In the cross-sectional study, venous blood was drawn from 73 consecutive untreated CHB patients and 21 healthy controls (HC) at Nanfang Hospital (Guangzhou, China). According to the guidelines of the European Association for the Study of Liver Diseases (16), these CHB patients were divided into three groups of study subjects: immune-tolerant (IT) carriers (IT, n = 24), immune activation (IA, n = 27), and inactive carriers (IC, n = 22). The characteristics of all participants, with respect to demographic, biochemical, and virological features, are listed in Table 1.

Forty-one HBeAg-positive CHB subjects who had participated in a prospective clinical trial of telbivudine (600 mg/day, clinical trial number: NCT00962533) (17) and another 24 patients from a prospective clinical trial of Peg-IFN-α-2a (180 μg/week, clinical trial number: NCT01086085) (18) in Nanfang Hospital were also studied.

All patients provided written documentation of informed consent to enter the study. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Ethical Committee of Nanfang Hospital.

Serum HBsAg, HBeAg, anti-HBs, anti-HBe, and anti-HBc levels were quantitatively analyzed by the ARCHITECT i2000SR system (Abbott Ireland Diagnostics Division, Sligo, Ireland). Quantification of serum HBV-DNA was assayed by the COBAS TaqMan HBV Test (Roche Molecular Diagnostics, Pleasanton, CA, USA), which has a detection limit of 12 IU/ml or 69.84 copies/ml.

Peripheral blood mononuclear cells were separated on Ficoll-Histopaque (BD Biosciences, Shanghai, China) density gradients and routinely cryopreserved as previously described (19). Thawed PBMCs were stained with anti-α-CD14-APC, anti-α-CD19-PE-Cy7, anti-α-CD3-APC-Cy7, and anti-α-CD56-FITC antibody (BD Biosciences, San Jose, CA, USA) for phenotype staining, while staining with 7AAD-PerCp (BD Biosciences) excluded dead cells. The stained cells were used for cell subset sorting on a BD FACSAria III (BD Biosciences). CD56⁺ NK cells were sorted from fresh PBMCs isolated from HCs (n = 12) and CHB patients (n = 12) with CD56 MicroBeads and MACS separation columns (Miltenyi Biotech, Shanghai, China), according to the manufacturer’s protocol. This purification protocol resulted in >95% purity of the selected cells, as determined by flow cytometry analysis (FACS) using anti-α-CD3-APC-Cy7 and anti-α-CD56-FITC antibody.

Total cellular RNA was isolated and purified from PBMCs or cell subsets using the miRNeasy miRNA isolation kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The cDNA synthesis was performed with 100 ng total miRNA using the miScript Reverse Transcription Kit (Qiagen). The expression of miR-155 genes was determined by real-time RT-PCR, which was performed with miScript SYBR Green PCR Kit (Qiagen) by using commercially available qPCR primer (Qiagen) on a LightCycler 480 (Roche Diagnostics International, Rotkreuz, Switzerland) according to the manufacturer’s instructions. The expression levels of each gene in the PBMCs were presented as values normalized against 10⁶ copies of U6 small nuclear RNA (RNU6B) transcripts. The miRNA in serum was extracted by miRNeasy Serum/Plasma Kits (Qiagen) and was synthesized as above. In addition, C. elegans miR-39 was added to serum samples as an internal control according to manufacturer’s recommends. Then quantitative real-time RT-PCR analysis for each gene in serum was performed as it was for the PBMCs, and the results were presented as values normalized against 10⁶ copies of C. elegans miR-39 miRNA transcripts. It was noteworthy that only 48 patients had enough corresponding serum retained to accomplish serum miR-155 detection.

Natural killer cells among the PBMCs were identified by flow cytometry analysis as the CD3⁻/CD56⁺ specimens. Expression of several activating (NKG2D, NKp46) and inhibitory (NKG2A, KIR3DL1/s1, Tim3) NK receptors, as well as secondary signals required for IFN-γ production (CD16) and early activation molecules (CD69) were examined by labeling with monoclonal antibody as described above.

Fresh isolated NK cells were cultured at a density of 5 × 10⁵ cells/ml in RPMI 1640 (Life Technologies-Thermo Fisher Scientific Corp, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, Grand Island, New York, NY, USA), 2 mM l-glutamine (Invitrogen, Carlsbad, CA, USA) and antibiotics/penicillin–streptomycin (Life Technologies, Carlsbad, CA, USA). NK cells were stimulated with IL-12 (10 ng/ml), IL-15 (10 ng/ml), and IL-18 (100 ng/ml) for 24 h as previously described (20). After 18 h of stimulation, 10 µg/ml of brefeldin A (BD Biosciences) was added to each well and incubated for the remaining 6 h for intracellular staining, while anti-α-CD107a-PE was added at the same time. Then the cultured cells were stained with LIVE/DEAD Fixable Near-IR Dead Cell Stain kit (life Technologies) to exclude dead cells and then the remaining viable cells were exposed to anti-α-CD3-PerCp, anti-α-CD56-APC, anti-α-CD3-PE-Cy7, anti-α-CD56-PerCp, anti-α-CD16-FITC, and anti-α-IFN-γ-PE, anti-α-TNF-α-FITC anti-α-Granzyme A-PE, anti-α-Perforin-APC, anti-α-Granzyme B-APC, or isotype antibody (BD Biosciences) for phenotype or intracellular staining, while with Fix/Perm A/B (Invitrogen) fixed and broken membranes. The stained cells were analyzed on a BD Canto II flow cytometer (BD Biosciences, San Jose, CA, USA).

The hsa-miR-155-5p micrON™ miRNA mimic (miR-155 mimic) and miRNA mimic Ncontrol (miR-control) were synthesized by RiboBio (Guangzhou, China). CD56⁺ NK cells were sorted from fresh PBMCs isolated from IA patients (n = 6) with CD56 MicroBeads and MACS separation columns as described above. Then 5 × 10⁵ purified NK cells were resuspended in 0.5 ml of Gene Pulser electroporation buffer (Bio-Rad, CA, USA) with 100 nM miR-155 mimic or a miR-control, and were then transferred into Gene Pulser micropulser electroporation cuvettes (Bio-Rad). Transient transfection of the resuspended cells was performed using the Gene Pulser Xcell electroporation system (Bio-Rad), according to the manufacturer’s protocol. The transfected NK cells were immediately rescued after transfection in pre-warmed complete RPMI 1640 medium in 48-well culture plates, then incubated at 37°C and 5% CO₂ for 3 days before analysis. Then the cells were stimulated and stained as above.

The separated NK cells (effector, E) were seeded onto 96-well plates at ratios of 5:1 with 1 × 10⁴ K562 (target cells-1, T-1), or at ratios of 2:1 with 1 × 10⁴ HepG2.2.15 (T-2), and then incubated with IL-12 (10 ng/ml), IL-15 (10 ng/ml), and IL-18 (100 ng/ml) for 18 h. The cytolytic activity of the NK cells was determined using a Cytotoxicity LDH Assay kit-WST (Dojindo Molecular Technologies, Japan) according to the manufacturer’s instructions.

Continuous data were shown as medians (minimum-maximum). Two-group comparisons were evaluated by the Mann–Whitney U-test, Wilcoxon signed-rank test, Chi-squared test, and paired t-tests. The Kruskal–Wallis H test was used when comparing more than three groups. Repeated measures analysis was used to compare changes in miR-155 expression during treatment. Differences in continuous variables were evaluated by paired Student’s tests. The correlations between variables were analyzed using the Spearman’s rank order correlation coefficient. Receiver-operating characteristic (ROC) curves were constructed to predict a CR to telbivudine treatment. All of the tests were two-tailed and a p-value of <0.05 calculated by SPSS Statistics 21.0 and GraphPad Prism 5.0 was considered statistically significant.

For the first step, we examined the expression of miR-155 in PBMCs and serum samples from all CHB patients by real-time RT-PCR. The miR-155 levels were lower in PBMCs from CHB patients than from HCs (Figure 1A, p = 0.005). The median miR-155 levels in PBMCs were downregulated in both the IT and IC groups compared to the HC group (Figure 1A, HC vs IT, p < 0.001; HC vs IC, p = 0.01). Interestingly, the miR-155 expression was higher in IA patients with abnormal ALT levels (Figure 1A, IA vs IT, p = 0.01). At the same time, miR-155 serum levels during various phases of chronic HBV infection were similar to those occurring in PBMCs (Figure 1B, HC vs IT, p = 0.023; IA vs IT, p = 0.013). Although serum miR-155 expression in CHB patients had a tendency to be lower than in the HC group, these differences were not statistically significant (Figure 1B). No correlation was found between miR-155 expression in PBMCs and serum miR-155 levels (Figure S1A in Supplementary Material).

We then evaluated the association between ALT levels, HBV DNA, and miR-155 mRNA expression in PBMCs and serum of the subjects with chronic HBV infections. As shown in Figure 1C, Spearman analysis revealed that there was a positive correlation between miR-155 levels in PBMCs and ALT levels (r² = 0.121, p = 0.003), but no correlation was observed between miR-155 levels in PBMCs and HBV DNA loads (Figure S1B in Supplementary Material, p = 0.281). However, there was no correlation between miR-155 levels in serum and ALT (Figure 1D, p = 0.269). As TLRs had been reported to mediate miR-155 expression and function during HBV infection (14), we also examined the relationship between miR-155 and TLR2 and found there was a high positive correlation between them (Figure S1C in Supplementary Material, r² = 0.305, p < 0.001).

In contrast to miR-155, miR-146a is a negative modulator of innate and adaptive immunity and its upregulation in CHB causes impaired T cell function, which may contribute to immune defects and immunopathogenesis that develops during chronic HBV infection (21). We also examined the miR-146a expression in PBMCs and serum samples of CHB. Although miR-146a and miR-155 had a positive correlation in PBMCs (Figure S2A in Supplementary Material), there was no significant difference of miR-146a expression in PBMCs and serum from all CHB patients or patients with different phases of chronic HBV infection (Figures S2B,C in Supplementary Material).

Chronic HBV infection is a complicated process, especially in the IA phase. Abnormal ALT levels generally reflect the inflammation activity accompanied with immune activation in these patients. According to EASL guideline, under normal circumstances, only IA patients, but not IT and IC, need to be treated. However, impaired immune function in IA patients could not eliminate HBV completely, which result in HBV persistence. To further determine the effect of immune dysfunction of IA patients on miR-155 expression, we measured miR-155 levels by real-time RT-PCR in subpopulation of PBMCs from HBV-infected subjects. Compared with non-infected HCs, although the miR-155 levels of IA patients were comparable in whole PBMCs, their expression still were significantly decreased in CD56⁺ NK cells (p = 0.041), but not in CD3⁺ T cells, CD14⁺ monocytes, or CD19⁺ B cells (Figure 2A). As mentioned above, SOCS1 and SHIP1 are two direct targets of miR-155, validated in many cell types. We next examined the expression of these two genes in NK cells from HC and IA. The mRNA level of SOCS1 was much higher in IA patients than in HCs (Figure 2B, p = 0.004), while the mRNA levels of SHIP1 were not significantly different (Figure 2C). SOCS1 and SHIP1 were reported to control IFN signaling and cytokine production in NK cells. Consistently, NK cells isolated from IA patients displayed decreased IFN-γ (Figure 2D, p = 0.027) and relatively ordinary TNF-α (Figure 2E, p = 0.343) secretion compared with NK cells from HCs after stimulation with IL-12, IL-15, and IL-18. These results suggested that reduced expression of miR-155 in NK cells might contribute to lower cytokine production in activated NK cells. As defects in activation and antiviral function of NK cells having been described in patients with chronic HBV infection (22, 23), we also determined several activating (NKG2D, NKp46) and inhibitory (NKG2A, KIR3DL1/s1, Tim3) NKRs, as well as secondary signal required for IFN-γ production (CD16), early activation molecules (CD69), and cytotoxicity factors (Perforin, CD107a, Granzyme A, Granzyme B) in NK cells between IA patients and HCs. We found that there were similar lower perforin secretion (Figures S3A,B in Supplementary Material, p = 0.009) and NKG2D expression (Figures S3A,F,G in Supplementary Material, %, p = 0.002, MFI, p = 0.001) of NK cells in IA patients than HCs, while the secretion of CD107a, Granzyme A, and Granzyme B (Figures S3A,C–E in Supplementary Material) and the expression of NKG2A, KIR3DL1/s1, Tim3, CD16, CD69 had no significant differences in IA patients and HCs (data not shown).

To confirm the role of miR-155 in cytokine production in NK cells after activation, we examined the IFN-γ and TNF-α production after miR-155 overexpression in NK cells from IA patients. miR-155 mimics or miR-controls were transfected into NK cells by electronic transduction and about a threefold increase in the miR-155 level was detected by real-time RT-PCR analysis. Compared to miR-controls, electronic transduction of NK cells with the miR-155 mimics resulted in an increase of IFN-γ⁺ and TNF-α⁺ NK cells after stimulation with IL-12, IL-15, and IL-18 (Figure 3A). The frequency of IFN-γ⁺ (p = 0.033) and TNF-α⁺ (p = 0.025) NK cells were significantly higher in the miR-155 mimic group than in the miR-control group (Figures 3B,C). Moreover, miR-155 mimics transfected NK cells displayed a decrease in target gene SOCS1 expression (Figure 3D, p < 0.001). Meanwhile, to further analysis the effect of miR-155 in NK cells activity and cytotoxicity, we also determined the aforementioned factors. Regrettably, miR-155 upregulation would not increase the secretion of perforin (Figures S4A,C in Supplementary Material) and expression of NKG2D (Figures S4G–I in Supplementary Material) or other factors (Figures S4A,D–F in Supplementary Material) and expression of NKG2D, but only elevated the expression of CD69 after IL-12, IL-15, and IL-18 stimulation (Figures S4A,B in Supplementary Material). Consistent with the results above, overexpression of miR-155 did not promote the cytolytic activity of NK cells on K562 or HepG2.2.15 target cells (Figures S4J,K in Supplementary Material). These results suggested that miR-155 might act as a positive regulatory molecule for NK cell-mediated cytokine production, but might not have an effect on the cytotoxicity of NK cells.

Considering that miR-155 may be beneficial for cytokine-mediated antiviral function, we investigated whether miR-155 expression was associated with the antiviral response to nucleoside analog therapy. The miR-155 levels in PBMCs were examined in 41 treatment naïve HBeAg-positive IA patients who received telbivudine treatment. They were divided into complete response (CR, n = 18) and non-complete response (NCR, n = 23) groups according to their treatment outcomes at week 52. Patients in the CR group had a normalized ALT level and HBeAg seroconversion and achieved a reduction of the serum HBV DNA level to <300 copies/ml at week 52, while those in the NCR group had either a serum HBV DNA level >300 copies/ml or were positive for HBeAg at week 52 (24). The baseline clinical features of CR and NCR patients were comparable, and on-treatment HBV-DNA, HBsAg, and HBeAg were significantly lower in CR than NCR at week 12, week 24, and week 52 (Table 2). Interestingly, miR-155 expression in PBMCs from the CR group at the initiation of treatment was significantly higher than in those from the NCR group (Figure 4A, p = 0.007). A ROC curve was generated to assess the usefulness of baseline miR-155 levels to predict a CR at week 52. The optimal cut-off value for the miR-155 expression in PBMCs was 5,135.35/10⁶ RUN6B copies. This indicated that sensitivity for detection of a CR was 72.2% with a specificity of 69.6% (Figure 4B). These results suggested that a higher level of miR-155 expression at baseline might be conducive to the immune control of HBV during nucleoside analog treatment. In the meantime, we also examined the expression of miR-146a in PBMCs between CR and NCR groups at baseline, unfortunately, there was no any differences between the two groups (Figure S5A in Supplementary Material).

We also examined the dynamic change of miR-155 expression during telbivudine therapy. Available PBMCs from 20 IA patients (including 10 CR and 10 NCR) were followed prospectively for analysis and exhibited a significant reduction of miR-155 expression at week 12 (p = 0.032) and week 52 (p = 0.017, Figure 4C). As analyzed separately, miR-155 expression in NCR patients showed a steady trend with no obvious decrease (Figure 4D). However, CR patients displayed reduced miR-155 expression but reached statistically significant differences only at week 52 (Figure 4E, p = 0.003). Moreover, we observed no significant difference of miR-146 expression after the start of therapy between the CR and NCR groups (data not shown).

Additionally, we tested the baseline and dynamic changes of miR-155 expression in PBMCs of IA patients during peg-IFN-α-2a therapy. A total of 24 treatment naïve IA patients who received 48 weeks of peg-IFN-α-2a treatment were divided into a sustained virological response (SVR, n = 7) group and a non-sustained virological response (NSVR, n = 17) group according to their treatment responses at week 72. Subjects in the SVR group had undergone HBeAg seroconversion and had achieved a serum HBV DNA levels <1,000 copies/ml, while NSVR group had either serum HBV DNA levels >1,000 copies/ml or were positive for HBeAg. The baseline clinical features of SVR and NSVR patients were comparable, and HBV DNA and HBsAg were significantly lower in the SVR group than the NSVR group at weeks 48 and 72 (Table 3). However, unlike in the telbivudine treatment groups, there were no significant differences of miR-155 expression at the baseline of therapy (Figure 5A, p = 0.3408). Moreover, all the patients exhibited gradually decreasing miR-155 expression during treatment (Figure 5B). By comparing the SVR and NSVR groups, we found that miR-155 expression in both decreased gradually after starting treatment and showed no significant differences during the course of treatment (Figures 5C,D). Meanwhile, we also examined the expression of miR-146a in PBMCs between SVR and NSVR groups at baseline, but there was still no any differences between the two groups (Figure S5B in Supplementary Material), too.