Potential conflict of interest: J.L.H. is on the speakers' bureau of, and has received research support grants from, Bristol-Myers Squibb, Novartis, GlaxoSmithKline, and Roche. N.V.N. is an employee of Novartis.
This study was supported by grants from the Major Science and Technology Special Project of China (2012ZX10002-003 and 2011CB946100) and the Natural Science Foundation of China (30830091 and 81270025).
Address reprint requests to: Jinlin Hou, M.D., Department of Infectious Diseases and Hepatology Unit, Nanfang Hospital, Southern Medical University, No. 1838 North Guangzhou Avenue, Guangzhou 510515, China. E-mail: firstname.lastname@example.org; fax: +86-20-87719653; or Yu Zhang, Ph.D., Department of Immunology, Peking University Health Science Center, No. 38 Xue Yuan Road, Beijing 100191, China. E-mail: email@example.com; fax: +86-10-82801436.
Given the clinical significance of hepatitis B e antigen (HBeAg) seroconversion in chronic hepatitis B virus (HBV) infection, it is critical to elucidate the mechanisms regulating this process. In the present study, we found that the frequency of circulating chemokine (C-X-C motif) receptor 5 (CXCR5)+CD4+ T cells was higher in patients who had achieved HBeAg seroconversion in both cross-sectional (P < 0.001) and longitudinal (P = 0.009) studies. These cells were able to produce a significantly higher level of intracellular interleukin 21 (IL-21) after stimulation with HBV peptides in patients with telbivudine-induced HBeAg seroconversion (P = 0.007). Furthermore, sorted CXCR5+CD4+ T cells from HBeAg seroconverters boosted a higher frequency of antibody against hepatitis B e antigen (anti-HBe)-secreting B cells in coculture assay (P = 0.011). Of note, the increase in frequency of anti-HBe-secreting B cells was abrogated by soluble recombinant IL-21 receptor-Fc chimera (P = 0.027), whereas exogenous recombinant IL-21 enhanced this effect (P = 0.043). Additionally, circulating CXCR5+CD4+ T cells shared similar phenotypic markers, and were positively correlated in frequency with, splenic follicular T helper cells. Conclusion: Circulating CXCR5+CD4+ T cells, by producing IL-21, may have a significant role in facilitating HBeAg seroconversion in patients with chronic HBV infection. (Hepatology 2013;58:1277–1286)
Hepatitis B e antigen (HBeAg) seroconversion is defined as the loss of HBeAg from serum with the concomitant appearance of T-cell-dependent antibody (Ab) against HBeAg (anti-HBe). Both spontaneous and treatment-induced HBeAg seroconversions significantly reduce the risk of disease progression.[1, 2] Therefore, major treatment guidelines have adopted HBeAg seroconversion as a primary endpoint for antiviral therapy in patients with HBeAg-positive chronic hepatitis B (CHB).[1, 3] However, not all patients with CHB undergo HBeAg seroconversion while taking antiviral therapy, even though viral replication has been suppressed to a very low level for a long time. Given the clinical significance, but relatively low rate, of treatment-induced HBeAg seroconversion, it is critical to elucidate the mechanisms regulating this process.
Ab production by B cells against protein antigens is usually dependent on help from CD4+ T cells. There is a CD4+ T-cell subset in B-cell follicles, named follicular T helper (Tfh) cells, that is defined by expression of chemokine (C-X-C motif) receptor 5 (CXCR5), inducible costimulator (ICOS), programmed cell death 1 (PD-1), and high expression levels of interleukin (IL)-21 and B-cell lymphoma 6. This CD4+ T-cell subset is specialized for helping B cells to develop into Ab-producing cells in germinal centers. However, it is difficult to obtain lymphoid tissue from patients for research purposes. So, a surrogate strategy is required for studying Tfh-cell responses in humans. CXCR5+CD4+ T cells in peripheral blood may serve such a purpose. Although CXCR5 is transiently expressed in activated T cells, sustained expression is largely restricted to Tfh cells.[6, 7] Several studies have demonstrated that circulating CXCR5+CD4+ T cells shared some properties with Tfh cells.[8-10] Moreover, CXCR5+CD4+ T cells derived from both circulation and germinal centers potently induce Ab production during coculture with B cells in vitro.[7, 9, 11] In this regard, analysis of circulating CXCR5+CD4+ T cells may facilitate the investigation of Tfh cells.
To date, little is known about the role of circulating CXCR5+CD4+ T cells in patients with chronic hepatitis B virus (HBV) infection. In the present study, we investigated the phenotypes and functions of circulating CXCR5+CD4+ T cells in patients with chronic HBV infection and explored the relationship between circulating CXCR5+CD4+ T cells and HBeAg seroconversion.
Patients and Methods
One hundred and two patients with chronic HBV infection were recruited at Nanfang Hospital (Guangzhou, China) for the cross-sectional study. Patients were classified into immune tolerant carrier (IT; n = 20), HBeAg-positive CHB (n = 47), and inactive carrier (IC; n = 35) groups according to American Association for the Study of Liver Diseases guidelines. Thirty-eight healthy controls (HCs) were enrolled. Forty-two patients with HBeAg-positive CHB from Nanfang Hospital who participated in a clinical trial of telbivudine were studied longitudinally. Twenty milliliters of heparinized blood were collected at week 0, 12, 24, and 52 after starting telbivudine treatment. Subjects were classified into either a complete response (CR; n = 16) group, if they had undergone HBeAg seroconversion and achieved serum HBV DNA level less than 300 copies/mL by week 52, or a noncomplete response (NCR; n = 26) group, if serum HBV DNA was reduced, but HBeAg remained positive. All patients in both groups achieved normal alanine aminotransferase (ALT) levels by week 52. Fifty milliliters of heparinized blood were taken for in vitro studies from another 20 CHB patients enrolled in the same clinical trial after 52 weeks of telbivudine therapy. Patients were divided into CR (n = 10) and NCR (n = 10) groups according to the aforementioned criteria. In addition, spleen tissues and 5 mL of matched heparinized blood were obtained from 10 patients who underwent splenectomy resulting from HBV-related liver cirrhosis-induced hypersplenism.
Exclusion criteria for these studies were coinfection with hepatitis A virus, hepatitis C virus (HCV), hepatitis D virus, hepatitis E virus, and human immunodeficiency virus. Patients with primary biliary cirrhosis, primary hepatocellular carcinoma, and autoimmune diseases were also excluded. These studies were conducted according to Declaration of Helsinki guidelines and were approved by the ethical committee of Nanfang Hospital. Written informed consent was obtained from all subjects.
Serological Assays and HBV-DNA Assays
Serological assays and HBV DNA quantitation assays were performed as previously described. The lowest detection limit for HBV DNA is 300 copies/mL. The normal range for serum ALT level is 0-40 U/L.
Cell Cultures and Reagents
Cells were stimulated in vitro with the following reagents: phorbol-12-myristate-13-acetate (PMA; Sigma-Aldrich, St. Louis, MO); ionomycin (Sigma-Aldrich); brefeldin A (BFA; BD Biosciences, San Jose, CA); HBV peptides (a mixture of twenty-six 18-mer-peptides overlapped by 8 or 10 residues and covering the full C open reading frame of the HBV; Sigma-Aldrich); recombinant IL-2 (rIL-2; PeproTech, Rocky Hill, NJ); recombinant IL-21 receptor-Fc chimera (rIL-21R-Fc; R&D Systems, Minneapolis, MN); recombinant immunoglobulin (Ig)-G1 (rIg-G1; R&D Systems); recombinant IL-21 (rIL-21; PeproTech); recombinant HCV core protein (rHCV; American Research Products, Inc., Waltham, MA); pokeweed mitogen (Sigma-Aldrich), recombinant hepatitis B core antigen (rHBcAg; amino acids 1-183; ProSpec, East Brunswick, NJ); and recombinant HBeAg (rHBeAg; containing 10 precore residues at its N-terminus and 1-149 residues from the end of precore to its C-terminus; ProSpec). Given that the HBV precore and core regions have a low level of variability, the overall results of the study should not be significantly affected by the potential, but limited, mismatch between the genotype of the infecting viruses for individual patients and that of the detection reagents for immune analyses (Supporting Methods).
Peripheral blood mononuclear cells (PBMCs) were isolated and stored as previously described. Spleen tissues were mechanically dispersed and lymphocytes were isolated by Ficoll-Hypaque density gradient centrifugation. PBMCs and spleen-derived lymphocytes were stained with fluorescence Abs at room temperature for 20 minutes and analyzed on a BD FACSCanto II flow cytometer (BD Biosciences).
Intracellular Cytokine Staining
Intracellular cytokine staining after stimulation with PMA/ionomycin was performed as previously described.[12, 13] To determine the frequency of HBV-specific IL-21-producing CXCR5+CD4+ T cells, thawed PBMCs were cultured with or without HBV peptides (4 µg/mL) for 72 hours, and BFA (1 µg/mL) was added for the last 12 hours of culture. A response was considered positive if the percentage of IL-21-producing CXCR5+CD4+ T cells exceeded that of medium-only control (background) by 0.35% and was at least two-fold above the background (Supporting Methods).
In Vitro Ab Production and Enzyme-Linked Immunospot Assay
Circulating T cells (CXCR5+CD4+ or CXCR5−CD4+) and autologous B cells (CD19+) were sorted from either CR or NCR patients by a BD influx cell sorter (BD Biosciences). HBV-specific Ab production was assessed using the enzyme-linked immunospot (ELISPOT) assay, as previously described,[14-16] with modifications. Briefly, the purified T and B cells (105 cells of each/well) were cocultured in the presence of rIL-2 (10 ng/mL) and stimulated with either rHBeAg (5 μg/mL) or rHBcAg (5 μg/mL) for 5 days. Subsequently, supernatants were collected for measurement of IL-21 by enzyme-linked immunosorbent assay (ELISA), cells were harvested and transferred into 96-well nitrocellulose plates (Millipore, Billerica, MA) precoated with either rHBeAg (10 μg/mL), rHBcAg (10 μg/mL), or rHCV (10 μg/mL), and cultured in the presence of pokeweed mitogen (5 μg/mL) and rIL-2 (10 ng/mL) for another 48 hours. Subsequently, plates were sequentially incubated with biotinylated anti-human IgG and IgM (2 μg/mL;, Mabtech AB, Nacka Strand, Sweden), streptavidin-alkaline phosphatase (Mabtech AB), and 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium substrate (Invitrogen, Carlsbad, CA). Spots were counted by Immuno-SpotS4 Macro Analyzer (Cellular Technology, Inc., Santa Monica, CA). An ELISPOT response was considered positive if the number of spots in the HBV antigen-stimulated cultures exceeded that of the HCV antigen control by 5. This cut-off value was set as the mean plus 2 standard deviations of the number of spots in cultures without antigen stimulation. To test the role of IL-21 in Ab production, rIL-21R-Fc (10 μg/mL) or rIL-21 (50 ng/mL) was included in the culture in some cases as described.
The concentration of IL-21 was quantitated in duplicate wells using a commercial human IL-21 ELISA kit (Bender MedSystems GmbH, Vienna, Austria) in accord with the manufacturer's instructions.
Fresh PBMCs (1 × 106 cells/mL) were labeled with carboxyfluorescein succinimidyl ester (CFSE; 1.5 µM; Molecular Probes, Eugene, OR) and resuspended at 106 cells/mL in the medium; labeled cells were cultured with rHBeAg, rHBcAg, or rHCV (5 μg/mL) for 7 days or with medium only as a negative control. At the end of culture, cells were harvested and stained with α-CD4/allophycocyanin and α-CXCR5/PerCP-Cy5.5. The proliferation rate of CXCR5+CD4+ T cells was expressed as the percentage of cells that diluted CFSE intensity at least once at time of harvest.[18, 19]
Data are expressed as median (range). Mann-Whitney's U test, Wilcoxcon's signed-rank test, and the chi-square test were used when two groups were compared. Kruskal-Wallis's H test was used when more than two groups were compared. A receiver operating characteristic (ROC) curve was constructed to identify the optimal cut-off value for predicting HBeAg seroconversion to treatment. Correlations between variables were assessed with Spearman's rank-order correlation coefficient. All statistical analyses were based on a two-tailed hypothesis test with a significance level of P < 0.05.
Circulating CXCR5+CD4+ T Cells Is Expanded in Chronic HBV Infection
To find out whether chronic HBV infection could drive CXCR5+CD4+ T-cell differentiation, frequencies of circulating CXCR5+CD4+ T cells in CD4+ T cells were measured in patients with chronic HBV infection and HCs (Fig. 1A; Supporting Table 1). A significantly higher frequency of CXCR5+CD4+ T cells was observed in patients with chronic HBV infection, relative to the HC group (15.58 [6.61-28.87] versus 11.97 [7.63-16.62]%; P < 0.001; Fig. 1B). To illustrate the presence of HBV-specific cells in the overall increased CXCR5+CD4+ T-cell population in chronic HBV infection, we examined IL-21 production by these cells in response to HBV peptide stimulation. Although only a small fraction of CXCR5+CD4+ T cells could secrete IL-21, their representation was significantly higher in the chronic HBV infection group than that in the HC group (0.79 [0.00-3.42] versus 0.00 [0.00-0.38]%; P < 0.001; Fig. 1C). Likewise, a small, but definite, fraction of CXCR5+CD4+ T cells from chronic HBV infection patients proliferated in the presence of either rHBeAg or rHBcAg, relative to negative controls (P < 0.001; Fig. 1D). Consequently, both overall and HBV-specific CXCR5+CD4+ T cells were expanded in chronic HBV infection.
Higher Frequency of Circulating CXCR5+CD4+ T Cells Is Associated With HBeAg Seroconversion
Frequencies of CXCR5+CD4+ T cells were analyzed in both cross-sectional (Supporting Table 1) and longitudinal (Supporting Table 2) studies to investigate the association between circulating CXCR5+CD4+ T cells and HBeAg seroconversion. Cross-sectional data showed that the frequency of CXCR5+CD4+ T cells in the IC group who had achieved HBeAg seroconversion (20.92 [12.30-28.87]%) was significantly higher than the IT (11.58 [8.82-14.50]%; P < 0.001) and CHB (13.60 [6.61-23.43]%; P < 0.001) groups (Fig. 2A). Longitudinal data showed that the frequency of CXCR5+CD4+ T cells in the CR group was significantly higher than the NCR group (P = 0.009; Fig. 2B) during 52 weeks of telbivudine therapy. An increase in frequency of CXCR5+CD4+ T cells at week 12, relative to week 0, which was defined as “increasing pattern,” was found in the majority (14 of 16) of CR patients, but in only half (13 of 26) of NCR patients. The difference was statistically significant (P = 0.014; Fig. 2C). An ROC curve was generated, which demonstrated that the change in frequency of CXCR5+CD4+ T cells at week 12, relative to week 0, was predictive of HBeAg seroconversion at week 52 (P = 0.032; Fig. 2D). In addition, the change in frequency of CXCR5+CD4+ T cells between week 12 and week 0 was negatively correlated with the change in concentration of serum HBeAg between weeks 12 and 0 (r = −0.358; P = 0.020; Fig. 2E).
These results suggest that a high frequency of circulating CXCR5+CD4+ T cells is associated with HBeAg seroconversion in both cross-sectional and longitudinal study, and its dynamic changes during the first 12 weeks of antiviral treatment may provide a clue on HBeAg seroconversion.
Higher Frequency of HBV-Specific IL-21 Producing CXCR5+CD4+ T Cells Can Be Detected in HBeAg Seroconverters
Intracellular cytokine staining was performed to examine the profile of cytokine production by CXCR5+CD4+ T cells from patients with chronic HBV infection or HC subjects (Supporting Table 1) after stimulation with PMA/ionomycin or HBV peptides (Fig. 3A). Among CXCR5+CD4+ T cells, PMA/ionomycin stimulation generated more IL-21-producing cells (7.94 [5.95-12.85]%) than IL-17- (1.25 [0.33-6.50]%; P = 0.001), IL-4- (1.17 [0.52-3.12]%; P = 0.001), or IFN-γ-secreting cells (5.28 [2.31-7.96]%; P = 0.019). This is in contrast to the CXCR5−CD4+ T-cell population, which predominantly contained IFN-γ-secreting cells (Fig. 3B). In the cross-sectional study, there was a higher frequency of IL-21+CXCR5+CD4+ T cells in the IC than IT or CHB groups after stimulation with either PMA/ionomycin (Fig. 3C) or HBV peptides (Fig. 3D). More important, a significantly higher frequency of HBV-peptide-stimulated IL-21+CXCR5+CD4+T cells was detected in the CR group than NCR group after 24 (P = 0.005) or 52 weeks (P = 0.002) of antiviral treatment (Fig. 3E). Notably, such a difference appeared to be specific for IL-21-producing cells, because the IFN-γ+CXCR5+CD4+ T cells stayed at the same level between the CR and NCR groups over the course of antiviral therapy (data not shown).
Circulating CXCR5+CD4+ T Cells From Patients With HBeAg Seroconversion Promotes Anti-HBe Production by Autologous B Cells In Vitro in an IL-21-Dependent Manner
The capacity of CXCR5+CD4+ T cells to promote Ab production by autologous B cells in response to HBV-specific antigens was investigated by ELISPOT assay in both the CR and NCR groups (Supporting Table 3; Fig. 4A). Given that the frequency of HBV-specific Abs producing B cells was rather low, pokeweed mitogen was included in the final stage to boost Ab production after 5 days of incubation with HBV antigens alone. Data showed that coculture of autologous B cells with CXCR5+CD4+ T cells resulted in significantly higher frequencies of both anti-HBe-secreting and anti-HBc-secreting B cells than coculture with CXCR5−CD4+ T cells in most settings (Fig. 4B,C). Most remarkably, frequency of anti-HBe-secreting B cells in coculture of CXCR5+CD4+ T and B cells from CR patients was significantly higher than that from NCR patients (16.00 [0.00-28.00] versus 1.50[0.00-12.00] spot-forming units [SFU]/105 B cells; P = 0.011; Fig. 4B).
To verify whether IL-21 is involved in anti-HBe production, the concentration of IL-21 in the supernatant of the coculture was quantitated by ELISA. There was a significantly higher level of IL-21 in the coculture with CXCR5+CD4+ T cells than in the coculture with CXCR5−CD4+ T cells after stimulation with rHBeAg in both the CR (P = 0.007) and NCR (P = 0.013) groups (Fig. 4D). There was also a trend of elevated levels of IL-21 in coculture of CXCR5+CD4+ T and B cells from subjects in the CR relative to the NCR group (P = 0.075; Fig. 4D).
Further investigation showed that blockade of IL-21 activity in the coculture with soluble rIL-21R-Fc resulted in suppression of anti-HBe production (10.50 [8.00-20.00] versus 1.50 [0.00-7.00] SFU/105 B cells; P = 0.027; Fig. 4E). In contrast, addition of rIL-21 to the coculture led to an enhancement of anti-HBe production (3.50 [1.00-6.00] versus 7.00 [2.00-9.00] SFU/105 B cells; P = 0.043; Fig. 4E).
Collectively, these results suggest that the CXCR5+CD4+ T-cell population in HBeAg seroconverters may be more competent to support anti-HBe production by B cells, and IL-21 is the primary factor involved in this process.
Frequency of Circulating CXCR5+CD4+ T Cells Positively Correlates With Spleen-Derived ICOS+PD-1+CXCR5+CD4+ T Cells
CXCR5 is known to be highly expressed on Tfh cells. It would be interesting to explore how closely the circulating CXCR5+CD4+ T cells resemble Tfh cells present in lymphoid tissues. To this end, the expression of additional markers typically associated with Tfh cells were measured in circulating CXCR5+CD4+ and CXCR5−CD4+ T cells. Significantly higher percentages of ICOS-expressing, PD-1-expressing, and IL-21-expressing cells were detected in the CXCR5+CD4+ T-cell population, relative to the CXCR5−CD4+ T-cell population, in patients with CHB (P < 0.001; Fig. 5A). Next, the phenotypes of circulating and spleen-derived CXCR5+CD4+ T cells from patients who underwent splenectomy resulting from HBV-related liver cirrhosis-induced hypersplenism were directly compared. Though they both expressed high levels of CD45RO, circulating CXCR5+CD4+ T cells had a relatively lower expression of ICOS and PD-1. Moreover, the two populations showed differential expression of CCR7 and CD69, consistent with their different anatomical positioning (Fig. 5B). Nevertheless, the frequency of circulating CXCR5+CD4+ T cells was found to be positively correlated with that of splenic ICOS+PD-1+CXCR5+CD4+ T cells (r = 0.723; P = 0.018; Fig. 5C).
These findings suggest that circulating CXCR5+CD4+ T cells are closely related to Tfh cells of lymphoid organs; analysis of these cell subsets may be helpful to assess the situation of Tfh cells in patients with chronic HBV infection.
HBeAg seroconversion is an important milestone in the natural history of chronic HBV infection and has significant clinical implications for management of patients with HBeAg positive CHB during antiviral treatment. The immune mechanisms involved in HBeAg seroconversion are still not fully understood. The data in the present study support the hypothesis that CXCR5+CD4+ T cells contribute to this process. First, a high frequency of circulating CXCR5+CD4+ T cells was associated with HBeAg seroconversion in both cross-sectional and longitudinal investigations. Second, the change in HBeAg levels that occurred in the first 12 weeks of telbivudine treatment was negatively correlated with the change in frequency of CXCR5+CD4+ T cells. Third, CXCR5+CD4+ T cells from patients who achieved HBeAg seroconversion were able to promote anti-HBe-production by autologous B cells in ELISPOT assay.
Although expression of CXCR5 is a common, early event in CD4+ T-cell activation, sustained expression is largely restricted to Tfh cells.[6, 7] In fact, several studies have demonstrated that circulating CXCR5+CD4+ T cells shared some properties with Tfh cells.[8-10] The present study revealed a positive correlation between frequencies of circulating CXCR5+CD4+ T cells and spleen-derived Tfh cells. In addition, we showed that circulating CXCR5+CD4+ T cells had significantly higher levels of ICOS, PD-1, and IL-21 expression than CXCR5−CD4+ T cells, which is the characteristic of Tfh cells. More importantly, in comparison with CXCR5−CD4+ T cells, circulating CXCR5+CD4+ T cells exhibited increased potency to support the production of anti-HBe and anti-HBc during coculture with autologous B cells. These observations suggest that the circulating CXCR5+CD4+ T-cell population from chronic HBV infection subjects has the potential to induce HBV-related Ab production during HBeAg seroconversion and possesses a specialized function in supporting humoral immune response in humans.
The mechanisms by which circulating CXCR5+CD4+ T cells promote HBV-related Ab production remain to be defined. Our previous study showed that high serum IL-21 levels after 12 weeks of antiviral treatment independently predicted HBeAg seroconversion in patients with CHB. Furthermore, IL-21 has a role in the clearance of HBV antigens in a mouse model of human hepatitis B and could promote anti-HBe secretion by human B cells in vitro, which is in accord with the notion that IL-21 has a well-established role in B-cell proliferation and differentiation. In the present study, we demonstrated that increased frequencies of HBV-specific IL-21+CXCR5+CD4+ T cells in patients with HBeAg seroconversion. More significantly, addition of IL-21 to coculture of T and B cells markedly promotes anti-HBe production, whereas blockade of IL-21 showed an inhibiting effect. Taken together, these data support that IL-21 represents a major mediator for the function of CXCR5+CD4+ T cells in the induction and maintenance of HBeAg seroconversion.
HBeAg acts as an immunomodulatory protein during HBV infection. A high amount of HBeAg is believed to induce T-cell tolerance or hyporesponsiveness. In line with this, we found that the proliferative capacity of circulating CXCR5+CD4+ T cells was inversely related to the concentration of rHBeAg in vitro (data not shown). In addition, we demonstrated that the frequency of CXCR5+CD4+ T cells was negatively correlated with the concentration of serum HBeAg at week 12, relative to week 0, during antiviral treatment. These observations suggest that the level of HBeAg is closely related to the frequency and function of circulating CXCR5+CD4+ T cells, which might contribute to the different characteristics of this cell subset during the natural history of a chronic HBV infection or response to antiviral therapy for CHB.
In summary, the present findings suggest that high frequency of circulating CXCR5+CD4+ T cells may promote HBeAg seroconversion in patients with chronic HBV infection, and IL-21 produced by CXCR5+CD4+ T cells may represent an important mediator of this effect. Therapy that targets expansion of CXCR5+CD4+ T cells or IL-21 release may be beneficial for the treatment of chronic HBV infection.
The authors express their sincere thanks to Prof. Antonio Bertoletti from the Singapore Institute for Clinical Science and Prof. Xiaoning Xu from the China Novartis Vaccine Research Center for their critical comments.