Fratricide among CD8 T Lymphocytes Naturally Infected with Human T Cell Lymphotropic Virus Type I

Transcription

1 Immunity, Vol. 13, , November, 2000, Copyright 2000 by Cell Press Fratricide among CD8 T Lymphocytes Naturally Infected with Human T Cell Lymphotropic Virus Type I Emmanuel Hanon,* Jane C. Stinchcombe, Mineki Saito,* Becca E. Asquith,* Graham P. Taylor, Yuetsu Tanaka, Jonathan N. Weber, Gillian M. Griffiths, and Charles R. M. Bangham* *Departments of Immunology and associated diseases. Although the cellular receptor for HTLV-I has been mapped to chromosome 17 (Sommerfelt et al., 1988), the molecule has not yet been identified. HTLV-I can infect a wide range of human cell types in vitro (Sommerfelt et al., 1988), but in vivo the virus is thought to be almost confined to the CD4 T lymphocyte Genito-Urinary Medicine and Communicable Diseases subset (Richardson et al., 1990). Furthermore, nearly all Imperial College School of Medicine, St Mary s Campus of the malignancies induced by HTLV-I are tumors of Norfolk Place CD4 T lymphocytes (Uchiyama, 1997). Infection of London W2 1PG CD8 T lymphocytes by HTLV-I has been suggested by Sir William Dunn School of Pathology PCR detection of proviral DNA, but this claim remained University of Oxford controversial because of the difficulty of excluding con- Oxford tamination of PCR reactions (Koyanagi et al., 1993; Cho United Kingdom et al., 1995). Department of Infectious Disease and Immunology A highly active virus-specific cytotoxic T lymphocyte Okinawa-Asia Research Center of Medical Science (CTL) response is found in a majority of HTLV-I-infected Faculty of Medicine individuals (Jacobson et al., 1990; Kannagi et al., 1992; University of the Ryukyus Parker et al., 1992, 1994; Bieganowska et al., 1999). Uehara-cho 207, Nishihara Therefore, infection of a specific cell subset by HTLV-I Okinawa raises the important possibility that it could become Japan susceptible to lysis by HTLV-I-specific CTLs (Barchet et al., 2000) or NK cells (De Vecchis et al., 1985) in vivo. In this study, we provide direct evidence that HTLV-I infects CD8 T lymphocytes (including HTLV-I-specific Summary CD8 T cells) in vivo and renders them susceptible to autologous CTLs ex vivo. The ability of CD8 T cells Infection and gene expression by the human T lympho- to kill neighboring CD8 T cells (fratricide) has been tropic virus type I (HTLV-I) in vivo have been thought previously reported in vitro using cell lines (Walden and to be confined to CD4 T lymphocytes. We show here Eisen, 1990; Su et al., 1993; Huang et al., 1999) or CTLs that, in natural HTLV-I infection, a significant propor- obtained from TCR transgenic mice and infected with tion of CD8 T lymphocytes are infected by HTLV-I. recombinant viruses (Raftery et al., 1999). These artificial Interestingly, HTLV-I-specific but not Epstein-Barr vi- systems may not be representative of a normal viral rus specific CD8 T lymphocytes were shown to be infection in vivo. The data presented here demonstrate infected. Furthermore, HTLV-I protein expression in that fratricide can occur between naturally infected naturally infected CD8 T lymphocytes renders them CD8 T lymphocytes and autologous CTLs obtained susceptible to fratricide mediated by autologous HTLVfrom HTLV-I-infected individuals. I-specific CD8 T lymphocytes. Fratricide among virus-specific CTLs could impair the immune control of HTLV-I and possibly other lymphotropic viruses. Results Introduction Human T cell lymphotropic virus type 1 (HTLV-I) is a human retrovirus that infects approximately million people worldwide. The majority (about 95%) of infected individuals develop no associated disease (asymptomatic carriers; ACs). However, in a small frac- tion of individuals, infection results in one of two types of disease: about 2% 4% develop adult T cell leukemia (ATL), and a further 2% 4% develop one or more of a range of inflammatory diseases. The most common of these conditions is HTLV-I-associated myelopathy/trop- ical spastic paraparesis (HAM/TSP), an inflammatory disease of the central nervous system. A knowledge of which cell types harbor the virus in infected individuals is important for understanding the biology of the virus and the pathogenesis of HTLV-I- To whom correspondence should be addressed ( c.bangham@ ic.ac.uk). Infection of CD8 T Lymphocytes by HTLV-I Short-term cultivation of PBMCs obtained from HTLV- I-infected patients is associated with increased Tax protein expression in CD4 T lymphocytes (Hanon et al., 2000). We have recently devised a sensitive flow cyto- metric technique to detect Tax expression within PBMCs (Hanon et al., 2000). The specificity of this assay is shown in Figure 1A. Detection of Tax expression in MT-2 cells, which are chronically infected with HTLV-I, showed 99.8% of positive cells. The use of this assay on PBMCs isolated from an uninfected patient showed 0.0% of Tax-expressing cells. In contrast, PBMCs isolated from an infected patient and processed in similar conditions showed 1.6% of Tax-expressing cells. As expected, no staining was observed using an irrelevant monoclonal antibody (mab) (Figure 1A). Using the same method, cells infected with a Vaccinia virus recombinant for Tax stained positive, while those infected with a wildtype Vaccinia virus were negative (data not shown). We used this new flow cytometric approach to better

2 Immunity 658 Figure 1. Infection of CD8 T Cells by HTLV-I (A) Expression of HTLV-I Tax in MT-2 cells, uninfected PBMCs, and PBMCs obtained from an HTLV-I-infected patient. The PBMCs were cultivated for 12 hr in vitro before being harvested. The Tax protein was detected with the Lt-4 mab, and an irrelevant mab was used as a control isotype. (B) Expression of HTLV-I Tax in naturally infected CD4 and CD8 T lymphocytes. PBMCs were isolated from an HTLV-I-infected patient (TAF) and either cultivated for 6 hr or processed fresh for the concomitant detection of Tax, CD4, and CD8 antigens. (C) Tax mrna detection by PCR analysis in purified CD8 T cells (96% purity) obtained from an HTLV-I-infected patient (TAF). To exclude the possibility of false-positive signal due to contamination of purified CD8 T cells by infected PBMCs, uninfected CD8 T cells were artificially contaminated with 10%, 5%, 2.5%, or 0% of infected PBMCs (TAF) and submitted to PCR analysis. As a control, -actin mrna expression was also investigated. showed a typical distribution (Bex and Gaynor, 1998) of Tax expression in the nucleus and perforin expression in lytic granules in the same cells (Figure 3C). Quantitative analysis of PBMCs obtained from three ACs of the virus and three HAM/TSP patients revealed in each case that a proportion of Tax-expressing lym- phocytes (from 4% to 28%) were CD8 T lymphocytes. The Tax protein is capable of transactivating both viral and cellular genes, including CD25. Therefore, to further confirm expression of Tax in CD8 T lymphocytes, we determined whether Tax expression induces CD25 upregulation in CD8 T lymphocytes. As predicted, the analysis showed that more than 55% of Tax-expressing cells are positive for CD25 after 12 hr of cultivation, while none were CD25 positive after 6 hr. Similar results have been obtained with PBMCs isolated from three different HTLV-I-infected patients (data not shown). Furthermore, the analysis of CD69 (early activation marker) and CD71 (proliferation marker) expression gave comparable results. The expression of another HTLV-I antigen (p24; characterize the phenotype of Tax-expressing PBMCs. Surprisingly, the analysis showed that, in addition to CD4 T lymphocytes, a fraction of CD8 T lymphocytes also expressed the Tax protein after 6 hr of cultivation ex vivo (Figure 1B). To confirm infection of CD8 T cells by HTLV-I, the expression of tax mrna was also investigated in purified CD8 T cells by the use of a semiquantitative reverse PCR. As shown in Figure 1C, expression of the tax mrna is detected in CD8 T cells obtained from an HTLV-I-infected patient (TAF). In contrast, tax mrna expression is hardly visible in uninfected CD8 T cells where 10%, 5%, 2.5%, or 0% of infected PBMCs (TAF) were added before cultivation. This excludes the possibility that the positive signal observed in purified CD8 T cells (96% purity) obtained from an HTLV-Iinfected patient (TAF) is due to contamination. As expected, the level of -actin mrna is similar in the different samples. Expression of Tax in cytotoxic CD8 T lymphocytes was also confirmed by fluorescence microscopic analysis of purified CD8 T lymphocytes, which

3 Fratricide among HTLV-I-Infected CD8 T Cells 659 Table 1. Expression of HTLV-I p24 Protein in CD8 T Lymphocytes Table 3. Expression of HTLV-I Tax Protein in Virus-Specific CD8 T Lymphocytes Expression of p24 (%) a % of Tax Expression a Patient Tax-Negative CD8 T Cells Tax-Positive CD8 T Cells EBV BMLF Specific HTLV-I Tax Specific Patient CD8 T Cells CD8 T Cells TAF TAN TAF 0.0 (1.3) b 2.6 (5.3) TAT TAC 0.0 (0.4) 7.6 (3.2) TAK ND (UN) 38.1 (1.8) a Cells were cultivated in the presence of 20 nm of Concanamycin HH 0.0 (0.8) 4.6 (4.0) A for 24 hr. HY 0.0 (0.3) 16.7 (0.1) quency of p24 expression among Tax-expressing CD8 a Cells were cultivated in the presence of 20 nm of Concanamycin A for 6 hr. HTLV-I Gag antigen) was assayed in CD8 T lymphob Percentage of tetramer-positive cells in CD8 T lymphocytes. cytes of three HTLV-I-infected patients (TAF, TAN, TAT). UN, undetectable; and ND, not determined. The analysis showed that a large proportion (from 26% to 44%) of Tax-positive CD8 T lymphocytes also expatients pressed the p24 protein after 24 hr of cultivation in vitro revealed that a proportion (2.6% to 38%) of (Table 1). Only 0.2% of Tax-negative CD8 T lympho- Tax specific CD8 T cells was expressing HTLV-I cytes expressed the p24 protein (Table 1). This fre- Tax after 6 hr of cultivation (Table 3). In contrast to Tax specific CTLs, Epstein-Barr virus BMLF T cells is similar to that previously observed in Taxeven specific CTLs did not show any Tax expression (0.0%), expressing CD4 T cells (Hanon et al., 2000). Altogether, in the presence of CMA (Table 3). these results demonstrate that, in addition to CD4 T lymphocytes, HTLV-I also infects CD8 T lymphocytes. Time Course Study of Tax Expression in CD8 T Lymphocytes CD8 T Lymphocytes Are Infected by HTLV-I In Vivo We have previously shown that infected CD4 T lympho- Although it is very unlikely, infected CD4 T lymphocytes cytes expressing the HTLV-I Tax protein become suscould produce virions and infect CD8 T lymphocytes ceptible to lysis by HTLV-I-specific CTLs (Hanon et al., during the short-term cultivation in vitro (Gessain et al., 2000) (CD56 NK cells did not contribute detectably 1990). To exclude this possibility, the expression of Tax to the lysis of Tax-expressing cells). This observation was also investigated in CD8 T lymphocytes cultivated suggested that HTLV-I-specific CTLs might also kill Tax- in the absence of CD4 T lymphocytes. Table 2 shows expressing CD8 T lymphocytes. A time course study of that neither the depletion of CD4 T lymphocytes nor Tax expression was therefore carried out using PBMCs the enrichment of CD8 T lymphocytes has a significant obtained from three HTLV-I-infected patients. Figure 2A effect on the frequency of Tax expression in CD8 T shows that Tax expression in CD8 T lymphocytes lymphocytes. reached a maximum at 6 hr and then decreased by more than 50% during the next 18 hr. In contrast to Tax expression, Preferential Infection of HTLV-I-Specific the percentage of CD8 T lymphocytes re- CD8 T Lymphocytes by HTLV-I mained constant throughout the time course (data not We wished to test the possibility that HTLV-I-specific shown), which excludes the possibility that the observed CD8 T lymphocytes could themselves become infected changes in Tax expression were due to a dramatic with HTLV-I. Since Tax is the dominant target epitope change in the frequency of CD8 T lymphocytes. recognized by HTLV-I-specific CTLs in HLA-A*0201 individuals (Jacobson et al., 1990; Kannagi et al., 1992; Par- Role of Cell-Mediated Cytotoxicity in the ker et al., 1992, 1994; Bieganowska et al., 1999), infection Disappearance of Tax-Expressing CD8 T of HTLV-I-specific CTLs was investigated using fluores- Lymphocytes In Vitro cent-labeled tetramers of HLA-A* microglobu- The perforin pathway has been shown to be essential for lin Tax peptide (Altman et al., 1996; Jeffery et al., HTLV-I-specific CTLs to mediate their cytotoxic function 1999). Flow cytometric analyses of five HTLV-I-infected against Tax-expressing lymphocytes (Hanon et al., 2000). Therefore, to determine whether the observed decrease in the frequency of Tax-expressing CD8 T Table 2. Expression of HTLV-I Tax Protein in CD8 T cells after 12 hr of cultivation is due to cell-mediated Lymphocytes cytotoxicity, PBMCs from three patients (TAU, TAN, TW) Tax Expression (%) were cultivated for 12 hr in the presence of 20 nm of concanamycin A (CMA), an inhibitor of the perforin- Patient PBMC Cultures CD8 T Cells CD8 T Cells dependent cytotoxic pathway (Kataoka et al., 1996). TAL Flow cytometric analyses showed that the frequency of CD4 depleted Tax-expressing CD8 T lymphocytes was respectively TAN , 1.7-, and 2.3-fold higher in CMA-treated cultures CD4 depleted TAZ ND 2.2 than in control cultures (Figure 2B). The requirement for CD8 enriched ND 2.8 cell cell contact was also investigated to further deter- ND, not determined. mine the role of cell-mediated cytotoxicity in the observed decrease of Tax expression. The frequency of Tax-

4 Immunity 660 expressing PBMCs obtained from these patients (TAU, TAN, TW) was compared in conventional cultures, where PBMCs are cultivated in round bottom wells and therefore establish frequent cell cell contacts, with other cultures, in which PBMCs are cultivated with the same volume of medium in flat bottom wells to reduce the frequency of cell cell contacts. Flow cytometric analyses made with PBMCs cultivated for 12 hr showed that the reduction of cell cell contacts is associated with an increase (2-, 1.5-, and 2-fold, respectively) in the frequency of Tax-expressing CD8 T lymphocytes (Figure 2B), which is similar to the increase caused by CMA treatment (Figure 2B). These observations exclude the possibility that a putative Tax-induced suicide process is implicated in the observed decrease of Tax expression (Figure 2A). Furthermore, these results are consistent with the possibility that Tax-expressing CD8 T lymphocytes are killed by HTLV-I-specific CTLs. Consistent with this hypothesis, the addition of an anti-mhc class I mab in culture of purified CD8 T cells also prevented the decrease in the frequency of Tax-expressing CD8 T cells, while an anti-mhc class II antibody did not (Figure 2C). Figure 2. Role of Cell-Mediated Cytotoxicity in the Disappearance of Tax-Expressing CD8 T Lymphocytes In Vitro (A) Time course study of Tax expression in naturally infected CD8 T lymphocytes. PBMCs were isolated from three HTLV-I-infected patients (TAU, TAN, TAF), cultivated for various times, and then processed for the detection of Tax, CD4, and CD8 antigens. (B) Effect of concanamycin A (CMA) and reduction of cell-to-cell contact on the frequency of Tax expression in CD8 T lymphocytes. PBMCs were isolated from an HTLV-I-infected patient, cultivated with or without CMA for 12 hr, and then processed for the concomi- tant detection of Tax and CD8 antigens. The effect of the prevention of cell cell contact has been investigated by comparing the frequency of Tax expression in CD8 lymphocytes cultivated in round bottom wells or in flat bottom wells. One representative experiment is shown. (C) Effect of anti-mhc class I or -MHC class II mabs on the frequency of Tax expression in CD8 lymphocytes. PBMCs were isolated from an HTLV-I-infected patient, cultivated with or without the mabs for 12 hr, and then processed for the concomitant detection of Tax and CD8 antigens. Fratricide among HTLV-I-Infected CD8 T Lymphocytes The decrease in the frequency of Tax-expressing CD8 T cells in culture requires close cell-to-cell contact and intact perforin granules. Consistent with these observations, confocal microscopic analysis (Figure 3) of cultures of purified CD8 T cells showed numerous CD8 T cells with perforin lytic granules polarized toward the point of contact with Tax-expressing CD8 T cells. Figures 3C 3F correspond to serial sections through a single group of cells showing a perforin-expressing T cell (boundaries shown by three large arrowheads) in contact with a cell that expresses both perforin and Tax. The granules of the cell that only expresses perforin are polarized toward the area of contact with the Taxexpressing cell. In contrast, the perforin granules of neighboring CD8 T cells (e.g., small arrowheads in [F]) and those of the Tax-expressing cell itself are dispersed. These morphological observations strongly support the possibility that Tax-expressing CD8 T cells are being killed by CTLs. In this case, there are two implications. First, the level of mortality in Tax-positive CD8 T lymphocytes should be specifically higher than in Tax-negative CD8 T lymphocytes. Second, the presence of CMA should specifically reduce this higher level of mortality in Tax-positive CD8 T lymphocytes. Therefore, to test our predictions, a CTL assay based on the selective propidium iodide incorporation by dead cells was used (Radosevic et al., 1990; Mattis et al., 1997). For this experiment, purified CD8 T lymphocytes isolated from five infected patients were cultivated in the presence or absence of 20 nm of CMA for 12 hr, harvested, and incubated with propidium iodide (PI) to label dead cells. The concomitant detection of HTLV-I Tax and PI by flow cytometry showed that, in the absence of CMA, the level of mortality in Tax-expressing CD8 T lymphocytes was indeed higher than in the Tax-negative population (Fig- ure 4). As predicted, in the presence of CMA, the level of mortality dropped in Tax-expressing CD8 T lympho-

5 Fratricide among HTLV-I-Infected CD8 T Cells 661 Figure 3. The Perforin Granules of CD8 T Cells Polarize toward the Site of Contact with Tax-Expressing CD8 T Cells Negatively purified CD8 T cells isolated from an HTLV-I-infected patient were cultivated for 12 hr and double immunolabeled with antibodies to detect both perforin (red) and Tax (green) expression. (A and B) Fluorescence and transmission images of the same cells showing perforin-containing granules (red) polarized toward the Tax-expressing cells (green). (C F) Serial sections through a single group of cells showing a perforin-expressing T cell (boundaries shown by three large arrowheads) in contact with a cell that expresses both perforin and Tax. The granules of the cell that only expresses perforin are polarized toward the area of contact with the Taxexpressing cell. In contrast, the perforin granules of neighboring CD8 T cells (e.g., small arrowheads in [F]) and those of the Taxexpressing cell itself are dispersed. One representative experiment out of three is shown. Bars represent 10 m. cytes while it slightly increased in the Tax-negative population (Figure 4). Altogether, these results demonstrate that HTLV-I-specific CTLs are capable of mediating fratricide against Tax-expressing CD8 T lymphocytes. Discussion Infection and gene expression by HTLV-I in vivo have been thought to be confined to CD4 T lymphocytes. We show here that, in naturally HTLV-I-infected patients, a significant proportion of CD8 T lymphocytes are also infected by HTLV-I. Interestingly, HTLV-I-specific but not Epstein-Barr virus-specific CD8 T lymphocytes were shown to be infected (Table 3). The reasons for this apparent preferential infection of HTLV-I-specific CD8 T cells are unknown. However, cell-to-cell contact (Miyoshi et al., 1981; Popovic et al., 1983) and cellular activation (Merl et al., 1984; Krichbaum-Stenger et al., Figure 4. Mortality Rates in Tax-Positive and -Negative CD8 T Lymphocytes CD8 T lymphocytes were negatively purified from five HTLV-I-infected patients (TAZ, TAN, TAU, HH, HY) and cultivated in the presence or absence of 20 nm of CMA for 12 hr. After being harvested, cells were incubated with propidium iodide (PI) to label dead cells and processed to detect concomitantly Tax and PI by flow cytometry.

6 Immunity ) greatly enhance the transmission of HTLV-I be- and enhances LCMV persistence in vivo (Zinkernagel et tween cells. Therefore, activation of HTLV-I-specific al., 1999). CD8 T cells by HTLV-I-infected cells (CD4 T cells and/ In conclusion, we show here that, in addition to CD4 or HTLV-I-infected dendritic cells) might increase their T lymphocytes, CD8 T lymphocytes including HTLV-Isusceptibility to HTLV-I infection. Dendritic cells can be specific CTLs are infected by HTLV-I. Furthermore, viral infected with HTLV-I in vitro (Makino et al., 1999), and protein expression within CD8 T lymphocytes renders Macatonia et al. (1992) estimated that 0.4% to 5.1% of them susceptible to fratricide mediated by virus-specific dendritic cells were infected with HTLV-I (as determined CTLs. This could specifically impair the immune control by in situ hybridization) in HTLV-I-positive individuals. of HTLV-I and therefore contribute to the persistence of Finally, in HIV infection, dendritic cells have been shown HTLV-I in vivo. A comparable mechanism might also to facilitate the transmission of the virus to CD4 T cells reduce the efficiency of the immune control of other by promoting their activation (Geijtenbeek et al., 2000). lymphotropic viruses, such as HIV-1, CMV, or HSV-1, Further work is required to investigate the interaction especially if the infection is selective for virus-specific between HTLV-I-specific CD8 T cells and their target T lymphocytes. cells. HTLV-I infection and protein expression could have Experimental Procedures important biological effects on CD8 T lymphocytes. HTLV-I infection has been shown to disrupt the function Patients and Cells Peripheral blood mononuclear cells (PBMC) were obtained from of CTLs in vitro (Popovic et al., 1984; Yarchoan et al., seven HTLV-I-infected patients with a clinical diagnosis of HAM/ 1986; Inatsuki et al., 1991). We also show that Tax protein TSP and four asymptomatic carriers of the virus. All patients gave expression in CD8 T cells is associated with upregula- informed consent. PBMCs were isolated on Histopaque-1077 tion of the activation markers CD25, CD69, and CD71. (Sigma, Dorset, UK) density gradient and washed three times with This might deregulate normal CD8 T lymphocyte biomicrobeads phosphate- buffered saline (PBS). A CD8 T cell isolation kit and CD4 logical functions. Furthermore, we demonstrate that (Miltenyi Biotec Ltd, Surrey, UK) were used according to viral protein expression in HTLV-I-infected CD8 T lymthe manufacturer s instructions to enrich or deplete the respective PBMC subpopulations. Cells were cultured at per milliliter phocytes renders them susceptible to fratricide mediin RPMI 1640 medium (GIBCO, Paisley, UK) supplemented with 10% ated by autologous CTLs (Figures 2, 3, and 4). Several fetal calf serum (FCS) (Sigma, Dorset, UK), 2 mm glutamine (GIBCO), observations support this conclusion. First, reduction 100 IU/ml penicillin (GIBCO), and 100 g/ml streptomycin (GIBCO). of cell-to-cell contact, inhibition of perforin production, In some experiments, 20 nm of concanamycin A (CMA) (Sigma), 5 or anti-mhc class I but not anti-mhc class II antibodies g/ml of anti-mhc class I mab (W6/32, IgG2a) (Serotec, Oxford, prevented the rapid decrease in the frequency of Tax- England), or 5 g/ml of anti-mhc class II mab (TÜ39, IgG2a) (Becton Dickinson UK Ltd.) was added to the culture medium. For the experiexpressing CD8 T cells observed in cultures ex vivo ment where the requirement of cell-to-cell contact has been tested, (Figure 2). Second, confocal microscopic analysis (Fig- the cells have been cultivated at per milliliter. ure 3) showed numerous CD8 T cells with perforin lytic granules polarized toward the point of contact with Tax- Flow Cytometry expressing CD8 T cells, which are morphological char- For the concomitant detection of Tax, CD4, CD8, and CD25 antigens, acteristics of the cell-mediated cytotoxicity process cells were fixed in PBS containing 2% paraformaldehyde (Sigma) (Berke, 1994). Finally, a flow cytometric CTL assay (Rawere for 20 min and resuspended in PBS at 4 C until use. Fixed cells washed with PBS containing 7% of normal goat serum (NGS) dosevic et al., 1990; Mattis et al., 1997) demonstrated (Sigma) and incubated with PC5-labeled anti-cd4, ECD-labeled a higher level of cell death in Tax-positive CD8 T lymanti-cd8, and PE-labeled anti-cd25 mabs (Beckman Coulter, Bedphocytes than in Tax-negative CD8 T lymphocytes. fordshire, UK) for 15 min at room temperature. The cells were then Furthermore, the presence of CMA specifically reduced washed and permeabilized with PBS containing 0.1% Triton X-100 the high rate of mortality in Tax-positive but not Tax- (Sigma) for 10 min at room temperature. Permeabilized cells were negative CD8 T lymphocytes (Figure 4). washed and resuspended in PBS/7% NGS containing an anti-htlv-i Three recent lines of evidence suggest that the class I Tax mab (Lt-4; IgG 3 ) (Lee et al., 1989) or an isotype control mab (Southern Biotechnology Associates, Birmingham, AL) for 20 min at MHC-restricted T cell response is highly active in HTLV-I room temperature. The cells were then washed twice and resusinfection in vivo. First, the dominant HTLV-I antigen recpended in PBS/7% NGS containing FITC-labeled goat F(ab ) 2 antiognized by virus-specific CTLs, Tax, is subject to posi- mouse IgG 3 serum (Southern Biotechnology Associates) for 20 min tive selection in healthy HTLV-I carriers (Niewiesk et al., at room temperature. Finally, the cells were washed twice and ana- 1994). Second, the class I allele HLA-A*02 is associated lyzed by flow cytometry on a Coulter EPICS XL (Beckman Coulter). with both protection against the HTLV-I-associated ingen), For the concomitant detection of Tax and p24 (HTLV-I gag anti- flammatory disease HAM/TSP and with a reduction in the cells were processed as described above for detection of Tax, but an anti-p24 mab (MAB8817; IgG 1 ) (Chemicon International, the provirus load of HTLV-I (Jeffery et al., 1999). Finally, Temecula, CA) was used in addition to the Lt-4 mab. The anti-p24 autologous HTLV-I-specific CTLs efficiently kill Tax- Mab was then detected with a PE-labeled goat F(ab ) anti-mouse 2 expressing CD4 T lymphocytes in freshly isolated and IgG 1 serum (Southern Biotechnology Associates). In addition, the unstimulated peripheral blood mononuclear cells (Ha- CD4 and CD8 membrane antigens were detected after Tax and p24 non et al., 2000). The occurrence of fratricide within labeling in the presence of 1% mouse serum to prevent crossreactions. HTLV-I-infected CD8 T lymphocytes might therefore impair the immune control of HTLV-I in vivo (Barchet et For the concomitant detection of Tax and cell mortality by flow cytometry, the cells were harvested and incubated for 10 min in the al., 2000), especially if the virus selectively infects HTLVpresence of 5 g per milliliter of propidium iodide (PI) (Sigma). Then I-specific T lymphocytes as suggested by our data (Tathe cells were washed twice with PBS, fixed in PBS containing 2% ble 3). It has been similarly suggested that CTL-medi- paraformaldehyde for 20 min, and resuspended in PBS at 4 C until ated lysis of LCMV-specific B lymphocytes, infected use. Fixed cells were then washed with PBS/7% NGS and processed with LCMV, impairs the antibody response to the virus as described above to detect the Tax protein.

7 Fratricide among HTLV-I-Infected CD8 T Cells 663 Analysis of PBMCs for the presence of Tax specific CD8 T Berke, G. (1994). The binding and lysis of target cells by cytotoxic lymphocytes was performed by using fluorescent-labeled tetramers lymphocytes: molecular and cellular aspects. Annu. Rev. Immunol. of HLA-A* microglobulin Tax peptide (Altman et al., 12, ; Jeffery et al., 1999). PBMCs were incubated with PBS con- Bex, F., and Gaynor, R.B. (1998). Regulation of gene expression by taining 5% FCS and Tax tetramer at 37 C for 30 min. The cells HTLV-I Tax protein. Methods 16, were then washed three times in ice-cold PBS, fixed in PBS containing 2% paraformaldehyde for 30 min at 4 C, and processed as Bieganowska, K., Hollsberg, P., Buckle, G.J., Lim, D.G., Greten, T.F., described above to detect the Tax, CD4, and CD8 antigens. Schneck, J., Altman, J.D., Jacobson, S., Ledis, S.L., Hanchard, B., et al. (1999). Direct analysis of viral-specific CD8 T cells with soluble HLA- A2/Tax11-19 tetramer complexes in patients with human Quantification of tax mrna T cell lymphotropic virus-associated myelopathy. J. Immunol. 162, RNA was isolated using High Pure RNA Isolation Kit (Boehringer Mannheim, Lewes, UK). cdna was then synthesized using 1st Strand cdna Synthesis Kit for RT-PCR AMV (Boehringer Mannheim). Cho, I., Sugimoto, M., Mita, S., Tokunaga, M., Imamura, F., and PCR reactions were carried out in 50 l containing cdna prepared Ando, M. (1995). In vivo proviral burden and viral RNA expression from 10 ng of RNA with primers for HTLV-I Tax or human -actin. in T cell subsets of patients with human T lymphotropic virus type- Tax primers were 5 -TCGCTGCCGATCACGATGCGTTTCC-3 and 1-associated myelopathy/tropical spastic paraparesis. Am. J. Trop. 5 -AACACGTAGACTGGGTATCC-3. Human -actin primers were Med. Hyg. 53, AAGAGAGGCATCCTCACCCT-3 and 5 -TACATGGCTGGGGTG De Vecchis, L., Graziani, G., Macchi, B., Grandori, C., Pastore, S., TTGAA-3. Cycle conditions for amplification of the tax sequence Popovic, M., Gallo, R.C., and Bonmassar, E. (1985). Decline of natuwere 1 cycle at 95 C for 5 min followed by 45 cycles at 95 C for 60 ral cytotoxicity of human lymphocytes following infection with hus, 58 C for 75 s, 72 C for 90 s, and 1 final cycle at 72 C for 10 min. man T-cell leukemia/lymphoma virus (HTLV). Leuk. Res. 9, Cycle conditions for amplification of the -actin sequence were 1 Geijtenbeek, T.B., Kwon, D.S., Torensma, R., van Vliet, S.J., van cycle at 95 C for 5 min followed by 30 cycles at 95 C for 60 s, 58 C Duijnhoven, G.C., Middel, J., Cornelissen, I.L., Nottet, H.S., Kewalfor 75 s, 72 C for 90 s, and 1 final cycle at 72 C for 10 min. In Ramani, V.N., Littman, D.R., Figdor, C.G., and van Kooyk, Y. (2000). preliminary experiments, the number of cycles was varied between DC-SIGN, a dendritic cell-specific HIV-1-binding protein that en- 25 to 50 to determine the appropriate number of cycles for a semihances trans-infection of T cells. Cell 100, quantitative PCR (data not shown). Gessain, A., Saal, F., Giron, M.L., Lasneret, J., Lagaye, S., Gout, O., Confocal Microscopy De The, G., Sigaux, F., and Peries, J. (1990). Cell surface phenotype Cells were adhered to glass slides coated with 50 g/ml poly-lcell and human T lymphotropic virus type 1 antigen expression in 12 T lysine (Sigma), fixed in PBS containing 2% paraformaldehyde lines derived from peripheral blood and cerebrospinal fluid of (Sigma) for 20 min, and incubated in PBS at 4 C until use. Fixed West Indian, Guyanese and African patients with tropical spastic cells were washed with PBS containing 7% of normal goat serum paraparesis. J. Gen.Virol. 71, (NGS) (Sigma) and permeabilized with PBS containing 0.1% Triton Hanon, E., Hall, S., Taylor, G., Saito, M., Davis, R., Tanaka, Y., Usuku, X-100 (Sigma) for 10 min at room temperature. Permeabilized cells K., Osame, M., Weber, J., and Bangham, C. (2000). Abundant Tax were washed and resuspended in PBS/7% NGS containing anti- protein expression in CD4 T cells infected with human T-cell leu- Tax mab (Lt-4; IgG 3 ) and anti-perforin mab ( G9; IgG 2b ) (Becton kaemia virus type I (HTLV-I) is prevented by cytotoxic T lymphocytes. Dickinson UK Ltd.) for 20 min at room temperature. The cells were Blood 95, then washed twice and resuspended in PBS/7% NGS containing Huang, J.F., Yang, Y., Sepulveda, H., Shi, W., Hwang, I., Peterson, FITC-labeled goat F(ab ) 2 anti-mouse IgG 3 serum (Southern Biotech- P.A., Jackson, M.R., Sprent, J., and Cai, Z. (1999). TCR-mediated nology Associates) and Texas red labeled goat anti-mouse IgG 2b internalization of peptide-mhc complexes acquired by T cells. Sciserum (Southern Biotechnology Associates) for 20 min at room temence 286, perature. Finally, the cells were washed twice and mounted in 90% glycerol/pbs containing 2.5% 1-4-diazabicyclo[2.2.2]octane (DABCO) Inatsuki, A., Yasukawa, M., and Kobayashi, Y. (1991). The effect of (Sigma) antibleaching reagent. Slides were then analyzed using the human T cell leukaemia virus type I infection on a herpes simplex Bio-Rad MRC 1024 confocal microscope. virus-specific CD8 cytotoxic T cell clone. Br. J. Haematol. 77, Acknowledgments Jacobson, S., Shida, H., McFarlin, D.E., Fauci, A.S., and Koenig, S. (1990). Circulating CD8 cytotoxic T lymphocytes specific for We thank the staff and patients of St Mary s Hospital. The authors HTLV-I px in patients with HTLV-I associated neurological disease. would also like to thank Graham Ogg (John Radcliffe Hospital, Oxford, Nature 348, UK) and Margaret Callan (John Radcliffe Hospital, Oxford, UK) Jeffery, K.J., Usuku, K., Hall, S.E., Matsumoto, W., Taylor, G.P., for providing fluorescent-labeled tetramers. This work was sup- Procter, J., Bunce, M., Ogg, G.S., Welsh, K.I., Weber, J.N., et al. ported by the Wellcome Trust (UK), the Royal Society (UK), and (1999). HLA alleles determine human T-lymphotropic virus-i (HTLV-I) the Fonds National Belge de la Recherche Scientifique (F.N.R.S., proviral load and the risk of HTLV-I-associated myelopathy. Proc. Belgium). E. H. is a Senior Research Assistant of the F.N.R.S. The Natl. Acad. Sci. USA 96, HTLV European Research Network supported by the European Kannagi, M., Shida, H., Igarashi, H., Kuruma, K., Murai, H., Aono, Community Biomed Program provided a forum for critical appraisal Y., Maruyama, I., Osame, M., Hattori, T., and Inoko, H. (1992). Target of this work. epitope in the Tax protein of human T-cell leukemia virus type I Received June 15, 2000; September 28, recognized by class I major histocompatibility complex-restricted cytotoxic T cells. J. Virol. 66, References Kataoka, T., Shinohara, N., Takayama, H., Takaku, K., Kondo, S., Yonehara, S., and Nagai, K. (1996). Concanamycin A, a powerful Altman, J.D., Moss, P.H., Goulder, P.R., Barouch, D.H., McHeyzerand Fas-based lytic pathways in cell-mediated cytotoxicity. J. Imtool for characterization and estimation of contribution of perforin- Williams, M.G., Bell, J.I., McMichael, A.J., and Davis, M.M. (1996). Phenotypic analysis of antigen-specific T lymphocytes. Science munol. 156, , Koyanagi, Y., Itoyama, Y., Nakamura, N., Takamatsu, K., Kira, J., Barchet, W., Oehen, S., Klenerman, P., Wodarz, D., Bocharov, G., Iwamasa, T., Goto, I., and Yamamoto, N. (1993). In vivo infection of Lloyd, A.L., Nowak, M.A., Hengartner, H., Zinkernagel, R.M., and human T-cell leukemia virus type I in non-t cells. Virology 196, Ehl, S. (2000). Direct quantitation of rapid elimination of viral antigenpositive lymphocytes by antiviral CD8( ) T cells in vivo. Eur. J. Krichbaum-Stenger, K., Poiesz, B.J., Keller, P., Ehrlich, G., Gaval- Immunol. 30, chin, J., Davis, B.H., and Moore, J.L. (1987). Specific adsorption of

8 Immunity 664 HTLV-I to various target human and animal cells. Blood 70, Lee, B., Tanaka,Y., and Tozawa, H. (1989). Monoclonal antibody defining tax protein of human T-cell leukemia virus type-i. Tohoku. J. Exp. Med. 157, Macatonia, S.E., Cruickshank, J.K., Rudge, P., and Knight, S.C. (1992). Dendritic cells from patients with tropical spastic paraparesis are infected with HTLV-1 and stimulate autologous lymphocyte proliferation. AIDS Res. Hum. Retroviruses 8, Makino, M., Shimokubo, S., Wakamatsu, S.I., Izumo, S., and Baba, M. (1999). The role of human T-lymphotropic virus type 1 (HTLV-1)- infected dendritic cells in the development of HTLV-1-associated myelopathy/tropical spastic paraparesis. J. Virol. 73, Mattis, A.E., Bernhardt, G., Lipp, M., and Forster, R. (1997). Analyzing cytotoxic T lymphocyte activity: a simple and reliable flow cytometry-based assay. J. Immunol. Methods 204, Merl, S., Kloster, B., Moore, J., Hubbell, C., Tomar, R., Davey, F., Kalinowski, D., Planas, A., Ehrlich, G., Clark, D., et al. (1984). Efficient transformation of previously activated and dividing T lymphocytes by human T cell leukemia-lymphoma virus. Blood 64, Miyoshi, I., Kubonishi, I., Yoshimoto, S., Akagi, T., Ohtsuki, Y., Shiraishi, Y., Nagata, K., and Hinuma, Y. (1981). Type C virus particles in a cord T-cell line derived by co-cultivating normal human cord leukocytes and human leukaemic T cells. Nature 294, Niewiesk, S., Daenke, S., Parker, C.E., Taylor, G., Weber, J., Nightingale, S., and Bangham, C.R. (1994). The transactivator gene of human T-cell leukemia virus type I is more variable within and between healthy carriers than patients with tropical spastic paraparesis. J. Virol. 68, Parker, C.E., Daenke, S., Nightingale, S., and Bangham, C.R. (1992). Activated, HTLV-1-specific cytotoxic T-lymphocytes are found in healthy seropositives as well as in patients with tropical spastic paraparesis. Virology 188, Parker, C.E., Nightingale, S., Taylor, G.P., Weber, J., and Bangham, C.R. (1994). Circulating anti-tax cytotoxic T lymphocytes from human T-cell leukemia virus type I-infected people, with and without tropical spastic paraparesis, recognize multiple epitopes simultaneously. J. Virol. 68, Popovic, M., Sarin, P.S., Robert-Gurroff, M., Kalyanaraman, V.S., Mann, D., Minowada, J., and Gallo, R.C. (1983). Isolation and transmission of human retrovirus (human t-cell leukemia virus). Science 219, Popovic, M., Flomenberg, N., Volkman, D.J., Mann, D., Fauci, A.S., Dupont, B., and Gallo, R.C. (1984). Alteration of T-cell functions by infection with HTLV-I or HTLV-II. Science 226, Radosevic, K., Garritsen, H.S., van Graft, M., de Grooth, B.G., and Greve, J. (1990). A simple and sensitive flow cytometric assay for the determination of the cytotoxic activity of human natural killer cells. J. Immunol. Methods 135, Raftery, M.J., Behrens, C.K., Muller, A., Krammer, P.H., Walczak, H., and Schonrich, G. (1999). Herpes simplex virus type 1 infection of activated cytotoxic T cells: induction of fratricide as a mechanism of viral immune evasion. J. Exp. Med. 190, Richardson, J.H., Edwards, A.J., Cruickshank, J.K., Rudge, P., and Dalgleish, A.G. (1990). In vivo cellular tropism of human T-cell leukemia virus type 1. J. Virol. 64, Sommerfelt, M.A., Williams, B.P., Clapham, P.R., Solomon, E., Goodfellow, P.N., and Weiss, R.A. (1988). Human T cell leukemia viruses use a receptor determined by human chromosome 17. Science 242, Su, M.W., Walden, P.R., Eisen, H.N., and Golan, D.E. (1993). Cognate peptide-induced destruction of CD8 cytotoxic T lymphocytes is due to fratricide. J. Immunol. 151, Uchiyama, T. (1997). Human T cell leukemia virus type I (HTLV-I) and human diseases. Annu. Rev. Immunol. 15, Walden, P.R., and Eisen, H.N. (1990). Cognate peptides induce selfdestruction of CD8 cytolytic T lymphocytes. Proc. Natl. Acad. Sci. USA 87, Yarchoan, R., Guo, H.G., Reitz, M.J., Maluish, A., Mitsuya, H., and Broder, S. (1986). Alterations in cytotoxic and helper T cell function after infection of T cell clones with human T cell leukemia virus, type I. J. Clin. Invest. 77, Zinkernagel, R.M., Planz, O., Ehl, S., Battegay, M., Odermatt, B., Klenerman, P., and Hengartner, H. (1999). General and specific immunosuppression caused by antiviral T-cell responses. Immunol. Rev. 168,