Human T-cell leukemia virus type-I (HTLV-I)-specific T-cell responses detected using three-divided glutathione-S-transferase (GST)-Tax fusion proteins

Journal of Immunological Methods 313 (2006) 61–73www.elsevier.com/locate/jim

Research paper

Human T-cell leukemia virus type-I (HTLV-I)-specific T-cellresponses detected using three-divided glutathione-S-transferase

(GST)-Tax fusion proteins

Kiyoshi Kurihara a, Yukiko Shimizu a, Ayako Takamori a, Nanae Harashima a,Misato Noji a, Takao Masuda a, Atae Utsunomiya b,

Jun Okamura c, Mari Kannagi a,⁎

a Department of Immunotherapeutics, Tokyo Medical and Dental University, Medical Research Division,1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan

b Department of Hematology, Imamura Bun-in Hospital, Kagoshima 890-0064, Japanc Institute for Clinical Research, National Kyushu Cancer Center, Fukuoka, 811-1395, Japan

Received 2 November 2005; received in revised form 16 March 2006; accepted 16 March 2006Available online 4 May 2006

Abstract

Insufficient T-cell response to human T-cell leukemia virus type-I (HTLV-I) is a potential risk factor in adult T-cell leukemia(ATL). We established an assay system for detecting HTLV-I-specific T-cell response by using recombinant glutathione-S-transferase (GST) proteins fused with HTLV-I Tax protein that was divided into three portions, Tax-A, -B, and -C, corresponding tothe N-terminal, central and C-terminal regions, respectively. When splenocytes from rats immunized with plasmids encoding TaxcDNA were incubated with these recombinant proteins, strong interferon gamma (IFN-γ-producing responses occurred againstGST-Tax proteins but not against control GST proteins. No such Tax-specific responses were observed in splenocytes from naiverats. Cocktails of oligopeptides corresponding to the Tax-A, -B, and -C regions also induced IFN-γ-producing responses whenincubated with splenocytes from immunized rats, but required higher amounts of antigens and there were a shorter periods ofsustained T-cell responses than with GST-Tax protein-based assay. Although splenocytes from immunized rats predominantlyreacted against GST-Tax-B protein, they failed to react with peptide cocktails corresponding to the Tax-B region, likely because themajor epitope was interrupted in the initially prepared series of peptides. Using a newly prepared peptide series we found thatsplenocytes predominantly reacted with a peptide located in the Tax-B region that overlaps with a previously identified cytotoxic Tlymphocytes (CTL) epitope of this rat strain. Using this system, we examined peripheral blood mononuclear cells (PBMC) from anATL patient who underwent complete remission following hematopoietic stem cell transplantation (HSCT). PBMC from thispatient produced a significant Tax-specific T-cell response predominantly against GST-Tax-A protein. This is consistent with theprevious finding that this patient exhibited a strong HLA-A2-restricted CTL response to Tax 11–19 epitope, which is located in theTax-A region. This study provides a diagnostic tool, useful for monitoring HTLV-I-specific T-cell immunity in patients and forsurveying HTLV-I-carriers to identify an immunological group at high risk for ATL development, regardless of their human

Abbreviations: HTLV-I, human T-cell leukemia virus type-I; ATL, adult T-cell leukemia; GST, glutathione-S-transferase; IFN-γ, interferongamma; CTL, cytotoxic T lymphocytes; PBMC, peripheral blood mononuclear cells; HSCT, hematopoietic stem cell transplantation; HLA, humanleukocyte antigen; MHC, major histocompatibility complex; ELISA, enzyme linked immunosorbent assay.⁎ Corresponding author. Tel.: +81 3 5803 5798; fax: +81 3 5803 0235.E-mail address: [email protected] (M. Kannagi).

0022-1759/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.jim.2006.03.013

mailto:[email protected]

http://dx.doi.org/10.1016/j.jim.2006.03.013

62 K. Kurihara et al. / Journal of Immunological Methods 313 (2006) 61–73

leukocyte antigen (HLA) types. It is also useful for predicting the location of T-cell epitopes, which may be applicable in futurevaccine strategies.© 2006 Elsevier B.V. All rights reserved.

Keywords: HTLV-I; Cellular immune response; Anti-tumor immunity; Viral infection; Epitope

1. Introduction

Adult T-cell leukemia (ATL) (Uchiyama et al., 1977),is a consequence of uncontrolled clonal expansion ofhuman T-cell leukemia virus type-I (HTLV-I)-infected T-lymphocytes (Poiesz et al., 1980; Hinuma et al., 1981;Yoshida et al., 1982), which are highly resistant to anti-tumor chemotherapy (Shimoyama, 1991). 1% to 5% ofHTLV-I-infected subjects develop ATL after an incuba-tion time of 40 years or more (Tajima, 1990; Arisawaet al., 2000). Another small portion of HTLV-I carriersdevelop inflammatory disorders, such as HTLV-I-asso-ciated myelopathy/tropical spastic paraparesis (Gessainet al., 1985; Osame et al., 1986), but most other infectedindividuals are asymptomatic.

Immunological studies suggest that insufficient hostT-cell response against HTLV-I may be an immunolog-ical risk factor for ATL. This is based on the infrequencyof HTLV-I-specific cytotoxic T lymphocytes (CTL) in-duction in vitro from ATL patients (Kannagi et al., 1984;Jacobson et al., 1990; Kannagi et al., 1994; Arnulf et al.,2004), and the association of elevated proviral load withinsufficient T-cell immunity in animal model of oralHTLV-I infection (Hasegawa et al., 2003), which is themajor route of mother-to-child HTLV-I-transmission inhumans (Hino et al., 1985; Nakano et al., 1986). The roleof HTLV-I-specific T-cell immunity in anti-tumor sur-veillance was also supported by the fact that HTLV-ITax, a crucial viral protein for HTLV-I leukemogenesis(Yoshida, 2001), serves as an immunodominant targetantigen for HTLV-I-specific CTL (Jacobson et al., 1990;Kannagi et al., 1991; Parker et al., 1992; Elovaara et al.,1993). Our previous study indicated that HTLV-I Tax-specific CTL response was strongly activated in ATLpatients who obtained complete remission after hemato-poietic stem cell transplantation (HSCT) from human leu-kocyte antigen (HLA)-identical donors (Harashima et al.,2004). In a rat model, Tax-targeted vaccines can induceanti-tumor immunity and eradicate HTLV-I-infectedlymphomas in vivo (Ohashi et al., 2000; Hanabuchiet al., 2001). These findings suggest that HTLV-I-spe-cific T-cell immune responses may contribute to con-trolling HTLV-I-infected cell growth in vivo.

However, no extensive survey of HTLV-I-specific T-cell response amongHTLV-I carriers has been conducted

to identify a low-responder group including a high-riskgroup for ATL. One of the reasons for the poor status ofthe immunological survey among HTLV-I carriers is theunavailability of handy methods to measure HTLV-I-specific T-cell responses, applicable to any subject. SinceT-cells recognize various small peptide fragments pre-sented by the major histocompatibility complex (MHC),which are heterogeneous among individuals, the com-bination of the target epitopes recognized by T-cells dif-fers widely among individuals.

The most ideal antigen to detect HTLV-I-specific T-cell response is autologous HTLV-I-infected cells, whichpossess autologous MHC presenting suitable HTLV-I an-tigen fragments (Kannagi et al., 1983). However, estab-lishment of autologous HTLV-I-infected cell linesrequires a long-term, and sometimes unsuccessful, cul-turing process.

Oligopeptides, corresponding to T-cell epitopes, canbe a good antigen, because they can bind MHC to auto-logous antigen presenting cells. Several T-cell epitopeshave been found in HTLV-I proteins (Kannagi et al.,1992; Parker et al., 1992; Utz et al., 1992; Elovaara et al.,1993; Koenig et al., 1993). However, because of theheterogeneity of human populations, the limited numb-ers of epitopes found are not sufficient for a wide survey.In order to cover all the potential epitopes, synthesis ofenormous numbers of serial oligopeptides will be re-quired, which is expensive. So far, cocktails of a series ofoverlapping oligopeptides are the only available methodto detect HTLV-I-specific T-cell response in HLA-un-known patients (Goon et al., 2004).

MHC/peptide tetramer is another tool to quantifyantigen-specific CTL in peripheral blood. In fact, Tax-specific T-cell populations in the peripheral blood ofHTLV-I-infected individuals can be demonstrated by usingHLA-A0201/Tax 11–19, HLA-A2402/Tax 301–309,HLA-A1101/Tax 88–96, and HLA-A1101/Tax 272–280tetramers (Harashima et al., 2004, 2005). However,this is not suitable for a wide immunological surveyeither, because of the limited number of available tet-ramers, which is far below the genetic heterogeneity inhumans.

In this study, we attempted to establish a protein-based assay system for detecting HTLV-I-specific T-cellresponse, by using recombinant Tax proteins fused to

63K. Kurihara et al. / Journal of Immunological Methods 313 (2006) 61–73

glutathione-S-transferase (GST). We divided HTLV-ITax protein into three portions, expecting to avoid theinfluence of multiple biological activities of Tax, andto roughly estimate the location of T-cell epitopes ingenetically heterogeneous individuals. By using thissystem, we detected HTLV-I Tax-specific T-cell re-sponses both in rats and humans, and demonstrated thatthe GST-Tax portions predominantly inducing interferongamma (IFN-γ)-producing T-cell response were theregions where the major CTL epitopes were previouslyidentified to be located. This study provides a usefuldetection system for HTLV-I Tax-specific T-cell re-

Fig. 1. Construction and preparation of glutathione-S-transferase (GST)-Tax(open boxes) amplified using BamHI-linkered sense and EcoRI-linkered antisTax cDNA in the plasmid DNA (pβMT-2 Tax) used as a template. Each PCR2T vector for transforming DH5α competent cells. (B) GST-Tax fusion proteGlutathione Sepharose 4B-column followed by gel-filtration by using HPLCexclusion chromatography were collected. The indicated pattern is of GST-TaGST-Tax-B (middle, 40.9 kDa) and GST-Tax-C (right, 41.1 kDa) fusion proteanti-GST antibody were visualized by the enhanced chemiluminescence met

sponses applicable for general populations of HTLV-I-infected individuals.

2. Materials and methods

2.1. Construction of expression vectors for GST-Taxfusion protein

Partially overlappedDNA fragments Tax-A, Tax-B, andTax-C, corresponding to the N-terminal, central, and C-terminal regions of HTLV-I Tax, respectively, were ampli-fied from plasmid containing wild type Tax cDNA

fusion proteins. (A) Location of DNA fragments, Tax-A, -B, and -Cense primer sets (arrowhead). Gray box indicates full length of HTLV-Iproduct was subcloned into EcoRI/BamHI multicloning site of pGEX-ins obtained from bacterial inclusion bodies were affinity purified on a. The peak fractions (number twelve and thirteen) separated on a sizex-A monitored at 280 nm. (C) The size of GST-Tax-A (left, 41.0 kDa),ins were confirmed by Western blotting. The bands reacting with goathod.


(GenBank accession no. M10085) (pβMT-2 Tax) by usingBamHI-linkered sense and EcoRI-linkered antisense pri-mers (Fig. 1). The primer sets used were Bam-TS (5′-TTGGATCCATGGCCCACTTCCCAGGGTT-3′) andEco-TAR (5′-TTGAATTCCGTGCTGCCCAAGGGTGGGTT-3′) for Tax-A, Bam-TBS (5′-TTGGATCCTCCCCCTTCCGAAATGGATA-3′) and Eco-TBR (5′-TTGAATTCCGAACGGAAGGAGGCCGTTTT-3′) for Tax-B,and Bam-TCS (5′-TTGGATCCCCCGTCACGCTAACAGCCTG-3′) and Eco-TR (5′-TTGAATTCCTCAGACTTCTGTTTCTCGGA-3′) for Tax-C. pβMT-2 Tax andits control plasmid pHβAPr.1-neo vector were kindly pro-vided by Kayoko Matsumoto (Osaka Red Cross BloodCenter, Osaka, Japan) (Matsumoto et al., 1994). Proteinexpression was controlled under the human β-actin promoter in these vectors. Each PCR product wassubcloned into EcoRI andBamHI site of pGEX-2T (Amer-sham Biosciences), a cloning vector to express proteinfused to GST.

2.2. Purification of GST-Tax fusion proteins

Competent cells (DH5α; TOYOBO Co., Ltd., Osaka,Japan) transformed by heat shock method with pGEX-2Tvector containing each Tax fragment, were cultured in2xYT medium (3.2 g of Bactotryptone, 2.0 g of BactoYeast Extract, 1 g of NaCl in 200 mL) and 50 μg/mL ofAmpicillin. Protein synthesis was induced by addition of0.1 mM IPTG (Wako, Osaka, Japan) for 6 h at 37 °C. Thecells were subsequently centrifuged, and suspended in asonication buffer (50 mM Tris–HCl, 1 mM EDTA,50 mM NaCl and 1% Triton-X), supplemented with aprotease inhibitor mixture (PMSF, aprotinin, leupeptin,and pepstatin, Sigma-Aldrich Co. St. Louis, Mo.), fol-lowed by sonication (3 times 1 min). The cell extractswere centrifuged at 10,000 rpm for 10min at 4 °C, and thepellets containing inclusion bodies were solubilized byfreshly prepared denaturation buffer (8 M Urea, 50 mMTris–HCl, 1 mMEDTA, 50 mMNaCl and 10 mMDTT).After incubation at 4 °C for 20 min, the solubilizedfraction cleared by centrifugation was dialyzed graduallyfor refolding of proteins, and finally solubilized in TEbuffer (50 mM Tris–HCl, 1 mM EDTA, 50 mM NaCl)containing 1 mM of DTT. The solution was applied to aGlutathione Sepharose 4B-affinity column and the GST-fused Tax proteins bound to the column were eluted byglutathione buffer (50 mM Tris–HCl pH 8.0, 20 mMreduced glutathione, 200 mM NaCl), and further purifiedby size exclusion chromatography (TSKgel G2000SWXL,TOSOH, Tokyo, Japan) using HPLC (BioCAD SPRINTPerfusion chromatography systems, PerSeptive Biosys-tems, Massachusetts). The finally purified protein was

stored at −80 °C at 1 mg/mL concentration. The proteinconcentration was determined by a protein assay kit (Bio-Rad, Tokyo, Japan) using BSA as a standard. ControlGST protein was obtained from bacterial culturesupernatant as per the manufacturer's instructions, andpurified by Glutathione Sepharose 4B-affinity and sizeexclusion chromatography.

2.3. Western blotting

GST-Tax fusion proteins were separated by SDS-PAGE, and transferred to a PVDF membrane. Afterincubation with blocking buffer (2% BSA, 10 mM Tris–HCl, 100 mMNaCl and 0.1% Tween 20), the membranewas incubated with goat anti-GST antibody (AmershamBiosciences, Little Chalfont Buckinghamshire EnglandUK), and then with horseradish peroxidase-conjugatedanti-goat IgG antibody (Amersham). Finally, the bandsreacting with antibodies were visualized by the enhancedchemiluminescence method (Amersham), and recordedby a charge-coupled device camera (FUJIFILM, Tokyo,Japan).

2.4. Synthetic HTLV-I Tax peptides

A total of 35 serial synthetic peptides, Tax P1 throughTax P35, corresponding to amino acid sequence of theentire HTLV-I Tax protein were prepared as describedpreviously. Each oligopeptide consists of 15 amino acidsand overlaps 5 amino acids with the next oligopeptide(Table 1). These peptides were grouped into three pools,Tax P1–P12, Tax P13–P22, and Tax P23–P35, approx-imately corresponding to HTLV-I Tax-A, Tax-B and Tax-C domains, respectively. Stock solutions prepared at1.2 mg/mL for Tax P1–P12, 1 mg/mL for Tax P13–P22,and 1.3 mg/mL for Tax P23–P35 were stored at −20 °Cuntil use.

In some experiments, 13 newly synthesized peptides,Tax P11a through Tax P23a (Thermo Electron GmbH,Ulm, Germany) covering the entire Tax-B region, wereused. Each of these oligopeptides consists of 15 aminoacids and overlaps 10 amino acids with previouslysynthesized peptides, Tax P12 through Tax P24, res-pectively (Table 1). These 13 peptides were used eitherindividually or grouped into one pool, Tax P11a–P23a.The purity (N80%) and identity of these peptides wereassessed by HPLC and mass spectrometry.

2.5. Animals

Inbred female F344/N Jcl-rnu/+ rats (F344 n/+; 4-week-old) were purchased from Clea Japan, Inc. (Tokyo,

Table 1Tax regions covered by Tax oligopeptides and GST-Tax proteins prepared

Initial series of Tax oligopeptides Additional series of Tax oligopeptides GST-Tax proteins(amino acid position)

Peptide ID Amino acid position Peptide ID Amino acid position

Tax P1 1–15 377777777777777777775

Tax P2 11–25Tax P3 21–35Tax P4 31–45Tax P5 41–55Tax P6 51–65 GST-Tax-ATax P7 61–75 (1–127)Tax P8 71–85Tax P9 81–95Tax P10 91–105Tax P11 101–125Tax P12 111–135 Tax P11a 106–120 3

777777777777777777775

Tax P13 121–145 Tax P12a 116–130Tax P14 131–145 Tax P13a 126–140Tax P15 141–155 Tax P14a 136–150Tax P16 151–165 Tax P15a 146–160Tax P17 161–175 Tax P16a 156–170 GST-Tax-BTax P18 171–185 Tax P17a 166–180 (113–237)Tax P19 181–195 Tax P18a 176–190Tax P20 191–205 Tax P19a 186–200Tax P21 201–215 Tax P20a 196–210Tax P22 211–225 Tax P21a 206–220Tax P23 221–235 Tax P22a 216–230 3

777777777777777777775

Tax P24 231–245 Tax P23a 226–240Tax P25 241–255Tax P26 251–265Tax P27 261–275Tax P28 271–285Tax P29 281–295Tax P30 291–305 GST-Tax-CTax P31 301–315 (224–353)Tax P32 311–325Tax P33 321–335Tax P34 331–345Tax P35 341–353


Japan). Rats were treated at the experimental animalfacilities in Tokyo Medical and Dental University underthe experimental protocol approved by the Animal CareCommittee of Our University.

2.6. DNA vaccination into rats

Inoculations of DNA-coated gold particles intorats by using Helios Gene Gun System (BioRad)were performed as previously described (Ohashi etal., 2000). Briefly, rats were anesthetized withketamine, and their fur was completely removed byusing a commercial depilatory agent. Gold particlescoated by pβMT-2 Tax or pHβAPr.1-neo plasmidplaced in a cartridge were accelerated by pressurizedhelium gas for penetration through cell membranes

and multiple layers of cells in the epidermis. Theconcentrations of DNA and gold were 1 μg/shot.Immunization was performed thrice with a 1-weekinterval, and 10 shots were given per immunization.One-week after the final immunization, ratswere sacrificed, and spleen cells were cryopreserved in−80 °C until use.

2.7. Human peripheral blood mononuclear cells (PBMC)

An acute ATL patient (Pt. #37-13), who receivedHSCT (Harashima et al., 2004) and underwent a com-plete remission, donated the peripheral blood underinformed consent at 3 years and 9 month after HSCT.PBMC were isolated by using Ficoll-Paque (AmershamPharmacia, Uppsala, Sweden) density centrifugation,


and cryopreserved in liquid nitrogen until use. CD4+ orCD8+ cells-depleted human PBMC fraction was puri-fied by negative selection using MACS (Miltenyi Bio-tec, Bergisch Gladbach, Germany). Briefly, cells weresuspended in MACS buffer (PBS containing 0.5% FBSand 2 mM EDTA), and labeled with FITC-conjugatedmouse anti-CD4 or CD8 antibodies (IgG1; BD Phar-mingen Co., San Diego, Calif.) and subsequently reac-ted with goat anti-mouse IgG microbeads (MiltenyiBiotec). After extensive washing, cells were separatedon a column that was placed in the magnetic field of aMACS separator. Unbound cells to the columns wereeluted. Purity of eluted cell fractions was analyzed usinga flow cytometer.

Fig. 2. Tax-specific T-cell responses in rats against syngeneic cell lines. Insplenocytes from F344 n/+ rats immunized with plasmids containing Tax cDNin the presence of various syngeneic stimulator cells. Splenocytes (1×105/without (white bars) or with the same number of formalin-fixed syngeneic GIFN-γ production in culture supernatants was measured by ELISA on dayincubated, then [3H]-thymidine incorporated in the splenocytes was measuexpressed as the mean±standard deviation. Similar results were obtained in

2.8. Cell lines

FPM1, a HTLV-I-infected rat T-cell line previouslyestablished from an F344 n/+ rat (Ohashi et al., 1999) wasmaintained in RPMI 1640 medium (Sigma-Aldrich) con-taining 10% heat-inactivated fetal bovine serum (FBS)(Sigma-Aldrich), 100 U/mL of penicillin, 100 μg/mL ofstreptomycin, and 250 ng/mL of amphotericin B (10%FBS-RPMI) supplemented with 5.5×10−5 M of 2-mercaptoethanol. G14 (Ohashi et al., 2000) is an inter-leukin (IL)-2 dependent HTLV-I negative CD8+ T-cellline derived from a F344 n/+ rat. G14-Tax (Ohashi et al.,2000) is a stable transfectant of G14 with HTLV-I Tax-expressing plasmids. G14 and G14-Tax were maintained

terferon gamma (IFN-γ) production (A) and cell proliferation (B) ofA (pβMT-2 Tax) or its control plasmid (pHβAPr.1-neo) were measuredwell) from rats immunized with indicated plasmids were co-cultured14 (gray bars), G14-Tax (black bars) or FPM1 (striped bars) cells, and4 and 7. Similar sets of the splenocytes and the stimulator cells werered on day 4 as described in Materials and methods. The results areanother sets of independent experiments.


in the presence of 10 U/mL of recombinant human IL-2(Shionogi Pharmaceutical Co., Osaka, Japan).

2.9. Proliferation assay

Whole spleen cells (1–2×105/well) from DNA-vaccinated rats were co-cultured with 1%-formalin-fixed syngeneic cell lines (1×105/well) or GST-Tax fu-sion proteins in 96-well U-bottom plates in triplicate at37 °C for 4 to 6 days, then pulsed with 37 kBq/well of[3H]-thymidine for an additional 16 h.Cellswere harvestedwith a Micro 96 Harvester (Skatron, Lier, Norway), and[3H]-thymidine incorporated into cells was measured in amicro plate β-counter (Micro Beta Plus; Wallac, Turku,Finland).

Fig. 3. Detection of Tax-specific T-cell responses in rats by using GST-Tax fuwith plasmids containing Tax cDNA (pβMT-2 Tax) or its control plasmid (pHgray bars), a mixture of GST-Tax-A, -B, and -C (total of 3.75 μg; black bars), GGST-Tax-C (1.25 μg; grid bars) fusion proteins, and IFN-γ production (A) an2. The results are expressed as the mean±standard deviation. Similar results

2.10. Cytokine production assay

Human PBMC or rat splenocytes (1–2×105/well)were incubated with various antigens in a 96-well U-bottom plates for 4 days and IFN-γ produced in culturesupernatant was measured by Human or Rat IFN-γ en-zyme linked immunosorbent assay (ELISA) kit (BIO-SOURCE, Inc., Calif.). Human PBMC were culturedin 10% FBS-RPMI, whereas rat splenocytes were cul-tured in 10% FBS-RPMI containing 2-mercaptoethanol.Half of the medium was renewed at harvest and thecytokine was also measured on day 7 and 10. Absor-bances were measured at 450 nm using a microplatereader (Bio Rad) and analyzed with Microplate ManagerIII software.

sion proteins. Splenocytes (2×105/well) from F344 n/+ rats immunizedβAPr.1-neo) were cultured without (white bars) or with GST (2.5 μg;ST-Tax-A (1.25 μg; striped bars), GST-Tax-B (1.25 μg; dotted bars) or

d cell proliferation (B) were measured as described in the legend of Fig.were obtained in another sets of independent experiments.


3. Results

3.1. Preparation of GST-Tax fusion proteins

Since HTLV-I Tax serves as a dominant CTL targetantigen in both humans and rats, we preparedrecombinant GST-HTLV-I Tax fusion proteins asantigens to detect HTLV-I-specific T-cell responses inHTLV-I-infected individuals. In order to reduce multiplebiological effects of Tax itself on responder cells, andalso to aid identifying reactive Tax regions, we divided

Fig. 4. Different patterns of T-cell responses detected by GST-Tax and pep(2×105/well) from pβMT-2 vaccinated rats were co-cultured with various confor 4 days, and IFN-γ concentration in the culture supernatants was measuredeach GST Tax protein and Tax peptide. Final concentrations of proteins (A) wwell for a mixture of GST-Tax-A, -B, and -C that were calculated to contain a seach GST-Tax-A, -B, or -C protein. Final concentrations of peptides (B) wereP13–P22, and 1.625, 6.5, and 26 μg/well for Tax P23–P35. Tax P1–P12, Tregions of GST-Tax-A, -B, and -C proteins, respectively. The insert in B (squpeptide mixture Tax P11a–P23a, covering the entire Tax-B region as describe±standard deviation. Similar results were obtained in another sets of indepe

the wild type Tax into three portions, N-terminal (Tax-A), central (Tax-B) and C-terminal (Tax-C). Tax-Aoverlaps with Tax-B for forty-five nucleotides, whileTax-B and -C overlap for forty-two nucleotides (Fig.1A). GST-Tax fusion proteins extracted from bacterialinclusion bodies were affinity-purified by a GlutathioneSepharose 4B-column, and the peak fractions (numbertwelve and thirteen), separated on a size exclusionchromatography, were collected (Fig. 1B). After finalpurification, the size and expression of GST-Tax fusionproteins were confirmed by Western blotting by using

tide-based immunoassay. IFN-γ-producing responses of splenocytescentrations of GST-Tax fusion proteins (A) or Tax peptide mixtures (B)by ELISA. The triangle markers indicate the concentration gradient ofere 0.156, 0.625, and 2.5 μg/well for GST, 0.234, 0.938, and 3.75 μg/imilar range of molarities of GST, or 0.078, 0.312, and 1.25 μg/well for1.5, 6, and 24 μg/well for Tax P1–P12, 1.25, 5, and 20 μg/well for Taxax P13–P22, and Tax P23–P35 correspond approximately to the Taxare) represents the result of IFN-γ production against newly preparedd in the Materials and methods. The results are expressed as the meanndent experiments.


anti-GST monoclonal antibody (Fig. 1C). These puri-fied GST-Tax fusion proteins were used in followingexperiments.

3.2. Detection of Tax-specific T-cell responses in rats byusing GST-Tax fusion proteins

To investigate T-cell responses against GST-Tax fusionproteins, we used rats vaccinated with pβMT-2 Tax con-taining Tax cDNA by gene gun as donors for immunecells. Tax-specific T-cell response in these rats was firstconfirmed by IFN-γ production and [3H]-thymidine in-corporation in response to formalin-treated cells of syn-geneic rat cell lines (Fig. 2). Rats inoculated with emptyvector (pβAPr.1-neo) served as controls, splenocytesfrom rats inoculated with Tax cDNA-containing vectorbut not from control rats produced significant levels ofIFN-γ against Tax-expressing G14-Tax and FPM1 cells,but not to control G14 cells (Fig. 2A). Similarly, T-cellproliferative responses were observed in the splenocytesfrom immunized rats against Tax-expressing stimulatorcells (Fig. 2B). Subsequently, we examined T-cellresponses of these rats against GST-Tax fusion proteins.As shown in Fig. 3A, IFN-γ production against mixture ofGST-Tax-A, -B, and -C proteins was clearly detectable atday 4, and markedly increased at day 7 in rats inoculatedwith pβMT-2Tax, but not in control rats. Among the threeportions of Tax, the central domain (Tax-B) elicited thestrongest response. The IFN-γ production against controlGST protein was very low. Significant levels of pro-liferative T-cell response occurred in splenocyte culturesfrom immunized rats against the mixture of GST-Tax-A,-B, and -C proteins and GST-Tax-B, but to a lower extent

Fig. 5. Epitope mapping of primary splenocytes from immunized rats. Freshwere incubated with each Tax peptide indicated (1 μg/well), and IFN-γ producThe results are expressed as the mean±standard deviation.

against GST-Tax-A and -C. Low levels of backgroundproliferation by GST proteins were observed in both con-trol and immunized rat splenocytes (Fig. 3B). PredominantT-cell responses against Tax-B region was similarlyobserved in all three Tax-DNA-inoculated rats tested(data not shown). These results indicated that GST-Taxfusion proteinswere processed and presented by rat splenicantigen-presenting cells to be recognized by Tax-specificT-cells.

3.3. Comparison of T-cell responses detected by GST-Tax and peptide-based immunoassays

It has been shown that cocktails of Tax and Envoligopeptides can be used to detect HTLV-I-specific res-ponses (Goon et al., 2004). Therefore, we compared GST-Tax protein-based assays with peptide-based assays todetect T-cell responses. We prepared a total of 35 syntheticTax peptides, each of which consisted of 15 amino acidswith 5 amino acids overlapping with neighboring peptides,covering the entire sequence of HTLV-I Tax. Thesepeptides were grouped into the three pools, Tax P1–P12,Tax P13–P22 and Tax P23–P35, approximately corre-sponding to theHTLV-I Tax-A, Tax-B and Tax-C domains,respectively.

As shown in Fig. 4A, when splenocytes from Tax-DNA vaccinated rats were cultured with various con-centrations of individual GST-Tax (0.078–1.25 μg), or amixture of GST-Tax-A, -B, and -C (0.234–3.75 μg)fusion proteins, dose-dependent IFN-γ production wasobserved. Maximal IFN-γ production was observedwhen the mixture of GST-Tax-A, -B, and -C was added.Among each GST-Tax fusion protein, the highest IFN-γ

ly prepared splenocytes (2×105/well) from Tax-DNA-immunized ratstion in the supernatants after 4 days of culture was measured by ELISA.


production was observed against Tax-B region, consis-tent with the results in Fig. 3. However, when similarconcentrations (0.0625–1.3 μg) of peptide mixtures ofTax P1–P12, Tax P13–P22 or Tax P23–P35 werecultured, IFN-γ production was hardly detectable (datanot shown). We subsequently applied larger amounts ofpeptide mixtures (1.5–26 μg) into the assay (Fig. 4B).Under these conditions, dose-dependent IFN-γ produc-tion was detected in the presence of the peptide mixtureTax P1–P12, indicating that higher amounts of peptideswere required to activate T-cell responses than GST-Taxfusion proteins. However, the response against TaxP13–P22 corresponding to Tax-B region was still hardlydetectable.

Fig. 6. Detection of Tax-specific T-cell responses in human. (A) Cryopreseremission after HSCTwere cultured with medium (white bars), GST (2.5 μg)or 1.25 μg of each GST-Tax-A (striped bars), -B (dotted bars), or -C (grid bars)ELISA after 4, 7, and 10 days of culture. (B) Whole PBMC, or CD4+ cellsprepared by negative selection on a magnetic column from the same patient (2(G, gray bars), or a mixture of GST-Tax-A, -B, and -C (3.75 μg) (T, black bars)ELISA after 4 days of culture. The results are expressed as the mean±stand

3.4. Coincidence of the location of major T-cell epitopeand GST-Tax portion inducing predominant T-cellresponse in rats

Since predominant T-cell response was observedagainst GST-Tax-B but not against peptide mixture TaxP13–P22 corresponding to Tax-B region, we addition-ally synthesized a new series (Tax P11a through P23a) of15-mer oligopeptides, each of which had 10 overlappingamino acids with previously synthesized peptides cor-responding to the entire Tax-B region (Table 1). Whenincubated with the mixture of new peptides (Tax P11a–P23a), splenocytes from Tax-DNA vaccinated rats pro-duced detectable levels of IFN-γ (Fig. 4B, insert). We

rved PBMC (2×105/well) from an ATL patient (#37-13) in complete(gray bars), a mixture of GST-Tax-A, -B, and -C (3.75 μg) (black bars),fusion proteins. IFN-γ production in the supernatants was measured by-depleted (CD4[−]) or CD8+ cells-depleted (CD8[−]) PBMC fractions×105/well) were cultured with medium (M, white bars), GST (2.5 μg)fusion proteins. IFN-γ production in the supernatants was measured byard deviation.


then used these new peptides individually to assess theepitope region, and found that the rat splenocytes reactedmost strongly with Tax P18a (176–190) (Fig. 5) Thisregion coincidentally includes previously identifiedmajor CTL epitope Tax 180–188 in this rat strain(Hanabuchi et al., 2001). These observation indicatedthat the poor response to the Tax P13–P22 mixture wasdue to an incomplete coverage of the T-cell epitope bythe neighboring peptides Tax P18 (171–185) and P19(181–195) in the mixture. Thus, the locations of T-cellepitopes estimated by the GST-Tax protein-based assaythoroughly matched those ascertained by the peptidemixture-based assay.

3.5. T-cell responses of HTLV-I infected human againstGST-Tax fusion proteins

Finally, we investigated human HTLV-I-specific T-cellresponses against GST-Tax fusion proteins. We usedcryopreserved PBMC from an ATL patient (Pt. #37-13),who underwent complete remission after allogeneicHSCT from an HLA-identical donor. We previously de-monstrated that HLA-A2-restricted CD8+ CTL responseagainst Tax 11–19 epitope was strongly activated in thispatient (Harashima et al., 2004). When the PBMC werecultured with or without GST-Tax-A, -B, or -C proteinsfor 4 days, these PBMCproducedmarked levels of IFN-γresponding to GST-Tax-A, and also to GST-Tax-C to alesser degree, but not to control GST protein (Fig. 6A).This is consistent with the fact that the major CTL epitopeTax 11–19 of this patient is located in the Tax-A region.To confirm that GST-Tax proteins were cross-presentedon MHC-I, purified CD4+ and CD8+ PBMC fractionsfrom the same patient were cultured with GST-Tax pro-teins. As shown in Fig. 6B, both CD4+ and CD8+ frac-tions showed IFN-γ production against GST-Tax proteinsas well as whole PBMC. IFN-γ production by CD8+

fractions was not because of the contamination of CD4+

cells, as 99.7% cells of this fraction was negative for CD4as confirmed by flow cytometry. Thus, the three-dividedTax protein-based assay established in the present studysuccessfully detected human T-cell response, and thepattern of responses against each proteinmatchedwith thelocation of the major CTL epitope.

4. Discussion

In the present study, we established an assay systemusing three-divided GST-Tax fusion proteins anddemonstrated HTLV-I-specific T-cell response in bothrats vaccinated with Tax-DNA and an HTLV-I-infectedhuman patient. This assay can be used regardless of in-

dividual heterogeneity in MHC, and is useful to estimatethe location of major T-cell epitopes.

Comparison of the T-cell response against GST-Taxproteins andmixed peptides showed some similarities anddifferences. In both assays, Tax-specific T-cell responsewas detected, although lower amounts of antigens wererequired for GST-Tax proteins than the peptide mix. Ratsplenocytes, which produced IFN-γ against GST-Tax-Bprotein, poorly reacted with the initially synthesized pep-tide mixture Tax P13–P22 corresponding to Tax-B(Fig. 4B), likely because themajor T-cell epitope naturallypresented by antigen-presenting cells was fragmented inthe initially synthesized peptide series. To avoid such falseignorance in a peptide-based assay, a large number ofoligopeptides should be prepared.

The kinetics of T-cell responses against GST-Taxproteins in vitro also differed from those against thepeptide mixture. IFN-γ production against GST-Tax pro-teins lasted for about 7–10 days, whereas that against thepeptide mixture peaked at day 4, and diminished there-after (data not shown). Thus, GST-Tax protein-basedassay can detect T-cell responses with smaller amounts ofantigen and keep the response for a longer period thanpeptide-based assay, although the peptides were indis-pensable to identify fine T-cell epitopes.

It is known that exogenous antigens are preferentiallypresented on MHC-II molecules in antigen-presentingcells and also cross-presented onMHC-I in a lesser degree(Watts and Amigorena, 2001). The splenocytes ofimmunized rats predominantly reacted with GST-Tax-Bprotein containing the region of Tax 180–188, which isthe major epitope recognized by Tax-specific CD8+ CTLof this rat strain. In the ATL patient, the PBMC pre-dominantly reacted with GST-Tax A that is containingTax 11–19, the major CTL epitope of this patient. Thesecoincidences suggested that exogenously added GST-Taxproteins were presented on not only MHC-II but alsoMHC-I via cross-presentation pathway (Guermonprezet al., 2003) (Houde et al., 2003). As a fact, purified CD8+

cells from the patient could react with exogenously addedGST-Tax proteins (Fig. 6B). The preferential IFN-γproduction to GST-Tax-A protein by whole PBMC thatcontains both CD4+ and CD8+ cells suggests that themajor CTL epitopes and helper T-cell epitopes might co-localize within Tax-A region. The precise mechanisms ofGST-Tax antigen-presentation remain to be clarified.

We previously reported that the PBMC from the ATLpatient tested in the present study at 6 months after butnot before HSCT exhibited a strong Tax-specific CD8+

CTL response against autologous HTLV-I-infected cells.This patient has been in complete remission over 3 yearsafter HSCT, and a similar CTL response against


autologous HTLV-I-infected cells was still observed at3.5 years after HSCT, just before the present study wasconducted. Here, we successfully detected Tax-specificT-cell response in this patient by using GST-Tax protein-based assay. Tax-specific T-cell response is a potentialindicator for the immunological status in ATL patientsafter anti-tumor therapy. GST-Tax-based assay can beapplicable to other ATL patients without time-consum-ing procedures to establish autologous cell lines.

In conclusion, we established a protein-based assaysystem that provides a useful tool tomonitor the HTLV-I-specific T-cell response. Previous findings indicated thatinsufficient HTLV-I-specific T-cell response is one of therisk factors of ATL, and that Tax-targeted vaccines mightpotentially induce prophylactic effects against ATLdevelopment. GST-Tax protein-based detection systemof T-cell response established in the present study isuseful for a wide survey to identify the immunologicalhigh-risk group of ATL in HTLV-I-infected individuals,and for roughly predicting the location of T-cell epitopeswhich may be applicable for vaccine strategies.

Acknowledgments

This study was supported by the grants from the Min-istry of Education, Science, Culture and Sports of Japan,and the Ministry of Health, Welfare, and Labour of Japan.

References

Arisawa,K., Soda,M., Endo, S.,Kurokawa,K.,Katamine, S., Shimokawa,I., Koba, T., Takahashi, T., Saito, H., Doi, H., Shirahama, S., 2000.Evaluation of adult T-cell leukemia/lymphoma incidence and itsimpact on non-Hodgkin lymphoma incidence in southwestern Japan.Int. J. Cancer 85, 319.

Arnulf, B., Thorel, M., Poirot, Y., Tamouza, R., Boulanger, E., Jaccard,A., Oksenhendler, E., Hermine, O., Pique, C., 2004. Loss of the exvivo but not the reinducible CD8+ T-cell response to Tax in humanT-cell leukemia virus type 1-infected patients with adult T-cellleukemia/lymphoma. Leukemia 18, 126.

Elovaara, I., Koenig, S., Brewah, A.Y., Woods, R.M., Lehky, T.,Jacobson, S., 1993. High human T cell lymphotropic virus type 1(HTLV-1)-specific precursor cytotoxic T lymphocyte frequenciesin patients with HTLV-1-associated neurological disease. J. Exp.Med. 177, 1567.

Gessain, A., Barin, F., Vernant, J.C., Gout, O., Maurs, L., Calender, A.,de The, G., 1985. Antibodies to human T-lymphotropic virus type-I in patients with tropical spastic paraparesis. Lancet 2, 407.

Goon, P.K., Igakura, T., Hanon, E., Mosley, A.J., Barfield, A., Barnard,A.L., Kaftantzi, L., Tanaka, Y., Taylor, G.P.,Weber, J.N., Bangham,C.R., 2004. Human T cell lymphotropic virus type I (HTLV-I)-specific CD4+ T cells: immunodominance hierarchy and preferen-tial infection with HTLV-I. J. Immunol. 172, 1735.

Guermonprez, P., Saveanu, L., Kleijmeer,M., Davoust, J., Van Endert, P.,Amigorena, S., 2003. ER-phagosome fusion defines anMHC class Icross-presentation compartment in dendritic cells. Nature 425, 397.

Hanabuchi, S., Ohashi, T., Koya, Y., Kato, H., Hasegawa, A., Takemura,F., Masuda, T., Kannagi, M., 2001. Regression of human T-cellleukemia virus type I (HTLV-I)-associated lymphomas in a ratmodel:peptide-induced T-cell immunity. J. Natl. Cancer Inst. 93, 1775.

Harashima, N., Kurihara, K., Utsunomiya, A., Tanosaki, R., Hanabuchi,S., Masuda, M., Ohashi, T., Fukui, F., Hasegawa, A., Masuda, T.,Takaue, Y., Okamura, J., Kannagi, M., 2004. Graft-versus-tax re-sponse in adult T-cell leukemia patients after hematopoietic stem celltransplantation. Cancer Res. 64, 391.

Harashima, N., Tanosaki, R., Shimizu, Y., Kurihara, K., Masuda, T.,Okamura, J., Kannagi, M., 2005. Identification of two new HLA-A*1101-restricted tax epitopes recognized by cytotoxic Tlymphocytes in an adult T-cell leukemia patient after hematopoi-etic stem cell transplantation. J. Virol. 79, 10088.

Hasegawa, A., Ohashi, T., Hanabuchi, S., Kato, H., Takemura, F.,Masuda, T., Kannagi, M., 2003. Expansion of human T-cell leu-kemia virus type 1 (HTLV-1) reservoir in orally infected rats:inverse correlation with HTLV-1-specific cellular immune re-sponse. J. Virol. 77, 2956.

Hino, S., Yamaguchi, K., Katamine, S., Sugiyama, H., Amagasaki, T.,Kinoshita, K., Yoshida, Y., Doi, H., Tsuji, Y., Miyamoto, T., 1985.Mother-to-child transmission of human T-cell leukemia virus type-I. Jpn. J. Cancer Res. 76, 474.

Hinuma, Y., Nagata, K., Hanaoka, M., Nakai, M., Matsumoto, T.,Kinoshita, K.I., Shirakawa, S., Miyoshi, I., 1981. Adult T-cellleukemia: antigen in an ATL cell line and detection of antibodiesto the antigen in human sera. Proc. Natl. Acad. Sci. U. S. A. 78,6476.

Houde,M., Bertholet, S., Gagnon, E., Brunet, S., Goyette, G., Laplante,A., Princiotta, M.F., Thibault, P., Sacks, D., Desjardins, M., 2003.Phagosomes are competent organelles for antigen cross-presenta-tion. Nature 425, 402.

Jacobson, S., Shida, H., McFarlin, D.E., Fauci, A.S., Koenig, S., 1990.Circulating CD8+ cytotoxic T lymphocytes specific for HTLV-IpX in patients with HTLV-I associated neurological disease.Nature 348, 245.

Kannagi,M., Sugamura, K., Sato, H., Okochi, K., Uchino, H., Hinuma, Y.,1983. Establishment of human cytotoxic Tcell lines specific for humanadult T cell leukemia virus-bearing cells. J. Immunol. 130, 2942.

Kannagi, M., Sugamura, K., Kinoshita, K., Uchino, H., Hinuma, Y.,1984. Specific cytolysis of fresh tumor cells by an autologouskiller T cell line derived from an adult T cell leukemia/lymphomapatient. J. Immunol. 133, 1037.

Kannagi, M., Harada, S., Maruyama, I., Inoko, H., Igarashi, H.,Kuwashima, G., Sato, S., Morita, M., Kidokoro, M., Sugimoto, M.,Funahashi, S.-I., Osame, M., Shida, H., 1991. Predominantrecognition of human T cell leukemia virus type I (HTLV-I) pXgene products by human CD8+ cytotoxic T cells directed againstHTLV-I-infected cells. Int. Immunol. 3, 761.

Kannagi, M., Shida, H., Igarashi, H., Kuruma, K., Murai, H., Aono, Y.,Maruyama, I., Osame, M., Hattori, T., Inoko, H., et al., 1992.Target epitope in the Tax protein of human T-cell leukemia virustype I recognized by class I major histocompatibility complex-restricted cytotoxic T cells. J. Virol. 66, 2928.

Kannagi, M., Matsushita, S., Shida, H., Harada, S., 1994. Cytotoxic Tcell response and expression of the target antigen in HTLV-Iinfection. Leukemia 8 (Suppl 1), S54.

Koenig, S., Woods, R.M., Brewah, Y.A., Newell, A.J., Jones, G.M.,Boone, E., Adelsberger, J.W., Baseler, M.W., Robinson, S.M.,Jacobson, S., 1993. Characterization of MHC class I restrictedcytotoxic T cell responses to tax in HTLV-1 infected patients withneurologic disease. J. Immunol. 151, 3874.


Matsumoto, K., Akashi, K., Shibata, H., Yutsudo, M., Hakura, A.,1994. Single amino acid substitution (58Pro–NSer) in HTLV-I taxresults in loss of ras cooperative focus formation in rat embryofibroblasts. Virology 200, 813.

Nakano, S., Ando, Y., Saito, K., Moriyama, I., Ichijo, M., Toyama, T.,Sugamura, K., Imai, J., Hinuma, Y., 1986. Primary infection ofJapanese infants with adult T-cell leukaemia-associated retrovirus(ATLV): evidence for viral transmission from mothers to children. J.Infect. 12, 205.

Ohashi, T., Hanabuchi, S., Kato, H., Koya, Y., Takemura, F.,Hirokawa, K., Yoshiki, T., Tanaka, Y., Fujii, M., Kannagi, M.,1999. Induction of adult T-cell leukemia-like lymphoproliferativedisease and its inhibition by adoptive immunotherapy in T-cell-deficient nude rats inoculated with syngeneic human T-cellleukemia virus type 1-immortalized cells. J. Virol. 73, 6031.

Ohashi, T., Hanabuchi, S., Kato, H., Tateno, H., Takemura, F., Tsukahara,T., Koya, Y., Hasegawa, A., Masuda, T., Kannagi, M., 2000.Prevention of adult T-cell leukemia-like lymphoproliferative diseasein rats by adoptively transferred Tcells from a donor immunizedwithhuman T-cell leukemia virus type 1 Tax-coding DNA vaccine.J. Virol. 74, 9610.

Osame, M., Usuku, K., Izumo, S., Ijichi, N., Amitani, H., Igata, A.,Matsumoto, M., Tara, M., 1986. HTLV-I associated myelopathy, anew clinical entity. Lancet 1, 1031.

Parker, C.E., Daenke, S., Nightingale, S., Bangham, C.R., 1992. Activated,HTLV-1-specific cytotoxic T-lymphocytes are found in healthyseropositives as well as in patients with tropical spastic paraparesis.Virology 188, 628.

Poiesz, B.J., Ruscetti, F.W., Gazdar, A.F., Bunn, P.A., Minna, J.D.,Gallo, R.C., 1980. Detection and isolation of type C retrovirusparticles from fresh and cultured lymphocytes of a patient withcutaneous T-cell lymphoma. Proc. Natl. Acad. Sci. U. S. A. 77,7415.

Shimoyama, M., 1991. Diagnostic criteria and classification ofclinical subtypes of adult T-cell leukaemia-lymphoma. A reportfrom the Lymphoma Study Group (1984–87). Br. J. Haematol.79, 428.

Tajima, K., 1990. The 4th nation-wide study of adult T-cell leukemia/lymphoma (ATL) in Japan: estimates of risk of ATL and itsgeographical and clinical features. The T- and B-cell MalignancyStudy Group. Int. J. Cancer 45, 237.

Uchiyama, T., Yodoi, J., Sagawa, K., Takatsuki, K., Uchino, H., 1977.Adult T-cell leukemia: clinical and hematologic features of 16cases. Blood 50, 481.

Utz, U., Koenig, S., Coligan, J.E., Biddison, W.E., 1992. Presentationof three different viral peptides, HTLV-1 Tax, HCMV gB, andinfluenza virus M1, is determined by common structural featuresof the HLA-A2.1 molecule. J. Immunol. 149, 214.

Watts, C., Amigorena, S., 2001. Phagocytosis and antigen presenta-tion. Semin. Immunol. 13, 373.

Yoshida, M., 2001. Multiple viral strategies of HTLV-1 for dysregula-tion of cell growth control. Annu. Rev. Immunol. 19, 475–496.

Yoshida, M., Miyoshi, I., Hinuma, Y., 1982. Isolation and character-ization of retrovirus from cell lines of human adult T-cell leukemiaand its implication in the disease. Proc. Natl. Acad. Sci. U. S. A.79, 2031.

Documents

Human T-cell leukemia virus type-I (HTLV-I)-specific T-cell responses detected using three-divided glutathione-S-transferase (GST)-Tax fusion proteins