|Home | About | Journals | Submit | Contact Us | Français|
The human polyomavirus BK virus (BKV) establishes a latent and asymptomatic infection in the majority of the population. In immunocompromised individuals, the virus frequently (re)activates and may cause severe disease such as interstitial nephritis and hemorrhagic cystitis. Currently, the therapeutic options are limited to reconstitution of the antiviral immune response. T cells are particularly important for controlling this virus, and T cell therapies may provide a highly specific and effective mode of treatment. However, little is known about the phenotype and function of BKV-specific T cells in healthy individuals. Using tetrameric BKV peptide-HLA-A02 complexes, we determined the presence, phenotype, and functional characteristics of circulating BKV VP1-specific CD8+ T cells in 5 healthy individuals. We show that these cells are present in low frequencies in the circulation and that they have a resting CD45RA− CD27+ memory and predominantly CCR7− CD127+ KLRG1+ CD49dhi CXCR3hi T-betint Eomesoderminlo phenotype. Furthermore, their direct cytotoxic capacity seems to be limited, since they do not readily express granzyme B and express only little granzyme K. We compared these cells to circulating CD8+ T cells specific for cytomegalovirus (CMV), Epstein-Barr virus (EBV), and influenza virus (Flu) in the same donors and show that BKV-specific T cells have a phenotype that is distinct from that of CMV- and EBV-specific T cells. Lastly, we show that BKV-specific T cells are polyfunctional since they are able to rapidly express interleukin-2 (IL-2), gamma interferon (IFN-γ), tumor necrosis factor α, and also, to a much lower extent, MIP-1β and CD107a.
In healthy individuals, the polyomavirus BK virus (BKV) establishes a latent, or “smoldering,” but asymptomatic infection. However, in immunocompromised individuals, the virus frequently escapes the normal immunological surveillance to become systemically active, after which it may cause severe pathology. BKV elicits interstitial nephritis of the allograft in about 5% of kidney transplant recipients, making it an important cause of graft failure and graft loss. In up to 30% of hematopoietic stem cell transplant (HSCT) recipients, the virus induces hemorrhagic cystitis, thereby significantly contributing to morbidity and length of hospitalization (1).
Currently, the main mode of therapy for patients suffering from BKV infection comprises reconstitution of the immunological antiviral response. In solid organ transplant recipients, this is achieved through tapering of the immunosuppressive medication. Unfortunately, this comes at the cost of increased allograft rejection, and in HSCT recipients this is an unattractive approach due to a considerable increase in the risk of graft-versus-host disease. So far, antiviral agents, such as cidofovir and leflunomide, have shown little effect on BKV replication in vivo (1). It is therefore crucial to develop new modes of therapy. In this regard, the normal T cell response was shown to be very important for keeping BKV at bay (2). Treatments that involve the infusion of autologous ex vivo-expanded virus-specific T cells, as well as newer vaccination strategies with autologous antigen-presenting cells pulsed ex vivo with viral antigen, are therefore promising candidates that could provide highly specific and effective modes of therapy (3, 4). It is well established that different T cell specificities give rise to different T cell phenotypes, which in turn is indeed also related to T cell function (5, 6). In this regard, little is known about the normal phenotype and function of BKV-specific T cells that are controlling BKV infection in healthy individuals, information that is necessary for the successful design of effective T cell therapies and vaccination strategies.
In the current study, we used fluorescent tetrameric HLA-A02 complexes presenting four different immunodominant BKV epitopes in order to visualize and characterize circulating BKV-specific CD8+ T cells. Phenotype and functional characteristics of these cells were analyzed in 5 healthy HLA-A02-positive adults. Furthermore, these BKV tetramers contain epitopes with a high degree of homology to the corresponding polyomavirus JC virus (JCV) epitopes, varying from a two-amino-acid difference to no difference at all. Indeed, cross-reactivity between the respective BKV and JCV tetramers was demonstrated (7–10). Moreover, antigen-presenting cells pulsed with BKV lysate can activate JCV-specific T cells and vice versa (8). Therefore, it is highly likely that the BKV-specific CD8+ T cells described in the current study are in fact also JCV-specific CD8+ T cells. Since it is well established that CD8+ T cell specificity correlates with phenotype, we compared the phenotypic characteristics of these BKV-specific CD8+ T cells to those of cytomegalovirus (CMV)-, Epstein-Barr virus (EBV)-, and influenza virus (Flu)-specific CD8+ T cells circulating in the same individuals to see how these phenotypes relate to each other (5). We found low frequencies of circulating BKV virion protein 1 (VP1)-specific CD8+ T cells that predominantly displayed an effector-memory cell phenotype. Furthermore, we observed that these cells were phenotypically distinct from circulating CMV- and EBV-specific CD8+ T cells, whereas they resembled Flu-specific T cells in several aspects. Lastly, BKV-specific T cells were polyfunctional with regard to their rapid expression of cytokines upon stimulation but did not express granzyme B.
We obtained peripheral blood mononuclear cells (PBMCs) from 25 buffy coats deriving from different HLA-A02-positive healthy blood donors, aged between 18 and 64 years, from Sanquin Blood Supply, Amsterdam, Netherlands. Corresponding plasma samples were not available for serologic or virologic examinations.
PBMCs were isolated using standard density gradient centrifugation, after which they were cryopreserved until the day of analysis.
All tetrameric complexes were obtained from Sanquin (Amsterdam, Netherlands). For BKV, four different immunodominant epitopes shared by the majority of BKV strains were selected including two BKV capsid protein VP1 epitopes, BKV VP1-derived AITEVECFL (VP1 p44) and BKV VP1 LLMWEAVTV (VP1 p108) (11), and two large T antigen protein (LTAg) epitopes, BKV LTAg LLLIWFRPV (LTAg p579) and BKV LTAg FLHCIVFNV (LTAg p410) (12, 13). These were incorporated in phycoerythrin (PE)-labeled HLA-A02 tetrameric complexes. The respective epitopes show a high degree of overlap with the corresponding epitopes of JCV and simian virus 40 (SV40): JCV VP1p36-44 SITEVECFL and SV40 VP1p46-54 SFTEVECFL, JCV VP1p100-108 ILMWEAVTL and SV40 VP1p110-118 ILMWEAVTV, JCV LTagp409-417 FLKCIVLNI and SV40 LTAgp408-416 FLKCMVYNI, and JCV LTAgp577-585 LLLIWFRPV and SV40 LTAgp577-585 LMLIWYRPV. Indeed, a high degree of cross-reactivity with the JCV VP1p36 and VP1p100 tetramers and the SV40 VP1p110 peptide has been demonstrated. For the detection of the other circulating antiviral CD8+ T cells, allophycocyanin (APC)-labeled HLA-A2 tetramers loaded with hCMV-pp65 protein-derived NLVPMVATV (CMV pp65 p495), influenza matrix protein 1-derived GILGFVFTL (Flu MP1 p58), and EBV BMLF-1 derived GLCTLVAML (EBV BMLF-1 p259) were used.
After thawing, PBMCs were labeled with carboxyfluorescein succinimidyl ester (CFSE). Thereafter, 2 × 10E6 PBMCs were incubated with one of four BKV peptides (1.25 × 10E−3 mg/ml) and recombinant human interleukin-2 (IL-2) (50 U/ml; Biotest, Solihull, United Kingdom) in culture medium that consisted of Iscove's modified Dulbecco's medium (IMDM), 10% human pool plasma (PAA, Piscataway, NJ, USA), penicillin/streptomycin, and β-mercaptoethanol (Sigma, Zwijndrecht, Netherlands) for the duration of 10 days. Extra IL-2 was added on day 3 and day 6 of culture. Controls comprised PBMCs cultured in medium with IL-2 alone, PBMCs cultured in medium containing the respective peptide alone, or PBMCs in culture medium alone without the addition of the respective peptide or IL-2. After culture, PBMCs were first incubated with the respective PE-labeled tetramer for 30 min in the dark at 4°C, after which they were incubated with the following antibodies: CD3-Alexa Fluor 700, CD8 PercP-eFluor 710, CD27 APC-Alexa Fluor 780 (eBioscience Inc., San Diego, CA, USA), and CD45RA PE-Cy7 (BD Biosciences, San Jose, CA, USA) for 30 min in the dark at 4°C. The Live/Dead fixable red cell stain kit (Invitrogen Ltd., Paisley, United Kingdom) was used in every staining to exclude dead cells from the analysis. Thereafter, measurements were done on the BD Biosciences' FACSCanto II flow cytometer. BKV-specific CD8+ CTLs were defined as being live CD3+ CD8+ tetramer+ events that had become CFSE dim due to cell division. Analyses were done using FlowJo software v9.6 (FlowJo, Ashland, OR, USA).
Before surface staining, up to 10 × 10E6 PBMCs were incubated with the respective tetramer for 30 min in the dark at 4°C. Live/Dead fixable red cell staining was utilized during all flow cytometry measurements and analyses in order to gate out dead cells. Monoclonal antibodies used for the determination of T cell phenotype and function included the following: CD3, V500; CCR7, PE-Cy7; CD45RA, PE-Cy7; CD45RA, fluorescein isothiocyanate (FITC); CD8, brilliant violet 421 (all from BD Biosciences); CD49d, PE-Cy7; CXCR3, Alexa Fluor 647; CD38, brilliant violet 421; CD8, Alexa Fluor 700; CD127, APC-Alexa Fluor 780; CD27, PerCP-eFluor 710; CD45RA, eFluor605; CD8, eFluor 605; CD27, APC-Alexa Fluor 780; CX3CR1, APC; HLA-DR, Alexa Fluor 700 (all from eBioscience); CXCR3, PE (R&D systems, Abingdon, United Kingdom); and KLRG1, Alexa Fluor 488 (14).
After extracellular staining, antibodies were fixed and the cells were permeabilized using fluorescence-activated cell sorter (FACS) Lysing solution and FACS Permeabilizing solution 2 (BD Biosciences) according to the manufacturer's instructions. Cells were subsequently incubated with a combination of the following intracellular antibodies for 30 min in the dark at 4°C: T-bet, brilliant violet 421 (BioLegend, San Diego, CA, USA; eomesodermin, PerCP-eFluor 710; eomesodermin, Alexa Fluor 647; Ki-67, PE-Cy7; granzyme B, Alexa Fluor 700 (BD Biosciences); and granzyme K, FITC (Immunotools, Friesoythe, Germany). All flow cytometry measurements for these experiments were done using BD Biosciences' LSRFortessa flow cytometer. Analysis of the results was performed with FlowJo software v9.6 (FlowJo, Ashland, OR, USA), always gating out duplets by using forward scatter width/height and sideward scatter width/height event characteristics.
Cytokine release after peptide or phorbol 12-myristate 13-acetate (PMA)/ionomycin stimulation was performed as described by Lamoreaux et al. (15). PBMCs were thawed and rested overnight in suspension flasks (Greiner-Bio-One, Alphen a/d Rijn, Netherlands) in RPMI supplemented with 10% fetal calf serum (FCS), penicillin, and streptomycin (culture medium). Two million cells per well were stimulated with PMA (10 ng/ml) and ionomycin (1 μg/ml) or with the viral peptides in culture medium in the presence of anti-CD107a FITC (eBioscience), αCD29 (TS2/16, 1 μg/ml), brefeldin A (10 μg/ml; Invitrogen), and GolgiStop (BD Biosciences) in a final volume of 200 μl for 4 h or 5 h (peptides) at 37°C and 5% CO2. Stimulations were performed in untreated, round-bottom, 96-well plates (Corning, Amsterdam, Netherlands). Subsequently, cells were incubated with the appropriate tetramers, followed by incubation with CD3 V500, CD8 brilliant violet 421, CD4 PerCP efluor 710 (Ebioscience), and Live/Dead fixable red cell stain for 30 min at 4°C. The cells were then washed twice, fixed, and permeabilized (Cytofix/Cytoperm reagent; BD Biosciences) and were subsequently incubated with the following intracellular monoclonal antibodies: gamma interferon (IFN-γ), APC-Alexa Fluor 750 (Invitrogen); tumor necrosis factor alpha (TNF-α,) Alexa Fluor 700; IL-2, APC; and anti-Mip-1β, PE-Cy7 (BD Biosciences) for 30 min at 4°C. Cells were washed twice and measured on an LSRFortessa flow cytometer and analyzed with FlowJo Version 9.6 software, always gating out dead cells and duplets.
Lymphocytes were gated based on the forward scatter and sideward scatter event characteristics. Duplets were then gated out as described above. Live/Dead fixable red cell staining was used to gate out dead or dying cells. CD3+ events were then gated, after which the expression of CD8 was plotted against tetramer-positive events (Fig. 1 and and2).2). Total CD3+ CD8+ (CD8+ T cells) events were then gated separately from total CD3+ CD8+ tetramer+ (virus-specific CD8+ T cells) events. Gates were drawn on total CD8+ T cells or a major CD8+ T cell population (naive, effector, or memory), after which they were copied to the virus-specific CD8+ T cells for the following markers: CD45RA was plotted against CD27 in order to define CD45RA+ CD27+ naive, CD45RA− CD27+ memory, and CD45RA+/− CD27− effector cell populations (16); the expression of CCR7 was determined by plotting the former against CD45RA, and gates were placed based on the double-positive naive CD8+ T cell population (not shown); CD127 was plotted against KLRG1, and gates were placed based on the CD127+ KLRG1− naive CD8+ T cell population (see Fig. 4); CD49d was plotted against CD45RA, and gates were placed based on the CD45RA+ CD49d− naive CD8+ T cell population (see Fig. 5); CD38 was plotted against HLA-DR, and gates were placed based on the double-negative CD8+ T cell populations (not shown). Ki-67 was plotted against CD38, and gates were placed based on the double-negative CD8+ T cell population (not shown); granzyme B was plotted against granzyme K, and gates were placed based on the double-negative naive CD8+ T cell population (see Fig. 7); T-bet was plotted against Eomes, and gates were placed based on the double-negative naive CD8+ T cell population (see Fig. 6). Gates for the following markers were placed based on their expression on total alive lymphocytes: CXCR3 was plotted against CD3 within total alive lymphocytes, and gates were placed based on the CXCR3hi CD3− natural killer cell population (see Fig. 5); PD-1 was plotted against CD3, and gates were placed based on the PD-1lo CD3− population (not shown).
The two-tailed Mann-Whitney test was used for analysis of differences between T cell populations. P values of less than 0.05 were considered statistically significant.
In view of the low frequencies in which BKV-specific CD8+ T cells are found in the circulation of healthy individuals, we first cultured CFSE-labeled PBMCs from 25 healthy HLA-A02-positive adults with four different BKV peptides, two derived from the capsid protein VP1, and two from the large T antigen (LTAg) protein, for 10 days in the presence of IL-2, in order to determine in which of these donors BKV-specific CD8+ T cells were detectable (7, 11). This yielded expanded BKV-specific CD8+ T cell populations that were readily detectable by flow cytometry in 15 of 25 subjects (60%). Seven of these were found to be positive for the VP1 p44 tetramer (28%), eight for the VP1 p108 tetramer (32%), five for the LTAg p579 tetramer (20%), and none for the LTAg p410 tetramer. Four subjects were positive for multiple BKV tetramers (16%) (Table 1). In contrast to the peptide/IL-2 condition, IL-2-only and peptide-only controls never yielded expanded BKV-specific CD8 T cell populations in this 10-day time window (data not shown).
The low frequencies of these cells in the circulation complicate a direct ex vivo detection with tetramers, and prior expansion of CD8+ T cells by culture will hamper a reliable determination of their phenotype and function. Therefore, we measured a substantial amount of PBMCs (~10 × 10E6) per antibody panel ex vivo only in those subjects whose PBMC fraction contained a particularly large BKV-specific CD8+ T cell population after 10-day expansion. As shown in Table 1, subjects 1, 2, 4, 6, 8, and 9 all had VP1 tetramer-positive populations that exceeded 1% of total CD8+ T cells. Indeed, BKV VP1-specific CD8+ T cells were detectable directly ex vivo in 5 of these 6 individuals (Fig. 1).
We then measured the expression of a wide variety of T cell markers in order to determine the phenotype of the BKV-specific CD8+ T cells. For comparison, the expression of the same markers was determined on HLA-A02 epitope-restricted CMV-, EBV- and Flu-specific CD8+ T cells circulating in the same five subjects (Table 2 and Fig. 2). As shown in Fig. 3A, we found that BKV VP1-specific T cells were CD45RA−, denoting them as antigen-experienced cells. Moreover, the majority of them expressed CD27 but lacked CCR7, marking them as being predominantly memory, and more specifically as “effector-memory” T cells (TEM) (Fig. 3C) (16, 17). The memory phenotype was supported by a high expression of CD127, the IL-7-receptor α chain that is important for maintaining memory cell homeostasis in the absence of antigen (Fig. 4A and andB)B) (18). The majority of BKV-specific T cells expressed killer cell lectin-like receptor G1 (KLRG1), an inhibitory receptor that was previously found to be highly expressed by TEM cells (Fig. 4A and andC)C) (14). However, as shown in Fig. 4C, KLRG1 expression levels varied notably between individuals. Flu-specific T cells were also predominantly CD45RA− CD27+ CCR7− TEM cells with a high expression of CD127 (Fig. 3 and and4B).4B). The expressions of KLRG1 by the different virus-specific T cells did not differ significantly from each other (Fig. 4C). In contrast, the expression of CD127 by CMV- and EBV-specific T cells was significantly lower than that in BKV- and Flu-specific T cells (Fig. 4B).
With regard to their activation status, BKV-specific T cells did not express the activation markers CD38 or HLA-DR or the marker of active proliferation, Ki-67 (data not shown). PD-1, a marker of T cell exhaustion or activation, was expressed by nearly 20% of BKV-specific T cells but did not differ significantly from total CD8+ T cells, CD8+ memory T cells, or the other virus-specific T cells (data not shown).
In conclusion, circulating BKV VP1-specific CD8+ T cells in healthy adults have a memory phenotype and can be considered to be in a “resting” state since they did not express activation markers and were not actively proliferating.
BKV is generally accepted to latently reside in epithelial cells lining the urogenital tract. However, in immunocompetent individuals, the virus has also been detected in the central nervous system, stool samples, and saliva (19–21). Since BKV-specific T cells are present in the circulation, one could theorize that they are trafficking to a site of (re)activation. Thus, we determined the expression of several specific T cell homing markers, in order to elucidate to which sites BKV-specific CD8+ T cells can migrate.
BKV-specific cells lacked expression of CD103, the alpha E integrin chain, a marker for tissue-resident T cells that is also expressed by a small proportion of circulating regulatory CD8+ T cells (22). The cells did highly express CD49d (Fig. 5B), the integrin α4 chain that mediates T cell migration through endothelial and epithelial barriers into various tissues and also through the blood-brain barrier (23, 24). However, since CD49d was also highly expressed on T cells specific for other viruses, this appeared not to be a specific trait of BKV-specific T cells (Fig. 5B and andD).D). BKV-specific T cells expressed high levels of CXCR3, a chemokine receptor which has been shown to mediate T cell homing to reactive lymph nodes, stressed epithelium, and sites of inflammation in a general sense (25, 26). Interestingly, Flu-specific T cells, but not CMV- and EBV-specific T cells, displayed a similarly high expression of CXCR3 (Fig. 5A and andC).C). BKV-specific T cells did not express the chemokine receptors CCR4 or CCR9 (data not shown).
The T-box transcription factors T-bet and eomesodermin (Eomes) are key regulators of CD8+ T cell differentiation and have been shown to cooperatively steer not only important CD8+ T cell functions such as the production of IFN-γ, expression of the IL-12 receptor, and expression of granzyme B but also the overall differentiation process of naive CD8+ T cells toward an effector or memory phenotype (27–31). Interestingly, the different virus-specific CD8+ T cell populations displayed distinct and specific patterns of T-bet and Eomes expression (Fig. 6). BKV-specific T cells displayed an intermediate T-bet expression with a markedly low expression of Eomes (Fig. 6). Similarly, Flu-specific T cells also displayed a T-betint Eomeslo expression, with almost no cells in the double-positive population. In contrast, CMV- and EBV BMLF1-specific T cells were highly expressing mainly both T-bet and Eomes (Fig. 6).
To extend the phenotypic analysis of BKV-specific T cells, we determined the expression of the cytotoxic effector molecules granzyme B and granzyme K. Interestingly, the immediate cytotoxic capability of BKV VP1-specific CD8+ T cells appears to be limited since we found them not to express granzyme B (Fig. 7A and andB).B). In contrast, CMV- and EBV-specific CD8+ T cells comprised a substantial population of granzyme B+ T cells (Fig. 7A and andB).B). Also, a small population of Flu-specific T cells expressed granzyme B; however, this was lower than for CMV- and EBV-specific T cells (Fig. 7A and andB).B). The virus-specific T cells did not differ from each other with regard to the expression of granzyme K (Fig. 7A and andCC).
Next, we wanted to determine whether the BKV VP1-specific CD8+ T cells are capable of producing cytokines immediately upon stimulation. Therefore, we stimulated the cells with PMA/ionomycin to subsequently measure the intracellular presence of IL-2, IFN-γ, TNF-α, MIP-1β, and CD107a. Due to stimulation-induced T cell receptor (TCR) internalization, there is an inverse correlation between the degree of stimulation and the ability to detect virus-specific T cells with tetramers. As a consequence, we were able to detect BKV-specific T cells only in the PBMC fractions of subjects 4 and 9 after PMA/ionomycin stimulation. In the other three subjects, the already few tetramer-positive events had become undetectable after stimulation (data not shown). Upon PMA/ionomycin stimulation, BKV VP1-specific CD8+ T cells from both subjects displayed a polyfunctional profile: nearly 11% of BKV VP1 p108-specific CD8+ T cells in subject 4 were able to rapidly express all five proteins (IL-2, IFN-γ, TNF-α, and to a lesser extent, also MIP-1-β and CD107a) upon stimulation. About 23% excreted a combination of four, and 52% excreted a combination of three of these (Fig. 8A and andB).B). In subject 9, about 4% of VP1p44-specific CD8+ T cells excreted all five proteins, 26% excreted a combination of four, and about 56% produced a combination of three of these proteins. In both subjects, the T cells most often produced the combination of IL-2, IFN-γ, and TNF-α (Fig. 8B).
We then repeated the experiment with cognate peptide stimulation. In order to determine the optimal peptide concentration at which tetramer-positive events were detectable while the cells were also still responding in the sense of cytokine production, we tested peptide concentrations ranging from 1,000 ng/ml to 0.05 ng/ml (data not shown). In subject 4, BKV-specific T cells were detectable and functionally active at 1 ng/ml. The higher concentrations (1,000, 100, and 10 ng/ml) rendered the detection of BKV-specific T cells impossible, again due to TCR internalization and their subsequent inability to bind the tetramer. In subject 9, BKV-specific T cells were detectable and functionally active at 1,000 ng/ml only. In both subjects, the stimulus provided by the lower concentrations proved to be insufficient to trigger cytokine production (data not shown). In contrast to the tetramer-negative CD8+ T cells, a considerable percentage of BKV-specific T cells in both subjects also produced the same proteins in response to their cognate peptide (Fig. 8A). The qualitative and quantitative differences in signaling induced by stimulation with PMA/ionomycin, which circumvents normal outside-in T cell signaling, and the stimulation with this specific concentration of cognate peptide and costimulation, given in this specific time window, offer a likely explanation for the relative differences between PMA/ionomycin- and peptide/costimulation-triggered cytokine release and degranulation.
In the current study, we show that BKV-specific CD8+ T cells are present in the circulation of healthy adults in very low frequencies. We show that the BKV VP1-specific CD8+ T cells have a phenotype of resting memory cells, predominantly being CD45RA− CD27+ CD127+ KLRG1+ CCR7− TEM cells that were not activated and did not actively proliferate. Furthermore, these T cells highly express CD49d and CXCR3. Interestingly, they expressed intermediate levels of T-bet but very little or no Eomes. Also, the direct cytotoxic potential of BKV-specific T cells appears to be limited since they are devoid of granzyme B in this resting state. Lastly, stimulation induced rapid excretion of inflammatory cytokines. As such, upon activation these BKV-specific CD8+ T cells are capable of rapidly mobilizing the immunological response by sending out various alarm signals.
The extremely low frequencies of BKV-specific CD8+ T cells in the circulation fit the finding that BKV is normally not detectable in this compartment and is residing at distal sites of viral latency, such as the kidney (32). Indeed, in a minority of healthy persons BKV is episodically shed in the urine (32, 33). Unfortunately, due to their extremely low frequencies in the circulation, we could not perform an ex vivo characterization of BKV LTAg-specific CD8+ T cells. LTAg is a nonstructural protein that is expressed only during the early phase of viral replication and during nonpermissive infection (1). The dominance of VP1-specific T cells over LTAg-specific T cells is in line with LTAg-specific antibody serum titers being lower than VP1-specific antibody titers (34, 35). The discrepancy between VP1 and LTAg expression during the viral replication cycle and, as such, the variations in antigen exposure could explain the differences in population size of VP1-specific and LTAg-specific T cells.
The function of BKV VP1-specific CD8+ T cells in the circulation might be dual: preventing BKV from entering the circulation, which is a common occurrence in severely immunocompromised individuals, and patrolling the circulation for local or distal BKV activity (1). Indeed, BKV-specific cells expressed high levels of CXCR3, known to mediate T cell homing to reactive lymph nodes, stressed epithelium, and sites of inflammation (24–26). Also, a small proportion expressed the central memory marker CCR7, which allows cells to migrate to resting lymph nodes. Furthermore, due to their high expression of CD49d these cells have the capacity to migrate through the blood-brain-barrier, pointing to the central nervous system as a latent niche of BKV infection, as was proposed previously (20). Considering that BKV-specific T cells also recognize epitopes of the JCV, high CD49d expression might be a likely explanation for the association between treatment with the anti-CD49d monoclonal natalizumab and the emergence of progressive multifocal leukoencephalopathy, a severe JCV-mediated neurological disorder (36, 37). However, the similarly high expression of CD49d by all the other circulating virus-specific CD8+ T cells perhaps rather emphasizes the neurotropism of the polyomaviruses, as well as the importance for BKV- or polyomavirus-specific T cells to be present in the central nervous system (23).
It has been demonstrated quite extensively that different pathogens can drive the differentiation of selective CD8+ T cell subsets (5, 6, 38, 39). Indeed, we show here that BKV-specific CD8+ T cells have a phenotype that is distinct from that of CMV- and EBV-specific T cells circulating in these same donors, while sharing some phenotypic aspects with Flu-specific T cells, such as the high expression of CD127 and CXCR3 and the low expression of Eomes. The latter struck us as particularly interesting since both BKV- and Flu-specific CD8+ T cells comprise mainly CD127+ memory cells and Eomes was proposed to be a key inducer of a memory CD8+ T cell phenotype (40). Also, it is remarkable that memory T cells, specific for viruses with two entirely different modes of infection, resemble each other in so many aspects, adding to the model of T cell differentiation as previously proposed by Appay et al. (5). Perhaps such phenotypic similarities could be explained by both T cell populations having abstained from specific extracellular signals for an extended period of time, knowing that BKV is normally not found in the circulation while Flu is thought to be cleared from the body after infection. Previously, we showed that circulating virus-specific T cells differ significantly from their counterparts residing in lymph nodes or the lungs, demonstrating how the tropism of the virus is also an important determinant of the T cell phenotype (41, 42). Local BKV-specific CD8+ T cells could very well display more of an effector phenotype, perhaps readily expressing granzyme B, in order to confer a more specialized defense at sites where the immunogenic pressure is higher. Indeed, if these circulating TEM cells are recruited to a site of BKV reactivation, they may generate or reinforce the local T cell-mediated defense. Therefore, it is crucial to also assess the characteristics of tissue-resident BKV-specific T cells, for example, those residing in the urogenital tract and/or the central nervous system. Lastly, considering the crucial role of CD4+ helper T cells in clearing viral infection, as well as the existence of cytotoxic CD4+ T cells, knowledge about the BKV-specific CD4+ T cell phenotype and function should be a primary research focus in this field (43). Unfortunately, this is currently hindered by specific technical limitations, among which, importantly, is the lack of available major histocompatibility complex (MHC) class II tetramers.
In conclusion, these data shed light on the phenotype and function of circulating BKV-specific CD8+ T cells in healthy adults. This information allows for a future comparison with BKV-specific CD8+ T cells circulating in immunocompromised individuals and their tissue-resident counterparts. These data may prove to be helpful for the future design of new immunotherapeutic options for the treatment of BKV-associated disease.
We thank Marjan J. Tempelmans Plat-Sinnige, Gijs van Schijndel, Ester M. M. van Leeuwen, Si La Yong, Simone H. C. Havenith, Daan J. aan de Kerk, and Nelly van der Bom-Baylon for their technical help and useful discussions.
Published ahead of print 17 July 2013