ABSTRACT
Objective Seroclearance of hepatitis B surface antigen (HBsAg) remains a challenge to the functional cure of HBV infection due to an HBV-specific immune tolerance that prevents virus eradication.
Methods In this clinical study, chronic hepatitis B patients treated with the standard nucleotide analog (NA) Adefovir Dipivoxil (ADV) or treated with ADV and IFN-α were immunized with granulocyte-macrophage colony-stimulating factor (GM-CSF) in combination with Human HBV Vaccine.
Results This THRIL-GM-Vac approach (THRee Injections of Low-dose GM-CSF followed by one dose of the HBV Vaccine) has achieved at least a 10-fold reduction of HBsAg level in 10 out of 23 (43.5%) patients tested. Moreover, two out of the ten patients eventually reached seroclearance. Thus THRIL-GM-Vac could break HBV-specific tolerance and boost HBsAg-specific cellular immunity. Consequently, the THRIL-GM-Vac approach could provide effective therapy for patients bearing medium-to-low levels of HBsAg.
Conclusion Our findings carry the clinical significance of showing a novel way to break immunotolerance that can tackle the persistent infection manifested as generated by chronic HBV burden in chronic hepatitis B (CHB) patients.
What is already known about the subject?Chronic hepatitis B virus (CHB) patients can be controlled with antiviral therapies while rarely cured or functionally cured. In addition to the standard anti-viral treatment, immunotherapeutic approaches may break the HBV established immunotolerance and, in turn, mount effective activation or reactivation of host anti-HBV immunity and can be achieved in HBV-related animal models. However, achieving such a functional cure in CHB patients has been challenging.
What are the new findings?
We demonstrated in this study that some levels of HBsAg seroclearance in CHB patients were treated with a 3-day treatment of GM-CSF followed by one HBV vaccine for six treatments within one year in addition to the standard of anti-viral treatment.
Notably, dynamics of HBsAg declined in responders were correlated with dynamics of HBsAg-specific Treg cells and reciprocally associated with the dynamical changes of HBsAg-specific IFN-γ +CD4+ (Th1) and IFN-γ +CD8+(Tc1) T cells over the course.
How might it impact on clinical practice in the foreseeable future?
It demonstrates in CHB patients that breaking Treg-immunotolerance can significantly be correlated with reducing the level of HBsAg in the ‘responders’ by immunotherapy.
Such correlation can be used to predict outcomes of the efficacy of an immunotherapeutic treatment prior to completing its treatments.
The evidence of HBsAg seroclearance in CHB patients by the presented immunotherapeutic approach will also shed light on improving such approach or similar strategies and further benefit those unmet medical needs.
INTRODUCTION
Loss of serum hepatitis B surface antigen (HBsAg) is the prevailing definition of a functional cure of HBV infection, as it provides the most sensitive marker indicating an efficient immune response against HBV genome expression[1, 2]. However, it is a challenge to achieve due to the occurrence of immunological tolerance during the viral persistence[3, 4].
Modulating regulatory mechanisms contribute to the establishment of immune tolerance, including T cell exhaustion and induction of suppressive regulatory cells [4-6]. Especially, Tregs dominate over effector T cells such as CD4+ T helper cells and CD8+ cytotoxic T cells and leave HBsAg expression uncontrolled in HBV infected hepatocytes [7, 8]. Therefore, breaking immune tolerance must be prioritized to eradicate serum HBsAg.
Granulocyte-macrophage colony-stimulating factor, GM-CSF, has long been considered an immune boosting agent and used as a hypothetical adjuvant for anti-viral vaccines, including HBV vaccines[9-11]. However, the reports of GM-CSF administration have shown inconsistent immunomodulatory effects, even including immune suppression in certain instances[12, 13]. In our previous studies, seeking to preferentially harness GM-CSF’s immune-boosting capacity, we scrutinized various immunization schedules and doses and developed an optimal protocol, designated as THRIL-GM-Vac (THRee Injections of Low-dose GM-CSF with one dose of Human HBV Vaccine). In a transgenic and an AAV-HBV-infected murine model, we demonstrated a high level of HBsAg clearance accompanied by anti-HBsAg antibodies by the use of THRIL-GM-Vac [14]. Furthermore, THRIL-GM-Vac consistently enhanced CD8 T cell-mediated immunity. There were particularly robust enhancements of HBsAg-specific IFN-γ-producing CD8 T cells, reductions of Treg cells, and tolerance was overcome [15, 16]. This study examined if such THRIL-GM-Vac treatment could also achieve a similar therapeutic benefit and HBsAg clearance in CHB patients.
RESULTS
Patients Characteristics
In total, 79 CHB patients were enrolled and randomized into 3 Groups: 26 patients in ADV (Group 1), 26 patients in ADV+IFN-α (Group 2), 27 patients in ADV+IFN-α+THRIL-GM-vac (Group 3); 26, 24, and 23 patients respectively completed the study treatment and follow-up period. The baseline characteristics of the patients are summarized in Table 1.
Enrollment for Each Group
Assessment of the HBsAg seroclearance under treatment with by the THRIL-GM-Vac in additions to ADV and IFN-α.
As the key parameter for a functional cure, serum HBsAg clearance during the treatment and follow-up period was monitored according to the scheme presented in Figure 1.
Groups 1 and 2 were treated under a standard of care (SOC) protocol for treatments, in which Group 1 patients were treated with an oral dose of adefovir NA (ADV) everyday oral dose, and Group 2 were treated with oral ADV plus IFNα injection (ADV+IFN-α) every day. Group 3 was a protocol added-on to the SOC with the above immunization schedule, ADV+IFN-α plus THRIL-GM-Vac treatment (ADV+IFN-α+ THRIL-GM-vac).
None of the patients in Group 1 attained HBsAg seroclearance. One patient (4.5%) in Group 2 who had a low HBsAg baseline showed seroclearance at week 24, and two patients (8.7%) in Group 3 showed seroclearance at week 24 and week 36, respectively, as shown in Figure 2a.
To better characterize the treatment effects on serum HBsAg, the dynamics of HBsAg levels were normalized by subtracting HBsAg levels at week 0 and shown in Figure 2b and 2c.
Patients with HBsAg levels decreased by up to 1-log10 and 2-log10 from week 0 baseline levels were identified. Treatment with ADV in Group 1 not only failed to cause any such instances, defined as HBsAg seroclearance. It also had little effect on the levels of antigen in general; a decrease of 1-log10 HBsAg level only occurred in one patient, and no patients had an HBsAg level decrease of 2-log10 or more (Fig. 2b, 2c). Notably, the combined treatment with IFN-α and ADV treatments (Group 2) was able to lower HBsAg levels; 36.4% (8 out of 22) patients showed a 1-log10 decrease, and 4.5% (1 out of 22) patients had a decrease of more than 2-log10 in HBsAg level. Not surprisingly, the most significant reduction of HBsAg was observed in the presence of THRIL-GM-Vac (Group 3), where 43.5% (10 out of 23) patients showed more than a 1-log10 decrease and 21.7% (5 out of 23) showed more than 2-log10 decrease of HBsAg level, including two patients who reached HBsAg seroclearance (Fig. 2b, 2c). Thus, these results confirmed that THRIL-GM-Vac treatment could increase the occurrence of HBsAg seroclearance when added to treatment with ADV and IFN-α.
THRIL-GM-Vac breaks the immune tolerance of HBsAg
Regulatory T cells are a crucial tolerance determinant for persistent HBV infection, and HBsAg is the key tolerance inducer. Previously, in HBV-transgenic and AAV-HBV-infected murine models, THRIL-GM-Vac was shown to generate an immune response profile in which Treg cells were down-regulated. Herein, HBsAg-specific CD4+FoxP3+ Treg cells were isolated from the patient’s PBMC and analyzed during the treatment and in the follow-up period. The ADV-only treatment (Group 1) had little impact on antigen-specific Treg cells (Fig. 3a), whereas, under ADV+IFN-α combined treatment (Group 2), the numbers of Treg cells decreased, but not to a statistically significant degree (Fig. 3b). Under THRIL-GM-Vac treatment (Group 3), antigen-specific Treg cells significantly decreased (Fig. 3d).
To better capture the consequential changes of HBsAg level after the three different treatments and their relationship to Foxp3+Treg changes, we first identified ‘responders’ and ‘non-responders’. Responders were defined as patients with reduced HBsAg levels (HBsAg decrease > 1-log10), and patients with maintained HBsAg levels (HBsAg decrease < 1-log10) were defined as ‘non-responders’. The correlation of their HBsAg levels with changes in antigen-specific Treg levels over time is shown in Figures 3b-3d. The decreases of HBsAg in ‘responders’ correlated with the decreases in Treg cells after treatment, suggesting that the loss of HBsAg may be a consequence of breaking immune tolerance. Much smaller decreases in Tregs were observed in non-responders compared to responders, consistent with the decrease in Tregs being related to the decrease in HBsAg. In addition, in Group 3, Treg cells from the non-responders also decreased. (Fig. 3d), indicating that THRIL-GM-Vac is potent enough to break the Treg-mediated tolerance in all patients, although additional factors may also contribute to the HBsAg clearance in this group.
The decreased anti-viral immunity due to HBV tolerance includes decreased activities of CD4+ helper T cells and CD8+ cytotoxic T cells [17]. We tested whether the present breaking of HBV tolerance and the associated Treg-decrease were reciprocally shadowed by increased levels of HBsAg epitope-specific CD4+ and CD8+ Teff cell activation. The effector-cell populations were defined as shown in Figure 4a. The IFN-γ response was significantly activated in Group 3, in both CD4+ T cells (Fig. 4b) and CD8+ T cells (Fig. 4c), supporting the notion of immune activation after the tolerance break.
The IFN-γ and antigen data from responders and non-responders are plotted side by side in Figures 4d to 4i to reveal relationships between HBsAg reduction and anti-viral cytokine dynamics. There was only one responder in the ADV-treatment group (Group 1), and this patient had more robust Th1 and Tc1 responses than the other patients in the group (Figs. 4d and 4g), suggesting that his/her anti-viral immune responses, rather than the anti-viral drug action, affected viral replication during the ADV treatments; perhaps the patient was a ‘spontaneous responder’ rather than a drug responder. In Group 2, responders showed more robust Th1 and Tc1 responses (Fig. 4e, 4h), but antigen clearance was not significantly different compared with the non-responders. Significant correlations between HBsAg decrease and robust Th1 and Tc1 responses were observed in Group 3, in which responders had up-regulation of Th1 even in the early period (week 4 to week 24), followed by sequential activation of Tc1 from week 12 through to the end of treatment at week 48. Meanwhile, HBsAg levels fell most markedly from week 4 until the end of treatments. These strong correlations under treatment with THRIL-GM-Vac suggested that anti-viral CMI, mediated by robust stimulation of both Th1 and Tc1 cells, was responsible for HBsAg clearance.
A major safety concern in CHB immunotherapeutic treatments is the potential risk of liver damage. Therefore, we also monitored dynamic changes of ALT in each patient throughout the study. As shown in Figure 5, ADV treatments did not affect ALT levels, reflecting its incompetence for immune activation. Some patients in the ADV+IFN-α group (Group 2) showed ALT increases to high but manageable levels after treatments, indicating mild liver damage. Group 3 showed nearly doubled ALT levels on average at week 24, and the levels declined afterward, suggesting that immunotherapy with THRIL-GM-Vac raised the ALT levels. This may reflect a beneficial effect of the targeted clearance of infected HBV hepatocytes by antigen-specific immune killing.
Taken together, these results suggest that efficient HBsAg clearance by adjunctive THRIL-GM-Vac treatment correlated with the breaking of tolerance through activation of antigen-specific cell-mediated immunity to kill infected cells and that this incurred limited liver damage.
THRIL-GM-Vac favored HBsAg clearance in patients with medium-to-low baseline antigen levels
Our clinical experiences have suggested that the baseline level of HBsAg may dictate the treatment efficacy. For example, patients who had reached low HBsAg baseline levels under NA treatments did not benefit from IFN-α treatment. This association was also observed in Group 2 patients treated with the ADV+IFN-α in this study. Therefore, we plotted the relationship of the baseline levels of HBsAg to the maximum values of HBsAg change, Treg change, Th1 change, and Th2 change during the treatments (Fig. 6).
As seen in Figures 6a to 6c, plotting the maximum changes in HBsAg levels against the baseline levels, the non-responders tend to cluster to the upper-left corner in each panel, consistent with an association between a higher antigen load and a lower reduction of load in response to treatments. The responders are spread towards the diagonal HBsAg clearance line (HBsAg < 0.07 IU/mL). The ADV+IFN-α group (Group 2; Fig. 6b) has only one patient with seroclearance, and the baseline level of this patient is moderately low at 0.14 IU/mL; the higher-baseline patients show limited antigen reduction responses. However, by adding the THRIL-GM-Vac into treatments (Group 3), two patients with medium-to-low baseline HBsAg (36.36 IU/mL and 29.32 IU/mL) reached seroclearance, and patients at higher baseline showed more HBsAg reduction towards the clearance line (Fig. 6c).
Higher circulating HBsAg level often triggers more severe immune exhaustion and impairments of anti-viral immune potential. Effective treatment has been shown to decrease Treg and up-regulated Th1 and Tc1 responses, especially in the medium-to-low baseline HBsAg patients (baseline level from 10 to 1000 IU/mL)[18]. To confirm that THRIL-GM-Vac could reverse immune impairment even in patients with higher HBsAg baselines, we plotted Treg reduction and anti-HBV cytokine increase against antigen baseline levels. As shown in Figures 6e-6k, the addition of IFN-α to ADV treatment (Group 2) slightly increased the proportion of patients with significant changes in immune responses compared to the ADV treatment alone. However, when the THRIL-GM-Vac was added to the ADV+IFN-α treatment regime, most patients in the Group 3 showed substantial improvements in cellular immune responses, irrespective of the magnitude of the antigen load at baseline(Fig. 6f-6l).
Taken together, these results demonstrate that an HBsAg seronegativity (cure) can be achieved in CHB patients by breaking immunotolerance through immunotherapeutic treatments with the THRIL-GM-Vac, which contains three treatments with GM-CSF followed by Human HBV Vaccine administration. Importantly, this novel THRIL-GM-Vac treatment was effective when added to a current standard procedure that uses anti-viral drug ADV and IFN-α treatments. The host cellular immune responses measured in this study supported the conclusion that down-regulation of tolerogenic Treg cells and robust enhancement of HBsAg-specific immune responses were probably the key mechanisms of enhanced therapeutic effect.
DISCUSSION AND CONCLUSION
Here we present our first efficacy report from a clinical investigation of a novel THRIL-GM-Vac treatment of HBV-infected patients in which the treatment was used as an adjunct to a standard NA and IFN-α therapy. This clinical study aimed to achieve HBV functional cure. Although we previously achieved nearly100% functional cure in murine models, this is the first study in CHB patients and we achieved substantial levels of HBsAg seroclearance with a strong correlation between HBsAg seroclearance and antigen-specific CMI. The main conclusions were as follows.
Firstly, major decreases in HBsAg were evident in the Group 3 patients who received THRIL-GM-Vac immunotherapy (Fig. 2). Breaking immunotolerance was evidenced by significant reductions of HBsAg specific Treg cells in the ‘responders’, and reciprocally by the significant enhancement of anti-HBV CMI, particularly in the responders, during the 48 weeks immunotherapy (Figs. 3 and 4). Although the antigen-specific CMI started to wane after week 48, the major mobilization of CMI could probably be extended with an approach using longer immunotherapeutic treatment and utilizing the ratio of Treg vs Teff as an efficacy biomarker.
Secondly, after the immunotherapeutic treatments were completed at week 48, HBsAg levels slowly rebounded even though Tregs stayed at low levels, (Figs. 2 and 3). This tendency was strongly associated with diminishing levels of both Th1 and Tc1 cells in responders (Figs. 3 and 4). The decline of CMI should be further examined in future trials with more immunization doses within one year or with extended duration to include multiple immunization courses. Meanwhile, the treatment may be optimized by developing new immunization schemes. For example, a sequential schedule with a new type of NA treatment may suppress viral replication further before the THRIL-GM-Vac treatments are initiated; this, in turn, may induce even higher levels of antigen-specific CD4+ Th1 cells and more efficient CD8+ Tc1 cells.
Third, we observed that some CHB patients responded to the THRIL-GM-Vac treatments strongly as “responders”, but some did not. This suggests that immune responses or immune sensitivity, to the treatments differed substantially between patients. Such differences could be pre-determined and an investigation of baseline immune profile could reveal surrogated marker(s) of patient suitability for this treatment approach. Such a respective study was carried out by using the patients from the study and findings are also submitted to this journal in a parallel report. The results indicated that the baseline immune profile is indeed an important parameter to predict the outcome and could be used to pre-screen so-called “sensitive” CHB patients for the treatments.
Fourth, our findings are consistent with the view that sequential waves of Th1 cells and Tc1 cells are essential for HBsAg reduction and clearance in responders. If further substantiated, this may contribute to the development of a generalized immune-surrogate endpoint development.
Last but not least, we found that the THRIL-GM-Vac treatment was particularly effective in reducing HBsAg levels in patients who had medium-to-low baseline levels. Through further protocol optimizations, a greater impact on patients with higher baseline levels may be achieved in the future.
MATERIALS AND METHODS
Drugs
Commercial products, Adefovir Dipivoxil Tablets at 10mg/tablet were manufactural produced under GMP compliant conditions and donated by Cisen Pharmaceutical Co., Ltd (Jinan, Shandong, China), Recombinant Human Interferon-α2b Injection at 5mIU/injection by Beijing Kawin Technology and human recombinant GM-CSF (Topleucon) at 75ul/dose were manufactural produced under GMP compliant conditions in E. Coli and donated by Amoytop Biotech (Xiamen, Fujiang, China). Human recombinant HBV vaccines 20ug/dose were manufactured and produced under GMP compliant conditions in the yeast system and donated to the study by Kantai Biologicals (Shenzhen, China).
Ethics statement
This clinical trial was registered in the Chinese Clinical Trial Registry (ChiCTR-TRC-13003254) as a multi-centered and randomized. However, we used frozen PBMCs of three cohorts patients from one of ten participating hospitals. This study was approved by the Research Ethics Committee of Huashan Hospital, Fudan University. Each participants in the study were enrolled in Huashan Hospital, Fudan University after given an informed written consent form and signed.
Patient enrollment and treatment schedules
All recruited CHB patients had been previously treated with anti-viral NA therapy and had reached a status of HBeAg-negative while remaining HBsAg-positive(HBsAg<5000IU/ml). Criteria for recruitment also included HBV DNA level (<100 copies/mL), normal blood alanine-aminotransferase test, and under consolidation therapy for more than one year. Patients were not enrolled for the study if they were pregnant or had concomitant human immunodeficiency virus infection, hepatitis C or D virus, primary biliary cholangitis, autoimmune hepatitis, Wilson’s disease, alcoholic or non-alcoholic steatohepatitis, or history of liver cirrhosis or hepatocellular carcinoma(HCC). Patients were randomly assigned into three groups: Group 1, adefovir NA treatment (ADV); Group 2, ADV plus IFN-α treatment (ADV+IFN-α); Group 3, ADV+IFN-α plus THRIL-GM-Vac treatment (ADV+IFN-α+THRIL-GM-vac).
ADV was administered orally every day (QD), IFN-α was administered as an intramuscular (i.m.) injection on alternate days (QOD), and THRIL-GM-Vac was administered as daily (QD) subcutaneous (s.c.) injections of GM-CSF for three consecutive days followed by an i.m. injection with Human HBV Vaccine at the same site on the fourth day on weeks 0, 4, 12, 24, 36, and 48, as shown in Figure 1
Biochemical, serological HBV analysis
Serum alanine aminotransferase activity (ALT) was determined by a fully automatic biochemical analyzer (Hitachi 7600, Tokyo, Japan). Serum HBsAg, HBsAb, and HBeAg were measured using enzyme immunoassay equipment (Cobas e411, Roche Diagnostics, Mannheim, Germany).
Stimulation of PBMC with HBsAg epitope in vitro
Peripheral blood mononuclear cells (PBMC) were isolated by Ficoll and resuspended in R10 with 10% FCS prior to stimulate with several HBsAg-specific epitope pools (peptide pools, 1 μg/mL, listed in supplementary table s1) as specific stimulators in the presence of anti-CD28 (100 ng/mL). Cells incubated with anti-CD3 (1 μg/mL) and anti-CD28 (100 ng/mL) were used as positive controls. The plates were incubated at 37°C in a humidified atmosphere with 5% CO2 for 72 h before the cells were labeled with specific monoclonal antibodies and subjected to flow cytometry analysis.
Flow cytometry detection
Fluorescent labeled antibodies used for flow cytometric analysis were: CD8 (SK1), CD4 (OKT4), FoxP3 (PCH101), IFN-γ (4S.B3), IL-2 (MQ1-17H12) from eBioscience, TNF-α (Mab11), IL-4 (MP4-25D2), IL-17 (BL168) from BioLegend. In cytokine detection, PMA (100 ng/mL) and ionomycin (1 mg/mL) were used as positive control stimulants. An LSRFortessa (BD Biosciences) was used for flow cytometry analyses.
Statistics
All results are presented as means ± Standard Error of Mean (SEM). For continuous variables, the Mann-Whitney U test was used for abnormal distribution, the two-tailed Student’s t-test for normal distribution, and homogeneity of variance. A P value of < 0.05 was deemed significant for all analyses.
Data Availability
All data produced in the present study are available upon reasonable request to the authors
AUTHOR CONTRIBUTIONS
BW, JMZ, HYJ, and LDY designed the clinical study and analyzed results. JMZ, HYJ, LDY managed the operation of trials. FFY, HC, LY managed patients screenings, samplings, recordings and clinical analysis. SG, GZ, WDZ, SRZ, XYZ, CFL, FY, XJ, XZ, and SJZ performed CMI experiments and analyzed data. SG, BW, JMZ wrote and edited the manuscript.
CONFLICT OF INTEREST
Authors declare that they have no competing interests.
ACKNOWLEDGMENTS
This work was partly supported by awards of the National Major Science and Technology Research Projects for the Control and Prevention of Major Infectious Diseases in China (2017ZX10202202, 2013ZX10002001, and 2013ZX10002001003) to Drs. JM Zhang, HY Jia, LD Yang, and B.Wang, and the National Natural Science Foundation of China (81871640, 82172255) to B. Wang. We thank Dr. Douglas Lowrie for his careful editing and proofreads and Dr. Yiwei Zhong at Shanghai Medical College of Fudan University for her technical assistance.