Graphical abstract

   graphic file with name fx1.jpg
   [53]Open in a new tab

Highlights

     * •
       Probiotics promote HBV clearance by producing spermidine
     * •
       Probiotics and spermidine enhance IFN-γ^+CD4^+ T cell function
       through autophagy
     * •
       Probiotics induce the decline of serum HBsAg in patients with low
       HBsAg levels
     __________________________________________________________________

   Commensal microbes are promising in the functional cure of HBV
   infection. Wang et al. demonstrate that probiotics and their metabolite
   spermidine effectively promote the decline of HBsAg and inhibit HBV
   replication via autophagy-enhanced IFN-γ^+CD4^+ T cell immunity,
   highlighting the therapeutic potential of probiotics and spermidine for
   the functional cure of HBV patients.

Introduction

   Globally, approximately 296 million people are chronically infected
   with hepatitis B virus (HBV), which poses a high risk of liver
   cirrhosis and hepatocellular carcinoma, making it as one of the most
   serious common infectious agents.[54]^1 In 2016, the World Health
   Organization called for the goal of eliminating viral hepatitis,
   defined as 95% reduction in HBV incidence and 65% reduction in
   mortality by 2030.[55]^2^,[56]^3 The ideal goal of antiviral treatment
   for HBV patients is to achieve a functional cure, which is
   characterized by the persistent suppression of HBV DNA and the loss of
   serum hepatitis B surface antigen (HBsAg).[57]^4^,[58]^5 Unfortunately,
   achieving the functional cure for chronic HBV infection is difficult.
   The natural history has a success rate of less than 1%, and current
   antiviral drugs yield no more than 5% success rate even after long-term
   treatment.[59]^6^,[60]^7 Dysfunctional or exhausted T cells have been
   demonstrated to restrict the strength of HBV clearance, and recovery of
   T cell immunity is believed to be the most promising strategy for
   functional cure.[61]^8 Stimulating T cell co-stimulatory molecule OX40
   and blockade of programmed cell death ligand 1 (PD-L1) synergistically
   promote the functional restoration of HBV-specific CD4^+ T cells,
   leading to the production of interferon (IFN)-γ and interleukin
   (IL)-21, which are essential for HBV clearance.[62]^9 Hoogeveen et al.
   also showed that functional cure is associated with robust responses by
   HBV-specific CD4^+ T cells.[63]^10 Therefore, it is important to
   develop exact and effective treatment strategies for restoring
   dysfunctional T cell immunity and achieve functional cure of HBV
   infection.

   The liver is the unique target organ that orchestrates HBV-mediated
   immune response and is coordinated by gut microbe-derived ligands or
   metabolites through the gut-liver axis.[64]^11 Altered and flexible
   profile of gut microbiota in HBV infection has been
   reported,[65]^12^,[66]^13^,[67]^14 and the establishment of gut
   microbiota determines the age-related HBV clearance, in which T cell
   response plays pivotal roles.[68]^15^,[69]^16 In the limited data of
   HBV patients, Ren et al. reported that fecal microbiota transplantation
   (FMT) presented the ability to induce the decline of HBV e-antigen
   (HBeAg) expression in persistent HBeAg-positive patients even after
   long-term antiviral treatment.[70]^17 Mechanistically, commensal
   microbes and their metabolites regulate the production of IL-22,
   IL-17A, and IL-21 to promote intestinal immune
   homeostasis.[71]^18^,[72]^19 However, multidrug-resistant organisms
   colonized in healthy donors’ feces during donation activities would
   pose a risk for FMT.[73]^20 Probiotics are widely used as a food
   supplement to assist in the treatment of many health or disease
   conditions.[74]^21 Currently, the molecular and immune mechanism by
   which probiotics promote HBV clearance remains unclear. Consequently,
   the key and specific microbes or metabolites that facilitate HBV
   clearance have yet to be identified.

   Here, we reported that probiotics (B. longum, L. acidophilus, and
   E. faecalis, BLE), widely used in clinical practice, promote HBV
   clearance in both specific-pathogen-free (SPF) mice and germ-free (GF)
   mice. Specifically, spermidine (SPD) derived from accumulated
   probiotics in the gut could transport to the liver and exhibit a
   similar anti-HBV effect to probiotics. Mechanistically, depletion of
   CD4^+ T cells or blockage of IL-21R or IFN-γ almost completely dampened
   probiotics-induced HBV inhibition, and SPD administration activated the
   CD4^+ T cell immunity through autophagy. In HBV patients undergoing
   antiviral treatment, probiotics showed the potential to induce the
   decline of serum HBsAg.

Results

Administration of probiotics BLE promotes HBV clearance and improves
intestinal commensal homeostasis and barrier function

   To evaluate the efficiency of probiotics (BLE) in HBV clearance,
   HBV-carrier mouse models were prepared by hydrodynamical injection
   (HDI) of adeno-associated virus/HBV1.2 (AAV/HBV1.2) plasmids in
   5-week-old C57BL/6 mice.[75]^22 BLE were administered by gavage daily
   starting from either 2 weeks prior to HDI (BLE Pre group) or after
   3 days post injection (BLE group) ([76]Figure 1A). As shown in
   [77]Figures 1B–1E, both BLE and BLE Pre significantly inhibited HBV
   replication, displayed as decreased levels of serum HBsAg, HBV DNA, as
   well as hepatic pregenomic RNA (pgRNA) and HBV RNA. Consistently, BLE
   administration significantly decreased the percentage of HBsAg-positive
   mice within 6 weeks post injection (wpi) ([78]Figure 1F). In addition,
   no liver injury was detected in mice treated with BLE ([79]Figures S1A
   and S1B). Since BLE and BLE Pre showed similar anti-HBV effects, we
   used orally daily administration of BLE as the routine probiotic
   treatment in the following experiments.

Figure 1.

   [80]Figure 1
   [81]Open in a new tab

   Supplementation with BLE promoted HBV clearance in mice

   (A) Study design: male five-week-old C57BL/6J mice were
   hydrodynamically injected with 6 μg of AAV/HBV1.2 plasmid and gavage
   fed with probiotics (BLE, PBS as control) daily starting two weeks
   before (Pre) or at 3 dpi of HDI (control, n = 10; BLE Pre, n = 10; BLE,
   n = 10).

   (B–E) Levels of serum HBsAg (B), HBV DNA (C), hepatic pgRNA (D), and
   HBV RNA (E) were detected by ELISA, qPCR, and northern blot,
   respectively.

   (F) The percentages of HBsAg-positive mice (HBsAg > 1 ng/mL) at the
   indicated time points were calculated.

   Data were expressed as mean ± SEM. The statistical tests were performed
   by two-way analysis of variance (ANOVA) with Dunnett’s multiple
   comparison test (B and C), t test (D), or log rank (F). ∗0.01 < p
   value < 0.05; ∗∗0.001 < p value < 0.01; ∗∗∗∗p value < 0.0001.

   Probiotics show well-defined improvement in gut microbiota. We then
   investigated the effect of BLE administration on microbiota profile in
   HBV-carrier mice. Principal coordinate analysis (PCoA) revealed
   significant shifts (p = 0.004) in intestinal microbiome composition
   between control and BLE mice ([82]Figure 2A). Taxonomic alpha diversity
   (Chao diversity) analysis showed a marked increase in species richness
   in BLE groups ([83]Figure 2B). Moreover, linear discriminant analysis
   effect size (LEfSe) analysis displayed that the microbiome of BLE mice
   was significantly enriched with families Prevotellaceae,
   Clostridiaceae, Akkermansia, and Ruminococcaceae ([84]Figure 2C), which
   have been suggested to contribute to the restoration of exhausted CD8^+
   T cell mediated by anti-PD-1 or CTLA-4
   antibodies.[85]^23^,[86]^24^,[87]^25 Further, qPCR analysis using
   bacterium-specific primers showed that the abundance of BLE in the
   feces significantly increased after probiotics supplementation,
   suggesting the accumulation of probiotics bacteria ([88]Figure 2D).

Figure 2.

   [89]Figure 2
   [90]Open in a new tab

   Accumulation of BLE contributed to gut microbiota homeostasis and HBV
   clearance

   (A–C) 16S rRNA sequencing was performed with feces collected at 2 wpi.
   Beta diversity comparison via PCoA analysis (A) and alpha diversity
   comparison via Chao index analysis (B) showed a difference between the
   two groups. Bray-Curtis distance algorithm was used for PCoA analysis.
   (C) LEfSe (linear discriminant analysis effect size) predictions for
   bacterial families found in fecal pellets of control (green) and BLE
   (red) mice were shown, with cutoffs of LDA < −2 or LDA > 2.

   (D) Fecal bacterial genomic DNA was extracted from control and BLE
   mice. The expression of B. longum, L. acidophilus, and E. faecalis was
   determined by qPCR.

   (E) Study design: male three-week-old C57BL/6J mice were treated with
   antibiotics water for two weeks and then hydrodynamically injected with
   6 μg of pAAV/HBV1.2 plasmid and fed with fecal microbiota from control
   and BLE mice daily starting at 3 dpi of HDI (FMT-control, n = 10;
   FMT-BLE, n = 10).

   (F–H) Levels of serum HBsAg (F), HBV DNA (G), and hepatic pgRNA
   (H) were detected by ELISA and qPCR.

   (I) The percentages of HBsAg-positive mice (HBsAg > 1 ng/mL) at the
   indicated time points were calculated.

   (J) Study design: male six-week-old C57BL/6J GF mice were
   hydrodynamically injected with 6 μg of pAAV/HBV1.2 plasmid and gavage
   fed with BLE daily starting at 3 dpi of HDI.

   (K) The expression of B. longum, L. acidophilus, and E. faecalis was
   determined by qPCR.

   (L) Levels of serum HBsAg, HBV DNA, and hepatic pgRNA were detected by
   ELISA and qPCR.

   Data were expressed as mean ± SEM. The statistical tests were performed
   by analysis of similarities (ANOSIM) (A), t test (B, D, H, K, and L),
   parametric factorial Kruskal-Wallis (KW) sum-rank test (C), two-way
   ANOVA (F and G), or log rank (I). ∗0.01 < p value < 0.05; ∗∗0.001 < p
   value < 0.01; ∗∗∗ 0.0001 < p value < 0.001; ∗∗∗∗p value < 0.0001.

   Intestinal microbiota plays a critical role in maintaining gut barrier
   and permeability, which is strongly associated with the development of
   chronic liver diseases.[91]^26 In accordance, we detected lower serum
   fluorescein isothiocyanate (FITC)-dextran concentration in BLE mice
   than in control mice ([92]Figure S1C), indicating improved intestinal
   barrier integrity in BLE mice. Consistently, the serum level of
   intestinal fatty acid binding protein (iFABP), a biomarker of
   enterocyte damage,[93]^27 significantly decreased in BLE mice
   ([94]Figure S1D). In contrast, there was higher expression of tight
   junction proteins, including Claudin 1, Occludin, and Tjp1, in small
   intestine (SI) tissues from BLE mice than those from control mice
   ([95]Figure S1E).

   It has been well documented that microbiota induces the production of
   IL-17A, IL-21, and IL-22 to improve gut integrity and
   homeostasis.[96]^18^,[97]^19 As expected, significantly higher
   expression of IL-17A, IL-21, and IL-22 at both the mRNA and protein
   levels was detected in SI tissues from BLE mice ([98]Figures S1F and
   S1G). Flow cytometric (FCM) analysis verified the marked increase in
   the frequency of T cells producing IL-17A, IL-21, and IL-22 in
   lymphocytes of small intestinal lamina propria, mesenteric lymph nodes
   (mLNs) and Peyer’s patches (PPs) from BLE mice than control mice
   ([99]Figures S1H–S1K). Collectively, these data suggest that
   accumulation of BLE improves intestinal commensal homeostasis and
   barrier functions in HBV-carrier mice.

Accumulation of BLE contributes to HBV clearance in mice

   To assess whether gut microbiota is involved in BLE-mediated HBV
   suppression, feces were collected from PBS- or BLE-treated HBV-carrier
   mice (donor mice). Then, FMT experiments were performed with recipient
   mice, which were gut-sterilized using a 14-day antibiotic mixture (ABX)
   treatment followed by HDI of HBV at the age of 5 weeks
   ([100]Figure 2E). As shown in [101]Figures 2F–2H, transplantation of
   fecal microbiota from BLE mice (FMT-BLE) greatly promoted HBV clearance
   in recipient mice, displaying obviously decreased levels of serum HBsAg
   and HBV DNA and reduced hepatic pgRNA expression when compared to mice
   that received control fecal microbiota (FMT-Control). Similarly, the
   percentage of HBsAg-positive mice in FMT-BLE mice was significantly
   lower than that in FMT-Control mice ([102]Figure 2I).

   To further validate the direct role of BLE in HBV clearance, we
   included GF mice to establish a HBV model by HDI, followed by
   probiotics administration ([103]Figure 2J). Consisting with results in
   SPF mice, supplementation with BLE largely increased the numbers of BLE
   in the feces from GF mice ([104]Figure 2K) and decreased levels of
   serum HBsAg, HBV DNA, and hepatic pgRNA ([105]Figure 2L). Taken
   together, these data suggest that the accumulation of BLE contributes
   to HBV clearance in mice.

IFN-γ and IL-21 produced by CD4^+ T cells are indispensable for BLE-induced
HBV clearance

   To further evaluate the role of individual immune subsets in
   probiotics-mediated HBV clearance, AsGM1 and neutralized antibodies
   against CD4, CD8, and TCRγ/δ were used in probiotics-treated
   HBV-carrier mice. As shown in [106]Figures 3A–3C, depletion of CD8^+
   T cells, γδT cells, or natural killer (NK) cells partially restored the
   reduced serum levels of HBsAg in BLE-treated mice. Notably, depletion
   of CD4^+ T cells almost completely abolished BLE-induced HBV clearance,
   suggesting that BLE promotes HBV clearance in a CD4^+ T cell-dependent
   manner. In line with the critical role of CD4^+ T cells, probiotics
   treatment led to augmentation of CD44^+CD62L^− effector CD4^+ T cells
   and CD44^+CD62L^+ central memory CD4^+ T cells in the spleen and liver
   ([107]Figure S2A).

Figure 3.

   [108]Figure 3
   [109]Open in a new tab

   Depletion of CD4^+ T cells and inactivation of IL-21 or IFN-γ abrogated
   the probiotics-initiated anti-HBV effects

   (A–C) Mice were treated i.p. with 100 mg neutralization antibody or
   isotype every 3 days from 0 to 6 wpi (control, n = 7; BLE, n = 7;
   BLE+αCD4, n = 9; BLE+αCD8, n = 5; BLE+αTCRγ/δ, n = 7; BLE+AsGM1, n =
   5). Serum levels of HBsAg (A), HBV DNA (B), and liver pgRNA (C) were
   detected by ELISA and qPCR. HBV-carrier mice were treated with BLE or
   control PBS.

   (D and E) RT-qPCR analysis of transcription factors (D) and cytokines
   (E) in liver tissues.

   (F) FCM analysis of percentages of hepatic T-bet^+ Th1 cells,
   PD-1^+CXCR5^+ Tfh cells, and RORγt^+ Th17 cells among CD4^+ T cells.

   (G) IL-21 and IFN-γ in liver homogenates were analyzed by ELISA.

   (H and I) Percentages and MFI of IL-21 (H) and IFN-γ (I) in hepatic
   T-bet^+Th1 cells.

   (J–L) Antibody-mediated blockade of IL-21R and IFN-γ abolished the
   effect of probiotics-induced HBV clearance. Mice were treated i.p. with
   100 mg neutralization antibody (anti-IL-21R and anti-IFN-γ) or isotype
   every 3 days from 0 to 6 wpi (control, n = 7; BLE, n = 7; BLE+αIL-21R,
   n = 4; BLE+αIFN-γ, n = 7). Serum levels of HBsAg (J), HBV DNA (K), and
   liver pgRNA (L) were detected by ELISA and qPCR.

   Data were expressed as mean ± SEM. The statistical tests were performed
   using two-way ANOVA with Dunnett’s multiple comparison test (A, B, J,
   and K) or t test. ns, no significance; ∗0.01 < p value < 0.05;
   ∗∗0.001 < p value < 0.01; ∗∗∗0.0001 < p value < 0.001; ∗∗∗∗p
   value < 0.0001.

   Growing evidence has demonstrated the involvement of different CD4^+ T
   subsets in anti-HBV immunity.[110]^28^,[111]^29 To gain deeper insight
   into the role of CD4^+ T subsets in BLE-mediated HBV clearance, we
   assessed hepatic mRNA expression of signature transcription factors and
   cytokines of Th1, Th17, follicular helper T (Tfh), and regulatory T
   (Treg) cells. Quantitative reverse-transcription PCR (RT-qPCR) revealed
   a significant increase in the expression of Eomes, Tbx21, Il-21, and
   Il12, while expression of Ifng but not Bcl-6, Rorc, Foxp3, Il17-a, and
   Tgfb showed no significant change in BLE mice ([112]Figures 3D and 3E).
   Consistently, FCM showed a significant increase of T-box expressed in T
   cell (T-bet)^+ Th1 cells, but not CXCR5^+PD-1^+ Tfh cells and ROR-γt^+
   Th17 cells, in the livers from BLE-treated mice ([113]Figure 3F). In
   accordance, BLE treatment led to the accumulation of IL-21^+CD4^+
   T cells and IFN-γ^+CD4^+ T cells in the liver and spleen
   ([114]Figures S2B–S2E). In addition, ELISA revealed higher levels of
   IL-21 and IFN-γ in liver homogenate in BLE mice ([115]Figure 3G). Tfh
   cells have been known as one of the main sources of IL-21 and
   IFN-γ.[116]^30 Further analysis revealed that the expression of IL-21
   and IFN-γ significantly increased in Th1 cells, but not in Tfh cells,
   after BLE treatment ([117]Figures 3H, 3I, [118]S2F, and S2G). It has
   been reported that IL-21 signaling enhances cytotoxicity and IFN-γ
   production in NK and CD8^+ T cells.[119]^31 Consistently, BLE treatment
   increased the production of IFN-γ in splenic CD8^+ T cells
   ([120]Figures S2H and S2I), indicating the potential contribution of
   CD8^+ T cells to BLE-induced HBV clearance. In addition, BLE treatment
   also led to the augmentation of hepatic and splenic IFN-γ^+CD4^+
   T cells in GF mice ([121]Figures S2J and S2K).

   Considering the critical role of IFN-γ and IL-21 in determining HBV
   persistence,[122]^16^,[123]^32 we therefore determined whether the
   clearance of HBV in probiotics-treated mice is mediated by the
   up-regulation of IL-21 or IFN-γ. As shown in [124]Figures 3J–3L, the
   treatment of neutralized antibodies against either IL-21R or IFN-γ
   completely abolished the anti-HBV effect mediated by probiotics
   treatment. In summary, these results illustrate that the probiotics
   administration enables the host to clear HBV by enhancing the
   production of IL-21 and IFN-γ in CD4^+ Th1 cells.

SPD produced by BLE promotes HBV clearance

   The microbiota-derived metabolites serve as important signals
   contributing to gut homeostasis and host immunity. We then investigated
   whether probiotics affect the microbial metabolism of the host.
   Phylogenetic investigation of communities by reconstruction of
   unobserved states (PICRUSt) analysis of 16S rRNA sequencing data
   indicated the enrichment of metabolism-related pathways in BLE-treated
   mice ([125]Figure S3A; [126]Table S1). In accordance, the gas
   chromatograph coupled with a Pegasus HT time-of-flight mass
   spectrometer (GC-TOF-MS) detected the alteration of metabolites in
   feces from BLE-treated mice. Pathway enrichment analysis using
   MetaboAnalyst 5.0 ([127]https://www.metaboanalyst.ca) showed that
   multiple metabolic pathways were significantly enriched in BLE mice,
   and 15 of them were consistent with the results of PICRUSt analysis
   ([128]Figures S3A and S3B; the same pathways marked in red). Totally,
   untargeted metabolomic analysis showed that 38 metabolites were
   significantly different between control and BLE mice ([129]Figures 4A
   and [130]S3C). Among them, oligosaccharides (1-kestose, raffinose,
   maltotriose), amino acids (N-Acetyl-L-leucine, glutamic acid), and
   polyamine (SPD) were the most varied enriched metabolites. Recently,
   studies showed that SPD modulates CD4^+ T cell differentiation and
   enhances vaccine-induced T cell function.[131]^33^,[132]^34 Further
   targeted metabolomics analysis of polyamines (including SPD, spermine,
   and putrescine) demonstrated that SPD and putrescine rather than
   spermine were accumulated in the livers of BLE mice, and the
   concentration of SPD was higher than that of putrescine
   ([133]Figure S3D). More importantly, the concentration of SPD was
   accumulated in the feces, peripheral serum, and livers from BLE-treated
   GF mice ([134]Figure 4B). These results indicate that BLE-produced SPD
   may translocate from the intestine to the liver through the blood
   circulation.

Figure 4.

   [135]Figure 4
   [136]Open in a new tab

   BLE-derived spermidine promoted HBV clearance

   (A) Untargeted metabolomic analysis of feces from control and BLE mice
   by gas chromatograph system coupled with a Pegasus HT time-of-flight
   mass spectrometer (GC-TOF-MS). Heatmap of differentially expressed
   metabolites in mice fed with BLE were shown (control, n = 9; BLE, n =
   9).

   (B) Targeted metabolomics analysis of SPD in feces, peripheral serum,
   and livers from GF mice.

   (C) Extracellular SPD concentrations in cultures of separate E.
   faecalis and mixed BLE in gifu anaerobe medium (GAM) medium
   supplemented with 2 mM arginine under different pH values.

   (D–F) HBV-carrier mice were fed with drinking water (control, n = 8) or
   3 mM SPD-containing water (SPD, n = 8). Levels of serum HBsAg (D), HBV
   DNA (E), and liver pgRNA (F) were detected by ELISA or qPCR.

   (G and H) FCM analysis of percentages and MFI of hepatic/splenic
   IL-21^+CD4^+ T cells (G) and IFN-γ^+CD4^+ T cells (H).

   (I) IL-21 and IFN-γ in liver homogenates were analyzed by ELISA.

   Data were expressed as mean ± SEM. The statistical tests were performed
   using t test and two-way ANOVA (D and E). ∗0.01 < p value < 0.05;
   ∗∗0.001 < p value < 0.01; ∗∗∗∗p value < 0.0001.

   It has been recently reported that bioactive polyamine can be produced
   by a hybrid system involving the collaboration of two bacterial groups
   and triggered by environmental acidification. In this system, E.
   faecalis, one of the three probiotic strains in BLE, functions as
   agmatine deiminase system to promote the production of polyamine by the
   collaboration of bacteria that have an acid resistance system composed
   of arginine decarboxylase (AdiA) and arginine-agmatine antiporter
   (AdiC), such as E. coli and B. longum.[137]^35 We thus hypothesized
   whether the augmented SPD in the liver of BLE mice was attributed to
   the accumulation of BLE. To address this, we first used traditional
   bacterial culture methods to obtain culture supernatant and detect SPD
   concentration. We found that cultured BLE produced significantly higher
   level of SPD in the arginine-containing medium at pH 5.0 than at pH 7.0
   ([138]Figure 4C). Moreover, administration of BLE produced markedly
   higher levels of SPD in the feces, liver tissues, and serum of
   antibiotics (ABX)-treated mice, while single E. faecalis was not
   effective to produce SPD both in vitro and in vivo ([139]Figures 4C and
   [140]S4A–S4C). Accordingly, in comparison to control and single
   E. faecalis-administrated mice, BLE administration significantly
   decreased the serum levels of HBsAg, HBV DNA, and hepatic pgRNA
   ([141]Figures S4D–S4F). Similar results were observed in FCM analysis
   of IFN-γ^+CD4^+ T cells in mice livers ([142]Figure S4G). Thus, BLE
   effectively produces SPD and enhances anti-HBV immunity, with
   E. faecalis being able to produce SPD with the cooperation of B. longum
   and L. acidophilus.

   We then further investigated the role of SPD in HBV clearance.
   Expectedly, SPD administration alone significantly decreased the levels
   of serum HBsAg, HBV DNA, and hepatic pgRNA in HBV-carrier mice
   ([143]Figures 4D–4F). In addition, SPD administration did not cause
   liver injury ([144]Figures S5A and S5B). Taken together, these results
   imply that BLE-derived SPD accumulated from the gut to the liver by
   peripheral circulation, which contributes to HBV clearance in mice.

SPD improves intestinal homeostasis and enhances Th1 cell immunity through
autophagy

   To further explore the mechanism underlying SPD-mediated HBV clearance,
   gut integrity and T cell immunity were evaluated. As shown in
   [145]Figures S5C–S5E, mice fed with SPD displayed a notable elevation
   in intestinal Claudin1 expression, accompanied by a clear reduction in
   serum levels of FITC-dextran and iFABP. This finding aligns with a
   previous report showing that SPD regulates the intestinal epithelial
   barrier function.[146]^36 Additionally, SPD feeding led to a great
   increase in the expression of intestinal IL-17A, IL-21, and IL-22 and
   the rise in the percentages and MFI of intestinal IL-17A^+CD4^+
   T cells, IL-21^+CD4^+ T cells, and IL-22^+CD4^+ T cells
   ([147]Figures S5F–S5J). Furthermore, consistent with the results in
   BLE-treated mice, we detected a marked increase in the frequency of
   IL-21^+CD4^+ T cells and IFN-γ^+CD4^+ T cells in livers and spleens
   from SPD-treated mice ([148]Figures 4G and 4H). ELISA confirmed the
   significantly increased levels of IL-21 and IFN-γ in liver homogenate
   from SPD mice ([149]Figure 4I), suggesting that SPD enhances CD4^+
   T cell immunity in HBV-carrier mice.

   It has been shown that SPD, a natural autophagy-promoting polyamine,
   exerts multiple functions, including extending lifespan and immune cell
   activation, in an autophagy-dependent manner.[150]^37 Therefore, we
   investigated whether SPD enhances CD4^+ T cell function via regulating
   the autophagy pathway. To address this issue, in vitro splenic cells
   isolated from OT-II mice were stimulated with ovabumin (OVA[323-339])
   peptide with or without SPD. FCM confirmed that SPD enhanced the
   production of IFN-γ in OVA-stimulated OT-II CD4^+ T cells in a
   dose-dependent manner ([151]Figure 5A). Concurrently, SPD
   dose-dependently promoted the accumulation of autophagosomes in OT-II
   CD4^+ T cells ([152]Figure 5B), while the autophagy inhibitor
   3-methyladenine (3-MA, 1 mM) further enhanced the SPD-induced
   accumulation of autophagosomes (SPD+3-MA, [153]Figure 5C), suggesting
   the activation of autophagy in SPD-treated CD4^+ T cells. Notably, 3-MA
   markedly decreased the SPD-induced augmentation of IFN-γ in CD4^+
   T cells ([154]Figure 5D). In line with the notion that SPD acts as the
   only known substrate for the hypusination of eukaryotic translation
   initiation factor 5A (eIF5A), which is essential for the synthesis of
   the autophagy transcription factor EB (TFEB),[155]^38 treatment with
   GC7 (SPD+GC7), a specific inhibitor of eIF5A hypusination, decreased
   the accumulation of SPD-induced autophagosomes ([156]Figure 5C) and
   completely damaged SPD-induced accumulation of IFN-γ in CD4^+ T cells
   ([157]Figure 5D). Consistently, SPD induced autophagosome accumulation
   in hepatic CD4^+ T cells from HBV-carrier mice ([158]Figure 5E). More
   importantly, SPD significantly promoted the accumulation of
   autophagosomes and IFN-γ in CD4^+ T cells derived from peripheral blood
   of HBV patients, while treatment of GC7 largely damaged this induction
   ([159]Figures 5F and 5G). Taken together, these data imply that SPD
   enhances CD4^+ T cell function in an autophagy-dependent manner.

Figure 5.

   [160]Figure 5
   [161]Open in a new tab

   SPD enhanced CD4^+ T cell function through autophagy in vivo and
   in vitro

   (A–D) OT-II CD4^+ T cells were prepared by culturing splenic
   mononuclear cells isolated from OT-II mice with OVA[323-339] peptide
   for 3 days. OT-II CD4^+ T cells were then treated for 24 h with
   increasing concentrations of SPD (A and B). 3-MA (1 mM) and GC7 (10 μM)
   were used to inhibit autophagy (C and D). IFN-γ levels (A and D) and
   autophagosome levels (B and C) in CD4 ^+ T cells were measured by FCM.

   (E) Hepatic mononuclear cells were isolated from HBV mice, and levels
   of autophagosome in CD4 ^+ T cells were measured by FCM.

   (F and G) PBMCs isolated from HBV patients were stimulated with SPD
   and/or GC7 for 24 h. Autophagosome levels (F) and IFN-γ levels (G) in
   CD4^+ T cells were measured by FCM.

   Data were expressed as mean ± SEM. The statistical tests were performed
   using t test. ns, no significance; ∗0.01 < p value < 0.05; ∗∗0.001 < p
   value < 0.01; ∗∗∗ 0.0001 < p value < 0.001; ∗∗∗∗p value < 0.0001.

BLE accelerates the decline of serum HBsAg in human flora-associated mice and
HBV patients with antiviral treatment

   To address whether the probiotics would cause similar alterations in a
   more natural microbiome and evaluate the pre-clinical therapeutic
   potential of BLE and SPD in HBV clearance, we constructed the human
   flora-associated (HFA) model by reconstituting the microbiota of GF
   mice with fecal samples from HBV patients. Three weeks later, the HFA
   mice were injected with AAV/HBV1.2 by HDI and treated with either BLE
   or SPD, PBS as control ([162]Figure 6A). As shown in
   [163]Figures 6B–6D, both BLE and SPD showed significantly higher
   capacity of HBV clearance in HFA mice. Compared with PBS-treated HFA
   mice, BLE/SPD-treated HFA mice showed remarkably decreased levels of
   serum HBsAg, HBV DNA, and hepatic pgRNA. Both BLE and SPD significantly
   promoted the accumulation of autophagosomes in hepatic CD4^+ T cells
   ([164]Figure 6E) and enhanced the production of IFN-γ in hepatic and
   splenic CD4^+ T cells ([165]Figures 6F and [166]S6A). More importantly,
   both BLE and SPD significantly enhanced the production of IFN-γ in
   hepatic and splenic HBV-specific CD4^+ T cells, indicating the
   potential contribution of BLE and SPD to HBV-specific immunity
   ([167]Figures 6G, [168]S6B, and S6C).

Figure 6.

   [169]Figure 6
   [170]Open in a new tab

   Probiotics accelerated HBsAg clearance in HFA mice and HBV patients

   (A) Study design: male three-week-old C57BL/6J GF mice were fed with
   fecal microbiota from HBV patients daily for 3 weeks and then
   hydrodynamically injected with AAV/HBV1.2 plasmid and fed with BLE or
   SPD for 6 weeks (control, n = 4; BLE, n = 4; SPD, n = 4).

   (B–D) Levels of serum HBsAg (B), HBV DNA (C), and hepatic pgRNA
   (D) were detected by ELISA and qPCR.

   (E and F) Levels of autophagosome (E) and IFN-γ (F) in CD4^+ T cells
   were measured by FCM.

   (G) MHC II-HBcAg[128-140]-Tetramer was used to evaluate HBV-specific
   CD4^+ T cells. Percentages of IFN-γ in hepatic HBV-specific CD4^+
   T cells were measured by FCM.

   (H) A total of 20 eligible patients who remained persistently positive
   for HBsAg following >6 months of ongoing nucleos(t)ide analogs- or
   PegIFNα2b-based antiviral therapy were enrolled; among them, 10
   patients received BLE therapy (control, n = 10; BLE, n = 10).

   (I) HBsAg levels in each patient before and after BLE treatment were
   analyzed and compared with HBV patients without BLE treatment.

   (J and K) Levels of autophagosome (J) and IFN-γ (K) in CD4 ^+ T cells
   from PBMCs were measured by FCM.

   Data were expressed as mean ± SEM. The statistical tests were performed
   using two-way ANOVA with Dunnett’s multiple comparison test (B and C)
   or t test. ∗0.01 < p value < 0.05; ∗∗0.001 < p value < 0.01; ∗∗∗
   0.0001 < p value < 0.001.

   We therefore conducted the preliminary observation for clinical effects
   of BLE combined with antiviral treatment in HBV patients with
   persistent low level of serum HBsAg (less than 2,000 IU/mL), which were
   usually considered as advantaged population trend for functional
   cure[171]^39 ([172]Tables S2 and [173]S3; [174]Figure 6H). The dynamic
   profiles of serum HBsAg levels before and after BLE treatment have been
   illustrated ([175]Figure S7A). Compared with control patients, patients
   with BLE treatment for 6 months showed an obvious decline of serum
   HBsAg and HBeAg ([176]Figures 6I and [177]S7B). In detail, all 10
   patients presented a declined trend of serum HBsAg level, and 3 of 10
   patients even obtained the successful loss of serum HBsAg after the
   treatment (patients B5, B6, B7). Notably, patient B5 and B6 maintained
   the state of HBsAg negative for 6 months ([178]Figure S7A). Expectedly,
   BLE supplement significantly promoted the accumulation of
   autophagosomes and IFN-γ expression in CD4^+ T cells from peripheral
   blood mononuclear cells (PBMCs) ([179]Figures 6J and 6K). Although this
   study was a preliminary observation with small number of patients,
   these results clearly suggest the therapeutic potential of BLE toward
   the functional cure of HBV infection.

SPD synergically enhances the anti-HBV effect of entecavir in mice

   Based on the aforementioned data proving the role of SPD in T cell
   activation and HBV clearance, we next determined whether
   co-administration of SPD could enhance the therapeutic effect of
   entecavir (ETV), the most widely used medication that efficiently
   inhibits HBV DNA replication but not HBsAg production.[180]^40 As shown
   in [181]Figures 7A–7D, compared to ETV monotherapy, which only
   suppressed HBV DNA level and showed minimal effects on HBsAg and pgRNA,
   the combination therapy (ETV+SPD) group demonstrated a more pronounced
   reduction in HBsAg and pgRNA levels, while displaying comparable
   inhibitory effects on HBV DNA. In accordance with the promotion of Th1
   immunity in BLE-treated mice, RT-qPCR detected significantly increased
   expression of Il21, Ifng, Il12, Tbx21, and Eomes in mice with
   co-administration of ETV with SPD but not in mice with ETV monotherapy
   ([182]Figure 7E). Furthermore, compared with ETV monotherapy,
   combination therapy significantly increased the production of IFN-γ in
   splenic and hepatic CD4^+ T cells ([183]Figures 7F and 7G). Taken
   together, these data demonstrate that SPD greatly enhances the anti-HBV
   therapeutic effect of ETV by promoting antiviral immunity and suggest
   the therapeutic potential of SPD toward the functional cure of HBV
   infection.

Figure 7.

   [184]Figure 7
   [185]Open in a new tab

   Combination of SPD promoted the HBV inhibition in HBV-carrier mice

   (A) Male five-week-old HBV-carrier mice were administered with
   entecavir (ETV) or co-administered with 3 mM SPD water (ETV+SPD)
   combined with ETV (control, n = 5; SPD, n = 5; ETV, n = 5; ETV+SPD, n =
   5).

   (B–D) Levels of serum HBsAg (B), HBV DNA (C), and liver pgRNA (D) were
   detected by ELISA or qPCR.

   (E) RT-qPCR analysis of transcription factors and cytokines in liver
   tissues.

   (F and G) FCM analysis of percentages and MFI of splenic (F) and
   hepatic (G) IFN-γ^+CD4^+ T cells.

   Data were expressed as mean ± SEM. The statistical tests were performed
   using t test. ns, no significance; ∗0.01 < p value < 0.05; ∗∗0.001 < p
   value < 0.01; ∗∗∗ 0.0001 < p value < 0.001.

Discussion

   Functional cure of HBV infection is considered an achievable
   therapeutic goal, given that it is almost universally observed after
   adult infection. However, the current anti-HBV drugs, namely
   nucleos(t)ide analogs as well as peginterferon-α (PegIFNα), cured fewer
   than 1% in the natural history of chronic HBV infection.[186]^41 The
   high burden of viral antigens that promote T cell exhaustion with
   T cell dysfunction is the key barrier to functional cure.[187]^8 In
   this study, we demonstrate the critical role of probiotics in enhancing
   CD4^+ T cell function and promoting the clearance of HBV. Furthermore,
   probiotics showed the potential to accelerate the decline of serum
   HBsAg levels in HBV patients with common anti-HBV drugs. Probiotics are
   widely used as food supplement to assist in the treatment of many
   health conditions,[188]^21 and no liver injury was detected in mice
   treated with BLE in our studies. Our data here provide a safe and
   effective combination intervention strategy for chronic HBV infection.

   Scientific evidence supports the important roles of probiotics in
   alleviating the symptoms of several digestive system diseases.[189]^21
   A large number of studies also have demonstrated the disturbed gut
   microbiota in HBV patients.[190]^42 In this study, using 16S rRNA
   sequencing, we showed that probiotics administration alters microbiome
   compositions and enriches commensal diversity in HBV-carrier mice.
   Further, qPCR analysis using bacterium-specific primers showed the
   accumulation of BLE in probiotics-treated mice. Then, FMT experiments
   and GF mice were used to verify that accumulation of BLE contributes to
   HBV clearance. Therapeutic strategies to cure persistent HBV infection
   almost invariably include approaches to enhance antiviral T cell
   immunity.[191]^43 Our data suggest that BLE-induced HBV clearance is
   achieved by provoking CD4^+ T cell function. Depletion of CD4^+
   T cells, as well as blocking IL-21 and IFN-γ, the well-known effector
   molecules of CD4^+ Th1 cells, almost abolished BLE-initiated HBV
   clearance. Further expression analysis of signature transcription
   factors of different CD4^+ T cell subsets demonstrated the augmented
   expression of Tbx21 and Eomes but not Bcl-6, Rorc, and Foxp3 in
   probiotics-treated mice, suggesting the involvement of Th1 cells in
   probiotics-mediated HBV clearance. Although T-bet was also involved in
   Tfh cell development and germinal center generation, and expression of
   IL-21 and IFN-γ is detected in Tfh cells,[192]^44 FCM analysis showed
   no alteration in either IL-21^+ Tfh or IFN-γ^+ Tfh cells in BLE-treated
   mice. Meanwhile, BLE treatment did not influence the production of
   HBsAb ([193]Table S4). These results suggest that Th1 but not Tfh cells
   were the main effector cells contributing to BLE-induced HBV clearance.

   Another finding of our study is that we have demonstrated that SPD is
   the key metabolite responsible for HBV clearance and is produced by
   BLE. Firstly, PICRUSt analysis of 16S rRNA sequencing data and
   untargeted metabolomic analysis showed that the metabolism-related
   pathways, especially amino acid metabolism, were significantly enriched
   in mice with probiotics treatment. In particular, D-glutamine and
   D-glutamate metabolism and glutamic acid were the most abundant altered
   amino acid-related pathways and metabolites. It has been reported that
   glutamine metabolism promotes effector T cell generation and
   function,[194]^45 and glutamic acid is a highly effective regulator in
   initiating T cell-mediated immune responses.[195]^46 Moreover,
   beta-alanine metabolism, arginine and proline metabolism, and
   glutathione metabolism pathway contain the common metabolite SPD, which
   also enhanced the IFN-γ production of CD8^+ T cells.[196]^33 Secondly,
   targeted metabolomics analysis identified the enrichment of SPD in
   feces and livers from BLE-treated SPF mice. Moreover, single
   administration of SPD led to similar effects as probiotics in
   inhibiting HBV replication as well as improving intestinal homeostasis
   and Th1 cell function. Importantly, BLE supplementation increased SPD
   content in feces, peripheral serum, and livers and promoted HBV
   clearance in GF mice. Previous studies reported that SPD can be
   produced by many other microorganisms, such as Prevotellaceae,
   Clostridiaceae, and Ruminococcaceae.[197]^47 It needs to be further
   investigated whether BLE-induced alteration of other SPD-producing
   strains also participates in HBV clearance in SPF mice.

   Accumulated data demonstrated the effect of SPD on immunity, while its
   immunomodulatory effect was context
   dependent.[198]^34^,[199]^37^,[200]^38 SPD shifts Th17 cell
   polarization toward Tregs, while it did not affect the production of
   IFN-γ of naive CD4^+ T cells under Th1-polarizing conditions.[201]^34
   Matsumoto et al. reported increased IFN-γ and augmented fecal polyamine
   in probiotics-treated patients with autoimmunity.[202]^48 Our studies
   confirmed that SPD administration enhanced CD4^+ T cell function and
   promoted HBV clearance in vivo. SPD, a natural autophagy-promoting
   polyamine, exerts multiple functions, mainly through autophagy. It acts
   as the only known substrate for the hypusination eIF5A, which is
   essential for the synthesis of the autophagy transcription factor
   TFEB.[203]^38 In vitro studies showed that both autophagy inhibitor
   3-MA and eIF5A-hypusine inhibitor GC7 significantly decrease
   SPD-induced IFN-γ accumulation in CD4^+ T cells. Therefore, our work
   revealed that the SPD-autophagy pathway enhanced IFN-γ^+CD4^+ T cell
   immunity. However, other metabolites, especially oligosaccharides
   (1-kestose, raffinose, maltotriose), which are used as prebiotics to
   regulate the immune system,[204]^49^,[205]^50 were also up-regulated
   after probiotics treatment. Therefore, more work is required to
   determine the key metabolism-related pathways and master metabolites
   contributing to HBV clearance.

   The seroclearance or seroconversion of HBsAg has been identified as the
   key marker for HBV cure in clinics. However, the spontaneous conversion
   rate of HBsAg is as low as less than 1% in the natural
   population[206]^7 and as high as no more than 30% even in patients with
   HBsAg levels <1,000 IU/mL.[207]^51 Notably, our preliminary
   observational study reported that probiotics accelerate the decline of
   serum HBsAg levels in patients receiving anti-HBV drugs. Importantly, 3
   of 10 patients obtained the successful loss of serum HBsAg after the
   treatment with PegIFNα2b and tenofovir in combination with probiotics.
   Although the sample size is small, the exciting data still showed the
   great potential of probiotics in the loss of HBsAg. A combination of
   probiotics or SPD and antiviral treatment might be the potential
   therapeutic strategy for HBV functional cure.

   In conclusion, SPD derived from BLE effectively promotes HBV clearance,
   which is likely to associate with more effective IFN-γ^+CD4^+ Th1
   cells. Our study has offered valuable insight into the immune mechanism
   of microbiota-mediated HBV clearance. Furthermore, it provides possible
   cues to use probiotics and SPD combination therapy with anti-HBV drugs
   in the clinical treatment of chronic HBV infection.

Limitations of the study

   Our study has some limitations. We used mice with hydrodynamic
   injection of AAV/HBV1.2 as the mouse model to perform HBV clearance
   studies. Although this is a commonly used HBV model, hydrodynamic
   injection induces the release of damage-associated molecular patterns
   upon hepatocellular death. This process is in sharp contrast to the
   stealthy nature of HBV, which typically evades from the innate immune
   detection machinery during natural infections.[208]^52 Besides,
   although the exciting data have been obtained in human beings,
   perspective and randomized controlled trial is required to confirm the
   therapeutic potential of BLE as well as the identified metabolite in
   the context of HBV inhibition. Moreover, although both BLE and SPD
   contributed to the function of HBV-specific CD4^+ T cells, whether they
   could suppress other pathogens by modulating host immunity needs to be
   further verified.

Resource availability

Lead contact

   Further information and requests for resources and reagents should be
   directed to and will be fulfilled by the lead contact, Chunhong Ma
   (machunhong@sdu.edu.cn).

Materials availability

   All unique/stable reagents generated in this study are available from
   the [209]lead contact upon request and upon completion of Materials
   Transfer Agreement.

Data and code availability

     * •
       16S rRNA sequencing data in this study have been deposited in the
       National Center for Biotechnology Information (NCBI) database under
       BioProject accession number SRA: PRJNA700493. Metabolomics data in
       this study have been deposited in the National Metabolomics Data
       Repository (NMDR) under Study ID ST003536.
     * •
       This paper does not report original code.
     * •
       Any additional information required to reanalyze the data reported
       in this paper is available from the [210]lead contact upon request.

Acknowledgments