with either M1 or BMLF1 (Figure (Figure4B).4B). Similarly, 11% of the different clones analyzed killed both M1- and BMLF1-pulsed target cells in a 51chromium release cytotoxicity assay (Figure (Figure4C).4C). The number of functionally cross-reactive clones varied with the technique used for their detection and reflected a similar ratio of BMLF1-responders to M1-responders (4:1) as seen in the assessment of IFN-γ production by the double-tetramer+ population within a polyclonal T cell line (1.5:1) (Figure (Figure2A).2A). These results definitively show that individual T cell clones can recognize and respond to both M1 and BMLF1, 2 epitopes that share little sequence similarity.

Figure 4
Cross-reactive clones are heterogeneous in their response to M1 versus BMLF1. (A) An outline of the experimental design used to clone cross-reactive T cells from healthy donor D-002 is shown. (B) CD8+ T cells incubated for 20 hours with either ...

Cross-reactive cells participate in the lymphoproliferation that defines IM syndrome.

Since we were able to detect cross-reactive cells that recognize M1 and BMLF1 in bulk culture and at a clonal level, we next determined whether these cross-reactive cells participated in the overzealous CD8+ T cell response that defines EBV-associated IM. We noticed that, despite the large expansion of EBV-specific cells, the frequency (as percentage of CD8+ T cells) of M1-specific cells in IM patients (mean, 0.20%; range, 0.02–0.49%) was similar to that in healthy influenza-immune donors (mean, 0.26%; range, 0.09–0.79%). The maintenance of a resting-state frequency suggested that at least a subset of M1-specific cells were proliferating in response to infection because non–cross-reactive memory cells should be diluted out by the proliferation of virus-specific cells. In fact, the average number of M1-specific cells (per ml of blood) was a very significant 4-fold higher in IM patients (0.004 × 106 per ml) compared with healthy donors (0.001 × 106 per ml) (P = 0.02) (Figure (Figure5A).5A). Of 8 IM patients, 5 had a higher than average number of M1-specific cells at presentation, and this number decreased over the course of the infection with contraction kinetics similar to the BMLF1 response and overall lymphoproliferation (Figure (Figure5,5, B–D). These observations suggested that the M1-specific population within each of these 5 IM patients contained T cells that were cross-reactive with an EBV-derived antigen.

Figure 5
IM patients have an augmented number of M1-specific cells in their bloodstream. PBMCs were isolated from 8 healthy donors or from 8 patients experiencing IM. Blood from IM patients was collected at various points after presentation with symptoms of IM. ...

We next costained freshly isolated CD8+ T cells from IM patients with M1- and BMLF1-loaded tetramers. Double-positive cross-reactive cells were prominent in 2 patients, E1101 and E1178 (Figure (Figure6B6B and Supplemental Figure 2). The percentage as well as the total number of cross-reactive cells shifted with this active infection, including a considerable increase at day 22 (0.003 × 106 M1+ BMLF1+ cells per ml of blood) after presentation of patient E1101’s clinical symptoms (Figure (Figure6,6, A and B). This translated to as many as 1/3 of the T cells specific to the immunodominant BMLF1 epitope being cross-reactive with M1. Fewer tetramer-defined cross-reactive T cells were detected during patient E1178’s infection, but there was a discernable increase in frequency at days 12 and 34 (0.0002 × 106 M1+ BMLF1+ cells per ml of blood) after presentation with IM (Supplemental Figure 2). These data support our hypothesis that cross-reactive T cells contribute to lymphoproliferation during IM. Although the BMLF1-specific T cell population represented only a minor proportion (0.4–1.9%) of the total CD8+ T cell pool in the blood, it remains possible that additional EBV-derived antigens simultaneously activate cross-reactive memory T cells with specificity for antigens other than influenza M1. Interestingly, both of these patients presented with symptoms of IM but with differences in severity. Patient E1101, who had the higher frequency of M1+ BMLF1+ cross-reactive cells, presented with severe (grade 5 on a scale of 1–5) symptoms and signs of IM, including fatigue, sweats, chills, sore throat, nausea, myalgia, lymphadenopathy, pharyngitis, and stomatitis. Notably, reversion of the CD4:CD8+ T cell ratio, a hallmark of an active viral infection, was not evident until day 34 after presentation. In contrast, patient E1178, who had a lower frequency of cross-reactive cells, presented with only moderate (grade 2–3) symptoms and signs of IM, including fatigue, loss of appetite, and only mild hepatosplenomegaly. For this patient, CD4:CD8+ reversion was observed on days 0 and 6, with a second episode of reversion again at day 27 after presentation, suggesting fluctuations in disease course. More patients will need to be studied to determine whether the correlation of cross-reactivity with M1 and disease severity will hold, but here we document high levels of cross-reactive T cells associated with severe IM pathology.

Figure 6
Acute EBV infection augments the number of cross-reactive cells that recognize M1 and BMLF1. CD8+ T cells were isolated ex vivo from patient E1101 at various time points after presentation with symptoms of IM. (A) The total number of antigen-specific ...

Skewing of the M1-specific Vβ17+ TCR repertoire during IM.

The M1-specific memory TCR repertoire is organized in a conserved pattern, as unique TCR subclones using the variable (Vβ) gene family Vβ17 can be ordered in a distinct hierarchy based on their joining gene (Jβ) usage, where Jβ2.7 dominates (used by 55–62% of subclones), followed by Jβ2.3 (used by 10–20% of subclones), and then, often, Jβ2.1, Jβ2.5, Jβ1.1, or Jβ1.2 at lower, more variable frequencies (Y.N. Naumov et al., unpublished observations). We investigated whether the M1-specific repertoires of E1101 and E1178 were skewed from this pattern and thereby reflective of cross-reactive TCR-mediated clonal expansions. Due to limited blood samples, we were unable to sort and sequence the TCRs of the M1-specific cells directly ex vivo. Rather, we generated M1-specific T cell lines from these IM patients, and Vβ analyses indicated that their M1-specific repertoires were focused on the Vβ17 family, similar to that previously described for healthy individuals (data not shown) (3033). However, when we sequenced the Vβ17+ subclones within the cell lines derived from both patients, they did not follow the highly conserved organizational pattern observed in healthy influenza-immune donors. At day 22 after presentation, the time point when the number of cross-reactive T cells was highest ex vivo, the Jβ2.3 family was overrepresented (30%) while the normally dominant Jβ2.7 family was vastly underrepresented (10%) within the M1+ Vβ17+ repertoire of patient E1101 (Figure (Figure6,6, B and C). The skewing of the repertoire was even more pronounced at day 165 (Jβ2.3, 37%; Jβ2.7, 3%), again a time when cross-reactive cells could easily be detected ex vivo (Figure (Figure6,6, B and C). When the M1-specific population was presumably more stable at 1 year after presentation, the repertoire was still skewed from that typical of a resting memory state but appeared to be slowly reverting, as the number of subclones using Jβ2.3 (25%) declined and those using Jβ2.7 (14%) were better represented (Figure (Figure6C).6C). Similarly, the M1-specific repertoire of patient E1178 was skewed during the acute phase of the immune response to EBV but in a way that was different from that of patient E1101. In the case of patient E1178, we found that the Jβ1.2 family was overrepresented within the M1+ Vβ17+ repertoire (50% at days 19 and 174), illustrating the point that a crossreactive TCR repertoire may be unique to each individual and therefore not easily predicted (Supplemental Figure 2). This perturbation in the influenza M1–specific TCR repertoire of both patients with IM is highly consistent with the concept that cross-reactive M1-specific T cells are expanded in the host during acute EBV infection in an antigen-driven manner by EBV-derived epitopes such as BMLF1.


We show here that cross-reactive T cells specific to a previously encountered virus could be major contributors to the overzealous CD8+ T cell response that defines IM. In 5 out of 8 patients, influenza M1–specific CD8+ T cells participated in EBV-induced lymphoproliferation. Of these 5 patients, 2 had dramatically skewed M1-specific TCR repertoires and increased levels of clearly identifiable, tetramer-defined, cross-reactive CD8+ T cells capable of recognizing the 2 dissimilar epitopes influenza M1 and EBV-BMLF1. Based on our ability to culture these cross-reactive cells from 3 out of 8 healthy donors with previous exposure to both viruses, these cross-reactive cells are maintained in memory and their functional responses to either antigen can include cytotoxicity and the production of MIP-1β, IFN-γ, and TNF-α. Cross-reactive T cells may play a major role in the development of IM, and the diversity of their functions may contribute to the severity of the syndrome. These studies examined only one cross-reactive population while it is likely that infection with EBV, a virus with the potential to encode hundreds of epitopes, could reactivate many memory T cell populations yet undefined. As demonstrated here, the identification of cross-reactive T cells can be complicated by their ability to recognize alternative peptides having little sequence similarity to their native ligand, their strict growth requirements in vitro, and the sensitivity of the different techniques used to detect them. These are all challenges for future elucidations of individual cross-reactive T cell responses and their potential impact on the outcome of IM.

Our work suggests that acute EBV infection activated influenza M1–specific CD8+ T cells. These cross-reactive M1-specific T cells are most likely memory cells for the following reasons: (a) by this age, everyone is immune to influenza A virus; (b) almost all HLA-A2+ individuals develop an M1-specific response; and (c) costaining with M1- and BMLF1-loaded tetramers showed that these cross-reactive cells brightly stained with the M1-loaded tetramer, suggestive of a high avidity interaction typical of antigen-specific memory. The activation of cross-reactive memory cells coincident with the development of IM disease pathology is highly analogous to the examples of heterologous immunity we have observed in mouse models, where memory T cell responses to a prior infection with an unrelated virus altered the host’s immune response to a subsequent infection and caused a marked deviation in disease course (34).

It has recently been shown that acute HIV infection can up-regulate the expression of activation markers such as CD38, HLA-DR, and Ki67 on memory cells specific to influenza A virus, EBV, and CMV, but the role of the TCR in this activation was not determined (35). Our work suggests that acute EBV infection can activate influenza-specific memory cells through a TCR-dependent mechanism. The expansion of M1-specific memory cells was evident in only 5 of 8 patients with IM despite the fact that all probably had memory to M1 and all would have been influenced by any cytokine-mediated, or bystander, activation. When possible, we also looked for the expansion of a second memory T cell population, specific for CMV-pp65. Only 2 IM patients proved to be CMV seropositive, E1155 and recent enrollee E1238. The frequency of pp65-specific T cells in patients E1155 and E1238 was low during massive, EBV-induced lymphoproliferation. At day 0, E1155 and E1238 had a pp65-specific T cell frequency of 0.2% and 0.7% respectively while, by 41–50 days after presentation, those frequencies climbed to 0.6% and 1.1% respectively. These data would suggest that the pp65-specific memory populations of these 2 patients did not contain T cell clones cross-reactive with EBV and were therefore initially diluted out by the extensive proliferation of EBV-specific T cells.

For 2 of the 5 patients with higher M1 frequencies (E1101 and E1178), we determined that BMLF1 was at least 1 of the EBV-derived antigens recognized by cross-reactive T cells. Although both patients shared this particular pattern of cross-reactivity, their responses remained unique. Vβ17+ subclones using Jβ2.3 preferentially expanded in patient E1101 while those using Jβ1.2 preferentially expanded in patient E1178. Thus, the M1-specific TCR repertoire of both patients was notably skewed from that known to be conserved among healthy HLA-A2+ individuals (Y.N. Naumov et al., unpublished observations). These data are further suggestive of antigen-driven clonal expansions because a bystander activation mechanism would drive the expansion of all clones and would maintain the conserved repertoire organization. The fact that cross-reactive T cells specific for M1 and BMLF1 were observed in 2 but not all patients is probably reflective of these clonal differences, known as the private specificity of each individual TCR repertoire. In support of this, cross-reactive T cell responses involving lymphocytic choriomeningitis virus (LCMV) and vaccinia virus (VV) showed that only 50% of VV-challenged LCMV-immune mice mounted a strong response to a specific LCMV epitope, NP205–213, and this was a function of the private specificity of the LCMV memory TCR repertoire of each mouse (22). The existence of T cell responses unique to the individual adds to the complexity in resolving the importance of cross-reactive responses in human infections.

Previous studies showed that the extent of T cell activation and proliferation correlated with the severity of IM (7, 36). We suggest that the magnitude of the T cell response represents the combined efforts of cross-reactive memory and primary T cells. Our limited clinical data indicate that the number of cross-reactive cells specific for M1 and BMLF1 may correlate with disease severity, but further investigation is necessary to confirm this observation. The severity of IM might be influenced by both the number of activated cross-reactive cells and their effectiveness at clearing virus. There is recent evidence for the exacerbation of human disease by the activation of cross-reactive influenza-specific memory CD8+ T cells during acute HCV infection. Like EBV, acute HCV infection is often asymptomatic, but when clinical symptoms manifest, they are likely caused by the immune response. HCV encodes an epitope, NS31073–1081, that can activate influenza NA231–239–specific memory cells (14). However, unlike EBV-BMLF1 and influenza M1, which are only 33% similar in sequence, HCV-NS3 and influenza NA are 78% similar in sequence. In congruence with our observations during EBV infection, the activation of these cross-reactive NA-specific memory cells enhanced the magnitude of the CD8+ T cell response to HCV and resulted in severe disease pathology (37). Despite strength in numbers, these HCV-specific T cells were unable to sufficiently clear the virus, and the patients developed persistent HCV infections (37). Cross-reactive cells may be inefficient at interacting with infected host cells due to a low avidity for the alternative ligand, but they may still produce pathology-generating cytokines. Their presence may interfere with the primary T cell response by preventing access to the infected cells or changing patterns of T cell immunodominance, thereby prolonging resolution of the infection.

Our work in animal models has suggested that cross-reactive T cells can induce pathological conditions despite their ability to clear virus. LCMV-specific memory cells lower the titer of VV delivered intranasally but in doing so also alter the disease pathology from pulmonary edema to bronchiolitis obliterans (34, 38, 39). The cytokines, notably IFN-γ, secreted by these activated memory cells may have played a major role in the development of this immunopathology. In the present study, we showed that cross-reactive cells specific to M1 and BMLF1 secreted several cytokines in a hierarchal pattern whereby most secreted MIP-1β, fewer secreted IFN-γ, and even fewer secreted TNF-α. Similar functional hierarchies were observed in studies with freshly isolated HIV- or CMV-specific CD8+ T cells from healthy donors with these persistent infections as well as during the primary immune response to EBV in IM patients (4042). The mechanism behind T cell functional heterogeneity has been extensively studied by varying the quantity of TCRs engaged and by varying the quality of ligands used to engage them (28, 4345). Our work presents an opportunity to apply this knowledge to T cell cross-reactivity involving 2 natural viral epitopes in order to understand how cross-reactive T cells mediate the development and severity of IM. Interestingly, both MIP-1β and TNF-α levels are known to be elevated in the serum and tonsils of IM patients compared with healthy controls (46, 47). MIP-1β can be readily secreted because it is preformed and stored within human CD8+ T cells (48). Even a low avidity interaction may stimulate its release from the cell. This is logical because MIP-1β broadens the immune response by recruiting other immune cells to the site of infection (49). The increase in number of activated immune cells may enhance IM severity. Fewer of the cross-reactive cells secreted TNF-α following peptide stimulation, suggesting that a higher avidity interaction may be required to initiate its production. However, an overall increase in the number of responding T cells capable of secreting TNF-α at the site of infection could be harmful to the host and promote the clinical symptoms of IM (50, 51). In fact, the high production of TNF-α, possibly by cross-reactive memory cells, has been implicated in the immunopathogenesis associated with dengue hemorrhagic fever (52). Thus, depending on the cross-reactive specificity pattern and private specificity of the TCR repertoire, cross-reactive memory T cells activated by EBV may function and modulate the disease outcome of each individual very differently. Identifying these cross-reactive patterns will be a challenge for the future.

In conclusion, our data suggest that cross-reactive memory T cells participate in the massive lymphoproliferation that characterizes EBV-associated IM and may influence disease severity. For the purposes of this study, we focused on detection and activation of M1-specific memory cells, but EBV, a virus with a large genome that encodes numerous different proteins, has the potential to generate epitopes reactive with many different memory T cell populations. The cross-reactive pattern that emerges is influenced both by an individual’s unique history of infection and by the private specificity of the TCR repertoire responding to each of those infections. Overall, this demonstration of cross-reactivity involving 2 immunodominant epitopes derived from 2 of the most common human viruses among people that share the most common MHC class I haplotype in North America highlights the potential importance of cross-reactive T cells in human disease states.



Influenza A virus–immune patients between the ages of 18 and 23 with acute EBV infection were volunteers from the University of Massachusetts (UMass) Student Health Services (Amherst, Massachusetts, USA). HLA typing was performed using the Lymphotype Class I system (Biotest) and an Olerup SSP kit (GenoVision). Acute EBV infection was confirmed by a monospot test and the detection of capsid-specific IgM in patient sera. Positive staining with HLA-A2 tetramers loaded with influenza M1 was used as an indication that these individuals had been exposed to influenza A virus in the past. Patients provided up to 8 blood samples (50 ml each) starting at presentation with IM (day 0), then weekly for the following 6 weeks, and then again at 1 year. Healthy donors between the ages of 24 and 50 were volunteers from the research community at UMass Medical School (Worcester, Massachusetts, USA). HLA status and immunity to EBV and influenza A virus were assessed using monoclonal antibody (BB7.2; BD) and tetramer stains, respectively. Previous exposure to EBV was confirmed by the detection of capsid-specific IgG in donor sera. Donors provided up to 3 blood samples (60 ml each). This study was approved by the Human Studies Committee at UMass Medical School, and all subjects participating in our study gave informed consent.

Blood preparation and bulk T cell culture.

PBMCs were isolated using Ficoll-Paque plus (Amersham Biosciences) and were stained with anti-CD8+ microbeads before being positively selected using the Miltenyi Biotech MACS system. CD8+ lymphocytes were plated at 2.5 × 105 per ml together with peptide-pulsed (1 μM), irradiated (30 Gy) T2 cells (CRL-1992; ATCC) at 5 × 104 per ml in 4 ml total volume per well of a 12-well plate. T cell lines were fed media (AIM-V [Gibco; Invitrogen Corp.] supplemented with 14% human AB serum [Nabi Biopharmaceuticals], 16% MLA-144 supernatant [ref. 53], 10 U/ml rIL-2 [BD], 1% L-glutamine [Gibco; Invitrogen Corp.], 0.5% β-mercaptoethanol [Sigma-Aldrich], and 1% HEPES [HyClone]) every 3–4 days and were restimulated with T2 cells weekly.

HLA-A2–restricted peptides.

The following peptides were synthesized to greater than 90% purity by Biosource: EBV-BMLF1280–288 (GLCTLVAML), influenza A virus-M158–66 (GILGFVFTL), tyrosinase (YMNGTMSQV), and HIV-gag77–85 (SLYNTVATL).

MHC class I tetramers.

A detailed description of the protocol used by the tetramer facility at UMass Medical School has been previously published (25). Tetramers were assembled using the above peptide sequences for EBV-BMLF1 and influenza M1 and were conjugated to PE (Sigma-Aldrich), APC (CALTAG Laboratories), or Quantum Red (Sigma-Aldrich). Tetramers assembled with HIV-gag or tyrosinase (Immunomics) were used as negative controls, and nonspecific staining was never observed.

Extracellular and intracellular staining.

Cells were plated at 106 per well and washed with staining buffer (PBS, 2% FCS, and 1% sodium azide). Tetramers were incubated at room temperature for 40 minutes and were washed off. Cells were either fixed with FACS Lysing Solution (BD) or permeabilized using Cytofix/Cytoperm (BD) according to the manufacturer’s instructions. The following monoclonal antibodies were used: anti–IFN-γ (B27; BD), anti–MIP-1β (D21-1351; BD), and anti–TNF-α (mAb11, eBioscience). Isotype control antibodies did not stain positive.

T cell cloning.

T cell lines were costained in 2% FCS/PBS with tetramers as described above. Double tetramer-positive cells were sorted using BD Vantage into 96-well plates at 1 cell per well. Each well contained 105 irradiated donor-specific CD4+ T cell blasts and 2 × 103 of a 1:1 mixture of irradiated T2 cells pulsed with BMLF1 or M1 peptides. Clones were fed media every 3–4 days and restimulated with T2 cells at day 7; their specificity was tested on day 14.


Our protocol was adapted from a previously published method using Mabtech reagents (54). The APCs were K562 cells stably transfected with HLA-A2.1 and pulsed with 50 μM peptide for 3 hours at 37°C, then washed of free peptide. Clones were tested using split-well analysis; 5 μl of each clone was loaded per well per peptide-loaded cell type. Plates were stained with streptavidin-horseradish peroxidase (Mabtech), diluted 1:100, and Nova Red Substrate (Vector Laboratories) according to the manufacturers’ instructions, then read manually.

51Chromium release assay.

K562/HLA-A2.1 cells (54) served as the targets and were pulsed with 100 μM peptide for 1 hour at 37°C and then for an additional hour with 100 μCi 51Cr per 106 cells. Targets were washed 3 times and plated at 5 × 103 per well. Clones were tested using split-well analysis; 30 μl of each clone was loaded per well per target type. Plates were incubated for 8 hours at 37°C, and the supernatants were analyzed with a MicroBeta TriLux scintillation counter (PerkinElmer).

Sequencing the TCR CDR3 β.

RNA was isolated from 106 cells derived from T cell lines using TRIzol (Invitrogen Corp.) according to the manufacturer’s instructions. Using 0.5–1 μg of RNA, cDNA synthesis was performed with 2 μM poly-T (Integrated DNA Technologies) and Superscript III reverse transcriptase (Invitrogen Corp.) according to manufacturer’s protocol. The cDNA was amplified with 1.6 μM each of Vβ17-specific (nomenclature of Arden et al.; ref. 55) and constant gene–specific (Cβ-specific) primers (56) using the protocol recommended by Applied Biosystems with AmpliTaq DNA polymerase. PCR products were ligated into the pCR4-TOPO vector (Invitrogen Corp.) and were used to transform TOP10 chemically competent cells (Invitrogen Corp.) according to the manufacturer’s protocol. Individual colonies were picked for overnight cultures. DNA was isolated with QIAprep miniprep kits (QIAGEN), and complementarity-determining region 3β (CDR3β) was sequenced at the UMass Nucleic Acid Facility (Worcester, Massachusetts, USA) using universal primers.

Supplementary Material

Supplemental data:


We thank Rose Cicarelli, Robin Brody, and Joyce Pepe for their technical assistance. This work is supported by a Worcester Foundation grant, Deutsche Forschungsgemeinschaft fellowship CO 310/1-1, and NIH grants AI-49320, AI-42845, and DR-32520 and immunology training grant 5 T32 AI-07349-16. The contents of this publication are solely the responsibility of the authors and do not represent the official view of the NIH.


Nonstandard abbreviations used: Cβ, constant region of TCR β-chain; CDR3, complementarity-determining region 3; EC50, effective concentration eliciting 50% of the maximum response; IM, infectious mononucleosis; Jβ, joining region of TCR β-chain; LCMV, lymphocytic choriomeningitis virus; MIP-1β, macrophage inflammatory protein-1β; Vβ, variable region of TCR β-chain; VV, vaccinia virus.

Conflict of interest: The authors have declared that no conflict of interest exists.


1. Hoagland RJ. Infectious mononucleosis. Prim. Care. 1975;2:295–307. [PubMed]
2. Young LS, Rickinson AB. Epstein-Barr virus: 40 years on. Nat. Rev. Cancer. 2004;4:757–768. [PubMed]
3. Rea TD, Russo JE, Katon W, Ashley RL, Buchwald DS. Prospective study of the natural history of infectious mononucleosis caused by Epstein-Barr virus. J. Am. Board Fam. Pract. 2001;14:234–242. [PubMed]
4. Auwaerter PG. Infectious mononucleosis in middle age. JAMA. 1999;281:454–459. [PubMed]
5. Axelrod P, Finestone AJ. Infectious mononucleosis in older adults. Am. Fam. Physician. 1990;42:1599–1606. [PubMed]
6. Rickinson, A.B., and Kieff, E. 2001. Epstein-Barr virus. In Fields virology. D.M. Knipe and P.M. Howley, editors. Lippincott Williams & Wilkins. Philadelphia, Pennsylvania, USA. 2575–2627.
7. Silins SL, et al. Asymptomatic primary Epstein-Barr virus infection occurs in the absence of blood T-cell repertoire perturbations despite high levels of systemic viral load. Blood. 2001;98:3739–3744. [PubMed]
8. Andersson J, et al. Acyclovir treatment in infectious mononucleosis: a clinical and virological study. Infection. 1987;15(Suppl. 1):S14–S20. [PubMed]
9. Torre D, Tambini R. Acyclovir for treatment of infectious mononucleosis: a meta-analysis. Scand. J. Infect. Dis. 1999;31:543–547. [PubMed]
10. Brehm MA, et al. T cell immunodominance and maintenance of memory regulated by unexpectedly cross-reactive pathogens. Nat. Immunol. 2002;3:627–634. [PubMed]
11. Mongkolsapaya J, et al. Original antigenic sin and apoptosis in the pathogenesis of dengue hemorrhagic fever. Nat. Med. 2003;9:921–927. [PubMed]
12. Nilges K, et al. Human papillomavirus type 16 E7 peptide-directed CD8+ T cells from patients with cervical cancer are cross-reactive with the coronavirus NS2 protein. J. Virol. 2003;77:5464–5474. [PMC free article] [PubMed]
13. Selin LK, Nahill SR, Welsh RM. Cross-reactivities in memory cytotoxic T lymphocyte recognition of heterologous viruses. J. Exp. Med. 1994;179:1933–1943. [PMC free article] [PubMed]
14. Wedemeyer H, Mizukoshi E, Davis AR, Bennink JR, Rehermann B. Cross-reactivity between hepatitis C virus and influenza A virus determinant-specific cytotoxic T cells. J. Virol. 2001;75:11392–11400. [PMC free article] [PubMed]
15. Welsh RM, Selin LK, Szomolanyi-Tsuda E. Immunological memory to viral infections. Annu. Rev. Immunol. 2004;22:711–743. [PubMed]
16. Burrows SR, Khanna R, Burrows JM, Moss DJ. An alloresponse in humans is dominated by cytotoxic T lymphocytes (CTL) cross-reactive with a single Epstein-Barr virus CTL epitope: implications for graft-versus-host disease. J. Exp. Med. 1994;179:1155–1161. [PMC free article] [PubMed]
17. Misko IS, et al. Crossreactive recognition of viral, self, and bacterial peptide ligands by human class I-restricted cytotoxic T lymphocyte clonotypes: implications for molecular mimicry in autoimmune disease. Proc. Natl. Acad. Sci. U. S. A. 1999;96:2279–2284. [PubMed]
18. Strang G, Rickinson AB. Multiple HLA class I-dependent cytotoxicities constitute the “non-HLA-restricted” response in infectious mononucleosis. Eur. J. Immunol. 1987;17:1007–1013. [PubMed]
19. Tomkinson BE, Maziarz R, Sullivan JL. Characterization of the T cell-mediated cellular cytotoxicity during acute infectious mononucleosis. J. Immunol. 1989;143:660–670. [PubMed]
20. Bousso P, et al. Individual variations in the murine T cell response to a specific peptide reflect variability in naive repertoires. Immunity. 1998;9:169–178. [PubMed]
21. Cibotti R, et al. Public and private V beta T cell receptor repertoires against hen egg white lysozyme (HEL) in nontransgenic versus HEL transgenic mice. J. Exp. Med. 1994;180:861–872. [PMC free article] [PubMed]
22. Kim SK, et al. Private specificities of CD8 T cell responses control patterns of heterologous immunity. J. Exp. Med. 2005;201:523–533. [PMC free article] [PubMed]
23. Gotch F, Rothbard J, Howland K, Townsend A, McMichael A. Cytotoxic T lymphocytes recognize a fragment of influenza virus matrix protein in association with HLA-A2. Nature. 1987;326:881–882. [PubMed]
24. Callan MF. The evolution of antigen-specific CD8+ T cell responses after natural primary infection of humans with Epstein-Barr virus. Viral Immunol. 2003;16:3–16. [PubMed]
25. Catalina MD, Sullivan JL, Bak KR, Luzuriaga K. Differential evolution and stability of epitope-specific CD8(+) T cell responses in EBV infection. J. Immunol. 2001;167:4450–4457. [PubMed]
26. Steven NM, et al. Immediate early and early lytic cycle proteins are frequent targets of the Epstein-Barr virus-induced cytotoxic T cell response. J. Exp. Med. 1997;185:1605–1617. [PMC free article] [PubMed]
27. Trautmann L, et al. Dominant TCR V alpha usage by virus and tumor-reactive T cells with wide affinity ranges for their specific antigens. Eur. J. Immunol. 2002;32:3181–3190. [PubMed]
28. Valitutti S, Muller S, Dessing M, Lanzavecchia A. Different responses are elicited in cytotoxic T lymphocytes by different levels of T cell receptor occupancy. J. Exp. Med. 1996;183:1917–1921. [PMC free article] [PubMed]
29. Lawson TM, et al. Functional differences between influenza A-specific cytotoxic T lymphocyte clones expressing dominant and subdominant TCR. Int. Immunol. 2001;13:1383–1390. [PubMed]
30. Lawson TM, et al. Influenza A antigen exposure selects dominant Vbeta17+ TCR in human CD8+ cytotoxic T cell responses. Int. Immunol. 2001;13:1373–1381. [PubMed]
31. Lehner PJ, et al. Human HLA-A0201-restricted cytotoxic T lymphocyte recognition of influenza A is dominated by T cells bearing the V beta 17 gene segment. J. Exp. Med. 1995;181:79–91. [PMC free article] [PubMed]
32. Moss PA, et al. Extensive conservation of alpha and beta chains of the human T-cell antigen receptor recognizing HLA-A2 and influenza A matrix peptide. Proc. Natl. Acad. Sci. U. S. A. 1991;88:8987–8990. [PubMed]
33. Naumov YN, Naumova EN, Hogan KT, Selin LK, Gorski J. A fractal clonotype distribution in the CD8+ memory T cell repertoire could optimize potential for immune responses. J. Immunol. 2003;170:3994–4001. [PubMed]
34. Welsh RM, Selin LK. No one is naive: the significance of heterologous T-cell immunity. Nat. Rev. Immunol. 2002;2:417–426. [PubMed]
35. Doisne JM, et al. CD8+ T cells specific for EBV, cytomegalovirus, and influenza virus are activated during primary HIV infection. J. Immunol. 2004;173:2410–2418. [PubMed]
36. Williams H, et al. Analysis of immune activation and clinical events in acute infectious mononucleosis. J. Infect. Dis. 2004;190:63–71. [PubMed]
37. Urbani S, et al. Heterologous T cell immunity in severe hepatitis C virus infection. J. Exp. Med. 2005;201:675–680. [PMC free article] [PubMed]
38. Chen HD, et al. Memory CD8+ T cells in heterologous antiviral immunity and immunopathology in the lung. Nat. Immunol. 2001;2:1067–1076. [PubMed]
39. Selin LK, Varga SM, Wong IC, Welsh RM. Protective heterologous antiviral immunity and enhanced immunopathogenesis mediated by memory T cell populations. J. Exp. Med. 1998;188:1705–1715. [PMC free article] [PubMed]
40. Appay V, et al. HIV-specific CD8(+) T cells produce antiviral cytokines but are impaired in cytolytic function. J. Exp. Med. 2000;192:63–75. [PMC free article] [PubMed]
41. Callan MF, et al. CD8(+) T-cell selection, function, and death in the primary immune response in vivo. J. Clin. Invest. 2000;106:1251–1261. [PMC free article] [PubMed]
42. Gillespie GM, et al. Functional heterogeneity and high frequencies of cytomegalovirus-specific CD8(+) T lymphocytes in healthy seropositive donors. J. Virol. 2000;74:8140–8150. [PMC free article] [PubMed]
43. Betts MR, et al. The functional profile of primary human antiviral CD8+ T cell effector activity is dictated by cognate peptide concentration. J. Immunol. 2004;172:6407–6417. [PubMed]
44. Madrenas J, Germain RN. Variant TCR ligands: new insights into the molecular basis of antigen-dependent signal transduction and T-cell activation. Semin. Immunol. 1996;8:83–101. [PubMed]
45. Sloan-Lancaster J, Allen PM. Altered peptide ligand-induced partial T cell activation: molecular mechanisms and role in T cell biology. Annu. Rev. Immunol. 1996;14:1–27. [PubMed]
46. Foss HD, et al. Patterns of cytokine gene expression in infectious mononucleosis. Blood. 1994;83:707–712. [PubMed]
47. Nakayama T, et al. Selective induction of Th2-attracting chemokines CCL17 and CCL22 in human B cells by latent membrane protein 1 of Epstein-Barr virus. J. Virol. 2004;78:1665–1674. [PMC free article] [PubMed]
48. Zaitseva M, et al. Human peripheral blood T cells, monocytes, and macrophages secrete macrophage inflammatory proteins 1alpha and 1beta following stimulation with heat-inactivated Brucella abortus. Infect. Immun. 2001;69:3817–3826. [PMC free article] [PubMed]
49. De Rosa SC, et al. Vaccination in humans generates broad T cell cytokine responses. J. Immunol. 2004;173:5372–5380. [PubMed]
50. Vassalli P. The pathophysiology of tumor necrosis factors. Annu. Rev. Immunol. 1992;10:411–452. [PubMed]
51. Xu L, et al. Cutting edge: pulmonary immunopathology mediated by antigen-specific expression of TNF-alpha by antiviral CD8+ T cells. J. Immunol. 2004;173:721–725. [PubMed]
52. Rothman AL, Ennis FA. Immunopathogenesis of Dengue hemorrhagic fever. Virology. 1999;257:1–6. [PubMed]
53. Rabin H, et al. Spontaneous release of a factor with properties of T cell growth factor from a continuous line of primate tumor T cells. J. Immunol. 1981;127:1852–1856. [PubMed]
54. Britten CM, et al. The use of HLA-A*0201-transfected K562 as standard antigen-presenting cells for CD8(+) T lymphocytes in IFN-gamma ELISPOT assays. J. Immunol. Methods. 2002;259:95–110. [PubMed]
55. Arden B, Clark SP, Kabelitz D, Mak TW. Human T-cell receptor variable gene segment families. Immunogenetics. 1995;42:455–500. [PubMed]
56. Maslanka K, Piatek T, Gorski J, Yassai M, Gorski J. Molecular analysis of T cell repertoires. Spectratypes generated by multiplex polymerase chain reaction and evaluated by radioactivity or fluorescence. Hum. Immunol. 1995;44:28–34. [PubMed]
57. Chothia C, Boswell DR, Lesk AM. The outline structure of the T-cell alpha beta receptor. EMBO J. 1988;7:3745–3755. [PubMed]

Articles from The Journal of Clinical Investigation are provided here courtesy of American Society for Clinical Investigation