Presentation of human T cell leukemia virus type 1 (HTLV-1) Tax protein by dendritic cells: the underlying mechanism of HTLV-1-associated neuroinflammatory disease

Abstract

HTLV-1 is the etiologic agent of a debilitating neurologic disorder, HAM/TSP. This disease features a robust immune response including the oligoclonal expansion of CD8⁺ CTLs specific for the viral oncoprotein Tax. The key pathogenic process resulting in the proliferation of CTLs and the presentation of Tax peptide remains uncharacterized. We have investigated the role of APCs, particularly DCs, in priming of the anti-Tax CTL response under in vitro and in vivo conditions. We investigated two routes (direct vs. indirect) of Tax presentation using live virus, infected primary CD4⁺/CD25⁺ T cells, and the CD4⁺ T cell line (C8166, a HTLV-1-mutated line that only expresses Tax). Our results indicated that DCs are capable of priming a pronounced Tax-specific CTL response in cell cultures consisting of naïve PBLs as well as in HLA-A*0201 transgenic mice (line HHD II). DCs were able to direct the presentation of Tax successfully through infected T cells, live virus, and cell-free Tax. These observations were comparable with those made with a known stimulant of DC maturation, a combination of CD40L and IFN-γ. Our studies clearly establish a role for this important immune cell component in HTLV-1 immuno/neuropathogenesis and suggest that modulation of DC functions could be an important tool for therapeutic interventions.

Introduction

HTLV-1 causes a number of abnormalities—the most prominent, being ATL and HAM/TSP. More than 20 million people worldwide are infected by HTLV-1 [1]; endemic regions include southern Japan, the Caribbean, Central and South America, the Middle East, and Africa [2]. Although a majority of the infected individuals remain asymptomatic carriers, 0.5–3% develop HAM/TSP [3], which is a slowly progressive, demyelinating disease of the CNS, similar to multiple sclerosis but with a defined, causative agent and marked by weakness, stiffness, and paralysis of the legs [4, 5]. Both diseases are more common in women, and both are associated with an activated immune response with infiltrating lymphocytes and macrophages in the areas of demyelination and axonal dystrophy [4,5,6,7,8]. The majority of the humoral and cellular immune responses during HAM/TSP is directed to the HTLV-1 transactivator protein Tax, which is oncogenic and has been studied extensively with respect to regulating cellular and viral gene expression [9,10,11,12,13,14].

One of the most striking features of the cellular immune response in patients with HAM/TSP is the highly elevated numbers of HTLV-1-specific CTLs in the blood and cerebrospinal fluid, which are lower or absent in asymptomatic carriers [15,16,17,18,19,20,21]. A number of HTLV-1-associated antigens, including HTLV-1, Pol, Env, and Gag, as well as Rex protein can be targets of the CTL response [22]. In addition, responses to Rof and Tof epitopes in certain patients with HAM/TSP, as well as in some asymptomatic carriers, have been detected [22, 23]. Despite these responses to various HTLV-1-associated epitopes, CTL activity is restricted to the Tax protein predominantly, and in HLA-A*02⁺ patients, the Tax_11–19 peptide (LLFGYPVYV) has been defined as an immunodominant epitope [16, 24,25,26,27,28]. The importance of this hyperactivated, Tax-specific CTL response in determining proviral load is unclear. Some studies have shown that there is a zero or positive correlation between proviral load and CTL frequency, suggesting that the CTL response is ineffective at controlling proviral load [19, 21, 29,30,31,32]. In contrast, there is conflicting evidence that CTLs may help to reduce proviral load [8, 33, 34]. As a result of this complexity, the significance of the CTL response in HAM/TSP remains elusive, and inferring CTL control by the frequency of antigen-specific CD8⁺ T cells could be misleading. Thus, a careful estimation of the functionality of CTLs is required.

The ability of a CTL response to control a viral infection is affected by CD8⁺ T cell attributes such as T cell receptor avidity, specificity, and cellular activation, as well as APC attributes such as efficiency of epitope processing and presentation and the cell maturation state. The role of APCs with respect to HTLV-1 pathogenesis in general and the Tax-specific CTL response in particular remains uncharacterized. The current theory postulates CD4⁺/CD25⁺ T cells as being the major stimulus for hyperactivated Tax-specific CD8⁺ T cell responses observed in HAM/TSP disease [35]. However, this theory is based on a study performed by using pre-primed responder CD8⁺ T cells from patients. The nature of the CD4⁺/CD25⁺ T cells preferentially infected with HTLV-1 is unknown. In a study to identify Tax_11–19 peptide-HLA-A*201 complexes on CD4⁺/CD25⁺ T cells, Yamano et al. [35] reported that HTLV-1-infected CD4⁺/CD25⁺ cells are not suppressive but are stimulatory for HTLV-1 Tax-specific proliferation of CD8⁺ T cells, suggesting that CD4⁺/CD25⁺ cells infected with HTLV-1 may lack T regulatory cell function in HAM/TSP patients.

DCs are among the most potent APCs, with the ability to stimulate naïve T cells, and are key regulators of the immune system, linking stimulatory and inhibitory components of normal immunity. In addition to the conventional, endogenous route of class I antigen presentation, DCs have the unique ability to acquire exogenous antigens from virus-infected primary target cells (live or apoptotic), soluble proteins, and immune complexes and cross-present them to naïve CD8⁺ T cells [36]. DCs can also acquire antigens from infectious and fusion-competent, nonreplicating viruses [37,38,39,40]. In patients with HAM/TSP, a high frequency of HLA-A2-restricted, Tax-specific CTLs correlates positively with virus burden and Tax expression, indicating the existence of continuous antigen presentation [18, 19, 21, 41]. Therefore, the main focus of this study was to determine the maximal route of Tax-specific CTL activation targeting epitope 11–19, as the affinity of this epitope for HLA-A2 was shown to be unusually high [24], and the Tax_11–19-specific response is immunodominant frequently in patients with HAM/TSP [16, 25,26,27]. Using in vitro and in vivo priming assays, we demonstrated that DCs could drive an efficient Tax-specific CTL response in naïve PBLs from the normal donor as well as in HHD II mice. In addition, infected CD4⁺/CD25⁺ T cells alone did not generate as great a Tax-specific CTL response as when DCs were added to the culture. To the best of our knowledge, this is the first comprehensive study dissecting the participation of DCs in the induction of the Tax-specific cellular immune response, underlying the immuno- and neuropathogenesis of HAM/TSP.

MATERIALS AND METHODS

Cell lines and antibodies

The HTLV-1-transformed T cell line C8166 (kind gift from Dr. Shao-Cong Sun, Pennsylvania State University College of Medicine, Hershey, PA, USA) was cultured in RPMI 1640 (Mediatech, Herndon, VA, USA). All growth media were supplemented with 10% heat-inactivated FBS (Hyclone, Logan, UT, USA) with penicillin (100 U/ml) and streptomycin (100 mg/ml; Mediatech). Cells were maintained at 37°C in 5% CO₂ at 90% relative humidity. The unconjugated p19 antibody was obtained from ZeptoMetrix Corp. (Buffalo, NY, USA), and anti-Tax mAb (LT4) was generously provided by Dr. Yuetsu Tanaka (Kitasato University, Kanagawa, Japan). Both antibodies were color-conjugated using an allophycocyanin labeling kit (Dojindo Molecular Technologies, Rockville, MD, USA). The sources of other antibodies and reagents are described below with their respective use.

In vitro generation and culture of MDDCs

Highly purified MDDCs were generated from the PBMCs of healthy HTLV-1-negative individuals as described previously [42]. PBMCs were isolated from heparinized blood by Ficoll-Paque Plus (Amersham Biosciences, Uppsala, Sweden) density gradient centrifugation and tested for the HLA haplotype using a FITC-conjugated anti-HLA-A2 antibody (clone BB7.2, BD Biosciences, San Jose, CA, USA). Monocytes were allowed to adhere to the bottom of six-well plates, and nonadherent PBLs were separated by washing and frozen in 10% DMSO RPMI media for future use. The purity of monocytes was always >95%, as assessed by flow cytometry (FACScan, BD Biosciences). The monocyte-enriched adherent cell population was cultured in 1% normal human plasma (Sigma-Aldrich, St. Louis, MO, USA) in the presence of rhGM-CSF (100 IU/ml; PeproTech, Rocky Hill, NJ, USA) and rhIL-4 (300 IU/ml; PeproTech) for 5 days at 37°C and 5% CO₂. Cells were provided with fresh cytokines every other day. DC differentiation and immature phenotype were confirmed by multicolor FACS analyses for DC surface markers (CD40, CD80, CD83, CD86, HLA-DR, and DC-specific ICAM-grabbing nonintegrin) in conjunction with Lin-1 (BD Biosciences). The negative controls for the FACS analyses were FITC/PE/allophycocyanin/PE-Cy5-labeled, irrelevant antibodies and were used to establish plot boundaries.

In vitro priming assay

Immature DCs from HLA-A2⁺ donors were harvested on Day 5 of differentiation and were subjected to maturation by exposure to a combination of human CD40L (0.5 μg/ml, PeproTech) and IFN-γ (1000 U/ml, eBioscience, San Diego, CA, USA), Tax protein (25 nM), or cell-free HTLV-1 virus (125 ng, Advanced Biotechnologies, Columbia, MD, USA) for 2 days to induce phenotypic and functional maturation. The dose of Tax protein and virus concentration was predetermined by titration. Cells were washed to remove maturation stimuli and confirmed for their mature phenotype based on the CD83 positivity and the up-regulation of costimulatory molecules. Immature and mature DCs were pulsed with Tax_11–19 peptide (50 μg/ml, Alpha Diagnostic, San Antonio, TX, USA) for 2 h at 37°C and washed in serum-free medium. Peptide-pulsed DCs were then cocultured with the autologous, naïve PBLs (1:10 DC:T cell) in the presence of IL-2 (50 U/ml, R&D Systems, Minneapolis, MN, USA) and IL-15 (2.5 μg/ml, PeproTech) for a period of 4 weeks, providing another round of stimulation after 2 weeks. The priming of the Tax-specific CTL response was estimated by the quantitation of IFN-γ-producing and Tax_11–19-specific CD8⁺ T cells. All assays were performed in duplicate and repeated at least three times using cells from different donors.

IFN-γ ELISPOT assay and pentamer staining

At the completion of the stimulation period as described above, CD8⁺ T cells from each experimental group were separated from the total PBLs using an EasySep CD8-negative selection kit, per the manufacturer’s instructions (StemCell Technologies, Vancouver, BC, Canada). Isolated CD8⁺ T cells were divided into two parts. One part was subjected to the IFN-γ ELISPOT assay (BD PharMingen, San Diego, CA, USA), and the other was analyzed for the presence of Tax_11–19 pentamer-positive cells. Briefly, 1 × 10⁵ CD8⁺ T cells were serially diluted (in duplicate) by a factor of three in complete RPMI medium within an ELISPOT plate that had been coated previously with anti-hIFN-γ capture antibody. After overnight in-plate peptide stimulation, the spot-forming cells were developed and counted using an ELISPOT plate reader. For the pentamer staining, an allophycocyanin-conjugated anti-CD8 antibody (Caltag/Invitrogen, Carlsbad, CA, USA) was used in conjunction with the PE-conjugated Tax_11–19 pentamer (ProImmune, Bradenton, FL, USA). A total of 100,000 events were collected for each sample and gated to include the CD8⁺ population.

In vivo priming of the Tax-specific CTL response

Animal studies were performed according to the appropriate institutional review committee guidelines. Line HHD II mice were developed to express chimeric human (α1, α2) and mouse (α3) HLA-A2.1 heavy chain, covalently linked to human β₂-microglobulin light chain [43]. These mice are engineered to be H-2D^b and β₂-microglobulin double-knockouts, resulting in mice that express the chimeric human HLA-A2.1 molecule but lack surface expression of H-2D^b and H-2K^b. To perform in vivo priming, we immunized HHD mice with 0.1 mg of the Tax_11–19 peptide emulsified in IFA plus 0.14 mg hepatitis B virus core helper peptide 128–140 (TPPAYRPPNAPIL) s.c. at the base of the tail. After 3 weeks, mice received a booster immunization, and 7 days later, splenocytes were restimulated in vitro with Tax_11–19 peptide-pulsed (100 nM), syngeneic bone marrow-derived DCs [44]. Responder cells were tested on Day 6 of culture for their ability to produce IFN-γ following 6 h of stimulation with 1 μg/ml peptide or plate-bound anti-CD3 in the presence of 1 μg/ml brefeldin A. Cells were surface-stained with anti-mouse CD8 antibody and intracellularly with an anti-mouse IFN-γ antibody using the Cytofix/Cytoperm kit (BD Biosciences). Fifty thousand events were collected and gated to include the CD8⁺/IFN-γ⁺ T cell population.

Direct presentation of Tax

MDDCs were infected by exposing cells to cell-free, live HTLV-1 for 48 h. Cells were then harvested and washed with serum-free media. Infection of DCs was confirmed as described previously using allophycocyanin-conjugated Tax_Lt4 and p19 antibodies. Uninfected cells were used as a comparison control. Using flow cytometry, 50,000 events were collected and gated to include the Lin-1⁻ population. Prior to coculturing, Tax-specific CD8⁺ T cells (obtained from line HHD II mice) were labeled with CFSE, per the manufacturer’s instructions (Caltag/Invitrogen). Once labeled, Tax-specific CTLs were cocultured with MDDCs. Cells were harvested on Days 1, 3, and 5 and stained with a PE-Cy5-conjugated anti-mouse CD8 antibody and PE-conjugated Tax_11–19 pentamer and analyzed by flow cytometry. Fifty thousand events collected for each sample were gated to include the CD8⁺/Tax_11–19⁺ population.

Cross-presentation of Tax from the infected CD4⁺/CD25⁺ T cells

HLA-A2⁺ CD4⁺/CD25⁺ T cells were isolated from HLA-A2⁺ normal donor PBMCs using a human CD4⁺ T cell enrichment kit (StemCell Technologies), followed by a human CD25-positive selection kit (StemCell Technologies). Isolated cells were resuspended at a concentration of 2 × 10⁶ cells/ml in complete RPMI and treated with PHA (2 μg/ml). The cells were then exposed to 125 ng/ml HTLV-1-purified cell-free virus for 48 h. Upon confirmation of infection by flow cytometry using p19 and Tax antibodies, the PHA-treated cells were washed and cocultured with CFSE-labeled CTL clones (obtained from line HHD II mice) in the absence and presence of MDDCs, prepared from the same HLA-A2-positive donor. The coculture was supplied with the 5-μM antiretroviral agent AZT to block viral replication and to avoid direct antigen presentation [45]. Additionally, CD40L was added in the coculture. Cells were harvested on Days 1, 3, and 5 and stained with PE-Cy5-conjugated anti-mouse CD8 antibody and PE-conjugated Tax_11–19 pentamer and analyzed by flow cytometry. Fifty thousand events collected for each sample were gated to include the CD8⁺/Tax_11–19⁺ population.

Cross-presentation of Tax from the live and apoptotic C8166 cells

For the live cell experiment, Tax-specific CD8⁺ T cells (also obtained from line HHD II mice) were labeled with CFSE and cocultured with Tax-expressing CD4⁺ T cell line C8166 at a 1:1 ratio in the presence or absence of DCs as above. For the apoptotic cell experiment, C8166 cells were first incubated in complete media with 150 μg mitomycin C at 37°C for 90 min to inhibit cell proliferation. These cells were then washed three times in complete media to remove mitomycin C and cocultured as per the live cell experiment. Cells were harvested on Days 1, 3, and 5 and stained with PE-Cy5-conjugated anti-mouse CD8 antibody and PE-conjugated Tax_11–19 pentamer and analyzed by flow cytometry. Fifty thousand events collected for each sample were gated to include the CD8⁺/Tax_11–19⁺ population.

RESULTS

DCs could prime a Tax-specific CTL response in naïve PBLs and in line HHD II mice

Previous studies [35] have demonstrated that CD4⁺/CD25⁺ cells are preferentially infected with HTLV-1 and are the primary cells providing stimulus toward the induction and maintenance of a Tax-specific CTL response in patients with HAM/TSP. However, these studies used pre-primed T cells from a limited number of patients with HAM/TSP and did not consider the involvement of APCs. Therefore, we found it important to perform our studies with naïve T cells from normal donors and determine whether DCs could indeed prime these cells toward a Tax-specific immune response. To determine this, we adopted an in vitro priming strategy as described by Huang et al. [46] for the development of DC-based vaccines. To identify individual IFN-γ-producing cells, Tax_11–19-primed CD8⁺ T cells were incubated in IFN-γ-coated plates in the presence of peptide. As shown in Figure 1, control wells that did not receive Tax_11–19 peptide exhibited no response, indicating the specificity of the assay. Compared with immature DC-pulsed samples, CD40L/IFN-γ-matured, DC-pulsed samples (positive control) demonstrated a greater response (24 compared with 185 spot-forming cells/10⁵ total CD8⁺ T cells; approximately an eightfold increase). Tax-matured DCs exhibited a similar effect, with 197 spot-forming cells, whereas HTLV-1-matured DCs resulted in a much greater response, with 282 spot-forming cells, approximately a 12-fold increase compared with the immature DCs (Fig. 1). Similar results were obtained by the MHC pentamer analyses as shown in Figure 2. The HTLV-1-matured DCs demonstrated the greatest expansion of Tax_11–19-specific CD8⁺ T cells (3723) compared with the other variables (572, 808, and 609 for the immature, CD40L/IFN-γ-matured, and Tax-matured DCs, respectively; Fig. 2). The negative control pentamer showed no positive staining in Tax_11–19-pulsed samples, and DCs prepared from HLA-A2⁻ donors did not result in the priming of a Tax_11–19-specific response (data not shown). Collectively, these results demonstrate that DCs are capable of stimulating the induction and proliferation of HTLV-1-specific T cells through the presentation of Tax_11–19peptide in HLA-A2⁺ individuals. In both assays, DCs and/or T cells obtained from the HLA-mismatched donors did not result in the proliferation of Tax_11–19-specific T cells from multiple replicates of this experimental set-up (data not shown), ruling out the possibility of the nonspecific response.

Figure 1.

Analysis of IFN-γ secretion by naïve CD8⁺ T cells through Tax_11–19 peptide stimulation by DCs. HLA-A2⁺ PBLs from the HTLV-1-seronegative donors were stimulated in vitro with Tax_11–19 peptide-loaded immature or mature DCs (CD40L/IFN-γ, Tax, or HTLV-1) for a period of 4 weeks. Control consisted of untreated, unpulsed DCs cocultured with naïve PBLs. Primed CD8⁺ T cells were isolated and stimulated overnight with the peptide within the microwell of an ELISPOT plate (BD PharMingen), which was precoated with anti-human IFN-γ antibody. Each variable was tested at least in duplicate. Numbers above bar graphs indicate the average number of spot-forming cells/10⁵ CD8⁺ T cells from the duplicate wells. Data are representative of two independent experiments performed in duplicates.

Figure 2.

Detection of Tax_11–19 pentamer-positive cells induced by peptide-loaded DCs. To examine the ability of DCs to induce the expansion of Tax_11–19⁺ CD8 T cells within a naïve population, PBLs from HTLV-1-seronegative donors were cocultured with immature or mature peptide-loaded DCs as described in Figure 1. Cells were then harvested and stained with allophycocyanin-conjugated anti-human CD8 antibody and PE-conjugated Tax_11–19 pentamer and analyzed by flow cytometry. A total of 100,000 events collected for each sample were gated to include the CD8⁺ population. Numbers indicate the absolute number of Tax_11–19⁺ CD8 T cells. Data are representative of three separate experiments.

To determine whether similar results could be obtained in vivo, parallel studies were performed in line HHD II transgenic mice. These mice express chimeric human (α1, α2) and mouse (α3) HLA-A2.1 heavy chain, covalently linked to human β₂-microglobulin light chain, but lack expression of the mouse MHC class I molecules. The T cells of the HHD mice undergo positive and negative selection using the human HLA-A*0201 molecule and thus recognize antigen in the context of this particular MHC. Consequently, these mice were able to mount a Tax_11–19 HLA-A*0201-restricted response. To prime mouse splenocytes, line II transgenics were immunized with a Tax_11–19 peptide, as described in Materials and Methods. The splenocytes were isolated and cocultured with peptide-loaded, syngeneic, matured DCs for 6 days prior to analysis. The results from the two representative mice (of a total of three harvested; Fig. 3) demonstrated that DCs pulsed with the Tax_11–19 peptide induced a higher IFN-γ response (2.15% and 8.75% CD8⁺/IFN-γ⁺ T cells from mouse 1 and mouse 2, respectively) when compared with those pulsed with a HLA-mismatched Gag_77–85 peptide (0.009% and 0.072% CD8⁺/IFN-γ⁺ T cells from mouse 1 and mouse 2, respectively). The response obtained with the Tax_11–19 peptide was comparable with that obtained with the anti-CD3 antibody used as a positive control (3.81% and 7.71% CD8⁺/IFN-γ⁺ T cells in mouse 1 and mouse 2, respectively). Thus, our in vivo results are comparable with those obtained with the in vitro priming studies. As a result of these studies, we were able to obtain Tax_11–19-specific CD8⁺ T cell clones, which provided us with a more efficient and convenient way to understand DC-mediated, Tax-specific responses in the subsequent studies of antigen presentation. Thus, these newly obtained Tax-specific clones were used in all follow-up studies to avoid a long experimental set-up required for the priming of naïve PBLs.

Figure 3.

Induction of Tax_11–19-specific CD8⁺ T cells in line HHD II mice, which were immunized with 0.1 mg of the Tax_11–19 peptide s.c. at the base of the tail.These mice were developed to express chimeric human (α1, α2) and mouse (α3) HLA-A2.1 heavy chain, covalently linked to human β₂-microglobulin light chain. These mice are engineered to be H-2D^b and β₂-microglobulin double-knockouts, resulting in mice that express the chimeric human HLA-A2.1 molecule but lack surface expression of H-2D^b and H-2K^b. After 3 weeks, mice received a booster immunization, and 7 days later, splenocytes were restimulated in vitro with Tax_11–19 peptide-pulsed (100 nM) syngeneic bone marrow-derived DCs or DCs stimulated with a HLA-mismatched peptide or plate-bound anti-CD3 (positive control). Responder cells were tested on Day 6 of culture for their ability to produce IFN-γ following 6 h of stimulation with 1 μg/ml peptide or plate-bound anti-CD3 in the presence of 1 μg/ml brefeldin A. Cells were surface-stained with anti-CD8 antibody and intracellularly with anti-IFN-γ antibody. Numbers in upper right corners represent the percentage of cells positive for CD8 and IFN-γ. Data are from two representative mice out of at least three mice harvested/group.

HTLV-1-infected DCs could lead to the proliferation of Tax_11–19-specific CD8⁺ T cells

We have shown previously that DCs exhibit maximum binding affinity to cell-free HTLV-1 as compared with other cell types tested [47], indicating that these cells could be the first type interacting with HTLV-1, particularly during mucosal exposure. DCs obtained from patients with HAM/TSP were found to be infected with HTLV-1 [45, 48] and can be infected easily in vitro by cell-free [45, 49] and cell-associated [50, 51] virus. Therefore, it is likely that infected DCs could process and present Tax antigen in the context of MHC class I, leading to the expansion of Tax-specific CD8⁺ T cells. To test this possibility, we exposed DCs to live virus for 48 h and then stained intracellularly for the presence of HTLV-1 core protein p19 and the expressed Tax protein. Compared with the uninfected cells, DCs exposed to virus were shown to be infected efficiently; 91% of cells stained positive for p19, and >50% of those demonstrated Tax expression (Fig. 4), which was critical to determine before performing the Tax presentation assay. As an additional control, we confirmed cell-free HTLV-1 infection of MDDCs by detecting the presence of HTLV-1 proviral DNA by the PCR analysis (Supplemental Fig. 1A) as well as the presence of viral core protein p19 in the culture supernatant using a p19 ELISA (Supplemental Fig. 1B). Infected DCs were then used to stimulate the proliferation of Tax_11–19-specific CD8⁺ T cell clones, which has been labeled with the proliferation marker CFSE and analyzed on Days 1, 3, and 5. As a positive control, we used mitomycin C-treated syngeneic mouse fibroblast cultures pulsed with the Tax peptide, which are used routinely to maintain the Tax_11–19-specific T cell clones in culture. As a negative control, we cultured CFSE-labeled, Tax-specific T cell clones alone and fibroblasts without peptide stimulation cocultured with the CFSE-labeled, Tax-specific T cell clones. In addition, we compared the proliferation of uninfected DCs with those infected. To calculate the number of generations of newly formed daughter cells following Day 1 (i.e., the extent of the Tax-specific CTL clone proliferation), the GMFI of CFSE of the Day-1 peak was divided by the GMFI of CFSE of each progressive peak so that the number of generations (n) = Day 1 GMFI/Day 3 GMFI, or n = Day 1 GMFI/Day 5 GMFI. As shown in Figure 5, CTLs alone (n=2.55 and 2.58 for Days 3 and 5, respectively) and fibroblasts without peptide (n=1.12 and 2.27 for Days 3 and 5, respectively) demonstrated modest proliferation. However, infected DCs (n=28.99 and 158.33 for Days 3 and 5, respectively) induced the proliferation of Tax-specific CD8⁺ T cells at a much higher rate than the peptide-pulsed fibroblasts starting from Day 1 leading to Day 5 (n=6.26 and 34.88 for Days 3 and 5, respectively) and uninfected DCs (n=12.87 and 12.72 for Days 3 and 5, respectively). Cells from the subsequent days were not analyzed as a result of the large increase in T cell numbers visibly apparent in the culture plates. As these studies were performed in the absence of CD4⁺ T cells (unlike priming studies with naïve PBLs), our results clearly indicate that one potential source of Tax antigen for DCs is self-infection and that infected DCs could drive the proliferation of the preprimed, Tax-specific CTLs, even in the absence of CD4 T cells. Nevertheless, to compare our observations directly with previous studies performed by Yamano et al. [35], we cocultured infected CD4⁺/CD25⁺ T cells from HLA-A2⁺ individuals with the Tax-specific CTL clones in the absence and presence of DCs. First, it was important to confirm infection of these cells, which were exposed to purified cell-free HTLV-1. As shown in Figure 6, the isolated cells demonstrated efficient infection via the analyses of p19 (99% positive cells) and Tax (97% positive cells) expression. The HTLV-1-infected CD4⁺/CD25⁺ T cells alone were able to induce moderate proliferation of the Tax-specific CTL clones (n=2.00 and 8.21 for Days 3 and 5, respectively); however, when DCs were added to the culture, the extent of proliferation was much higher (n=20.67 and 96.65 for Days 3 and 5, respectively; Fig. 7).

Figure 4.

Confirmation of HTLV-1 infection of DCs. MDDCs were exposed to purified cell-free HTLV-1 for 48 h, washed to remove free virus, fixed, permeabilized, and stained intracellularly with an allophycocyanin-conjugated anti-p19 and anti-Tax mAb (LT4) antibody. Fifty thousand events collected for each sample were gated to include the Lin-1⁻ population.

Figure 5.

Infected DCs can induce the proliferation of Tax-specific CTLs. Tax-specific CD8⁺ T cells were CFSE-labeled and cocultured with infected DCs. Cells were harvested on Days 1, 3, and 5 and stained with PE-Cy5-conjugated anti-mouse CD8 antibody and PE-conjugated Tax_11–19 pentamer and analyzed by flow cytometry. Fifty thousand events collected for each sample were gated to include the CD8⁺/Tax_11–19⁺ population. (Top) Tax-specific CTL clones (obtained from line HHD II mice used in experiments for Fig. 3) cultured alone (negative control 1). (Middle left) Mitomycin C-treated mouse fibroblast without Tax_11–19 peptide cocultured with Tax-specific CTL clones (negative control 2). (Middle right) Mitomycin C-treated mouse fibroblast pulsed (2 h) with Tax_11–19 peptide and cocultured with Tax-specific CTL clones (positive control). (Bottom left) Uninfected MDDCs cocultured with Tax-specific CTL clones. (Bottom right) Infected MDDCs cocultured with Tax-specific CTL clones. Histograms represent the relative amounts of CFSE expression remaining at each given time-point (Days 1, 3, and 5). Data are representative of three independent experiments.

Figure 6.

Confirmation of HTLV-1 infection in CD4⁺/CD25⁺ T cells. HLA-A2⁺ CD4⁺/CD25⁺ T cells were isolated and treated with PHA. Cells were then exposed to purified cell-free HTLV-1 for 48 h, washed to remove free virus, fixed, permeabilized, and stained intracellularly with an allophycocyanin-conjugated anti-p19 and anti-Tax mAb (LT4) antibody. Fifty thousand events collected for each sample were gated to include the CD4⁺/CD25⁺ population. The numbers indicate the percent-positive p19 or Tax-expressing cells.

Figure 7.

Direct and DC-mediated cross-presentation of Tax from infected CD4⁺/CD25⁺ T cells. Tax-specific CD8⁺ T cells (obtained from line HHD II mice used in Fig. 3) were CFSE-labeled and cocultured with infected CD4⁺/CD25⁺ T cells from Figure 6 in the absence (left) or presence of DCs (right). Cells were harvested on Days 1, 3, and 5 and stained with PE-Cy5-conjugated anti-mouse CD8 antibody and PE-conjugated Tax_11–19 pentamer and analyzed by flow cytometry. Fifty thousand events collected for each sample were gated to include the CD8⁺/Tax_11–19⁺ population. The experiment was run in duplicates.

DCs cross-present Tax antigen from live and apoptotic Tax-expressing CD4⁺ T cells

Exogenous viral antigens are known to be channeled into the DC endogenous pathway for class I presentation, a phenomenon known as cross-presentation [52, 53]. HIV antigens (RT_476–484, Nef_73–82, and Gag_20–28) have been shown to be cross-presented by DCs from infected CD4⁺ T cells [54]. This cell population is infected predominantly with HTLV-1 as well, and these cells have been shown to express higher levels of Tax mRNA in patients with HAM/TSP [19]. Moreover, the unique ability of DCs to cross-present immune complexes efficiently to boost the response allows these cells to surpass any baseline priming that may be observed as a result of direct antigen presentation from infected CD4⁺/CD25⁺ T cells. Therefore, we investigated the possibility of Tax cross-presentation by DCs by using the Tax-expressing (C8166) CD4⁺ T cell line as a source of Tax antigen for DCs. The C8166 cell line contains a mutated virus allowing Tax expression but not virus production [55, 56], thus providing a good system to avoid direct presentation from viral replication. To be able to compare the results of this experiment directly with those of our other experiments, these cells were first confirmed to be HLA-A2-positive by flow cytometry (Supplemental Fig. 2). In addition, we examined them for the expression of p19 and Tax as described in Materials and Methods. As expected, the C8166 cells expressed only Tax (Fig. 8A), confirming the suitability of this cell line for use in cross-presentation studies, considering virus and Tax independently. These cells were cultured with Tax_11–19-specific T cell clones in the absence or presence of autologous DCs without previous exposure to Tax or peptide. The coculture was also supplied with CD40L as a maturation agent; however, it is also possible that CD4⁺ T cells could have provided activation signals to DCs. Moreover, AZT was used to block viral infection, if any, and to avoid direct antigen presentation. As evident from Figure 8B, C8166 cells were able to induce expansion of Tax-specific T cell clones through direct presentation (n=3.05 and 81.06 for Days 3 and 5, respectively); however, the proliferative capacity of these clones in the presence of DCs was much higher (n=18.62 and 315.85 for Days 3 and 5, respectively).

Figure 8.

DC-mediated cross-presentation of Tax from live and dead CD4⁺ T cells. (A) C8166 cells were permeabilized and stained intracellularly with an allophycocyanin-conjugated p19 and Tax mAb (LT4) antibody. Fifty thousand events collected for each sample were gated to include the CD4⁺ population. (B) Tax-specific CD8⁺ T cells (obtained from line HHD II mice used in Fig. 3) were CFSE-labeled and cocultured with C8166 cells in the absence or presence of DCs. The coculture was also supplied with AZT (5 μM) and CD40L (0.5 μg/ml). Cells were harvested at Days 1, 3, and 5 and stained with a PE-Cy5-conjugated anti-CD8 antibody and PE-conjugated Tax_11–19 pentamer and analyzed by flow cytometry. Fifty thousand events collected for each sample were gated to include the CD8⁺/Tax_11–19⁺ population. (Left) Response observed with C8166 cells in the absence of DCs; (right) the same in the presence of DCs. Histograms represent the relative amounts of CFSE expression remaining at each given time-point. The experiment was repeated at least twice. (C) C8166 cells were first treated with mitomycin C and then cocultured as per the live cell experiment in B. Cells were harvested at Days 1, 3, and 5 and stained with a PE-Cy5-conjugated anti-CD8 antibody and PE-conjugated Tax_11–19 pentamer and analyzed by flow cytometry. Fifty thousand events collected for each sample were gated to include the CD8⁺/Tax_11–19⁺ population. (Left) Response observed with C8166 cells in the absence of DCs; (right) the same in the presence of DCs. Histograms represent the relative amounts of CFSE expression remaining at each given time-point. The experiment was repeated at least twice.

Finally, we wanted to determine whether limiting the proliferative capacity of C8166 cells would have an effect on the efficiency of cross-presentation by DCs. To accomplish this, the T cell line was treated with mitomycin C before it was subjected to coculture. As shown in Figure 8C, the C8166 cells retained some stimulation capability (n=5.03 and 17.62 for Days 3 and 5, respectively), albeit weaker than the live cells following mitomycin C treatment. These observations suggest that although the proliferation of C8166 cells was inhibited, they were still able to provide Tax antigen, probably as a result of the high levels of ongoing Tax expression. However, DC-mediated cross-presentation (n=22.56 and 315.85 for Days 3 and 5, respectively) was not affected by the mitomycin treatment of these CD4⁺ T cells (Fig. 8C).

DISCUSSION

The anti-Tax CTL response observed in patients with HAM/TSP is one of the most intriguing and known cellular immune responses against a chronic human viral infection. Instead of controlling proviral load, the anti-Tax immune response seems to follow it passively, exhibiting a positive (instead of negative) correlation between the two [30]. A high proviral load is a significant risk factor for the development of HAM/TSP [57]. As HTLV-1 does not readily produce detectable extracellular virions, viral burden is quantified as the integrated proviral DNA load, which remains relatively constant within a host over time [58] but can vary considerably between individuals, reaching as high as 40% of PBMCs infected in some patients [57]. The prevalence of HAM/TSP increases strikingly once the proviral load exceeds 1% with accompanying persistent viral (Tax) gene expression [59]. Infected cells from patients with HAM/TSP express higher levels of Tax than those from asymptomatic carriers [60]. Tax expression is a major predictor of HAM/TSP status and accounts for high proviral load by driving the proliferation of infected cells [61]. The variation in the efficacy of the anti-Tax CTL response is one of the largest determinants of interindividual variation in proviral load within the HAM/TSP patient population [62]. The large pool of Tax-specific CTLs has been shown to be responsible for the neuropathology associated with HAM/TSP through one of two mechanisms: autoimmunity [63] or bystander cell damage [64]. As a result of these observations, the overall significance of the Tax-specific CTL response remains questionable, and the key elements of its functionality are yet to be elucidated. The antiviral efficacy of a CTL response is known to be affected by many factors including the efficiency of epitope processing and presentation by DCs [65, 66]. However, the role of DCs with respect to HTLV-1 pathogenesis in general and the Tax-specific CTL response in particular remains uncharacterized, despite clinical evidence supporting involvement of DCs in HTLV-1 pathogenesis [48]. In this regard, HTLV-1 has been shown to infect DCs efficiently in vitro [67] and in patients with HAM/TSP [49]. Additionally, DCs obtained from patients with HAM/TSP exhibit a maturation state and induce rapid lymphocyte proliferation [67] in contrast to those obtained from patients with ATL, who exhibit defective maturation [68], thereby demonstrating direct association with HTLV-1 immunopathogenesis in two disease states. The progression to HAM/TSP is marked by the spontaneous proliferation of lymphocytes, levels of which are believed to reflect the severity of the disease [69]. Depletion of DCs from the patient’s PBMCs abolishes spontaneous proliferation of lymphocytes, whereas supplementing DCs, but not B cells or macrophages (other potent APCs), restores proliferation. Furthermore, spontaneous proliferation of lymphocytes can be blocked by mAb to MHC class II, CD86, and CD58, indicating a DC-dependent mechanism [48]. In addition, some studies about DCs in HIV-1 and HTLV-1 disease indicate that infection of DCs may play a critical role in the development of T cell abnormalities [70]. Our previous studies also demonstrated that the HTLV-1 Tax protein is capable of modulating DC maturation and function in a manner similar to the TLR ligand [71,72,73,74,75,76]. Following up on those observations, data presented here from the in vitro and in vivo priming experiments demonstrate clearly that DCs, once exposed to HTLV-1 or Tax, undergo activation and provide constant antigen presentation and costimulation, resulting in the intense proliferation of lymphocytes, which represents a major pathogenic pathway in HTLV-1-associated autoimmune neuroinflammatory disease. Based on our observations and those of other laboratories, we hypothesize that in the setting of infection, HTLV-1-infected CD4⁺ T cells provide stimulus to CD8⁺ T cells and activate DCs simultaneously by the up-regulation of CD40L on their surface (demonstrated previously in patients with HAM/TSP) [51], the binding of which on DCs results in the up-regulation of costimulatory molecules and in the secretion of inflammatory cytokines [77, 78]. These activated DCs then take charge and boost this immune response, providing constant antigen presentation and leading to the intense proliferation of Tax-specific CTLs, as seen in patients with the neuroinflammatory condition. Therefore, DCs are critical mediators of the Tax-specific immune response during the onset and progression of HAM/TSP, and the nature of DC/T cell interaction dictates the effectiveness of Tax-specific CTLs after primary infection.

In addition to the classic class I presentation of the endogenously processed antigens, DCs are known to take up cell-associated antigens from live and dead cells. Based on the available clinical information, cross-presentation of Tax is the major possibility underlying continuous stimulation of anti-Tax CTLs, as CD4⁺ T cells are the primary HTLV-1-infected cell population and express high levels of Tax in patients with HAM/TSP [19]. Epidemiologic observations also suggest a greater possibility for cross-presentation of Tax. HAM/TSP is more likely to develop after i.v. exposure to HTLV-1 [79, 80], whereas mucosal exposure is suggested to result in ATL [81]. DCs are more likely to become infected in the mucosa and CD4⁺ T cells in the periphery. This scenario is also consistent with the concept of Tax cross-presentation by DCs. However, a small fraction of blood DCs (0.4–5%) from HAM/TSP patients was found to be infected with HTLV-1 [48] and was shown to be susceptible to infection in vitro [45, 49, 50]. Therefore, we found it important to investigate direct presentation of Tax peptides by infected DCs along with the cross-presentation pathway. Our data confirmed that both possibilities exist with respect to the Tax presentation by DCs, as shown previously in other systems [54, 82,83,84]. Although the chances of DCs cross-presenting Tax from apoptotic cells in patients with HAM/TSP are slim, as cell death is rare during HTLV-1 infection, these results are significant with respect to the DC-mediated antigen presentation from the leukemic cells in patients with ATL. Thus, different routes may be operating in patients with ATL (direct) compared with those with HAM/TSP (indirect), resulting in the two immunologically distinct pathogeneses. Overall, these studies provide comprehensive analyses of DC/T cell interplay, underlying a chronic, virus-induced autoimmune neuropathogenesis, and open new avenues for therapeutic interventions in HTLV-1-associated neuroinflammatory disease. Although the actual significance of the anti-Tax CTL response remains to be determined, our results establish DCs as the central player in this process and suggest a bidirectional approach to determining the solution. Similar to what is being considered for other autoimmune diseases, the inhibition of DC maturation could be explored as an immunosuppressive strategy in HAM/TSP. In this regard, specific agents are known to target differentiation-activated but not resting human DCs, such as the CMRF-44 antibody [85]. A benzothiophene derivative has also demonstrated marked suppression of T cells in patients with HAM/TSP by interfering with DC maturation [86]. Conversely, it is possible that Tax-specific CTLs are effective in killing the target cells during early infection but become exhausted during chronic antigen stimulation, as is seen with other persistent viral infections such as HIV and hepatitis C and B viruses in humans and lymphocytic choriomeningitis virus in rodents [87,88,89]. In that scenario, based on our observations, therapeutic efforts could be directed toward reconstituting a patient’s immune status with a fully functional DC-mediated CTL response or reversing exhausted phenotypes, as has been suggested for other chronic infections [90] by blockade of PD-1-mediated inhibitory signals [91,92,93,94,95] or by IL-10R [87]. This strategy has also been shown to enhance the potential of therapeutic vaccination [94]. In this regard, PD-1 involvement has been suggested recently in HTLV-1 infection [96]. Overall, the information obtained from these studies will provide insights pertinent to translational research initiatives about the development of anti-HTLV-1 therapeutics and immunologic paradigms for other neuroinflammatory syndromes.

ACKNOWLEDGMENTS

These studies were supported by the United States Public Health Service/National Institutes of Health grant R01 AI077414-01 to P. J.