Human T cell lymphotropic virus 1 (HTLV-1) infects 20 million people worldwide [1]. While the majority of infected individuals are asymptomatic carriers (AC) of the virus, 5-10% of infected people develop either adult T cell leukemia/lymphoma (ATL) [2] or a chronic, progressive, neurological disease termed HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP) [3, 4]. HAM/TSP is characterized by perivascular inflammatory infiltrates in the brain and spinal cord, with a predominance of HTLV-1-specific CD8⁺ T cells [5, 6]. High frequencies of these effector cells have been demonstrated in peripheral blood with even higher frequencies in cerebrospinal fluid (CSF) of patients with HAM/TSP [7–9], and robust humoral responses against HTLV-1 antigens that can be detected in the CSF as well as the serum [4, 10]. While the cellular and humoral immune responses against HTLV-1 play crucial, protective roles in HTLV-1 infection, chronically activated immune responses have been suggested to underlie the pathogenesis of HAM/TSP [11]. Therefore, characterization of HTLV-1-specific immune responses may provide evidence of immune dysregulation during disease progression in HAM/TSP patients, and may help identify novel immunotherapeutic targets in HTLV-1-related diseases.

Despite strong HTLV-1-specific immune responses, HTLV-1 proviral loads are significantly elevated in HAM/TSP patients compared to AC [12]. Increased expression particularly of the trans-activating viral gene encoding HTLV-1 Tax induces the expression of various cellular genes, including IL-2, IL-15, and their receptors [13–16], which directly contributes to lymphocyte activation in HAM/TSP patients [9, 17]. A novel gene, HTLV-1 basic leucine zipper factor (HBZ), encoded by the minus strand of the HTLV-1 proviral genome has been identified [18]. HBZ mRNA is ubiquitously expressed in all ATL cells and promotes the growth and survival of the leukemic cells [19]. HBZ protein was found to inhibit Tax-mediated viral gene transcription from the 5’ LTR and to selectively suppress the classical NF-κB pathway [18, 20–23]. Previous in vivo studies also demonstrated that HBZ expression enhanced HTLV-1 infectivity, T cell proliferation and lymphoma [24–26]. Furthermore, HBZ mRNA expression was detected in HAM/TSP patients, and was correlated with proviral load and disease severity [27]. Since these findings suggested that HBZ has a critical role in HTLV-1 persistence and the development of ATL and HAM/TSP, it is important to define HBZ-specific immune responses in HTLV-1-infected individuals.

Recent evidence has shown that HBZ is an immunogenic protein recognized by HBZ-specific CTL clones [28, 29]. HBZ-specific CD8⁺ T cells are detected in AC and HAM/TSP patients, and HBZ-specific CTL clones were able to lyse naturally infected cells isolated from AC and HAM/TSP patients, but not ATL patients [28, 29]. Despite recent studies on HBZ-specific cellular immune responses, there are no reports on the humoral immune responses to HBZ. We recently reported that a luciferase immunoprecipitation system (LIPS), a highly sensitive, quantitative technology, could efficiently detect HTLV-1 antigen-specific antibody responses in serum of HTLV-1-infected individuals [30, 31]. Since the LIPS assay can detect antibody responses against multiple antigens, profiling of HTLV-1-specific antibody responses using LIPS demonstrated a differential pattern of antibody responses for HTLV-1 Gag, Env and Tax between HTLV-1-infected and uninfected subjects as well as between the AC and ATL and HAM/TSP patients [30, 31]. Here we optimized the LIPS assay for detection of immunoreactivity against HBZ, and first determined antibody responses against HBZ in HTLV-1-infected individuals.

The HBZ gene is constitutively expressed in HTLV-1-infected cells, ATL cells and PBMC of HTLV-1-infected individuals [19, 27, 34] and is thought to be involved in both regulation of viral gene transcription and T-cell proliferation [18, 20–23], suggesting that HBZ has a critical role in HTLV-1 infectivity and the development of HTLV-1-related diseases [24–27]. Therefore, it is important to define HBZ-specific immune responses in HTLV-1-infected individuals. In this study, we screened 436 serum/plasma samples from Jamaica and the United States, including NDs, ACs, ATL and HAM/TSP patients, and first defined the humoral immune response to HBZ in HTLV-1-infected individuals. The results demonstrated that antibody responses for HBZ were detected in 11.8% of HTLV-1-infected individuals. The frequency of antibody response for HBZ was low compared to a high frequency of antibody responses for Gag, Env and Tax (99.3%, 92.3% and 93.0%, respectively) as previously described [31]. Recently, it was reported that HBZ-specific CD8⁺ T cells were detected in HTLV-1-infected individuals, and HBZ-specific CTL clones were able to lyse naturally infected cells [28, 29]. Likewise, it was demonstrated that the predicted binding affinity of HLA molecules to HBZ peptides is significantly weaker than that of Tax peptides and that the frequency of HBZ-specific CD8⁺ T cells is significantly lower than the frequency of Tax-specific CD8⁺ T cells [29]. Our results support previous reports [28, 29] suggesting that HBZ is an immunogenic protein although HBZ-specific immune responses appear to be lower compared to the other HTLV-1 immunodominant proteins, Gag, Env and Tax. Moreover, HBZ-specific humoral immune responses did not show any association with HTLV-1-related disease outcomes, while HTLV-1-infected AC and ATL patients from HAM/TSP patients could be discriminated based on the differential antibody responses for Gag, Env and Tax [31]. ATL patients showed lower mean immunoreactivity against HBZ compared to HAM/TSP patients (Figure 1A, Table 2). This is consistent with our previous report that ATL patients demonstrated lower levels of antibody responses to all three HTLV-1 immunodominant proteins, Gag, Env and Tax, compared to HAM/TSP patients [31]. This may be a consequence of a more global immunosuppressed state in ATL than HAM/TSP [35]. Interestingly, the distribution of an HBZ-specific antibody response by ATL subtype demonstrated that immunoreactivity against HBZ might differentiate the three clinical subtypes of ATL patients. Among ATL patients, there was a significant difference in the HBZ-specific antibody responses between patients with the chronic and lymphoma subtypes (Figure 1D). The median survival of patients with ATL is 20 weeks; patients with the acute subtype survive for a median of 13 weeks, patients with the lymphoma subtype survive for a median of 20 weeks, and patients with the chronic subtype survive for a median of 25 weeks [36]. The diagnostic criteria and clinical classification of the chronic ATL subtype includes more absolute lymphocytes with T lymphocytosis, and is associated with a better prognosis compared to the acute and lymphoma ATL subtypes [37]. Therefore, our results suggest that a subset of ATL patients generate HBZ-specific immune response (of low magnitude), which may delay disease progression. Since previous reports did not show any significant differences in antibody responses for HTLV-1 Gag, Env and Tax among ATL subtypes [31], it is important to validate these observations with HBZ in a larger sample of ATL patients that includes the less common subtype, smoldering ATL. In addition, since high expression of HBZ mRNA has been reported in ATL cells [19], it would be of interest to correlate anti-HBZ antibody responses with levels of HBZ mRNA. Unfortunately, PBMCs from ATL patients in this study were not available for analysis.

The lack of correlation of antibody responses for HBZ with proviral loads or HBZ mRNA expression might partially be the result of virological properties of HBZ itself. It has been reported that HBZ mRNA was detectable in PBMCs of HAM/TSP and ATL patients after culture, but mainly remains retained in the nucleus more than in the cytoplasm [38]. In primary ATL cell lines, only the spliced form of HBZ protein was detected in the nuclear fraction [39]. These reports suggest that persistence of HTLV-1 might be a consequence of reduced HBZ translation or limited localization of HBZ protein, and perhaps of reduced exposure of infected cells to HBZ-specific host immune responses. In other retrovirus infections such as human immunodeficiency virus (HIV), the regulatory protein Tat plays an important role in viral infectivity and pathogenicity [40], however Tat-specific antibody responses are detected only in a small number of HIV-infected individuals [41]. Similarly, antibody responses for HBZ were only observed in a subset of HTLV-1-infected individuals. Genetic factors, such as HLA, may also play a role in generation of a specific immune response, but larger numbers of HTLV-1-infected individuals with HBZ-specific humoral immune response will be required to further characterize the humoral immune response against HBZ in HTLV-1-associated diseases.

Antibody responses for HTLV-1 Gag, Env and Tax were elevated both in serum and CSF of HAM/TSP patients (Figure 3), consistent with previous reports [10, 32]. By contrast, HBZ-specific antibody responses were only detected in CSF of HAM/TSP patients if there was a coincident serum antibody response against HBZ. These virus-specific antibodies in the CSF are either derived from the blood (leakage through the blood–brain-barrier) or alternatively, are synthesized locally within the CNS. Interestingly, the ratio of HBZ-specific antibody responses in CSF to serum was lower compared to ratios of antibody responses for HTLV-1 Gag, Env and Tax in CSF to serum (Figure 3). These results suggest that HBZ-specific antibody in the CSF is derived from the blood while antibodies for HTLV-1 Gag, Env and Tax are intrathecally synthesized. It has been previously reported that intrathecal antiviral antibody synthesis was confirmed by the presence of HTLV-1-specific antibodies and oligoclonal IgG in CSF [42–45] and that the lack of intrathecal antibody response to HTLV-1 in HAM/TSP correlates with higher proviral loads and worse outcome [46]. In the future it will be interesting to examine the relationship between CSF proviral load, and the presence or level of anti-HBZ antibody in CSF and to compare the clinical phenotype of HAM/TSP patients with and without CSF anti-HBZ antibodies.

Lastly, to further confirm the presence and the function of HBZ-specific antibody, we generated memory B cells producing HBZ-specific antibody from HAM/TSP patients with antibody response against HBZ. Since antigen-specific human memory B cells circulate at very low frequencies in peripheral blood, many researchers have relied on expansion and conversion of memory B cells into antibody-secreting cells by in vitro culture or development of alternative strategies [47–52]. In the present study, HBZ-specific antibody production was detected in memory B cells generated from all HAM/TSP patients who had antibody responses for HBZ in serum. As control, HBZ-specific antibody production was not detected in memory B cell cultures generated from HAM/TSP patients without HBZ-specific antibody response, while memory B cells generated from both HAM/TSP patients with and without HBZ-specific antibody response could produce antibodies specific to HTLV-I Gag, Env or Tax (data not shown). These results suggested that HTLV-1-specific memory B cells are maintained through an individual’s lifetime at levels that correlate with sustained serum antibody concentrations. Moreover, HBZ-specific antibody could significantly inhibit spontaneous lymphoproliferation of HAM/TSP patients without anti-HBZ response, including CD8⁺ T cell proliferation. HTLV-1 predominantly infects CD4⁺ T cells, but CD8⁺ T cells have also been shown to carry high proviral loads in HAM/TSP patients [53] and HBZ mRNA was detectable in both CD4⁺ and CD8⁺ T cells isolated from HAM/TSP patients (data not shown). It remains to be defined how HBZ-specific antibody inhibits CD8⁺ T cell proliferation in HAM/TSP patients. However, since humoral immunity is not limited to extracellular viral recognition but can neutralize a virus even within the cytosol of infected cells [54], HBZ-specific antibodies might therefore be able to interact with HBZ within HTLV-I infected cells. Further experiments including the nuclear retention and translation of HBZ mRNAs and the involvement of HBZ in lymphoproliferation of HAM/TSP patients would support these conclusions. Although we did not detect an inhibitory effect of HBZ-specific antibody on CD4⁺ T cell proliferation because this subset exhibits less spontaneous proliferation than CD8⁺ cells, the presence of antibody responses against HBZ was associated with less CD4⁺ T cell activation (frequency of CD25⁺ T cell subset) in HAM/TSP patients with anti-HBZ antibody responses. In CD8⁺ T cells, the frequency of CD25⁺ cells was also less in HAM/TSP patients with anti-HBZ antibody responses and approached statistical significance. These results demonstrated that HBZ-specific antibody responses have potentially beneficial roles in suppressing T cell activation in HAM/TSP patients. Since T cell activation is regulated through intricate molecular and immunological signaling networks, it will be of interest to determine how HBZ-specific antibody suppresses T cell activation of HAM/TSP patients.

In summary, this is the first report demonstrating the presence of a humoral immune response to HBZ in the context of HTLV-1-infection. Characterization of the immune response against each of the HTLV-1 viral antigens will further improve our knowledge of virus-host interactions and the pathogenesis of HTLV-1-related disorders.