HTLV-1 modulates the frequency and phenotype of FoxP3+CD4+ T cells in virus-infected individuals
© Satou et al.; licensee BioMed Central Ltd. 2012
Received: 24 December 2011
Accepted: 30 May 2012
Published: 30 May 2012
Skip to main content
© Satou et al.; licensee BioMed Central Ltd. 2012
Received: 24 December 2011
Accepted: 30 May 2012
Published: 30 May 2012
HTLV-1 utilizes CD4 T cells as the main host cell and maintains the proviral load via clonal proliferation of infected CD4+ T cells. Infection of CD4+ T cells by HTLV-1 is therefore thought to play a pivotal role in HTLV-1-related pathogenicity, including leukemia/lymphoma of CD4+ T cells and chronic inflammatory diseases. Recently, it has been reported that a proportion of HTLV-1 infected CD4+ T cells express FoxP3, a master molecule of regulatory T cells. However, crucial questions remain unanswered on the relationship between HTLV-1 infection and FoxP3 expression.
To investigate the effect of HTLV-1 infection on CD4+ T-cell subsets, we used flow cytometry to analyze the T-cell phenotype and HTLV-1 infection in peripheral mononuclear cells (PBMCs) of four groups of subjects, including 23 HTLV-1-infected asymptomatic carriers (AC), 10 patients with HTLV-1 associated myelopathy/tropical spastic paraparesis (HAM/TSP), 10 patients with adult T-cell leukemia (ATL), and 10 healthy donors. The frequency of FoxP3+ cells in CD4+ T cells in AC with high proviral load and patients with HAM/TSP or ATL was higher than that in uninfected individuals. The proviral load was positively correlated with the percentage of CD4+ T cells that were FoxP3+. The CD4+FoxP3+ T cells, themselves, were frequently infected with HTLV-1. We conclude that FoxP3+ T- cells are disproportionately infected with HTLV-1 during chronic infection. We next focused on PBMCs of HAM/TSP patients. The expression levels of the Treg associated molecules CTLA-4 and GITR were decreased in CD4+FoxP3+ T cells. Further we characterized FoxP3+CD4+ T-cell subsets by staining CD45RA and FoxP3, which revealed an increase in CD45RA−FoxP3low non-suppressive T-cells. These findings can reconcile the inflammatory phenotype of HAM/TSP with the observed increase in frequency of FoxP3+ cells. Finally, we analyzed ATL cells and observed not only a high frequency of FoxP3 expression but also wide variation in FoxP3 expression level among individual cases.
HTLV-1 infection induces an abnormal frequency and phenotype of FoxP3+CD4+ T cells.
Human T-cell leukemia virus type 1 (HTLV-1) is a delta type retrovirus, which causes leukemia of HTLV-1-infected CD4+ T cells, known as adult T-cell leukemia (ATL) [1–4], in 2 to 5 % of infected individuals. HTLV-1 is also associated with chronic inflammatory diseases [5, 6], including HTLV-1 associated myelopathy/tropical spastic paraparesis (HAM/TSP), uveitis, alveolitis , and dermatitis . It has been estimated that 20 million people are infected with HTLV-1 in the world. HTLV-1 has a characteristic proliferative strategy; HTLV-1 increases its copy number not via vigorous production of cell-free viral particle but mainly via proliferation of infected host cells, which contain the integrated HTLV-1 provirus in the host genome [9, 10]. Given the fact that HTLV-1 utilizes CD4+ T cells as the major host cell population, the pathogenesis by this virus may be due to abnormalities of CD4+ T cells in HTLV-1-infected individuals. However the precise characteristics of the putative CD4+ T-cell abnormality still remain to be elucidated.
In addition to viral structural proteins, such as Gag, Pol, and Env, HTLV-1 encodes several regulatory and accessory proteins, including Tax, Rex, p30, p12, and HTLV-1 bZIP factor (HBZ), which regulate viral gene expression or proliferation of infected host cells . After the HTLV-1 provirus is integrated into the host genome, the virus expresses these regulatory and accessory proteins to induce host cell proliferation or viral latency, resulting in persistent infection in vivo. Tax is known to influence various host cell-signaling pathways, for example activation of NF-κB, and to contribute to proliferation and survival of infected cells [11, 12]. Another viral gene, the HBZ, which is encoded in the minus strand of HTLV-1  and expressed constitutively in the infected host cells [14, 15], also contributes to proliferation of the infected cells [14, 16], dysregulation of differentiation and function of CD4+ T cells , and the pathogenesis of diseases such as T-cell lymphoma and chronic inflammatory diseases [17, 18]. On the other hand, viral protein expression induces the host immune response to eliminate the virus, which includes both antibody and cytotoxic T lymphocytes (CTL) against the viral antigens [19–21]. It has been reported that the CTL response against this virus determines HTLV-1 proviral load; yet, the host immune system cannot eliminate the HTLV-1 completely, which allows HTLV-1 to establish persistent infection in almost all infected individuals.
Recent studies have clarified the presence of various CD4+ T-cell subsets. CD4+ T cells can be divided into two major categories, effector T cells and regulatory T cells. Effector T cells induce the activation of immune responses by secreting pro-inflammatory cytokines whereas regulatory T cells, which express the transcription factor FoxP3 [22–24], suppress the immune response by both cell-contact dependent and independent mechanisms . As an example of cell contact dependent suppression, expression of the immune suppressive molecule CTLA-4 on the cell surface inhibits the activation of surrounding neighboring T cells . In addition, a recent report demonstrated that human FoxP3+CD4+ T cells were composed of three phenotypically and functionally different subsets according to the degree of FoxP3 expression and CD45RA expression, namely CD45RA+FoxP3low resting Treg cells (rTreg cells), CD45RA−FoxP3high activated Treg cells (aTreg cells), or CD45RA−FoxP3low non-suppressive T cells (FoxP3low non-Treg cells) . Both rTreg cells and aTreg cells have suppressive function, but FoxP3low non-Treg cells are not suppressive.
Previous studies have reported that the HTLV-1 provirus is enriched in effector/memory T cells [28, 29], and the phenotype of ATL cells shares certain characteristics with regulatory T cells based on the finding of FoxP3 expression [30, 31]. However there are few studies that systematically and specifically investigate which recently described CD4+ T-cell subset is infected by HTLV-1 in asymptomatic carriers (AC), HAM/TSP patients, and ATL patients. To elucidate this point, we analyzed peripheral mononuclear cells (PBMCs) from naturally HTLV-1-infected individuals, including AC, HAM/TSP, and ATL patients, by using multicolor flow cytometric analysis combined with the detection of the viral antigen Tax to identify the presence of HTLV-1 . We found the specific CD4+FoxP3+ T-cell subset is frequently infected with HTLV-1, which may allow the virus to achieve persistent infection in vivo, and should also contribute to the pathogenesis of the virus-associated diseases.
Characteristics of participants
Age, median years (IQR)
Male sex, no (%)
PVL median (IQR)
Consistent with previous reports that Tax expression is frequently silenced in ATL cells, Tax expression after ex vivo cultivation of ATL cells was not correlated with the proviral load (Figure 2C). The percentage of Tax positive cells tended to be lower than the proviral load even after ex vivo culture in AC and HAM/TSP patients, but we found that Tax positivity showed a significant correlation with the proviral load both in AC and HAM/TSP (r = 0.91 or 0.61, p = 0.00002 or 0.0334, respectively, Figure 2D and E). In order to investigate whether T-cell subset markers, including FoxP3 and CD45RA, are influenced by ex vivo cultivation, we analyzed their expression both before and after cultivation. The results showed that the frequency of FoxP3 or CD45RA was not significantly changed during ex vivo culture (Additional file 2: Figure S2). These findings collectively indicate the usefulness of this Tax detection system for this study.
FoxP3+ Treg cells play a crucial role in persistent infection and pathogenesis of chronic viral infection. Previous studies have suggested that Treg cells suppress virus-specific CD8+ T-cell effector functions in chronic human viral infections such as human immunodeficiency virus, hepatitis C virus and cytomegalovirus [34, 35]. Regarding this point, FoxP3+ Treg cells play a role in facilitating viral persistence. In HTLV-1 infection, the frequency of FoxP3+ cells is indeed correlated with the impairment of CTL activity against the viral antigen Tax in HAM/TSP patient . On the other hand, FoxP3+ Treg cells could prevent tissue damage caused by excessive immune response triggered by viral infection. In addition to these general roles of FoxP3+ Treg cells in chronic viral infection, FoxP3+ Treg cells should have some specific role in HTLV-1 infection, because FoxP3+ Treg cells are comprised in CD4+ T cells, which are a main host cell population of HTLV-1. Here we performed a comprehensive analysis of CD4+ T-cell subsets in individuals naturally infected with HTLV-1 and revealed that the frequency of HTLV-1 infection is positively correlated with the frequency of FoxP3+ T cells (Figure 1E). The increased FoxP3+ T cells themselves are frequently infected with HTLV-1 (Figure 3B), suggesting that HTLV-1 utilizes the FoxP3+ T cells as a host cell. What is the advantage for HTLV-1 to exist in FoxP3+ T cells? There are two possibilities for this preference. First, FoxP3+ T cells are known as hyper-proliferating cells in vivo with a doubling time of 8 days , which could contribute to clonal expansion of infected cells. Second, HTLV-1 can evade the host immune system by directly infecting this potentially immuno-suppressive cell population. Thus, HTLV-1-infection of FoxP3+ T cells should enable the virus to increase or maintain proviral load and achieve persistent infection.
How then does HTLV-1 infection target FoxP3+ T cells? This could be explained by the following two mechanisms. First, FoxP3+ T cells are known to contact with dendritic cells (DCs) frequently , which could increase the chance of de novo viral infection between DCs and FoxP3+ T cells. A recent study demonstrated that cell-free HTLV-1 efficiently infects DCs, and the infected DCs promote de novo infection of CD4+ T cells . This notion is consistent with the finding that effector/memory-type CD45RA− Treg cells, including FoxP3low non-Treg cells and FoxP3high aTreg cells, are more frequently infected with HTLV-1 than CD45RA+ rTreg cells (Figure 4E). Second, once FoxP3− T cells are infected with HTLV-1, HBZ should be expressed in the host cells. Since HBZ is recently reported to induce FoxP3 expression via enhancing TGF-β signaling pathway [17, 40], HTLV-1 infection is likely to convert FoxP3− cells into FoxP3+ cells. In addition, HTLV-1 has a cell-extrinsic effect on FoxP3+ cell generation. HTLV-1 infected cells secrete CCL22 via expression of Tax, which indirectly contributes to the generation and maintenance of HTLV-1 uninfected FoxP3+ cells [41, 42]. This would contribute to an increased number of HTLV-1-uninfected FoxP3+ cells.
Since FoxP3+ Treg cells play a crucial role in suppressing immune response, the increase of FoxP3+ cells observed in HTLV-1 infection may contribute to immunodeficiency, which is frequently observed in HTLV-1 infection . On the other hand, the high frequency of FoxP3+ T cells observed in HAM/TSP patients is paradoxical, because the pathogenesis of HAM/TSP is believed to be inflammatory. Therefore, we analyzed the phenotype of the increased FoxP3+ cells and observed that CTLA-4 and GITR expression of FoxP3+ T cells in HAM/TSP patient was significantly reduced compared to uninfected individuals (Figure 5C). A similar observation was reported previously that the expression level of FoxP3, GITR, or CTLA-4 mRNA in CD4+CD25+ T cells of HAM/TSP patients is lower than that of HD . That report used CD4+CD25+ as a marker of Treg cells, but CD4+CD25+ T cells contain not only FoxP3+ Treg cells but also FoxP3− activated T cells. Particularly the proportion of CD4+CD25+FoxP3−activated T cells is up-regulated in HAM/TSP patients, which is likely to reduce the proportion of FoxP3+ Treg cells in CD4+CD25+ T cells of HAM/TSP patients. Thus, the expression level of GITR or CTLA-4 in FoxP3+ T cells of HAM/TSP patients has not been elucidated yet. To avoid this concern, we utilized the multicolor flow cytometry, which enabled us to show that CTLA-4 and GITR were clearly down regulated in FoxP3+ T cells of HAM/TSP patients.
Then what is the underlying mechanism of this phenomenon? We reported recently that HBZ-Tg mice showed a pro-inflammatory phenotype in spite of the increase of Foxp3+ T cells , which is similar to HAM/TSP patients (Figure 1D). Treg associated molecules were also down regulated in Foxp3+ T cells of HBZ-Tg mice. Thus, HBZ-mediated FoxP3 dysfunction may play a role in the abnormality regarding FoxP3+ cells in HAM/TSP patients. It has been reported that Tax also contributes to the dysregulation of FoxP3+ Treg cells. Tax suppresses FoxP3 expression at transcriptional level , which alternatively or additionally could contribute to the abnormal phenotype of FoxP3+ cells. These findings collectively indicate that the increased FoxP3+ Treg cells were functionally impaired in HAM/TSP patients. Furthermore, FoxP3+CD4+ T cells in HAM/TSP patient contain an increased FoxP3+ non-Treg population (Figure 5F), which would contribute to the inflammatory phenotype of HAM/TSP via generation of pro-inflammatory cytokine-producing CD4+ T cells such as THAM cells  or exFoxp3 cells .
In the current study, we did not observed FoxP3 repression during Tax expression by ex vivo cultivation. This result seems to be inconsistent with a previous report that Tax represses FoxP3 expression . There are two possible explanation of this inconsistency. First, there is the difference of the ways to express Tax. In the previous study, the authors used transfection of plasmid that induces Tax expression by the CMV promoter. We used endogenous HTLV-1 provirus to express Tax. Therefore, the expression level of Tax in our current study should be much lower than that of the previous study. In addition, Tax expression was induced in a proportion of FoxP3+ cell in our current study. Second, there are differences in incubation time for Tax expression. In the previous study, the authors evaluated FoxP3 expression after 48 hours of transfection, whereas we evaluated FoxP3 expression within 24 hours after Tax expression.
High expression levels of CD25 are also well documented in HTLV-1 infection . Consistent with previous findings, CD25 expression is upregulated in FoxP3+ cells of HAM/TSP patient (Figure 5C). One determinant of the susceptibility to HAM/TSP is host genetic polymorphism such as MHC class 1, which influences the efficiency of CTL against HTLV-1 [48, 49]. HTLV-1-infected individuals who have HLA class I susceptible for HAM/TSP may allow high expression of Tax and/or HBZ, which could cause up-regulation of CD25 molecules in the FoxP3+ cell population (Figure 5C).
It is controversial whether ATL is a leukemia of FoxP3+ Treg cells or not. However, there is no a priori reason to assume that ATL cells must be exclusively derived from FoxP3+ Treg cells or non-Treg cells. Indeed, there are previous reports to support both possibilities. Some studies have reported that ATL cells have regulatory functions [50, 51], whereas other studies reported no regulatory function in ATL [52, 53]. We showed here that HTLV-1 is frequently detectable in CD4+FoxP3+ T cells (Figure 3B) in AC. More than half of ATL cells express FoxP3 (Figure 6), even though FoxP3 expression in ATL cells is variable as shown in the present and previous studies [30, 31]. These findings prompt us to propose an idea that more than a half of ATL cells are possibly derived from FoxP3+ Treg cells. We reported previously that HBZ expression is constitutively active but Tax expression is frequently silenced in ATL cells, which possibly contributes to high frequency of FoxP3+ ATL.
This study demonstrated that HTLV-1 infection induced the abnormality of frequency and phenotype of FoxP3+ T cells, suggesting that HTLV-1 has evolved a sophisticated strategy to achieve persistent infection by directly affecting the central regulator of the host immune system. HTLV-1-mediated dysregulation of FoxP3+ T cells is likely to be a critical cellular mechanism for the understanding HTLV-1 pathogenicity.
PBMCs were obtained from asymptomatic HTLV-1 infected carriers (n = 23), HAM/TSP patients (n = 10), ATL patients (n = 10), and age-matched healthy controls (n = 10). Characteristics of each group are presented in Table 1. ATL patients consist of 2 acute, 4 smoldering and 4 chronic types of ATL cases. Genomic DNA extracted from PBMCs was used to determine proviral load (PVL) as described previously . Briefly, PVL was quantified by real time PCR and calculated by using genomic DNA of TL-Om1, an ATL cell line with one copy of complete HTLV-1 provirus, as a standard of 100%. We defined AC with less than 2% of proviral load as AClow and AC with more than 2% of proviral load as AChigh. This study was conducted according to the principles expressed in the Declaration of Helsinki and approved by the Institutional Review Board of Kyoto University (844 and E-921). All patients provided written informed consent for the collection of samples and subsequent analysis.
The following antibodies were purchased from BD PharMingen; purified monoclonal antibody (mAb) for human CD3 (UCHT1), CD4 (RPA-T4), CD8a (RPA-T8), CD45RA (NI100) and CTLA-4 (BNI3). Purified mAbs for human CD25 (BC96), GITR (eBio AITR) and FoxP3 (236A/E7) were purchased from eBioscience.
PBMCs were isolated with Ficoll-Isopaque (GE Healthcare) gradient centrifugation. Flow cytometric analyses were carried out using a FACS CantoII with Diva Software (BD Pharmingen), and the data were analyzed by FlowJo software (Treestar). To discriminate dead cells, we used LIVE/DEAD Fixable Near-IR Dead Cell Stain Kit (Invitrogen). For cell surface staining, 106 cells were incubated with mAbs for 30 minutes at 4°C, and then analyzed. For intracellular staining, we used a human FoxP3 staining kit according to the manufacture’s protocol (eBioscience). To distinguish FoxP3+ and FoxP3− cell population clearly, we used isotype control according to the manufacture’s recommendation. To detect the viral antigen Tax, we cultured PBMCs from ACs or HAM/TSP patients for 12–18 hours and stained with monoclonal antibodies against FoxP3 or Tax (MI-73) , and then analyzed by flow cytometry.
To compare 2 groups when data were determined to have a Gaussian distribution, the Student t test was used. If data did not have a Gaussian distribution, the Mann–Whitney U test was used for unpaired data, and the Wilcox signed-ranks test was used for paired data. The AC group and HD did not differ significantly in sex or age, using chi-squared test and Mann–Whitney U test. Differences with P < 0.05 were considered to be statistically significant. Correlations were evaluated using Spearman’s rank correlation.
This study was supported by Grant-in-Aid for Scientific Research from Japanese Society for the Promotion of Science and Ministry of Health Labor and Welfare, a grant from Takeda Science Foundation, a grant from Naito Foundation. We thank Prof. Charles R.M. Bangham for critical reading of the manuscript and Ms. M. Nakashima for preparation of peripheral blood of patients. We are most grateful to the patients and healthy donors who participated in this study.