Introduction
Hepatitis B virus (HBV) is a hepatophilic virus that has infected approximately 2 billion people worldwide, of whom 3.5% have chronic and persistent infections, and are a major public health problem.1,2 Without effective treatment, chronic HBV infection can progress to cirrhosis or hepatocellular carcinoma accompanied by a high risk of liver failure.3,4 Currently, the main forms of antiviral therapy are nucleot(s)ide analogs (NAs) and pegylated interferon-α (PEGIFN-α). NAs are used widely for the treatment of chronic hepatitis B (CHB) because of their ease of use, good tolerability, and potent antiviral activity.5 However, long-term or even lifelong treatment is required because of the difficulty of obtaining durable immune control and high virological and clinical relapse rates after drug discontinuation. NA treatment can achieve the complete suppression of HBV viral replication, but a functional cure, i.e., HBV surface antigen (HBsAg) loss, is difficult to achieve with NA or IFN treatment alone. The incidence of hepatocellular carcinoma is four times higher in patients with virological suppression alone than in those with HBsAg loss.6
Thus, the functional cure rate for CHB needs to be improved. The combined use of PEGIFN-α and NAs, which have different mechanisms of action, has been found to improve the rate of HBsAg loss, but only to 9.1%.7 Thus, stronger immunomodulators need to be identified and applied to meet this goal.7
The HBV vaccine is often recommended as an alternative or complement to antiviral drugs for the treatment of CHB.8 In some patients with CHB, HBV vaccination alone effectively maintains alanine aminotransferase (ALT) normalization and HBV e antigen (HBeAg) serological conversion.9 In addition to preventing HBV infection by stimulating antibody production, HBV vaccines inhibit the replication of HBV DNA through a specific CD4+ T-cell-mediated immune response; furthermore, this approach is inexpensive and side effects are rare.10
The immunomodulator granulocyte-macrophage colony-stimulating factor (GM-CSF) is also commonly used clinically to treat CHB. It increases granulocyte production and promotes innate immunity, and may enhance vaccine effects. Compared with HBV vaccination alone, an appropriate combined vaccination and GM-CSF-based drug treatment regimen can increase antibody levels by 8–10 times, increase the immune memory response by up to 5–10 times, and double cytotoxic T lymphocyte action, suggesting that GM-CSF contributes to the reduction of HBsAg levels and promotion of HBsAg loss.11,12
Our group has developed an immunomodulatory/antiviral regimen consisting of IFN-α, adefovir plus GM-CSF, and HBV vaccination, whose application resulted in a 9.2% HBsAg loss rate in HBeAg-positive patients with CHB. This prospective multicenter randomized controlled study was conducted to evaluate the efficacy and safety of different combination regimens using the more efficient PEGIFN-α2b and tenofovir disoproxil fumarate (TDF) in patients with HBeAg-positive CHB that seroconverted after NA treatment, to contribute to the overall goal of identifying an optimal antiviral treatment regimen for CHB.
Methods
Study design and treatment
In this prospective multicenter randomized controlled study, patients were assigned randomly in equal numbers to three treatment groups for 48 weeks: TDF alone (control), PEGIFN-α2b + TDF, and PEGIFN-α2b + TDF + GM-CSF + HBV vaccine. All patients subsequently received TDF alone for an additional 24 weeks. The drug sources and doses were TDF (Brilliant Pharmaceutical B, Fuzhou, China), 300 mg orally once daily; PEGIFN-α2b (Pegberon; Amoytop Biotech, Xiamen, China), 180 µg injected subcutaneously once a week; GM-CSF (Topleucon; Amoytop Biotech), 75 µg injected subcutaneously on Wednesday, Thursday, and Friday of weeks 1, 4, 12, 24, 36, and 48, a total of 18 injections; and recombinant Chinese hamster ovary cell HBV vaccine containing 20 µg HBsAg (North China Pharmaceutical Company Ltd. Shijiazhuang, China), injected intramuscularly on Saturday in weeks 1, 4, 12, 24, 36, and 48, a total of six injections. The study was approved by the Ethics Committee of the First Affiliated Hospital, Zhejiang University School of Medicine (No. 2018-515).
Patients
The study was conducted at nine centers in China. The inclusion criteria were (1) age 18–65 years, (2) CHB and HBeAg positivity with maintenance of the HBV DNA level below the detection limit and HBeAg seroconversion for >1 year after NA therapy, and (3) provision of written informed consent. The exclusion criteria were (1) other viral infection (e.g., hepatitis A, C, D, or E virus or human immunodeficiency virus), (2) cirrhosis or Child-Pugh score ≥ 7 at the time of enrollment, (3) liver disease of another cause (e.g., autoimmune liver disease, alcoholic liver disease, non-alcoholic fatty liver disease, or drug-related hepatitis), (4) serum creatinine level higher than normal, (5) liver malignancy or alpha-fetoprotein level >100 ng/mL at the time of enrollment, (6) other malignancy, (7) history of important organ transplantation, (8) interferon (IFN) contraindication (e.g., autoimmune, endocrine, or psychiatric disease combined with a history of serious heart, brain, kidney, retina, or other important organ/tissue disease), and (9) allergy to IFN, NAs, GM-CSF, or the HBV vaccine.
Assessment
The enrolled patients were tested routinely to detect and quantify hepatitis B markers, HBV DNA, other viral hepatitis markers, routine blood laboratory and biochemistry parameters, autoimmune liver disease-related indicators, and thyroid function. Abdominal color Doppler ultrasound examinations were also performed. In the first and second weeks of the treatment period, routine blood and liver function tests were performed; in the fourth week, routine blood and liver function tests, hepatitis B marker detection, and HBV DNA quantification were performed. Thereafter, assessments were performed every 4 weeks until 48 weeks, and then every 12 weeks until 72 weeks. Routine blood and liver function tests were performed every 4 weeks, and hepatitis B marker detection, HBV DNA quantification, blood biochemistry, thyroid function testing, antinuclear antibody detection, alpha-fetoprotein measurement, and the assessment of other indicators were performed every 12 weeks.
Outcomes
The primary endpoints were the proportions of patients with HBsAg loss (<0.05 IU/mL) and HBsAg seroconversion (HBsAg level <0.05 IU/mL and HBV surface antibody level >10 IU/mL) at week 48. Secondary endpoints were HBsAg decline from baseline at weeks 12, 24, 36, 48, 60, and 72.
Sample size estimation
For sample size estimation, we assumed HBsAg loss rates of 5% in the control group and 20% in one or both of the other groups. With a type I error rate of 5%, 80% power, and 1:1:1 group allocation ratio, the expected absolute difference in risk between the PEGIFN-α groups and the control group was determined to be 15%. Considering a 15% dropout rate, we calculated that the final sample required was 291, with 97 per group.
Statistical analysis
Statistical analysis was performed using SPSS (version 26.0; IBM Corp., Armonk, NY, USA), and graphs were created using GraphPad Prism (version 9.0; GraphPad, San Diego, CA, USA). Categorical variables were reported as frequencies and percentages and were compared with chi-squared or Fisher’s exact tests. Continuous variables were compared between groups with the t-test and the Mann-Whitney U test for variables expressed as means and standard deviations or medians and interquartile ranges (IQRs). Kaplan-Meier survival analysis was used to estimate cumulative HBsAg negativity and serological conversion rates. Logistic regression analysis was used to analyze factors associated with HBsAg loss and seroconversion. Statistical significance was set at p<0.05.
Results
Patient characteristics
Of the 310 patients enrolled in this study and randomized to treatment groups, 23 were lost because of IFN intolerance and the 2019 coronavirus disease epidemic (Fig. 1). Data from the remaining 287 patients (control, n=100; PEGIFN-α2b + TDF, n=92; PEGIFN-α2b + TDF + GM-CSF + HBV vaccine, n=95) were included in the final analysis. Most (77.4%) patients were male, and the mean age was 38.68±8.91 years. The baseline characteristics of the patients are shown in Table 1. No significant differences in age, sex, HBsAg level, liver function, or routine blood parameters were observed among the groups. The main NA used in the past was entecavir (ETV), and previous antiviral treatment durations ranged from 17 to 81 months.
Table 1Baseline characteristics of patients with HBeAg-positive CHB and seroconversion after NA treatment
Variable | TDF, n=100 | PEGIFN-α2b + TDF, n=92 | PEGIFN-α2b + TDF + GM-CSF + HBV vaccine, n=95 | p-value |
---|
Age in years | 39.68±8.03 | 39.23±9.28 | 37.08±9.30 | 0.097 |
Sex, male | 76 (76.00%) | 71 (77.17%) | 75 (78.95%) | 0.885 |
BMI in kg/m2 | 23.00±2.76 | 23.38±3.36 | 22.76±3.14 | 0.390 |
HBsAg as log10IU/mL | 2.71±0.69 | 2.74±0.47 | 2.69±0.77 | 0.915 |
ALT in U/L | 23.75±9.43 | 24.00±8.75 | 23.80±8.40 | 0.980 |
AST in U/L | 24.17±7.01 | 24.01±8.24 | 23.43±7.66 | 0.780 |
GGT in U/L | 21.0 (13.0–28.0) | 20.0 (14.0–30.7) | 18.0 (15.0–28.0) | 0.657 |
TB in µmol/L | 14.79±6.57 | 15.04±8.34 | 13.25±5.49 | 0.155 |
ALB in g/L | 47.7 (46.0–50.0) | 48.2 (45.6–50.0) | 48.5 (46.8–50.0) | 0.294 |
WBC as ×109/L | 5.88±1.57 | 5.77±1.37 | 5.59±1.25 | 0.369 |
NEU as ×109/L | 3.48±1.23 | 3.53±1.39 | 3.33±1.01 | 0.508 |
RBC as ×109/L | 4.97±0.50 | 5.05±0.50 | 4.98±0.42 | 0.521 |
PLT as ×109/L | 208.67±54.08 | 194.48±55.31 | 198.90±50.17 | 0.469 |
History of previous NAs, n (%) | | |
Entecavir | 61 (61.00%) | 51 (55.43%) | 58 (61.05%) | 0.667 |
Adefovir | 25 (25.00%) | 19 (20.65%) | 13 (13.68%) | 0.137 |
Tenofovir | 31 (31.00%) | 36 (39.13%) | 39 (41.05%) | 0.302 |
Lamivudine | 14 (14.00%) | 16 (17.39%) | 17 (17.89%) | 0.726 |
Duration of NAs in months | 40 (20–65) | 48 (18–81) | 37 (17–64) | 0.446 |
HBsAg levels
In the control group, the mean HBsAg level decreased from 2.71±0.69 log10 IU/mL at baseline to 2.49±1.08 log10 IU/mL at 48 weeks and 2.44±1.06 log10 IU/mL at 72 weeks. In the PEGIFN-α2b + TDF group, the HBsAg levels at 0, 48, and 72 weeks were 2.74±0.47, 1.31±1.21, and 1.31±1.20 log10 IU/mL, respectively. In the four-drug group, these levels were 2.69±0.77, 0.93±1.12, and 1.09±1.19 log10 IU/mL, respectively (Fig. 2A). At weeks 12, 24, 36, 48, 60, and 72, the HBsAg level was higher in the control group than in the other two groups (all p<0.05; Fig. 2A). The decreases in the HBsAg level at 12, 24, 48, and 72 weeks in the control group (0.11±0.51, 0.19±0.59, 0.22±0.71, and 0.27±0.69 log10 IU/mL, respectively) were significantly lesser than those in the other two groups (all p<0.05), with no significant difference between the latter (Fig. 2B). During follow-up, the proportions of patients with HBsAg levels <3, <2, and <1 log10 IU/mL were significantly smaller in the control group than in the other two groups (all p<0.001), with no significant difference between the latter (Fig. 3).
HBsAg loss and serological conversion rates
At 48 weeks, the cumulative HBsAg negativity rates were 0.0% in the control group, 28.3% in the PEGIFN-α2b + TDF group, and 41.1% in the PEGIFN-α2b + TDF+ GM-CSF + HBV vaccine group (p<0.001; Fig. 4A). The cumulative HBsAg seroconversion rates at 48 weeks in these groups were 0.0%, 21.7%, and 33.9%, respectively (p<0.001; Fig. 4B). The rates at 72 weeks were similar (Fig. 4).
Safety
All three treatment regimens were generally tolerable, and no serious adverse events occurred (Table 2). Most events were related to PEGIFN-α2b, the most common being flu-like symptoms, including fever and malaise, followed by hair and weight loss. Neutropenia occurred in 80.43% (74/92) of patients in the PEGIFN-α2b + TDF group and 76.84% (73/95) of patients in the four-drug group. The proportions of patients with fever, malaise, hair loss, weight loss, neutropenia, thrombocytopenia, and abnormal thyroid function were significantly smaller in the control group than in the other two groups (all p<0.001), with no significant difference between the latter. All adverse reactions improved after PEGIFN-α2b discontinuation. No significant change in the fibrosis-4 (FIB-4) or aspartate aminotransferase to platelet ratio index (APRI) from baseline was observed in any group at 72 weeks (Table 3).
Table 2Adverse events occurring during follow-up
Variable | TDF, n=100 | PEGIFN-α2b + TDF, n=92 | PEGIFN-α2b + TDF + GM-CSF + HBV vaccine, n=95 | p-value |
---|
Fever | 0 (0.0%) | 47 (51.1%) | 51 (53.7%) | <0.001 |
Headache | 0 (0.0%) | 5 (5.4%) | 5 (5.3%) | 0.063 |
Fatigue | 4 (4.0%) | 46 (50.0%) | 48 (50.5%) | <0.001 |
Nausea | 0 (0.0%) | 3 (3.3%) | 2 (2.1%) | 0.214 |
Hair loss | 0 (0.0%) | 22 (23.9%) | 21 (22.1%) | <0.001 |
Weight loss | 0 (0.0%) | 12 (13.0%) | 14 (14.7%) | <0.001 |
Neutropenia | 1 (1.0%) | 74 (80.4%) | 73 (76.8%) | <0.001 |
Thrombocytopenia | 0 (0.0%) | 49 (53.3%) | 49 (51.6%) | <0.001 |
Thyroid dysfunction | 0 (0.0%) | 9 (9.8%) | 10 (10.5%) | <0.001 |
Table 3Changes in noninvasive fibrosis markers (FIB-4 index and APRI) at 48 and 72 weeks
Groups | TDF, n=100 | PEGIFN-α2b + TDF, n=92 | PEGIFN-α2b + TDF + GM-CSF + HBV vaccine, n=95 | p-valuea |
---|
FIB4 | | | | |
Baseline | 1.11±0.53 | 1.11±0.52 | 0.95±0.42 | 0.064 |
48 weeks | 1.06±0.47 | 2.71±2.25 | 1.82±1.06 | <0.001 |
72 weeks | 1.04±0.50 | 1.11±0.59 | 0.93±0.44 | 0.068 |
Change 72 weeks vs. baseline | −0.07±0.45 | 0.00±0.50 | −0.02±0.33 | 0.432 |
p-valueb | 0.128 | 0.991 | 0.570 | |
APRI | | | | |
Baseline | 0.34±0.15 | 0.35±0.18 | 0.32±0.15 | 0.431 |
48 weeks | 0.33±0.15 | 1.26±1.14 | 0.98±0.97 | <0.001 |
72 weeks | 0.32±0.17 | 0.37±0.20 | 0.36±0.20 | 0.163 |
Change 72 weeks vs. baseline | −0.02±0.17 | 0.03±0.20 | 0.04±0.18 | 0.059 |
p-valueb | 0.222 | 0.163 | 0.153 | |
Factors associated with HBsAg loss and serological conversion in patients treated with PEGIFN-α2b
Univariate and multivariate regression analyses showed that HBsAg loss was associated with baseline HBsAg levels <1,000 IU/mL [univariate p<0.05; odds ratio (OR) 0.47, 95% confidence interval (CI): 0.25–0.90, multivariate p=0.02)], peak ALT level in the first 12 weeks (OR 1.01, 95% CI: 1.00–1.01, both p=0.02), and GM-CSF and HBV vaccine use (univariate p=0.04; OR 2.17, 95% CI: 1.15–4.10, multivariate p=0.02). Age, sex, and BMI were not predictors of HBsAg negativity (Table 4). Similar results were obtained for HBsAg seroconversion, which was associated with peak ALT level in the first 12 weeks (OR 1.00, 95% CI: 1.00–1.01, p=0.05) and GM-CSF and HBV vaccine use (OR 2.07, 95% CI: 1.07–4.00, p=0.03) in multivariate analysis (Table 4).
Table 4Variables associated with HBsAg loss and seroconversion in patients receiving PEGIFN-α2b treatment
Variable | Univariate analysis
| Multivariate analysis
|
---|
OR (95% CI) | p-value | OR (95% CI) | p-value |
---|
HBsAg loss at 72 weeks | | | |
Age in years | 0.99 (0.95, 1.02) | 0.383 | | |
Sex, male | 0.79 (0.38, 1.66) | 0.534 | | |
BMI in kg/m2 | 1.01 (0.92, 1.10) | 0.888 | | |
Baseline HBsAg (<1,000 IU/mL vs. >1,000 IU/mL) | 0.54 (0.29, 1.00) | 0.049 | 0.47 (0.25, 0.90) | 0.022 |
Peak ALT in the first 12 weeks | 1.004 (1.000, 1.008) | 0.023 | 1.005 (1.001, 1.009) | 0.019 |
GM-CSF+HBV vaccine | 1.93 (1.05, 3.54) | 0.035 | 2.17 (1.15, 4.10) | 0.017 |
HBsAg seroconversion at 72 weeks | | |
Age in years | 0.98 (0.94, 1.01) | 0.198 | | |
Sex, male | 1.02 (0.48, 2.20) | 0.950 | | |
BMI in kg/m2 | 0.97 (0.88, 1.07) | 0.535 | | |
Baseline HBsAg (<1,000 IU/mL vs. >1,000 IU/mL) | 0.85 (0.45, 1.61) | 0.617 | | |
Peak ALT in the first 12 weeks | 1.003 (1.000, 1.007) | 0.058 | 1.004 (1.000, 1.007) | 0.050 |
GM-CSF+ HBV vaccine | 2.01 (1.05, 3.84) | 0.035 | 2.07 (1.07, 4.00) | 0.030 |
Discussion
NAs and PEGIFNs are the main antiviral therapeutic agents in this context, but neither acts directly on covalently closed circular DNA. The formation of this DNA in the nucleus is a fundamental step in the HBV lifecycle and provides a template for future virus generations.13 Due to the persistence of HBV covalently closed circular DNA, HBsAg loss is rarely achieved through spontaneous immune-mediated clearance or current therapies. NAs inhibit viral replication primarily by suppressing reverse transcription and have been used widely because they are easy to administer and well tolerated.14 However, they require long-term or even lifelong treatment because they do not exert immunomodulatory effects and due to low HBeAg seroconversion and negativity rates during treatment and high relapse rates after NA discontinuation.15,16 IFN has immunomodulatory and antiviral effects, and induces sustained HBeAg seroconversion and HBsAg negative conversion in some patients.17,18 The HBV vaccine has recently been recommended as a complement to antiviral therapy for patients with CHB. Its combination with the immune adjuvant GM-CSF can significantly enhance the HBV-specific host immune response, which has potential value for the reduction of HBsAg levels and promotion of HBsAg loss.8,12,19 A dose of GM-CSF as an adjuvant 24 h before HBV vaccine receipt significantly improved the serum conversion rate and serum protective antibody titer in individuals with poor serum conversion rates.11,20 Wang et al.12,21 reported that GM-CSF pretreatment once a day for 3 days before HBV vaccination eliminated HBsAg-positive hepatocytes compared with administration once or twice. National and international guidelines and several studies indicate that the addition of PEGIFN after NA use to suppress viral replication effectively reduces HBsAg levels and increases the rate of HBsAg clearance.22–24 This study included patients with CHB who were HBV DNA and HBeAg negative after NA treatment, and the currently recommended first-line antiviral drugs TDF and PEGIFN-α2b were used, with and without the HBV vaccine and adjuvant GM-CSF, to identify strategies and methods for the effective improvement of the functional CHB cure rate. At 72 weeks of follow-up, the cumulative HBsAg clearance and serological conversion rates were 28.3% and 21.7% in the PEGIFN-α2b + TDF group and 41.3% and 33.9% in the PEGIFN-α2b + TDF + GM-CSF + HBV vaccine group, and HBsAg clearance was not achieved in control group. Thus, the immunomodulatory/antiviral therapy improved the HBsAg loss rate significantly compared with NA monotherapy.
Reported HBsAg clearance and seroconversion rates in NA-experienced patients with CHB treated with PEGIFN-α2b and NAs for 48 weeks are as high as 50.93% and 48.15%, respectively,25 significantly higher than in the present study. A possible reason for the difference is a difference in the baseline HBsAg level; 71.30% of patients in the previous study had baseline HBsAg levels <500 IU/mL,26 and a low HBsAg level is a relevant factor for HBsAg negativity and serological conversion.27 Of patients with CHB treated with PEGIFN and NAs whose HBsAg levels were <1,500 IU/mL after treatment, 26.4% had HBsAg clearance and 18.7% had HBsAg serological conversion at week 48.22 We observed similar rates in the PEGIFN-α2b + TDF group in this study and higher rates in the four-drug group, which may be related to the enhancement of the HBV-specific immune response via the combined administration of the HBV vaccine and GM-CSF. Thus, the immunomodulatory/antiviral therapy was superior to NA monotherapy in terms of the reduction of the HBsAg level, with the addition of the HBV vaccine and GM-CSF further enhancing its efficacy.
Several studies have confirmed the association between low baseline HBsAg levels and HBsAg loss after treatment.26–28 In a clinical study conducted at Xi’an Jiaotong University, the baseline HBsAg and ALT levels during the first 12 weeks of treatment were predictors of HBsAg loss.22 In another study, the baseline HBsAg level and occurrence of 12-week ALT rebound were included in a simple scoring system that showed up to 0.78 and 0.81 efficacy for HBsAg clearance prediction in training and validation sets, respectively.29 The findings are consistent with the results of this study. The monitoring of changes in these indicators during follow-up and the timely adjustment of the dosing regimen and treatment course according to such changes are crucial.
The study has several limitations. First, as it was conducted with NA-treated patients with CHB and HBV DNA levels below the detection limit at baseline, HBV genotypes were not known. Thus, the effects of different HBV genotypes on the treatment response could not be examined. Second, as the observed trends of HBsAg decrease or even loss during treatment are not necessarily maintained in the long-term, future studies should be conducted with longer follow-up periods to better understand the changes in HBsAg kinetics and the maintenance of HBsAg clearance.
Conclusions
Among patients with CHB and undetectable HBV DNA levels and HBeAg seroconversion after NA treatment, especially those with high HBsAg levels, immunomodulatory/antiviral treatment regimens effectively reduced the HBsAg level and improved HBsAg clearance. The regimen including GM-CSF and HBV vaccine administration was most effective.
Abbreviations
- ALB:
albumin
- ALT:
alanine aminotransferase
- APRI:
aspartate aminotransferase to platelet ratio index
- AST:
aspartate aminotransferase
- BMI:
body mass index
- CHB:
chronic hepatitis B
- CI:
confidence interval
- CTL:
cytotoxic T lymphocyte
- FIB-4:
fibrosis-4
- GGT:
γ-glutamyl transpeptadase
- GM-CSF:
granulocyte-macrophage colony-stimulating factor
- HBeAg:
hepatitis B virus e antigen
- HBsAg:
hepatitis B virus surface antigen
- HBV:
hepatitis B virus
- IQR:
interquartile range
- NA:
nucleot(s)ide analog
- NEU:
neutrophil
- OR:
odds ratio
- PEGIFN-α:
pegylated interferon alpha
- PLT:
platelet
- RBC:
red blood cell
- TB:
total bilirubin
- TDF:
tenofovir disoproxil fumarate
- WBC:
white blood cell
Declarations
Acknowledgement
We thank all members of the China Chronic Hepatitis B clinical cure research group for their hard work and cooperation. We thank Medjaden Inc. for scientific editing of this manuscript.
Ethical statement
The study was approved by the Ethics Committee of the First Affiliated Hospital, Zhejiang University School of Medicine (No. 2018-515). All enrolled patients signed an informed consent form, and that the protocols conformed to the ethical guidelines of the latest version of the Declaration of Helsinki.
Data sharing statement
The datasets used to support the findings of this study are included within the article.
Funding
Ministry of science and technology of China (2017ZX10202202) and CAMS Innovation Fund for Medical Sciences (2019-I2M-5-045) and National Key R&D Program of China (2022YFC2304500).
Conflict of interest
The authors have no conflict of interests related to this publication.
Authors’ contributions
Study concept and design (YY, HJ), conduct of the literature search and writing of the manuscript (HJ, GY, JY), collection of patients’ samples and medical information (XZ, LY), data analysis and generation of the tables and figures (BW, JZ), execution of research (LB, XZ, KW, PZ, DY, YrZ, YY, YmZ, JG, CY, HC, YL, DX, LY, JL, JH, SZ and CJ), and obtained funding and critically revised the manuscript (YY, HJ). All authors have made a significant contribution to this study and have approved the final manuscript.