Clonal Expansion Within CD4⁺ and CD8⁺ T Cell Subsets in Human T Lymphotropic Virus Type I-Infected Individuals¹

Human T cell lymphotropic virus type I (HTLV-I)³ is endemic in a number of established geographic areas. While the vast majority of infected individuals in these areas are asymptomatic carriers (ACs), infection is associated with adult T cell leukemia (ATL), a fatal CD4⁺ T-lymphoproliferative disorder (1), and a nonfatal neurological disorder, HTLV-I-associated myelopathy/tropical spastic paraparesis (HAM/TSP) (2). In addition, a number of reports have described evidence of inflammatory involvement outside of the central nervous system in patients both with and without HAM/TSP. The associated clinical conditions include lymphocytic alveolitis, uveitis, Sjögren’s syndrome, arthropathy, and polymyositis (2, 3), and the interactions between infected cells and responding T cells in the various affected compartment(s) are considered to be key factors in disease development (4, 5). In all HTLV-I-infected individuals a relative and preferential tropism of HTLV-I for CD4⁺ T cells has been consistently demonstrated (6, 7), and it has been proposed that infection may result in a benign clonal expansion of HTLV-I-infected CD4⁺ T cells (8). This view has been supported by studies of Wattel and colleagues (9, 10, 11) who, using ligation-mediated inverse PCR amplification of the provirus with cellular flanking sequences, concluded that clonal expansion of the infected cells is common in all HTLV-I-infected individuals. Studies to evaluate clonal expansion of responding T cells in infected individuals have been somewhat limited. Cloning studies (12) and more recently an investigation using a PCR-based single-strand conformation polymorphism (SSCP) assay (13) have suggested the presence of oligoclonal accumulation of responding T cells in peripheral blood, with the latter study also showing human leukocyte Ag (HLA) class I-restricted CTL clones in patients with HAM/TSP. However, these studies were somewhat limited in that they were conducted on unfractionated peripheral blood lymphocytes, and no attempts were made to demonstrate oligoclonality within isolated T cell populations. In the present study we have employed a recently established multiplex PCR assay to analyze the complementarity-determining region 3 (CDR3) length variation in TCR Vβ transcripts to evaluate the clonal expansion within fractionated T cell subsets from HTLV-I-infected individuals with and without disease. In addition, we have determined the predicted amino acid sequences of the CDR3 regions by direct nucleotide sequencing methods to determine whether there is a bias in the motifs present in the expanded populations in different clinical conditions.

Four HTLV-I ACs, five patients with HAM, and one patient with ATL were examined. Uninfected controls were 60 normal individuals who were analyzed in a previous study using the identical multiplex PCR assay under identical conditions (17). Three clinically asymptomatic individuals with HTLV-II served as additional control subjects. These were two males, age 46 and 58, and one female, age 53.

HLA Ags were determined serologically by the standard National Institutes of Health microcytotoxicity test using sera that were standardized to the criteria and nomenclature of the 11th International Histocompatibility Workshop (14). Further HLA DNA typing of HLA class II DRB1* and DQB1* alleles were conducted by the PCR/RFLP method (15, 16). The results are shown in Table I.

PBMCs were obtained by centrifugation over Ficoll-Hypaque. Cells were washed twice in RPMI 1640 with 1% FCS and resuspended in PBS with 1% FCS at a concentration of 10 × 10⁶ cells/ml. Positive selection for CD8⁺ T cells was conducted by incubating the PBMC with anti-CD8 immunomagnetic beads (Dynal, Great Neck, NY) for 30 min at 4°C on a rotating shaker, as recommended by the manufacturer. The unbound cells were then incubated with anti-CD4 immunomagnetic beads for selection of CD4⁺ T cells. The cells bound to the beads were placed directly into RNAzolB (Biotex, Houston, TX) for isolation of total RNA.

Total RNA (2 μg) from the selected T cell populations was used for the first-strand cDNA synthesis using a TCR B-chain C region primer, 3CB3 (see Table II). The cDNA synthesis was performed with Moloney murine leukemia virus reverse transcriptase in buffer supplied by manufacturer (Life Technologies, Gaithersburg, MD) at 42°C for 1 h in a total volume of 120 μl. Specific primers used in RT-PCR for CDR3 of TCR BV-chains are shown in Table II. Specific combinations of two or three Vβ (BV)-specific forward primers (20 pmol each) were used for each multiplex PCR. Each reaction also contained 20 pmol of a reverse primer specific for the C region (3CB1), out of which 3 pmol were end labeled with ³²P using T4 kinase (Life Technologies). The BV primer combinations were selected on the basis of the location of each BV specific primer with respect to the end of the BV region. The 12 different reaction sets were as follows; A, TCRBV1, 18, 23; B, BV2 4, 8; C, BV3, 13S1; D, BV5S2, 5S1; E, BV6, 20; F, BV7, 22; G, BV9, 16; H, BV11, 12; I, BV15, 13S2; J, BV14, 17; K, BV19, 24; and L, BV10, 21. A representative multiplex PCR assay is shown in Fig. 1. A total of 10 μl of the cDNA was used for each PCR. Conditions for the PCR were as follow: After initial denaturation at 94°C for 5 min, 35 cycles were conducted in a DNA thermocycler (model 9600; Perkin-Elmer, Norwalk, CT) Each cycle consisted of denaturation at 94°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 1 min. PCR buffer conditions were 10 mM Tris-HCl (pH 8.3), 2 mM MgCl₂, and 50 mM KCL, with 20 pmol of each primer in a 100-μl reaction volume. After 35 cycles, an additional extension at 72°C for 10 min was conducted. Five microliters of the amplified products were loaded on a standard 6% polyacrylamide sequencing gel. Bands (spaced 3 bp apart) were visualized after overnight exposure to Kodak AR film (Rochester, NY). The radioactive bands were also analyzed and quantitated using a PhosphorImager (PhosphorImager SF; Molecular Dynamics, Sunnyvale, CA).

cDNAs were reamplified using a single TCR BV segment primer and an unlabeled 3CB1 primer. The amplified product was used as a substrate for DNA sequencing using the 3CB1 or BJ-R (reverse primer specific for the BJ region) primer, following purification over PCR Prep DNA purification columns (Promega, Madison, WI). Direct sequencing of the PCR products was conducted with fluorescent dideoxy terminators and analyzed on a model 373A Automated Sequenator (Applied Biosystems, Foster City, CA). Nucleotide sequences were further confirmed using specific BJ reverse primers (Table II).

Restriction in CDR3 length within a particular BV segment is visualized as a “dominant band” and based on its intensity with respect to the rest of the bands within the TCR BV-specific PCR product. The radioactive gels were scanned using a PhosphorImager, and specific criteria were used to define a band as “dominant” were that >50% of the counts within a V segment/family were contained this band. The clonality of the band was established by direct sequencing of the PCR products with the 3CB1 primer. Pilot studies demonstrated that when the counts within the dominant band were 50% or greater of the entire BV family, a single readable sequence was obtained using the 3CB1 primer. Sequencing with a specific BJ-R primer (selected on the basis of the first 3CB1 primed sequence) yielded a clearly readable sequence in an additional 35% of cases. Thus, in all, 82% of the cases in which the 50% criteria for dominance were met, direct sequencing of the PCR product yielded a single dominant sequence (17).

The multiplex PCR assay was used to assess the clonality of 26 TCR-BV families in the separated CD4⁺ and CD8⁺ T lymphocytes of five patients with HAM/TSP, four HTLV-I ACs, and one patient with ATL. Fig. 2 shows a typical autoradiogram of the PCR products analyzed by electrophoresis on a 6% polyacrylamide gel from a patient with HAM/TSP. In this individual, dominant bands of a single CDR3 length were observed in both the CD4⁺ and CD8⁺ T lymphocyte populations in a large number of different BV families. Table III summarizes the TCR BV families shown to be clonally expanded in both the CD4⁺ and CD8⁺ T lymphocyte populations in all 10 individuals studied. The results demonstrated that clonal expansion of both CD4⁺ and CD8⁺ T lymphocytes occurs in both asymptomatic carriers and patients with HAM/TSP. Moreover, while the total number of expanded clones in the CD8⁺ T lymphocyte population was much greater (3–10 clones) than that of the CD4⁺ T lymphocytes (1–3 clones), no appreciable differences were observed between asymptomatic carriers and patients with HAM/TSP. Similarly we were unable to observe differences in specific V-β expansion in patients with HAM/TSP compared with asymptomatic carriers. Even after stratification of HLA phenotype there was no evidence of expansion involving a unique T cell repertoire. Within the CD4 T cell populations, we observed clonal expansion in four of five patients with HAM/TSP involving a total of 18 different clones. In the asymptomatic carriers, clonal expansion was observed in two of four individuals and this involved 7 different clones (Table III). In one patient with ATL, we observed one specific band in the CD4 population, Vβ 7. In contrast, clonal expansion was not observed in any of the CD4⁺ T cell populations in three control HTLV-II-infected individuals. In addition we have shown in previous studies using the same PCR assay that clonal expansion within CD4⁺ T cell population of uninfected individuals occurs rarely, if at all (17). In the CD8 population, the T cell populations which did not exhibit clonal expansion were seven Vβs (3, 7, 10, 15, 17, 22, 23) in HAM/TSP and three Vβs (6, 7, 18) in HTLV-I carriers. Table III shows that overall Vβ utilization was extremely diverse, and no restricted or consistent pattern was obvious in either the asymptomatic carriers or HAM/TSP patients.

To determine whether the CDR3 of the TCR in patients with HAM/TSP may be restricted or conserved, we conducted extensive nucleotide sequencing of this region. Tables IV–VI showed a representative analysis of the predicted amino acid sequences of the CDR3 regions in asymptomatic HTLV-I carriers, patients with HAM/TSP, and in the one patient with ATL. The average length of the CDR3 region was not different among patients with HAM/TSP (10.66 in CD4, 10.48 in CD8), ATL (10 in CD4, 10.25 in CD8), and HTLV-I ACs (10.29 in CD4, 10.48 in CD8). PG or related sequence (PXG) motifs were more frequently observed in CD8 T cell populations of patients with HAM/TSP than in those of HTLV-I carriers. Eight of 29 clones (27.6%) had this sequence in HAM/TSP compared with 2 of 31 clones (6.5%) in HTLV-I carriers. However these differences were not of statistical significance when analyzed by Fisher’s exact test (χ² = 3.417, p = 0.06). It was interesting to note that 4 of 8 clones having a PG or related motif were from HAM/TSP patients bearing HLA-A2. LXG motifs were more frequently observed in HTLV-I carriers (6 of 38; 15.8%) than in HAM/TSP (3 of 35; 8.6%). JB utilization showed some conservation in that JB2.1 was detected in 50% of patients with HAM/TSP and in 29% of asymptomatic carriers.

In contrast to CD4⁺ T cells, it has been shown that CD8⁺ T cells and particularly CD45RO⁺, CD57⁺, and CD28⁻ populations may exhibit oligoclonality in normal individuals (17, 18, 19, 20). In a previous study that employed the same PCR-based assay used in our present study, we demonstrated that 72% of normal subjects had evidence of oligoclonality in the CD8⁺ T cell subset, with the mean number of clones per subject being reported to be 3.8 (17). Frequent oligoclonal expansion has also been observed in bone marrow transplant recipients and in patients with rheumatoid arthritis and multiple sclerosis (21, 22). With respect to HTLV-I infection, there have only been limited studies on the clonal expansion of peripheral blood lymphocytes. HTLV-I preferentially infects CD4⁺ T lymphocytes (6, 7), and it has been suggested that in many cases this may lead to a benign clonal expansion of the infected population(s) (8, 9, 10, 11). In addition, several studies have shown that in both ACs and in patients with HAM/TSP there is a vigorous CD8⁺ T lymphocyte CTL response (23, 24, 25, 26, 27, 28). One study on TCR variable regions conducted on Tax-specific CTL clones established from peripheral blood of HAM/TSP patients suggested a limited Vα and Vβ utilization (28), and sequence analysis of the CDR3 regions showed evidence of an oligoclonal expansion in responding T cells (12). A more detailed study on TCR Vβ regions based on analysis of cDNA from central nervous system lesions of HAM/TSP patients suggested that T cells containing restricted Vβ CDR3 motifs become activated upon HTLV-I infection (29). Recently, Höger et al. (13) clearly demonstrated the accumulation of T cell clones in patients with HAM/TSP using PCR-based SSCP methods which allowed the detection of conformational polymorphism of the CDR3 regions of TCR Vβ. This method detected T cell oligoclonality as labeled bands in a background smear pattern following SSCP gel electrophoresis. The numbers of reported clones ranged from 65 to 94 and were much higher in patients with HAM/TSP compared with ACs. However, these studies were limited in that no attempts were made to differentiate expansion in isolated CD4⁺ and CD8⁺ cell populations.

The CDR3 loops of both α- and β-chains are thought to carry the fine specificity of Ag recognition by T cells (30). Because nucleotide transferases add or remove nucleotides at the various V-D and D-J junctions during recombination, the CDR3 region of any V-J combination may vary in length by as many as 6–8 amino acids (31). A recently developed multiplex PCR approach that allows for the quantitative evaluation of the CDR3 length profile of each Vβ segment families was utilized in our study (17, 18, 19, 20, 21, 22). The detection of biased peaks compared with the expected symmetrical background distribution of the several bands can be regarded as evidence of clonality within the T cell repertoire, and Ag-driven clonal T cell expansion can be evaluated more conveniently as compared with the SSCP method described above. The major difference between the SSCP method and our procedure is related to the detection of clonality based on conformation vs the length of the CDR3 regions. In our studies, we could demonstrate the following. 1) Mono- or polyclonal expansion of CD4⁺ T cells was confirmed in two of four ACs and in four of five patients with HAM/TSP. In the one patient with ATL, a specific clone bearing Vβ 7 was detected in the CD4⁺ T cell subset. 2) Accumulated oligoclonality within the CD8⁺ T cell subset was evident in all subjects, including ACs (3–10 clones), HAM/TSP patients (3–8 clones), and an ATL patient (8 clones). The mean value of the number of clones was 6.8 clones per individual. No statistically significant differences in the average number of clones were observed between individuals with HAM/TSP and ACs. 3) The repertoire of the clonally expanded T cells was extremely diverse and there was no bias in Vβ gene usage. 4) There was no statistically significant restriction in the CDR3 motifs of the TCR Vβs. 5) In two patients with HAM/TSP, who were HLA-A2 positive, PG, or related motifs (PXG) in the Vβ CDR3 sequence were detected in four predominant clones.

The clonal expansion within CD4⁺ T cells observed in our study would seem to be related at least in part to the proliferation of cells as a result of virus infection. HTLV-I has a preferential and often exclusive tropism for CD4⁺ T lymphocytes and clonal expansion of HTLV-I-infected cells has been predicted by Southern blot hybridization assays (8, 32), and more recently by ligation mediated inverse PCR amplification methods (9, 10, 11). Our study provides the first direct demonstration of in vivo clonal expansion of infected cells within the fractionated CD4⁺ T cell subset. However, further studies that would require the isolation of individually expanded populations with subsequent molecular analysis to detect the presence of integrated HTLV-I provirus are required to definitely confirm the association of infection with expansion. In our experience, such studies are technically difficult in that the successful isolation of the expanded CD4⁺ T lymphocyte population requires very large quantities of peripheral blood mononuclear cells as starting material, and this is difficult to obtain from most patients. Moreover, these studies are also restricted by the limited availability of specific Vβ Abs that would be required to isolate the expanded population. However, at present we are attempting to overcome these difficulties and carry out such studies. The view that clonal expansion of CD4⁺ T lymphocyte populations is due to HTLV-I infection is also supported by the observations that expansion is only rarely observed in uninfected normal individuals and was not observed in any of three HTLV-II-infected individuals included in this study. In contrast to HTLV-I, HTLV-II has a preferential tropism for CD8⁺ T lymphocytes (34), and this would be consistent with the absence of clonal expansion of CD4⁺T lymphocytes in HTLV-II infection and its presence in HTLV-I infection.

The present studies also provide the first description of the frequent clonal T cell expansions within fractionated CD8⁺ cell population (6.8 clones per individual), which suggests there is Ag-driven polyclonal CD8⁺ cell expansion in HTLV-I-infected individuals. However, further studies will also be required to confirm this and to identify the Ags involved. While clonal expansion of CD8⁺ cell populations certainly occurs in normal uninfected individuals (17, 18, 19, 20), the number of expanded clones in the HTLV-I-infected group is higher than would be expected. The reason for the discordance in the numbers of clones (per subject) with the previously reported results using the PCR-based SSCP assay (13) (65–94 clones reported for HAM/TSP patients) is unknown but may partially result from the differences in the methods employed. Our results also differ from this study in that we could not demonstrate differences in the number of expanded clones between ACs and patients with HAM/TSP. However, direct comparison of the two studies is very difficult, as the latter study also did not differentiate between expansion of the CD4 and CD8 populations. Hara et al. (29) suggested the presence of restriction in CDR3 motifs of infiltrating lymphocytes in the spinal cord lesions of HAM/TSP patients as encephalitis-related motifs (LXG). We found no evidence for this. LXG motifs were only detected in 3 of 35 predominant clones identified in patients with HAM/TSP, but were also detected in ACs. PG or related motifs (PXG) of the TCR Vβ CDR3 regions, which previously were reported as structural motifs of the HTLV-I Tax-specific CTL clones restricted by HLA-A2, were detected in eight predominant clones in patients with HAM/TSP (four of these clones were also identified in HLA-A2-bearing patients). However, based on the innate tendency of the terminal deoxytransferase to increase the incorporation of glycine residues (33), the evaluation of CDR3 motifs involving G residues requires careful interpretation.

In HAM/TSP patients, both the high proviral DNA load (35, 36, 37, 38) and the extremely low levels of viral replication in PBLs (38) are accompanied by higher and active humoral immune responses (40– 42) and a higher frequency of HTLV-I-specific CD8⁺ CTLs compared with ACs (25). HLA haplotype analysis has been rewarding in pinpointing the role of cellular immune responses against HTLV-I in the pathogenesis of HAM/TSP (41) and a proliferative potency of the PBLs responding to HTLV-I in HAM/TSP patients has received attention as an in vitro equivalent of the cellular inflammation in the central nervous system of the patients (43, 44, 45, 46). Taken together, the data support the involvement and importance of anti-HTLV-I CTL activities in the pathogenesis of HAM/TSP (23, 24, 25). HTLV-I-specific CD8⁺ CTLs have been demonstrated directly in lymphocytes in the peripheral blood and CSF of HAM/TSP patients (23, 24, 25, 47) and in the peripheral blood of ACs (26, 27). Although the impairment of some nonspecific cellular immune activities has been reported in patients with HAM/TSP (48, 49, 50), a beneficial effect of immunosuppressive therapies in a majority of the patients is well established (51). The detection of oligoclonal expansions of CD8⁺ T cells in HTLV-I-infected individuals as demonstrated in the present study may preferentially reflect the presence of the HTLV-I-specific CTL clones in these subjects. In our comparative analysis, there was no significant difference in the TCR Vβ CDR3 length profile between HAM/TSP patients and ACs, suggesting that the presence of oligoclonal expansion of CD8⁺ T cells may be only one of the prerequisites for the development of disease.