Conceived and designed the experiments: PP CPM WHH CRR. Performed the experiments: PP CPM. Analyzed the data: PP CPM. Contributed reagents/materials/analysis tools: AL RLC RH EY LB MBL JLB. Wrote the paper: PP CPM WHH CRR.
The authors have declared that no competing interests exist.
West Nile virus (WNV) infection can result in severe neuroinvasive disease, particularly in persons with advanced age. As rodent models demonstrate that T cells play an important role in limiting WNV infection, and strong T cell responses to WNV have been observed in humans, we postulated that inadequate antiviral T cell immunity was involved in neurologic sequelae and the more severe outcomes associated with age. We previously reported the discovery of six HLA-A*0201 restricted WNV peptide epitopes, with the dominant T cell targets in naturally infected individuals being SVG9 (Env) and SLF9 (NS4b). Here, memory phenotype and polyfunctional CD8+ T cell responses to these dominant epitopes were assessed in 40 WNV seropositive patients displaying diverse clinical symptoms. The patients' PBMC were stained with HLA-I multimers loaded with the SVG9 and SLF9 epitopes and analyzed by multicolor flow cytometry. WNV-specific CD8+ T cells were found in peripheral blood several months post infection. The number of WNV-specific T cells in older individuals was the same, if not greater, than in younger members of the cohort. WNV-specific T cells were predominantly monofunctional for CD107a, MIP-1β, TNFα, IL-2, or IFNγ. When CD8+ T cell responses were stratified by disease severity, an increased number of terminally differentiated, memory phenotype (CD45RA+ CD27− CCR7− CD57+) T cells were detected in patients suffering from viral neuroinvasion. In conclusion, T cells of a terminally differentiated/cytolytic profile are associated with neuroinvasion and, regardless of age, monofunctional T cells persist following infection. These data provide the first indication that particular CD8+ T cell phenotypes are associated with disease outcome following WNV infection.
West Nile virus (WNV) was first reported in the eastern United States in 1999 and has since spread westward through the entire North American continent. Reported WNV cases peaked at about 10,000 in 2003 and WNV continues to represent a considerable public health threat. Among infected individuals, almost 50% report neurologic symptoms due to encephalitis or meningitis (
Most primary WNV infections are asymptomatic or exhibit a mild, flu-like illness that lasts no more than a few days. Successful resolution of primary WNV infection is associated with development of adaptive immunity to the virus. Some individuals undergo neurologic complications that can arise long after the acute phase of the infection has been resolved
The objective of this study was to delineate whether CD8+ T cell immunity to WNV was related to neuroinvasive disease. This included assessment of memory T cell phenotypes, i.e., CD45RA (a spliced isoform of the leukocyte common antigen), CD27 (a member of the TNF receptor superfamily), CD57 (an indicator of T cells with limited proliferation potential) and CCR7 (lymph node homing receptor), and polyfunctional T cell activity, i.e., production of more than one immune mediator that are known to be involved in control of other viral infections
Previously, we comparatively analyzed thousands of peptide ligands eluted from the HLA class I of infected cells, discovering six WNV-encoded, HLA-A*0201 restricted, peptide epitopes
Virus | Peptide | Sequence | Location | Protein |
WNV | SVG9 | SVGGVFTSV | 430–438 | Env |
WNV | RLD10 | RLDDDGNFQL | 78–87 | NS2b |
WNV | YTM9 | YTMDGEYRL | 518–526 | NS3 |
WNV | SLT9 | SLTSINVQA | 15–23 | NS4b |
WNV | SLF9 | SLFGQRIEV | 68–76 | NS4b |
WNV | ATW9 | ATWAENIQV | 862–870 | NS5 |
EBV | GLC9 | GLCTLVAML | 280–288 | BMLF-1 |
Our results demonstrated that the immunodominance of SVG9 and SLF9 that we previously showed by production of IFN-γ in vitro was also evidenced by direct staining of uncultured CD8+ T cells with multimers specific for SVG9 and SLF9. T cells specific for these two WNV epitopes and the control, EBV epitope GLC9 were detected in the PBMC of the WNV-infected but not uninfected participants (
Log10 frequency of p-MHC multimer positive events per 106 CD3+/CD8+ cells from WNV seropositive and WNV seronegative subjects are shown for (A) dominant epitopes and (B) non dominant WNV epitopes. All comparisons were made using a one-way ANOVA with a Tukey post-hoc test. In all cases, lines represent mean values.
In contrast to SVG9 and SLF9, the frequency of T cells specific for the four subdominant WNV epitopes (RLD10, SLT9, YTM9 and ATW9) did not differ between WNV seropositive and seronegative persons (p>0.05) (
Advanced age is a risk factor for WNV disease severity
Log10 frequency of p-MHC multimer positive cells specific for SVG9 (A) and (B) SLF9 WNV epitopes vs age. The linear regression model is shown with the 95% confidence bands. Log10 frequency of SVG9 (C) and SLF9 (D) p-MHC multimer positive cells were divided into three age groups: subjects ≤40, 41–59, and ≥60 at time of infection. For (C) data did not fit a normal distribution so a Kruskal- Wallis ANOVA was used with Dunn's method of comparative analysis and medians are shown.
In contrast to the association of age with increases in T cells specific for SVG9, subjects with neuroinvasive disease did not exhibit an elevated frequency of SVG9 or SLF9 specific memory T cells (data not shown). However it should be noted that in our cohort there was not a significant age-related association with neuroinvasive disease; 14% for subjects <40 years of age and 36% for those aged 60 and over (p = 0.61). Thus, the increase in SVG9 specific T cells with age observed in our cohort was independent of disease severity.
Little information is available pertaining to the surface phenotype of long-term memory, CD8+ T cells to WNV. A recent study showed that equal numbers of CD45RA+/− CD27− (effector phenotype) and CD45RA− CD27+ (memory phenotype) WNV-specific CD8+ T cells are found one month after infection
CD45RA | CD27 | CCR7 | CD57 | |
Tcm | − | + | + | − |
Tem | − | + | − | − |
Tdpem | + | + | − | − |
TemRA | + | − | − | + |
The surface phenotypes of memory T cells specific for SVG9 and SLF9 contained significant numbers of both Tem and Tdpem cells (
(A) Prevalent memory phenotypes of antigen experienced T cells are shown, based on expression of CD27, CD45RA, CCR7, CD57 as defined in
Because our cohort had no age-related association with disease severity, we independently examined relationships between WNV-specific CD8+ T cell phenotypes, disease severity, and age. First we tested T cell phenotype profiles between subjects who had neuroinvasive disease and those who did not. For SLF9 specific cells, the Tdpem subset was overrepresented in subjects who did not develop neuroinvasive disease (
(A) Memory phenotypes of antigen experienced T cells from donors with or without neuroinvasive symptoms are shown, based on expression of CD27, CD45RA, CCR7, CD57 as defined in
Although age did not correspond to disease severity in our cohort, we nevertheless investigated CD8+ T cell phenotypes among defined age groups. When T cell phenotypes were stratified by age, both the SVG9 and SLF9 specific Tdpem subtype (CD45RA+ CD27+ CCR7− CD57−) differed with age (
(A) Memory phenotypes of antigen experienced T cells from donors separated by age group are shown, based on expression of CD27, CD45RA, CCR7, CD57 as defined in
Having identified phenotypes unique to donors with neuroinvasive symptoms, we next investigated the functional capacity SVG9 and SLF9 specific T cells in these patients. The stimulation of PBMC with SVG9 and SLF9 generated responding T cells that were primarily monofunctional (51.1% for SVG9 and 71.0% for SLF9) (
PBMC from subjects with or without neuroinvasive disease were stimulated with the immunodominant SVG9 or subdominant SLF9 WNV peptide epitopes and functional responses were evaluated by multi-color flow cytometry. Percentages of CD8+ T cells expressing the markers shown were input in the SPICE software for display. Pies represent the distribution of responses by function for each group according to the color scheme below. Bars represent the level of response for each individual of the 32 different marker combinations. Asterisks indicate significant differences (p<0.05) by Wilcoxon signed -rank test.
The neuroinvasive cohort did demonstrate a significant increase in the cytolytic associated marker CD107a (
PBMC from donors with or without neuroinvasive disease were stimulated with the immunodominant SVG9 or subdominant SLF9 peptide WNV epitopes and functional responses were evaluated as described. Percentages of CD8+ T cells expressing the cytotoxicity surrogate marker CD107a were added for each of the 32 combinations shown in
Understanding the phenotype and functional status of T cells during a successful immune response, as well as during disease progression, is needed for the design of anti-viral immune interventions. The characterization of T cells following vaccination (YFV, vacccinia) or during chronic viral infection (HIV, EBV) indicates that particular T cell phenotypes and functional status mediate infection, yet T cell responsiveness varies from vaccine-to-vaccine and from virus-to-virus
Our most immediate objective was to examine T cell frequencies: Immune senescence is receiving considerable attention as a risk factor during acute WNV infections in that increasing age corresponds to greater disease severity
Although T cell frequencies for the dominant WNV/A*0201 epitope increased with age, a previous study
Disease outcome was found to be independent of T cell frequency, and the next logical step was to characterize the phenotype of responding T cells. Four T cell categories were tested (Tdpem, TemRA, Tcm, Tem), and the largest proportion of WNV specific T cells exhibited a Tdpem phenotype. This memory CD8+ T cell subset was recently identified as the predominant subset following vaccination with YFV17D
Neuroinvasive diseases including encephalitis, meningitis and meningoencephalitis, represent the most severe conditions following WNV infection, and the relationship of T cell phenotypes and neuroinvasive outcomes was examined. Most interesting was the observation that significantly more CD45RA was expressed on by the SVG9-specific CD8+ T memory cells of donors with neurological involvement. These T cells expressed CD57, the an indicator of low proliferation potential, and lacked expression of CD27, confirming their terminally differentiated TemRA phenotype
TemRA cells are most commonly observed during chronic infections where antigen persists after a primary infection with pathogens such as HIV and herpesviruses
Tdpem cells that follow YFV vaccination have a polyfunctional phenotype in that each cell produces >4 different immune mediators in addition to degranulation
In summary, we used six high-confidence WNV peptide epitopes to dissect memory T cells that remain following viral infection. T cells specific for two immunodominant WNV epitopes were detectable well after the infection was resolved and the frequency of these WNV specific T cells remained unchanged and, in some cases, increased with age; age and disease severity do not coincide with a dearth of immune reactivity. Surface phenotyping demonstrates that the majority of memory T cells have a Tdpem phenotype consistent with YFV17D vaccination, although T cells that respond to a natural WNV infection are predominatly monofuntional. In regards to neuroinvasive pathologies, T cells with a terminally differentiated phenotype represent a post-infection marker for neuroinvasive pathologies, implicating persistence of WNV antigens. Finally, age and neuroinvasive changes associated with WNV infection were specific to the dominant SVG9/A*0201 epitope complex and did not extend to the 5 other WNV epitopes presented by HLA-A*0201. Future studies of the acute immune responses that give rise to these memory T cell subsets are needed, and these studies may require both dominant and subdominant epitopes in order to identify immune response that coincide with disease progression.
Peripheral blood was collected from consenting subjects diagnosed with WNV in heparinized tubes in accordance with a protocol approved by the University of Pittsburgh IRB and the Research Ethics Board at McMaster University. For all donors, written, informed consent was obtained before enrolment into the study in accordance with a protocol approved by the University of Pittsburgh IRB and the Research Board at McMaster University.
Subjects were enrolled in the study following presentation of symptoms of WNV infection at local primary care clinics and detection of serum IgM by public health laboratories. HLA genotypes were determined using DNA sequence analysis at Pure Transplant Solutions (Oklahoma City, OK). Forty subjects expressing HLA A*0201 allele were selected for phenotype screening of WNV specific memory CD8+ T cells, using previously defined HLA A*0201 class I WNV specific multimers (
Subject ID | Sex | Age | Days post-diagnosis | Clinical condition |
77406 | M | 80 | 211 | Neuroinvasive |
05201 | M | 39 | 239 | Non-Neuroinvasive |
77411 | F | 48 | 125 | Non-Neuroinvasive |
55414 | M | 57 | 222 | Non-Neuroinvasive |
55407 | M | 63 | 210 | Non-Neuroinvasive |
77425 | M | 66 | 248 | Neuroinvasive |
77420 | M | 67 | 257 | Non-Neuroinvasive |
77403 | F | 61 | 256 | Non-Neuroinvasive |
77416 | M | 61 | 236 | Non-Neuroinvasive |
77413 | M | 37 | 230 | Neuroinvasive |
77407 | F | 23 | 216 | Non-Neuroinvasive |
55410 | M | 51 | 212 | Non-Neuroinvasive |
77409 | F | 81 | 249 | Non-Neuroinvasive |
77405 | M | 64 | 226 | Neuroinvasive |
55413 | M | 51 | 218 | Neuroinvasive |
55421 | M | 42 | 201 | Non-Neuroinvasive |
55415 | F | 43 | 240 | Non-Neuroinvasive |
77319 | F | 41 | 204 | Non-Neuroinvasive |
77330 | F | 45 | 190 | Non-Neuroinvasive |
77322 | F | 55 | 221 | Non-Neuroinvasive |
55303 | F | 53 | 222 | Non-Neuroinvasive |
77316 | M | 45 | 256 | Non-Neuroinvasive |
77303 | F | 57 | 236 | Neuroinvasive |
77327 | M | 47 | 271 | Non-Neuroinvasive |
77324 | M | 39 | 229 | Non-Neuroinvasive |
77311 | M | 64 | 231 | Neuroinvasive |
55319 | M | 47 | 216 | Non-Neuroinvasive |
77312 | F | 45 | 235 | Neuroinvasive |
77332 | F | 60 | >200 | Neuroinvasive |
77307 | F | 55 | 184 | Neuroinvasive |
IDSA19 | F | 64 | 552 | Non-Neuroinvasive |
IDSA28 | M | 20 | >200 | Non-Neuroinvasive |
IDSA12 | M | 61 | 86 | Non-Neuroinvasive |
IDSA30 | M | 57 | 556 | Non-Neuroinvasive |
IDSA31 | F | 37 | 237 | Non-Neuroinvasive |
IDB203 | M | 49 | 280 | Non-Neuroinvasive |
IDS104 | M | 41 | 362 | Non-Neuroinvasive |
IDS123 | F | 34 | >200 | Non-Neuroinvasive |
IDS41 | F | 62 | 327 | Non-Neuroinvasive |
IDB157 | M | 64 | 261 | Non-Neuroinvasive |
To determine the frequency of peptide-specific T cells in PBMC, cells were thawed on the same day of the assays, let recover for 2 hours at 37°C, then treated with 100 µg/ml of DNase I (Sigma-Aldrich), and washed twice in monoclonal wash buffer (HBSS with 0.01% NaN3, 1% FBS). After recovery, PBMC were incubated for 10 minutes at room temperature with individual WNV specific p-MHC multimers (ProImmune). Afterwards, cells were washed and then stained for 20 min at 4°C with monoclonal antibodies directed against cell surface antigens: CD3-APC-H7 and CD8 PerCP-Cy5.5 for T cell lineage; CD57-FITC, CD45RA- PE-Cy7 (all from BD Biosciences), CCR7-PE (R and D Systems) and CD27 PE-Cy5 (Beckman-Coulter) for T cell memory phenotyping; CD4, CD14, CD16 and CD19 were used to exclude cells binding nonspecifically to the p-MHC multimers (dump channel) and were labeled either with PE-TexasRed (Beckman-Coulter) or with V-450 (BD Biosciences); APC fluorotag (ProImmune) for p-MHC multimer identification. Cells were then washed and resuspended in stabilizing fixative (BD Biosciences) and analyzed on a FACS LSRII (BD Biosciences) within 24 hours. Color compensation was set up with every assay using mouse Ig binding beads (BD Biosciences). PBMC from HLA A*0201 WNV seronegative subjects were treated as described above and used as negative controls. HLA A*0201 restricted BMLF-1 EBV multimer (GLCTLVAML) was used as positive control. At least 100,000 events were collected from each subject. Data were analyzed with DIVA software (BD).
To identify and enumerate epitope-specific CD8+ cytotoxic T cells, we used peptide HLA *0201 multimeric complexes specific for six individual HLA A*0201-restricted WNV epitopes previously discovered by mass spectroscopy and validated by functional testing (
Cryopreserved PBMC from a total of 20 WNV infected donors (10 non-neuroinvasive and 10 neuroinvasive) were thawed for intracellular cytokine staining (ICS) and analyzed by flow cytometry in order to measure the expression of the following immune mediators: IFNγ, TNFα, IL-2, MIP-1β, and CD107a. Cells were rested in media overnight at 37°C and then washed. Aliquots of 106cells were then co-cultured with the immunodominant WNV Env peptide (SVG9, 10 µg/ml), anti-CD107a mAb FITC (BD Bioscience), GolgiStop and GolgiPlug (BD Bioscience) in a 96-well plate. The negative control was medium only and the positive control was cells stimulated with
The frequency of antigen specific, p-MHC reactive T cells does not follow a normal istribution
(TIF)
(TIF)
(TIF)
We thank Ayan Chakrabarti and Kim Stojka for technical assistance, and Dr. Giovanna Rappocciolo for helpful discussions. We also thank Dr James Lute of the Virology & Immunology Section PA Department of Health Bureau of Laboratories for determination of the WNV serologic status of the Idaho cohort.