Combination of pegylated interferon‐alpha and nucleos(t)ide analogue treatment enhances the activity of natural killer cells in nucleos(t)ide analogue experienced chronic hepatitis B patients

For NA‐experienced CHB patients who achieved low HBsAg levels, a combination of Peg‐IFN‐α and NA therapy is more likely to result in HBsAg decline and HBsAg clearance by increasing the activity of CD56briNK cells. And NA therapy may restore partial immune status in CHB patients through down‐regulating the expression of CCR4 and reducing the immunosuppressive effect of Treg cells.

Summary

A combination of pegylated interferon‐alpha (peg‐IFN‐α) and nucleos(t)ides analogue (NA) therapy can effectively reduce hepatitis B surface antigen (HBsAg), especially in NA‐experienced chronic hepatitis B (CHB) patients. However, the immune mechanism of this therapy is unclear. Forty NA‐experienced CHB patients were enrolled into this study. The frequencies of peripheral blood natural killer (NK) cells, dendritic cells (DCs), CD4⁺ T cells, CD8⁺ T cells, T helper (Th) cells, regulatory T cells (T_reg), B cells and follicular T helper (Tfh) cells were evaluated by flow cytometry. Seven of the 40 patients converted to peg‐IFN‐α combined with NA treatment, while the other 33 continued to NA therapy. The decrease in HBsAg was more pronounced in the combination treatment group, and only patients receiving combination treatment achieved HBsAg loss. The frequency and absolute number of CD56^bright NK cells in the combination treatment group increased significantly compared with the NA treatment group, whereas the CD56^dim NK cells were decreased. In the NA treatment group, the proportions of CD4⁺ T_N, CD8⁺ T_N, CD19⁺ B and cytotoxic T lymphocyte antigen‐4 (CTLA‐4)⁺CD4⁺ T cells were increased, while the proportions of CD4⁺ T_EM, CD8⁺ T_EM, CD25⁺CD4⁺ T_reg, CD25^highCD4⁺ T_reg, CD127^lowCD25⁺ T_reg, programmed cell death 1 (PD‐1)⁺CD4⁺ T, PD‐1⁺CD8⁺ T, CTLA‐4⁺CD8⁺ T, CCR4⁺CD25⁺ T_reg and CCR4⁺CD25^high T_reg cells were decreased after therapy. For NA‐experienced CHB patients who achieved low HBsAg levels, combination treatment is more likely to result in HBsAg decline and HBsAg clearance by increasing the activity of CD56^brightNK cells.

Introduction

Due to the prevalence of vaccination programmes and effective anti‐viral treatment, the high prevalence rate of chronic hepatitis B virus (HBV) infection is declining in some countries [1], but liver cirrhosis and hepatocellular carcinoma caused by chronic hepatitis B (CHB) remain some of the most serious global problems that threaten human health [2]. Nucleos(t)ide analogues (NA) and pegylated‐interferon‐alpha (peg‐IFN‐α) are widely used as first‐line anti‐viral drugs for patients with CHB [3]. Although NA can effectively inhibit HBV‐DNA replication, they have limited ability to scavenge hepatitis B surface antigen (HBsAg) (0–3% 1 year later) and require decades of continuous anti‐viral therapy [4, 5]. Peg‐IFN‐α has higher HBsAg clearance ability (3–7% 1 year later). However, due to the side effects of Peg‐IFN‐α, it is only suitable for a small population [6].

The immune system has a profound effect on chronic HBV infection and treatment of CHB. HBV evades immune system clearance by impairing the activity and function of immune cells, but Peg‐IFN‐α has the ability to inhibit viral replication and to activate immune responses [7]. In recent years, many studies have shown the potential HBsAg clearance and anti‐viral effects of Peg‐IFN‐α combined with NA, which revealed that combination treatments are superior to NA alone or Peg‐IFN‐α monotherapy [8, 9, 10, 11]. Studies have also confirmed that activity of natural killer (NK) cells is higher in CHB patients with HBsAg loss than in those without HBsAg loss after combined therapy [12, 13].

Our previous study showed that HBV‐DNA was inhibited and partial immune function had recovered in CHB patients after NA treatment: the frequency of CD56^bright NK cells in the peripheral blood was increased and the ratio of T helper 17 (Th17)/ regulatory T (T_reg) cells was recovered [14]. Based on the above experiment, the objective of our research was to determine whether Peg‐IFN‐α combined with NA treatment has a stronger immune effect than NA treatment in NA‐experienced CHB patients with low HBsAg levels. We also longitudinally discuss whether the immune function had recovered after long‐term NA anti‐viral therapy. Neither the first nor the second issue has been previously confirmed.

Material and methods

Patients

A total of 40 patients with CHB were recruited from the outpatient service of the Second Affiliated Hospital of Chongqing Medical University from July 2012 to February 2017. All patients with chronic HBV infection were confirmed to be positive for both HBsAg and detectable HBV virions for more than 6 months, and hepatitis B e antigen (HBeAg)‐positive or ‐negative before they took NA orally. Exclusion criteria were primarily the following: (1) patients who received other anti‐viral or immunomodulatory treatment before our study; (2) patients infected with other hepatitis viruses such as HAV, HCV and HDV as well as with HIV; (3) patients with autoimmune liver disease, alcoholic liver disease, decompensated cirrhosis or liver cancer; (4) those who met our research conditions but were followed‐up fewer than two times. Written informed consent was provided for all enrolled patients before our study, which conformed to the ethical guidelines of the Declaration of Helsinki and was approved by the Research Ethics Committee of the Second Affiliated Hospital of Chongqing Medical University, China.

The baseline was defined as when the serum HBV‐DNA was < 1000 copies/ml and serum HBsAg < 3000 IU/ml, with or without HBeAg loss and seroconversion, after receiving NA treatment. After reaching baseline, seven patients received a combination of subcutaneous injection of Peg‐IFN‐α (180 µg/week) and oral anti‐viral drugs (adefovir dipivoxil (ADV): 10 mg/day; lamivudine (LAM): 0·1 g/day; telbivudine (LdT): 600 mg/day; entecavir (ETV): 0·5 mg/day), and the other 33 patients continued with only oral anti‐viral drugs. Peg‐IFN‐α was replaced by oral anti‐viral drugs when serious adverse reactions occurred. An HBsAg level < 0·05 IU/ml was defined as HBsAg loss.

Sample collection

Peripheral blood samples were collected from each enrolled patient at each follow‐up time‐point after informed consent was provided, and the HBsAg levels, HBV‐DNA levels and biochemical indicators were detected in each sample.

Laboratory assessments

HBV‐DNA level was determined by a standard generic HBV‐DNA assay (ACON Biotech Co. Ltd, Hangzhou, China; dynamic range 1000 copies/ml to 1×10⁸ copies/ml). Serum levels of HBsAg, anti‐HBs, HBeAg and anti‐HBe were determined using a chemiluminescent microparticle immunoassay (Roche Diagnostics, Mannheim, Germany). Quantitative HBsAg and anti‐HBs levels were tested using a fully automated Architect instrument (Abbott Park, IL, USA), while the serum HBeAg and anti‐HBe concentrations were determined semiquantitatively. Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were assessed using an automatic biochemistry analyzer (Beckman LX‐20; Beckman, Brea, CA, USA). The haemoglobin (Hb) level and numbers of white blood cells (WBC), lymphocytes, neutrophilic granulocytes and platelets (PLT) were determined using a Sysmex XE‐5000. FibroScan (Echosens, Paris, France) was used to estimate the degree of liver fibrosis. The upper limit of normal value of ALT and AST was defined as 40 IU/l.

Flow cytometry analysis

Five millilitres of heparinized peripheral venous blood were collected at each time‐point. After the plasma was removed, the red blood cells were lysed using NH₄Cl lysis solution. In all, 10⁶ cells per tube were used for flow‐labelled antibody staining. The flow conjugated fluorescent antibodies used in our experiment are provided in the Supporting information, Table S1. Cells were classified according to our previous study [14]. Correction of non‐specific binding was performed using isotype‐matched antibodies. After staining for 30 min, the cells were washed with phosphate‐buffered saline containing 1% fetal bovine serum, and then centrifuged and resuspended in flow cytometry staining buffer. Flow cytometry analysis was performed using the FACS Canto II cell counter (Becton Dickinson, Franklin Lakes, NJ, USA) and FlowJo software version 6.1) (Tree Star, San Carlos, CA, USA).

Table 1.

Comparison of baseline characteristics between the two groups

Characteristics	Peg‐IFN‐α+ NA	NA	P‐value
Number	7	33
Male, n (%)	4 (57·1)	24 (72·7)	0·35
Age (years)	40·3 ± 5·9	44·1 ± 12·4	0·28
HBeAg‐positive, n (%)	2 (28·6)	8 (24·2)	0·57
Prior NA, n (%)
ADV	2 (28·6)	3 (9·1)
ETV	3 (42·8)	24 (72·7)
LAM	0	2 (6·1)
LdT	0	1 (3·0)
Combination	2 † (28·6)	3 ‡ (9·1)
Duration of prior NA treatment (years)
ADV	9·3 ± 3·0	5·0 ± 2·7	0·40
ETV	1·6 ± 0·3	0·8 ± 0·2	0·10
LAM		1·7 ± 0·4
LdT		0·5
Combination	2·1 ± 0·8	3·9 ± 2·0	0·60
HBV‐DNA (copies/ml)	Undetectable	Undetectable §
HBsAg (IU/ml)	718·8 ± 931·6	793·8 ± 663·6	0·31
Hb (g/l)	150·4 ± 19·6	156·2 ± 18·7	0·73
White blood cells (×109/l)	5·9 ± 1·7	6·6 ± 2·0	0·35
Lymphocyte (×109/l)	1·9 ± 0·5	1·8 ± 0·6	0·80
Neutrophil granulocyte (×109/l)	3·6 ± 1·5	4·2 ± 1·6	0·27
PLT (×109/l)	173·7 ± 46·1	167·8 ± 52·3	0·68
ALT (U/l)	22·8 ± 6·7	31·3 ± 28·5	0·95
AST (U/l)	24·7 ± 3·0	31·5 ± 18·6	0·63
FibroScan (Kpa)	5·2 ± 1·8	7·0 ± 2·5	0·17
One of the parents has CHB, n (%)	2 (28·6)	10 (30·3)	1·00
From urban areas, n (%)	4 (57·1)	18 (54·5)	1·00

Data are shown as mean ± standard deviation (s.d.) or proportions. Peg‐IFN‐α = pegylated‐interferon‐alpha; HBeAg = hepatitis B e antigen; HBsAg = hepatitis B surface antigen; NA = nucleos(t)ide analogue; ADV = adefovir dipivoxil; ETV = entecavir; LAM = lamivudine; LdT = telbivudine; Hb = hemoglobin; WBC = white blood cell; PLT = platelet; ALT = alanine aminotransferase; AST = aspartate aminotransferase; s.d. = standard deviation.

^†ADV was replaced by ETV in one patient; ADV was replaced with LdT in another patient.

^‡one patient was treated with ETV plus ADV; two patients were treated with LAM plus ADV.

^§Six patients had an HBV‐DNA load < 1000 copies/ml.

Statistical analysis

Statistical analysis was performed using spss 20.0 software (SPSS, Inc., Chicago, IL, USA) and GraphPad Prism version 7.0. Baseline characteristics were compared using non‐parametric statistical analysis for continuous variables and the c² test for categorical variables. Wilcoxon’s signed‐rank test was used for paired samples, and the Mann–Whitney U‐test was used for comparing the two groups. Analysis of the cumulative incidence of HBsAg loss was performed using the Gehan–Breslow–Wilcoxon test. Bilateral P‐values < 0·05 were considered statistically significant. Data are expressed as the median (range), ratio, absolute number, mean ± standard deviation (s.d.) or mean ± standard error of the mean (s.e.m.).

Results

Baseline patient characteristics

Of the seven combination treatment patients, add‐on treatment was used in five patients (one patient was treated with a combination of ADV and Peg‐IFN‐α; four patients were treated with a combination of ETV and Peg‐IFN‐α); two patients received sequential treatment (ADV and ETV were switched to Peg‐IFN‐α, respectively). The average time of Peg‐IFN‐α treatment for all patients was 36 weeks. Among the five patients who received add‐on treatment, one patient stopped NA therapy 1 month after the completion of Peg‐IFN‐α therapy, one patient discontinued NA therapy one month before the end of Peg‐IFN‐α treatment, one patient stopped NA and Peg‐IFN‐α simultaneously, and the remaining two patients continued NA treatment to present. No patients experienced viral rebound after stopping anti‐viral treatment. The other group of 33 patients was treated with NA. All patients had compensated liver disease. The basic clinical characteristics between the two groups are compared in Table 1. No significant differences were identified between the two groups (Table 1).

The combination treatment group was more capable of reducing and clearing HBsAg

HBsAg loss was considered the end‐point of clinical treatment for CHB. In our study, HBsAg loss occurred only in patients who received combination treatment, and three of seven patients had HBsAg loss. The changes in the HBsAg level from baseline to 120 weeks are shown in Fig. 1a. The decline in the mean HBsAg level in the combination treatment group was significantly more obvious than in the NA group from weeks 12 to 96 (P < 0·05). The cumulative estimate rate of HBsAg loss was 14·3, 28·5 and 42·8% for the combination treatment group at weeks 8, 36 and 48, respectively. The cumulative clearance rate of HBsAg was significantly higher in the combination treatment group than in the NA group (P < 0·001, Fig. 1b). HBV‐DNA was not detected in any patient from baseline to the end of the follow‐up in the combination treatment group. However, HBV‐DNA load fluctuated between 0 and 1000 copies/ml in the NA group (Supporting information, Fig. S1). Further analysis found that the absolute numbers of WBC, lymphocytes, neutrophilic granulocytes and PLT decreased significantly after Peg‐IFN‐α treatment in the combination treatment group (Fig. 1c–f).

Fig. 1.

Changes of hepatitis B surface antigen (HBsAg) in the combination treatment group and nucleos(t)ide analogue (NA) group during the entire treatment period. (a) Comparison of the degree of HBsAg decline in the combination treatment and NA groups. (b) Cumulative incidence of HBsAg loss between the combination treatment and NA groups. (c–f) Absolute number of white blood cells, lymphocytes, neutrophilic granulocytes and platelets in paired samples between baseline (BL) and the end of pegylated interferon‐alpha (peg‐IFN‐α) treatment (EOT) in the combination treatment group. Data are presented as mean ± standard error of the mean (s.e.m.). *P < 0·05, **P < 0·01, ***P < 0·001.

The frequency and absolute number of CD56^bright NK cells were clearly increased in the combination treatment group when compared with the NA group

As an important component of innate immunity, the proportions of circulation total NK (CD3⁻CD56⁺) and associated subpopulations, CD56^bright NK (CD3⁻CD56^bright) and CD56^dim NK (CD3⁻CD56^dim) cells, were analyzed (Supporting information, Fig. S2a). As shown in Supporting information, Fig. S2a,b, the proportion of CD56^bright NK cells was significantly increased while the proportion of CD56^dim NK cells was significantly decreased (P = 0·018) after Peg‐IFN‐α treatment in the combination group. By comparison of the two treatment groups, the combination treatment group had a significant increase in the frequency of CD56^bright NK cells and a significant decrease in the frequency of CD56^dim NK cells (week 24, P = 0·001; week 36, P = 0·017; week 48, P = 0·002) (Fig. 2c,d). However, longitudinal analysis of the NA group found no significant differences in the proportions of these three cell types (data not shown). As the blood cell counts mentioned above decreased after Peg‐IFN‐α treatment, the absolute numbers of NK cell subsets in our study were also analyzed. We found the absolute number of CD56^bright NK cells increased in the combination group after Peg‐IFN‐α treatment and was significantly higher in the combination than in the NA groups at week 48 (P = 0·007, Fig. 2e). The absolute number of CD56^dim NK cells, however, showed the opposite trend (Fig. 2f). These results indicate that combination therapy can enhance natural immune function and reduce HBsAg through increasing CD56^bright NK cells.

Fig. 2.

Analysis of natural killer (NK) cells in the combination therapy group (n = 7) and nucleos(t)ide analogue (NA) group (n = 33). (a,b) Comparison of the proportion of CD56^bright NK cells and CD56^dim NK cells in paired samples in the combination group between baseline (BL) and end of pegylated interferon‐alpha (peg‐IFN‐α) treatment (EOT), respectively. (c,d) Comparison of the proportion of CD56^bright NK cells and CD56^dim NK cells in the combination and NA groups during treatment, respectively. (e,f) Absolute number of CD56^bright NK and CD56^dim NK cells at baseline and week 48 between the two groups, respectively. Data are presented as mean ± standard error of the mean (s.e.m.). *P < 0·05, **P < 0·01, ***P < 0·001.

Dendritic cells (DCs) act as antigen‐presenting cells that activate adaptive immunity. In our study, the proportions of DCs, CD11c⁺ DCs and CD123⁺ DCs were detected by flow cytometry (Supporting information, Fig. S2b). No significant difference of DCs was seen between the two treatment groups (Supporting information, Fig. S2c). CD11c⁺ DCs showed no significant difference between the two treatment groups, despite the upward trend in the combination group (Supporting information, Fig. S2d). The frequency of CD123⁺ DCs was statistically higher in the NA group at week 24 (P = 0·022, Supporting information, Fig. S2e). Further analysis of longitudinal study revealed no significant differences in the three cell types in the NA group (data not shown).

Analysis of CD4⁺ and CD8⁺ T cell subsets in the two treatment groups

T cells have a non‐negligible status in the adaptive immune system, which primarily comprises CD4⁺ T cells and CD8⁺ T cells (Supporting information, Fig. S3a). We found no significant differences in any of the CD4⁺ T cell and CD8⁺ T cell subsets between the two treatment groups (Fig. 3b; Supporting information, Fig. S3b,e), except that the frequency of CD4⁺ central memory T cells (T_CM) in the combination group was higher than that in the NA group at week 48 (P = 0·026, Fig. 3a). In addition, longitudinal analysis of the NA group showed that the frequencies of CD4⁺ and CD8⁺ naive T lymphocyte cells (T_N) generally increased first and then decreased (Fig. 3c,d). Moreover, the proportion of CD4⁺ and CD8⁺ effecter memory T cells (T_EM) showed a general downward trend in the NA group (Fig. 3e,f). However, no obvious changes were seen in the effecter T cells (T_EFF) and T_CM subsets of CD4⁺ and CD8⁺ T cells in the NA group during treatment (data not shown).

Fig. 3.

Analysis of CD4⁺ and CD8⁺ T cell subsets in the combination group (n = 7) and nucleos(t)ide analogue (NA) group (n = 33). (a,b) Comparison of the frequencies of central memory and effector memory CD4⁺ T cells between the combination treatment group and NA group. (c–f) Longitudinal analysis of the changes in frequencies of naive CD4⁺, naive CD8⁺, effector memory CD4⁺ and effector memory CD8⁺ T cells in NA group. Data are presented as mean ± standard error of the mean (s.e.m.). #Compared with week 0 and P < 0·05; & = compared with week 4 and P < 0·05; *P < 0·05, **P < 0·01, ***P < 0·001.

Analysis of T_reg cell subsets in the two treatment groups

As a T cell subset with inhibitory function, T_reg cells play an essential role in maintaining immune homeostasis. The proportions of CD25⁺CD4⁺, CD25^highCD4⁺ and CD127^lowCD25⁺CD4⁺ T_reg cells were analyzed by flow cytometry in our study (Supporting information, Fig. S4). None of the cell subsets were significantly different between the two treatment groups, except that the proportion of CD127^lowCD25⁺CD4⁺ T_reg cells was significantly higher at week 24 in the combination group (P = 0·038, Fig. 4a–c). In a further longitudinal analysis of the NA group, the frequencies of all three cell subsets of T_reg cells increased transiently in the early treatment stage and decreased in the later treatment stage (Fig. 4d–f). The above results suggest that NA therapy reduced the immunosuppressive effect of T_reg cells.

Fig. 4.

Analysis of regulatory T cell (T_reg) subsets in the combination group (n = 7) and nucleos(t)ide analogue (NA) group (n = 33). (a–c) Comparison of the frequencies of CD25⁺, CD25^high and CD127^lowCD25⁺ T_reg cells between the combination group and NA group in sequence. (d–f) Longitudinal analysis of the changes in frequencies of CD25⁺, CD25^high and CD127^lowCD25⁺ T_reg cells in the NA group in sequence. Data are presented as mean ± standard error of the mean (s.e.m.). @ = Compared with week 0 and P < 0·05; α: compared with week 12 and P < 0·05; &: compared with week 24 and P < 0·05; #: compared with week 48 and P < 0·05; *P < 0·05, **P < 0·01, ***P < 0·001.

Analysis of Th cell subsets in the two treatment groups

Detection of Th cell subsets by flow cytometry revealed no significant differences in Th1, Th2, Th17, Th1/Th2, CD25⁺CD4⁺ T_reg/Th17 or CD127^lowCD25⁺CD4⁺ T_reg/Th17 cells between the two treatment groups (Fig. 5a–c; Supporting information, Fig. S5). However, longitudinal analysis of the NA group showed an increase in Th1 cells from baseline to week 120 (P = 0·048, Fig. 5d). We also found a slow decrease in Th17 cells and an irregular upward trend in CD127^lowCD25⁺CD4⁺ T_reg/Th17 cells throughout the NA treatment phase (Fig. 5e,f). However, no obvious trends were seen in Th2, Th1/Th2, CD25⁺CD4⁺ T_reg/Th17 or CD25^highCD4⁺T_reg/Th17 cells in the longitudinal analysis of the NA group (data not shown).

Fig. 5.

Analysis of T helper (Th) cell subsets in the combination group (n = 7) and nucleos(t)ide analogue (NA) group (n = 33). (a–c) Comparison of the frequencies of Th1, Th2 and Th17 cells between the combination group and NA group in sequence. (d–f) Longitudinal analysis of the changes in frequencies of Th1, Th17 cells and CD127^lowCD25⁺ regulatory T (T_reg)/Th17 cells in the NA group in sequence. Data are presented as mean ± standard error of the mean (s.e.m.). #Compared with 0 weeks and P < 0·05.

Analysis of follicular T helper (Tfh) cells and B cells in the two treatment groups

The interaction of Tfh cells and B cells is crucial for the humoral immune response. In our study, the frequencies of CXCR5⁺CD4⁺ Tfh, CD19⁺ B and CD27⁺CD38⁺CD19⁺ plasma cells showed a general upward trend and were slightly higher in the combination group; however, no obvious differences were found between the two groups (Fig. 6a–c,e). While the proportion of CD27⁺CD38^‐CD19⁺ memory B cells was lower in the combination than in the NA groups and showed statistical difference at week 12 (P = 0·038, Fig. 6d). In further longitudinal study of the NA group, except for the rising trend in CD19⁺ B cells from baseline to week 96 (P < 0·05, Fig. 6f), no obvious changes were observed in Tfh, plasma or memory B cells (data not shown). These data indicate that the humoral immune response was slightly improved after anti‐viral therapy.

Fig. 6.

Fluorescence‐activated cell sorting (FACS) analysis of follicular T helper (Tfh) and B cells in the combination (n = 7) and nucleos(t)ide analogue (NA) groups (n = 33). (a) Gating strategy of CXCR5⁺CD4⁺ Tfh cells and B cells. (b–e) Comparison of the frequencies of CXCR5⁺CD4⁺ Tfh, CD19⁺ B, CD38⁻CD27⁺CD19⁺ memory B and CD38⁺CD27⁺CD19⁺ plasma B cells between the combination and NA groups in sequence. (f) Longitudinal analysis of the changes in frequency of CD19⁺ B cells in the NA group. Data are presented as mean ± standard error of the mean (s.e.m.). #: compared with week 0 and P < 0·05; *P < 0·05.

Analysis of surface markers on immune cells in the two treatment groups

Tfh cells may help B cells to transition into antibody‐producing plasma cells and memory B cells by the binding of programmed cell death 1 (PD‐1), inducible co‐stimulator (ICOS) and CD40 ligand (CD40L) on Tfh cells to the ligands PD‐L1, ICOSL and CD40 on B cells. PD‐1, T cell immunoglobulin domain and mucin domain 3 (TIM3) and cytotoxic T lymphocyte‐associated antigen‐4 (CTLA‐4) on both CD4⁺ and CD8⁺ T cells have immunosuppressive functions. CCR4 acts as an activating molecule of T_reg cells, so we detected these surface molecules on relatively immune cells through flow cytometry (Supporting information, Fig. S6). After comparing the two treatment groups, no significant differences were found in any surface molecule on circulating immune cells (Supporting information, Fig. S7). Longitudinal analysis of the NA group found that both PD‐1⁺CD4⁺ and PD‐1⁺CD8⁺ T cells showed a declining trend from baseline to week 144 (P < 0·05, Fig. 7a,c). Additionally, CTLA‐4⁺CD4⁺ T cells showed a slow upward trend from baseline to week 120 (P < 0·05, Fig. 7b), while CTLA‐4⁺CD8⁺ T cells showed an opposite trend to that of CTLA‐4⁺CD4⁺ T cells from weeks 12 to 96 (P < 0·05, Fig. 7d). Furthermore, expression of CCR4 on CD25⁺ and CD25^highCD4⁺ T_reg cells showed a downward trend after a short‐term upward trend (Fig. 7e,f). No significant changes were observed in TIM3⁺CD4⁺ T cells, TIM3⁺CD8⁺ T cells or other cell surface molecules between the two treatment groups or in the longitudinal study of the NA group (data not shown). These results suggest that the immunosuppressive status of patients with CHB is ameliorated by NA treatment, while the increase in CTLA‐4⁺CD4⁺ T cells may be related to changes in the immune microenvironment, which is related to immune‐synaptic loss that inhibits proliferation of CTLA‐4.

Fig. 7.

Analysis of surface markers on CD4⁺ T, CD8⁺ T and regulatory T cells (T_reg) in the combination (n = 7) and nucleos(t)ide analogue (NA) groups (n = 33). (a,b) Longitudinal analysis of the expression changes of programmed cell death 1 (PD‐1) and cytotoxic T lymphocyte antigen‐4 (CTLA‐4) on CD4⁺ T cells in the NA group. (c,d) Longitudinal analysis of the expression changes of PD‐1 and CTLA‐4 on CD8⁺ T cells in the NA group. (e,f) Longitudinal analysis of the expression changes of CCR4 on CD25⁺ and CD25^high T_reg cells. Data are presented as mean ± standard error of the mean (s.e.m.). & = Compared with week 0 and P < 0·05; # = compared with week 12 and P < 0·05; π = compared with week 24 and P < 0·05; α = compared with week 48 and P < 0·05).

Discussion

Previous studies have compared the different efficacies of Peg‐IFN‐α combined with NA treatment and NA treatment alone in terms of virological and serological indicators, but each has adopted different experimental designs [8, 9]. Our study is based on CHB patients whose HBV‐DNA was inhibited and whose HBsAg levels were at a low level after treatment with NA. Because some immune functions are restored after NA treatment, such as those of CD8⁺ T cells, Peg‐IFN‐α can then be more effective in clearing virus from the blood and inside the liver [15]. Our results showed that HBV‐DNA in the combination treatment group was more stable and more difficult to rebound when compared with the NA group during the entire treatment period. Serum HBsAg has been shown to reflect the on‐treatment efficacy, and clearance of HBsAg can improve the long‐term prognosis [16, 17, 18]. It was also confirmed that Peg‐IFN‐α combined with NA therapy is currently the best treatment for reducing HBsAg level, especially in patients who received ‘switch to’ combination therapy [19]. Our results have confirmed that the decline of HBsAg level in patients receiving combination treatment was more pronounced than in those receiving NA treatment, which is consistent with the findings of a previous study [8]. More importantly, in our study, the rate of HBsAg loss was higher in the combination group than in the NA group in which HBsAg loss did not occur in any of the patients. Due to differences in sample size and baseline HBsAg levels, the rate of HBsAg loss was higher than that in another study [7]. Our research provides proof‐of‐concept that combination treatment has a stronger ability to reduce HBsAg level and results in a higher rate of HBsAg loss than NA treatment after HBV‐DNA, and HBsAg levels were reduced by NA anti‐viral treatment. However, it is still necessary to clarify which subgroup of patients is more likely to achieve HBsAg loss, thus providing a reliable basis for clinical practice.

It is also necessary to understand the advantages of combination therapy, such as immunological status. Previous studies have been devoted to the immunological changes in CHB patients who were treated with NA alone, Peg‐IFN‐α alone or combination treatment [7, 20, 21]. To explore the longitudinal changes in immune status between the combination treatment and the NA treatment groups, we examined innate and adaptive immune system‐associated immune cells and surface molecules. In patients with CHB, the cytotoxicity capacity of NK cells is retained but their ability to produce cytokines is reduced [22, 23]. In our study, the proportion of CD56^bright NK cells in the combination treatment group was significantly higher than that in the NA group, which was consistent with the experimental results of another study [13]; by contrast, the proportion of CD56^dim NK cells was significantly lower than that in the NA group. We detected a reduction in lymphocyte counts after Peg‐IFN‐α treatment; however, the absolute number of CD56^bright NK cells still increased, which indicates that a decreased lymphocyte count does not affect NK cell activity. For the abundant cytokine‐producing function of CD56^bright NK cells, combination treatment may help HBsAg clearance through restoring function of CD56^bright NK cells. We also analyzed the frequency of DCs subsets; no significant differences were found between the two groups or in the longitudinal analysis of the NA group. However, the frequency of CD11c⁺ DCs tends to increase in the combination treatment group, which may prompt us that the increased frequency of CD11c⁺ DCs may be beneficial in the reduction of HBsAg after combination treatment.

When infected with virus, the immune system promotes the proliferation and differentiation of CD4⁺ and CD8⁺ T lymphocytes to react to the infection [24]. While after persistent antigen stimulates, activated T lymphocytes are depleted and have low function [25]. In our study, no significant differences were observed in any subset of T cells between the two treatment groups. By analyzing the NA group, we found an increasing trend in the CD4⁺ and CD8⁺ T_N cells and a decreasing trend in the CD4⁺ and CD8⁺ T_EM cells. Our results indicate that depleted naive T cells are improved after anti‐viral therapy and that low levels of the T_EM cell phenotype may be associated with a reduction in HBV‐DNA stimulation after anti‐viral therapy.

According to surface molecules CD25, CCR6 and CXCR3, CD4⁺ T cells are divided into T_reg, Th1, Th2 and Th17 cell subsets [26], each of which has a different function and participates in immune regulation by secreting different cytokines [27]. A previous study has revealed an increasing frequency of T_reg and Th17 cells and corresponding cytokines in patients with CHB compared with healthy controls [28]. Other studies have found that effective anti‐viral treatment indirectly affects the immune system by down‐regulating the proportion of circulating CD127^lowCD25⁺ T_reg cells [28, 29]. We found no significant differences in T_reg and T_reg/Th17 cells between the combination treatment and NA groups, but the proportion of CD25⁺CD4⁺ T_reg, CD25^highCD4⁺ T_reg and CD127^lowCD25⁺CD4⁺ T_reg cells exhibited a significant downward trend in the NA group, which is consistent with the result of a previous study [30]. In addition, previous research has shown a decrease in the proportions of Th17 cells and an increase in the proportions of Th1 and Th2 cells after NA treatment [28], which is consistent with our longitudinal analysis of the NA group, except that we found no changes in Th2 cells. The findings above show that NA treatment may improve immune status by reducing the immunosuppressive effect of T_reg cells.

In terms of humoral immunity, previous study has showed that NA treatment did not significantly change the frequencies of Tfh and B cells in CHB patients with different immune status [14]. In our present study, except that memory B cells were higher in the NA group, no significant differences were found in the proportions of Tfh cells and B cells as well as their surface molecules in the combination treatment group when compared with the NA group. However, a slow increasing trend of Tfh cells was found in the combination treatment group and a slight upward trend of CD19⁺ B cells was found in both treatment groups. Our results indicate that humoral immunity improved after anti‐viral therapy, but it is still necessary to extend the follow‐up time and increase the sample size for further verification.

Regarding the expression of inhibitory surface molecules on immune cells, previous research has demonstrated that the expression of PD‐1 and CTLA‐4 on CD4⁺ T cells was increased in CHB patients when compared with healthy controls [31]. TIM3 is an important negative regulator in adaptive immunity, and it is also highly expressed in chronic viral infections [32]. In our study, no significant differences were found in any of the three inhibitory molecules between the two treatment groups, yet PD‐1 on both CD4⁺ and CD8⁺ T cells as well as CTLA‐4 on CD8⁺T cells in the NA group showed a downward trend, while CTLA‐4 on CD4⁺ T cells in the NA group showed an upward trend. These findings revealed that PD‐1 and CTLA‐4 expressed on T cells are associated with HBV‐DNA suppression and recovery of adaptive immune response. However, an increased frequency of CTLA‐4 on CD4⁺ T cells may be associated with a lack of receptors that block CTLA‐4 proliferation in immune synapses in patients with CHB. We also found that the expression of CCR4 on CD25⁺ and CD25^high T_reg cells decreased significantly during NA anti‐viral treatment. It was suggested that the decrease in T_reg cell immunosuppression is mainly caused by the decrease of CCR4 expression, but an accurate functional experiment requires the expression of forkhead box protein 3 (FoxP3) and transforming growth factor (TGF)‐β.

Our study has several limitations. First, the sample size of the study was small. Due to the small number of patients meeting the baseline criteria of this study and fewer patients were willing to receive Peg‐IFN‐α therapy, we were unable to easily expand the sample size. However, we obtained some interesting results from this small sample. Secondly, the gender of the patients in the two groups did not match well. However, the statistical analysis showed that the gender difference was not significant between the two groups. Therefore, whether the gender difference will affect the immune status of CHB patients after anti‐viral treatment needs to be further studied in future experiments. Despite these limitations, we obtained meaningful results through our careful clinical data analysis and experimental studies.

In conclusion, in NA‐experienced CHB patients with low levels of HBsAg and Peg‐IFN‐α combined with NA enhances immune activity primarily by increasing CD56^bright NK cell activity. By down‐regulating the expression of CCR4 and reducing the immunosuppressive effect of T_reg cells, NA therapy may restore partial immune status in CHB patients.

Disclosures

None.

Author contributions

P. H. and H. R. conceived and designed the study. X. P., L. Z., N. L., B. L., Z. C. and H. L.i conducted the experiments. X. P., L. Z., M. C. and M. P. analyzed the data. X. P., L. Z. and P. H. wrote the manuscript. All authors reviewed and approved the final version of the manuscript.

Supporting information

Fig. S1. Dynamic changes of HBV‐DNA load in the two groups. (a) Dynamic change of HBV‐DNA in the combination treatment group. (b) Dynamic change of HBV‐DNA in the NAs group.

Fig. S2. Fluorescence‐activated cell sorting (FACS) analysis of NK cells and DCs in the combination therapy group (n = 7) and NAs group (n = 33). (a) Gating strategy of NK cell subsets. (b) Gating strategy of DC cell subsets. (c‐e) Comparison of the proportions of DCs, CD11c⁺ DCs and CD123⁺ DCs in the combination group and NAs group during treatment in sequence. Data are presented as mean ± SEM. (*P < 0.05).

Fig. S3. Fluorescence‐activated cell sorting (FACS) analysis of CD4⁺ and CD8⁺ T cell subsets in the combination group (n = 7) and NAs group (n = 33). (a) Gating strategy of CD4⁺ and CD8⁺ T cell subsets. (b‐e) Comparison of the frequencies of naïve CD4⁺ T cells, naïve CD8⁺ T cells, effector CD4⁺ T cells, and central memory CD8⁺ T cells between the combination treatment group and NAs group during the treatment period. Data are presented as mean ± SEM. Abbreviations: T_N cell: naïve T lymphocyte; T_EM cell: effector memory T cell; T_CM cell: central memory T cell; T_EFF cell: effector T cell.

Fig. S4. Gating strategy in fluorescence‐activated cell sorting (FACS) of Treg cell subsets.

Fig. S5. Fluorescence‐activated cell sorting (FACS) analysis of Th cell subsets in the combination group (n = 7) and NAs group (n = 33). (a) Gating strategy of Th cell subsets. (b‐d) Comparison of Th1/Th2 cells, CD25⁺ Treg/Th17 cells, and CD127^lowCD25⁺ Treg/Th17 cells between the combination group and NAs group in sequence. Data are presented as mean ± SEM.

Fig. S6. Gating strategy in fluorescence‐activated cell sorting (FACS) of surface markers on CD19⁺ B cells, Tfh cells, CD4⁺ T cells, CD8⁺ T cells and Treg cells.

Fig. S7. Analysis of surface markers on Tfh cells, CD19⁺ B cells and CD4⁺ T cells in the combination group (n = 7) and NAs group (n = 33). (a‐f) Comparison of the frequencies of IL‐21R⁺CXCR5⁺CD4⁺Tcells, ICOSL⁺CD38^‐CD27⁺CD19⁺ memory B cells and CD40⁺CD38^‐CD27⁺CD19⁺ memory B cells, TIM3⁺CD4⁺T cells, CTLA‐4⁺CD4⁺T cells and CCR4⁺CD25^hiCD4⁺ Treg cells between the combination group and NAs group in sequence. Data are presented as mean ± SEM.

Table S1. Antibodies used in flow cytometry.