Enhanced Functions of Peripheral γδ T Cells in Chronic Hepatitis B Infection during Interferon α Treatment In Vivo and In Vitro

Background γδ T cells play an important role in infectious, autoimmune, or neoplastic diseases. Here, a study was conducted to investigate the dynamic changes in phenotype and function of peripheral γδ T cells in patients with chronic hepatitis B (CHB) during pegylated-interferon (pegIFN)-α treatment, and to explore their roles in IFN-α therapy. Methods Total 15 CHB patients with pegIFN-α therapy and 6 healthy controls (HC) were enrolled in this study. Flow cytometry was used for the study of frequency of peripheral γδ T cells, subtypes, effector or memory γδ T cells, and also the IFN-γ+, TNF-α+, CD107a+ or Granzyme B+ γδ T cells in 10 patients at week 0, 4, 8, 12, 24, 36 and 48 of treatment. Another 5 CHB patients and 6 HC were recruited for the γδ T cell isolation, and gene expression in γδ T cells was evaluated before or after IFN-α treatment in vitro. Results Although γδT cells decreased in CHB patients during pegIFN-α therapy, their capacities to produce TNF-α and to express CD107a were enhanced. More effector γδT cells (CD27-CD45RA+) were found in the response group than in non-response group. Furthermore, IFN-α boosted the expression of Mx2 and cytokine genes in γδT cells from CHB patients in vitro. Conclusion IFN-α could enhance the cytokine production or cytotoxicity potential of γδT cells in vivo and in vitro. The enhanced function of γδT cells might contribute to the effect of IFN-α treatment.


Introduction
Chronic hepatitis B (CHB) is a serious public health problem in Eastern Asia. The disease is caused by persistent infection with hepatitis B virus (HBV), and has many complex clinical manifestations [1,2]. To date, interferon (IFN)-α is one of the main antiviral drugs against HBV, which has potential advantages compared to the nucleoside analogs including a lack of drug resistance, a finite and defined treatment course, and a higher likelihood of hepatitis B surface antigen (HBsAg) clearance. After about 10 years of pegylated-interferon (pegIFN)-αbased therapy for hepatitis B, pegIFN-α has attracted increased interest due to its more effective antiviral activity than standard IFN-α [3,4].
It is well established IFN-α induces anti-viral proteins and mediates the immune response. First, after interacting with the cell surface receptors, IFN-α activates the JAK-STAT signaling pathway. This results in transcription of interferon-stimulated genes (ISGs) and expression of proteins (protein kinase R, 2-5A oligosynthetase/RNAse L system, or the Mx system), which induce antiviral activity. Second, and perhaps more importantly, IFN-α directly or indirectly modulates the immune system including activation of nature killer (NK) cells, B cells, CD8 + T cells, or production of some inflammatory cytokines [5,6].
In recent years, an increasing number of studies have shown that innate immune cells, especially NK cells, play an important role in the IFN-α treatment of viral hepatitis. Those studies have revealed that during IFN-α treatment, the specific T-cell response is suppressed but NK cells proliferate and are activated. Activation of NK cells in the early stage of IFN-α treatment is associated with the control of hepatitis C virus (HCV) infection [7,8].
γδ T cells, one type of innate immune cells, share many aspects of phenotype and function with NK cells. Similar to NK cells, γδ T cells express NKG2D, CD56 or other NK markers, exert strong cytotoxicity, and secrete different types of cytokines. As reported recently, γδ T cells have multifaceted immunological functions and might be a good immunotherapeutic tool for infectious diseases and tumors [9]. These cells display their activity either through directly affecting microbial growth, destroying the infected or transformed cells, or by modulating other immune functions, such as T helper (Th)1 immune response polarization, B cell activation, or dendritic cell maturation [10,11]. And Cimini's study has shown that IFN-α improves phosphoantigen-induced Vγ9Vδ2 T cell production of IFN-γ during chronic HCV infection [12].
In our previous study, we found that the proportion of total γδ T cells increased in the blood of CHB patients, but the frequency of Vδ2 T cells decreased [13,14]. In this study, we further investigated the change in γδ T cells in CHB patients treated with IFN-α, and explored the possible effect of γδ T cells in the treatment of CHB with IFN-α.

Materials and Methods Patients
Total fifteen treatment-naïve patients with CHB were selected from the Department of Infectious Diseases, Second Affiliated Hospital of Chongqing Medical University, China, during August 2011 to May 2013. All enrolled patients met the criteria for treatment with IFN-α, including serum alanine aminotransferase (ALT) level 80-400 U/L (2~10-fold the upper limit of normal), total bilirubin (TB) normal, HBV-DNA load 10 5 -10 8 copies/mL, and positive HBsAg and hepatitis B e antigen (HBeAg) for at least six months. Patients were excluded if they were co-infected with other hepatitis viruses or had other liver diseases, such as hepatocellular carcinoma, autoimmune disease, or alcoholic hepatitis. Six healthy age-matched controls (HCs) were also recruited. In these 15 CHB patients, ten of them were recruited for the dynamic observation of peripheral γδ T cells during IFN-α treatment for 48 weeks. The other five patients before IFN-α treatment and six HCs were blood sampled for γδ T cell sorting and IFN-α treatment in vitro. The characteristics of all these subjects are shown in Tables 1 and 2. Peg-IFN-α -2a (Pegasys; Hoffman-La Roche, Shanghai, China) was administered subcutaneously at a dose of 180μg once weekly for 48 weeks. At the treatment endpoint, the patients who had normal ALT level, loss of HBeAg, and >3 log 10 decrease in HBV-DNA were considered to be responders.
The trial was carried out with approval of the Ethics Committee of the Second Affiliated Hospital of Chongqing Medical University, and written informed consent was obtained from all participants.

Clinical evaluation and HBV-DNA quantification
HBsAg and HBeAg were measured by the Roche electrochemiluminescence method. HBsAg and HBeAg values were evaluated by a cut-off index (COI): a COI of over 1.0 indicated a positive response. Hepatitis B surface antibody, hepatitis B e antibody and hepatitis B core antibody were detected with a commercial enzyme immunoassay kit. Serum HBV-DNA levels were HBV-DNA(copies/mL) 7.7±0.6 0 determined by Roche real-time fluorescent quantitative polymerase chain reaction (PCR) (Lightcycler; Hoffman-La Roche,Swiss). The detection limit of this HBV-DNA assay was 1,000 copies/mL. All liver function parameters were evaluated using an automatic biochemical analyzer. The normal level of ALT, AST or TB was 6-48 IU/L, 10-45 IU/L, or 3-21 μmol/L.

TCR γδ T cell isolation
Blood from five CHB patients before IFN-α treatment and six HCs (clinical characters showed in Table 2) was collected to generate γδ T cells from the buffy coat, and to determine the gene expression of γδ T cells when stimulated with IFN-α in vitro. Peripheral blood mononuclear cells (PBMCs) were obtained from 30 mL fresh peripheral anticoagulated blood by Ficoll separation (Ficoll-Paque), and TCR γδ T cells were isolated from the PBMCs by positive selection using the TCR γδ T cell isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany). The isolations were performed according to manufacturer's instructions (Miltenyi Biotec, Bergisch Gladbach, Germany). Purity of the sorted TCR γδ T cells was determined by flow cytometry analysis. The results showed >95% purity of γδ T cells could be obtained in the acquired fractions.

Analysis of mRNA expression of γδ T cells stimulated with IFN-α by quantitative RT-PCR
The sorted γδ T cells were divided into two parts: one for IFN-α treatment, and the other for the controls. Cells were cultured in RPMI 1640 medium supplemented with 10% FCS, 100 U/mL penicillin and 100μg/mL streptomycin in 5% CO 2

Statistical analysis
All data were analyzed by SAS version 9.2 (SAS Institute, Cary, NC, USA). Repeated measures analysis of variance and simple-effects analysis were used to compare the change in proportion of total γδ T cells or the certain subtype of γδ T cells from responders and non-responders, or from the different treatment time points. A two-sided p value <0.05 was considered as Table 3. The gene-specific primers for quantitative RT-PCR. significant. For the comparison of gene expression in vitro, the gene expression from HCs was set as "1", and the significant change in gene expression from CHB patients was set as at least two fold (upregulation) or 0.5-fold (downregulation).

Results
The proportion of γδ T cells in CHB patients decreased during pegIFN-α therapy FACS gating and Figures of γδ T, Vδ1 or Vδ2 T cells were shown in Fig. 1A-D. The proportions of γδ T cells from responders or non-responders all declined consistently during treatment, from 12.9% or 11.8% before treatment to 6.6% or 7.2% at week 48 in treatment (Fig. 1E). These differences between responders and non-responders at each time point were not significant (p>0.05), but the proportion of γδ T cells at week 36 or 48 was significantly lower than that at week 0, 4, 8 and 12 (p<0.01). The proportion of Vδ1T cells from responders was higher than in non-responders at every time point, but no significant differences were shown between the two groups (p>0.05). The proportion of Vδ1 T cells from responders was 21.1% at baseline, then it decreased slightly to 17.3% at week 36, but it increased to 22.7% at week 48. (Fig. 1F).
There was no significant difference in the proportion of Vδ1 T cells between each time point in responders or non-responders (p>0.05). The proportion of Vδ2 T cells from responders decreased sharply from 75.8% before treatment to 64.0% at the end of treatment, (Fig. 1G) and the difference between week 48 and week 0, 4, 24 or 36 was significant (p<0.05). However, the differences between the responders and non-responders were not significant (p>0.05).
Responders had higher percentage of effector γδ T cells than nonresponders during pegIFN-α treatment CD45RA and CD27 molecules were used to investigate the effector or memory phenotype of γδ T cells. The γδ T cells could be divided as the naïve cells (CD45RA + CD27 + ), effector cells  Fig. 2A-C). The proportion of naïve γδ T cells increased from 32.1% to 40.7% in the responders and 39.2% to 48.1% in the non-responders during 24 weeks treatment, and then declined to 39.1% in the responders and 36.6% in the non-responders at week 48 (Fig. 2D). Although the proportion of naïve γδ T cells in the non-responders was higher than in the responders, no significant difference was found between the two groups or among the time points (p>0.05). The percentage of effector γδ T cells at the outset of treatment was 22.6% and 41.3% for non-responders and responders, respectively, and increased to 27.9% and 47.9% after 8 weeks, and finally dropped to 18.4% and 27.8% (Fig. 2E). The frequency of effector γδ T cells from responders was significantly higher than from non-responders at week 4 and 8 in treatment (p<0.05). The non-responders patients showed a significantly higher proportion of effector memory γδ T cells (Fig. 2F)   fluctuated during treatment (Fig. 2G), and no significant differences were shown between the two groups or among the time points (p>0.05).

Increased percentage of TNF-α-and CD107a-producing γδ T cells were found in CHB patients during pegIFN-α treatment
The most important properties of functional γδ T cells include cytokine production and cytotoxicity. Fig. 3A-D shows IFN-γ-, TNF-α-, Granzyme-B-and CD107a-producing γδ T cells, respectively. In the responders, the percentage of IFN-γ + γδ T cells declined from 70.5% at baseline to 39.2% at week 12, then increased to 66.3% at week 48 (Fig. 3E). There was a sharp decrease in the frequency of IFN-γ + γδ T cells in non-responders from 66.8% at 0 weeks to 27.8% at week 4, followed by an increase to 70.0% at the end of treatment. A significant difference in the percentage of IFN-γ + γδ T cells at week 4 was shown between responders and nonresponders (p<0.01). For each group, there was a significant difference between week 12 and the other times (p<0.01). The percentage of TNF-α + γδ T cells increased from 13.6% or 13.7% at week 0 to 36.0% or 42.6% at week 48 for responders or non-responders, respectively (Fig. 3F). There were no significant differences between these two groups at each point (p>0.05), but the frequency at 48 weeks was significantly higher than at other time points for each group (p<0.01). The percentage of Granzyme-B + γδ T cells from the responders peaked as 63.7% at week 36 (Fig. 3H), and it was significantly higher than that from non-responders (p<0.05). However, no significant differences were found between each time point in each group (p>0.05). As shown in Fig. 3G, the frequency of CD107a + γδ T cells increased markedly from 10.1% or 32.5% at week 0 to 82.4% at week 8 or 83.3% at week 12 in responders or non-responders, respectively. The frequency at week 0 or 24 from the non-responders was significantly higher than that from the responders (p<0.05).
Mx2 expression in IFN-α-treated γδ T cells from CHB patients showed significantly higher than that from HC We investigated changes in gene expression when γδT cells were directly treated with IFN-α. Some IFN-α-induced genes, Ifit1, Isg15, Mx1 and Mx2, were detected ( Fig. 4A-B). Among these genes, Isg15 and Mx1 expression in untreated-γδT cells from CHB patients was higher (but not significantly) than that from HCs. When compared to IFN-α-untreated γδT cells, expression of these four genes increased in the IFN-α-stimulated cells from HCs or CHB patients (data not shown). When compared to IFN-α-stimulated γδT cells from HCs, Mx2 gene expression from CHB patients showed a significant increase at 2.4-fold higher than that from HCs.
Elevated mRNA expression of Ifnγ, Tnfα, Il10, Il12 and Cxcl10 gene in γδ T cells from CHB patients after IFN-α treatment The expression of some cytokine or chemokine genes in γδ T cells was detected before and after IFN-α treatment (Fig. 4C-D). Before IFN-α stimulation, expression of Ifnγ, Tnfα, Il10 and Cxcl10 from CHB patients increased, but Il12a decreased compared to that of healthy controls. IFN-α enhanced expression of these genes in HCs and CHB patients, with a 2~4-fold increase in IFN-α-treated γδ T cells for Ifng, Tnfa, Il10 and Il12 expression, and about 300-fold increase for Cxcl10 expression. In comparison with IFN-α-stimulated γδT cells from HCs, Il10 and Cxcl10 gene expression from CHB patients displayed a significant increase compared with that from HCs.

Discussion
CHB is a serious disease, especially in Eastern Asian countries. The immune disorder in these patients is characterized as specific immunological inadequacy to HBV and nonspecific hyperimmune reaction to hepatocytes. As a result, persistent HBV infection can be accompanied by persistent liver inflammation [15,16]. To date, IFN-α is one of the most important anti-HBV treatments, whose mechanism includes a direct influence on HBV RNA, capsid, or covalently closed circular DNA stability in hepatocytes, and more importantly, it can exert an immunoregulatory function to improve CD4 + and CD8 + T cells, NK cells or other immune cells [6][7][8].
However, few studies have demonstrated the changes in γδT cells during IFN-α treatment.
γδT cells are one type of innate immune cells, and their functions include cytotoxicity, cytokine secretion, antigen presentation, and crosstalk with other immune cells. In the process of infection, autoimmune reaction, or tumor development, γδT cells are activated to stimulate, regulate, or even suppress the immune response [17][18][19]. γδT cells are mainly distributed in the epithelial tissues, such as the gut, skin, lung, or liver. The percentage of total γδT cells in peripheral blood is <10% of T cells. And the major (>70%) subtype of peripheral γδT cells is Vδ2T cells. However, in skin, gut or some other epithelial tissues, the subtype of Vδ1T cells constitutes the major population.
In the present study, dynamic changes in the percentage of γδT, Vδ1T or Vδ2T cells were detected. The percentage of total γδT cells decreased during treatment with IFN-in responders and non-responders. It was particularly interesting to find that all the percentages of Vδ1T cells during treatment in the responders were higher than in the non-responders. In our previous study, increased percentages of circulating Vδ1T cells and decreased Vδ2T cells were found in patients with CHB [13]. These data imply some important role of Vδ1T in anti-HBV immunity. Vδ1T cells were reported to recognize lipids presented by CD1c/d, MICA/B, or some stress-induced self-antigens. This differs from Vδ2T cells that uniquely respond to nonpeptide alkylphosphates [20,21]. Expansion of Vδ1T cells has been shown in infection with HIV-1 and cytomegalovirus, and expansion of Vδ2T cells occurs in infection with Cxcl10 was shown in C and D. All genes expression from peripheral γδ T cells was normalized to that of GAPDH and results of 5 CHB patients are presented with fold (Mean±SD) relative to the average of 6 HCs. The difference in the gene expression was significant when it was higher than 2-fold or lower than 0.5-fold.
doi:10.1371/journal.pone.0120086.g004 γδ T Cells and IFN-Alpha Therapy in CHB Patients mycobacterial species, Listeria monocytogenes, and Epstein-Barr virus. Certain antigens expressed on tissue cells might result in expansion of different subtypes of γδT cells [22,23].
According to the surface expression of CD45RA and CD27, human γδT cells can be divided into four subsets: naïve (CD45RA + CD27 + ), central memory (CD45RA -CD27 + ), effector memory (CD45RA -CD27 -), and effector cells (CD45RA + CD27 -) [24][25][26][27]. In our study, the circulating γδT cells were largely composed of naïve, effector memory, and effector cells. Few central memory γδT cells could be detected in the circulating γδT cells. Our results also showed a higher percentage of late-stage effector (CD45RA + CD27 -) γδT cells in the responders than non-responders. These results demonstrated that the γδT cells in the responders might have greater potential to respond to antigen and expand to the end stage of effector cells.
IFN-γ and TNF-α were the main cytokines produced in the activated γδT cells [28,29]. We found that the production of IFN-γ decreased during the first 12 weeks of treatment and then it gradually increased, and there was an increasing trend in TNF-α production throughout the treatment course. IFN-α could enhance RNA expression of IFN-γ gene in γδT cells in vitro. The results revealed that interaction of γδT cells with other immune cells could influence IFNγ production. It also implies that γδT cells are heterogeneous and different types of γδT cells have different patterns of cytokine production. So, the changing trend of IFN-γ production was not in accordance with that of TNF-α production.
CD107a (lysosomal-associated membrane protein-1) has been described as a functional marker of killer cell activity on CD8 + T, NK, γδT, or other immune cells [30]. During the first 12 weeks of IFN-α treatment, the percentage of CD107a + γδT cells increased dramatically from 10.1% (average) to 82.4% in the responders and 32.5% to 83.3% in the non-responders. Then, it decreased to 50.7% in the responders and 65.1% in the non-responders by the end of IFN-α treatment. Granzyme B is a caspase-like serine protease commonly found in the granules of cytotoxic lymphocytes or NK cells [31]. In our study, almost half of γδT cells could express Granzyme B from the beginning to the end of the treatment. Our results demonstrated that CD107a expression might be a sensitive marker for the cytotoxic activity of γδT cells when stimulated with IFN-α.
In addition, to observing the direct effect of IFN-α on γδT, the γδT cells were isolated from CHB patients and HCs for ex vivo experiments. Our results show IFN-α increased expression of all the observed genes. Isg15, Ifit1, Mx1 and Mx2 are among the best defined ISGs in immune cells [32]. We demonstrated that after IFN-α treatment, expression of Ifit1 and Mx2 in CHB patients increased very much and became higher than that in HCs, especially the significant increase for Mx2. Although MX1 and MX2 are closely related proteins, they exert different functions in different viral infections. Goujon et al. have shown that MX2 suppresses HIV-1 infection, but MX1 does not. In contrast, MX1 is an inhibitor of influenza A virus, but MX2 is ineffective [33]. For HBV infection, MX1 protein inhibits viral replication by interaction with hepatitis B core antigen [34]. Our results showed that the different pattern of Mx1 and Mx2 RNA expression in γδT cells from CHB patients before or after IFN-α stimulation.
γδT cells have a complicated cytokine profile. When activated, they secrete Th1 type cytokines (IFN-γ and TNF-α), or Th2 type cytokines [interleukin (IL)-4 and IL-10], and they produce IL-12 or other cytokines similar to antigen presenting cells. Also, chemokines are produced in γδT cells [35]. Our results showed that before IFN-α stimulation, expression of Ifng, Tnfa and Il10 increased, but the expression of Il12a and Cxcl10 decreased. Once stimulated with IFN-α, the Th1 and Th2 cytokines were all enhanced in CHB patients. These results suggested that the γδT cells from CHB patients were heterogeneous and had different functions. This study also showed that during PegIFN-α therapy, the function of γδ T cells was enhanced, and these cells might transfer into liver tissue and exert their effects. So it appeared that decreased percentage of γδ T cells and increased function could exist simultaneously.
In conclusion, we demonstrated that, although the proportion of total γδT cells decreased during IFN-α treatment, their functions of TNF-α production and CD107a expression were enhanced. The responders had more effector γδT cells than the non-responders had. In vitro, IFN-α boosted many ISGs in γδT cells. However, it was not the same in vivo, due to the complicated factors affecting the function of γδT cells.