For investigating in vitro effects of TLR8 stimulation, peripheral blood samples were available from CHB patients or HBV vaccinated healthy volunteers enrolled in HOPE cohort at University of Maryland, Baltimore. The study protocol was approved by the institutional ethical committee, and all subjects gave written, informed consent. Demographic and clinical details of patients are provided in Table 1 .

For investigating in vivo effects of TLR8 agonism, peripheral blood samples from previously completed phase 1b clinical trial (ACTRN identifier: 12617000235303) of a selective-TLR8 agonist (Selgantolimod (SLGN), GS-9688) in CHB patients (N=14) and in healthy subjects (13, 14) were available courtesy of Gilead Sciences. Paired samples from baseline (BL) and TLR8 single oral dose treatment (8 hours-post SLGN, 1.5 or 3 mg) time-points were used here.

The following TLR ligands (InvivoGen, each 1 µg/mL) were added into wells of 24-well plate containing 1-2 x 10⁶ PBMCs: Poly(I:C) HMW (TLR3), LPS-EK Ultrapure (lipopolysaccharide, TLR4), Imiquimod (IMQ) (R837, TLR7), Resiquimod (R848, TLR7/8), TL8-506 (TLR8), ssRNA40/LyoVec (TLR8) and CpG-ODN (TLR9). The concentration of 1 µg/mL is within the range of manufacturer recommendation and induced maximum cytokine response (not shown) (10, 12). Golgi plug (1 µl/mL, BD Biosciences) was added to block the cytokines release and cell culture continued for 18h. For some experiments, PBMCs were stimulated with agonists alone or in combination with recombinant human HBsAg subtype adw (10 µg/mL, Fitzgerld) (15) and PepMix HBV (LEP) Ultra (2 µg/mL, JPT) and cultured for 5 days at 37°C incubator. In experiments for transcription factor induction, cells were re-stimulated with Phorbol-12-myristat-13-acetate (PMA, 50ng/mL) and Ionomycin (Ion, 1 µg/mL) on day 4.

PBMCs were stained and analyzed by flow cytometry using standard methods (panels listed in Supplementary Table 1 ). Cells were acquired on a Cytek Aurora multi-color FACS machine and data analysis done by Flow Jo software (Tree Star, San Carlos, California, USA).

For examination of contribution of TLR8 signaling in HBV antigen-specific germinal center like (GC) T_FH induction, a cytokine-independent approach was followed (16, 17). It is known that formulation of antigen can influence the type and extent of activation markers induced in CD4⁺ T cells (18). To capture multiple activation markers induced in T_FH with TLR8 signaling, here we used combination of HBsAg intact protein and peptide, as is described for other antigens (16, 17). Approximately 1-2 x 10⁶ PBMCs were aliquoted into 24-well plate and stimulated with HBV envelope peptide (2 µg/mL, JPT) and recombinant human HBsAg subtype adw (10 µg/mL, Fitzgerald). Intact protein antigens or antigenic peptides are used for stimulating antigen specific CD4 T cell response. Staphylococcal enterotoxin B (SEB 1 µg/mL, Toxins Technology) served as positive control and AIM-V medium alone served as untreated negative control (UT). CXCR5-BV605 and CXCR3-PeCy7 antibodies (Biolegend, 1 μl each) were added directly into the cell culture well and also included in the staining panel. After 18h stimulation, PBMCs cultures were washed twice with 1X PBS and followed the FACS staining protocol as stated in the above method using AIM assay panel flow antibodies provided in Supplementary Table 1 . Cells were washed, fixed with 1% paraformaldehyde and acquired on the same day.

To examine TLR8-mediated differentiation of T_FH cells, CD14⁺ monocytes and naïve CD4⁺ T cells were isolated from autologous human PBMCs of CHB patients (n=4) by negative selection (Miltenyi Biotec). The following co-culture experiments were adopted from previously published methods (12). First, purified CD14⁺ monocytes (1 x 10⁶ cells/mL) were left untreated or treated with TLR8-specific agonists ssRNA40/Lyovec (1 µg/mL, InvivoGen) or TL8-506 (1 µg/mL, InvivoGen) for 1.5 h. Isolated autologous naïve CD4⁺ T cells were directly added into the culture at a ratio of 2:1 (monocyte:CD4⁺) along with SEB (1 µg/mL, Toxins Technology). After 6 days, cells were either processed for T_FH cell sorting for further downstream study or re-stimulated overnight (O/N) with PMA (phorbol-12-myristat-13-acetate, 50 ng/mL, Sigma) and ionomycin (Ion 1 µg/mL, Sigma) as positive control to effectively activate T_FH cells. After 3-4h, Golgi plug (1 µl/mL, Invitrogen) was added and culture continued for additional 18h to investigate T_FH cell proliferation and differentiation related markers by flow cytometry. The culture supernatants were recovered and tested for IL-12p40, IL-18 and IL-21 cytokine quantifications by Luminex multiplex immunoassay.

For T_FH sorting, cells recovered after 6 days of monocyte-naïve CD4 co-culture (n=4) were stained with anti-human PD-1-BV421, ICOS-BV510, CXCR5-BV605, CXCR3-PeCy7 and CD4-APC antibodies. T_FH cell sorting was carried out on BD FACS Aria III.

For naïve B-T_FH co-culture, naïve B cells were isolated (Miltenyi Biotec) from autologous PBMCs of CHB patients (n=4) and co-cultured with sorted T_FH cells (from above) at 2:1 (B:T_FH) in the presence of SEB (1 µg/mL) for 7 days after which T_FH-mediated differentiation of B cells into various subsets was analyzed by flow cytometry.

Purified CD14⁺ monocytes (1 x 10⁶ cells/mL) (N=4) were pre-treated with LEAF purified mouse IgG1, k isotype control, Ultra-LEAF purified anti-human IL-12p40 or Ultra-LEAF purified anti-human IL-18 neutralizing antibodies (each 10 µg/mL, Biolegend) for 2h at 37°C humidified incubator (19), followed by stimulation with TL8-506 or ssRNA40/Lyovec for 1.5h. For recombinant protein stimulation assay, isolated autologous naïve CD4⁺ T cells were treated with recombinant human IL-12 or recombinant human IL-18 (each 20 ng/mL, Biolegend) followed by co-culture experiments as described above.

PBMC from CHB patients obtained at baseline (BL, pre-TLR8) and TLR8 (8h post-Selgantolimod) single oral dose were tested. To examine HBsAg-specific antibody secreting B cells, cells were stimulated with polyclonal stimuli R848 (1 µg/mL, Mabtech) and recombinant human IL-2 (10 ng/mL, Mabtech) and cultured for 5 days to effectively induce memory B cell proliferation. B cell ELISPOT assay was conducted as previously described (6).

Quantification of IL-12p40, IL-18 and IL-21 in cell culture supernatants collected from co-culture experiments was analyzed by Luminex multiplex immunoassay (Luminex LX200 multianalyte system, Bio-Rad Corporation) according to manufacturer protocol.

Quantification of human class-switched immunoglobulin (IgG) was made in cell culture supernatants collected from naïve B-T_FH cell co-cultures using high sensitivity ELISA kit according to manufacturer’s protocol (XpressBio, Express Biotech International).

Statistical analyses were performed using one-way analysis of variance (ANOVA) or Kruskal-Wallis with Dunn’s multiple comparisons test. Paired samples between BL (pre) and TLR8 (8h-post) were analyzed by two-tailed paired, or the Wilcoxon signed-rank Student’s t test or non-parametric Mann-Whitney U test for comparisons between treatments. A p value of <0.05 was considered significant. Levels of significance are indicated by *: *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001 and ns (no significance).

TLR8 signaling induces IL-12 in monocytes, which supports T_FH differentiation in healthy individuals (12). We tested whether this pathway induces T_FH polarizing cytokines in samples from CHB patients. PBMCs isolated from CHB patients were stimulated with various TLR agonists for 18-24h or with medium alone (UT) as negative control. The ligation of TLR8 in monocytes by synthetic agonists, single stranded RNA40 (ssRNA40/LyoVec), TL8-506 and TLR7/8 ligand R848 induced proinflammatory cytokines IL-6, IL-12p40, IL-18, TNF-α in CD14⁺HLA-DR⁺ activated monocytes ( Figures 1A–D ) as analyzed by intracellular cytokine staining using flow cytometry. TLR8 agonist ssRNA40 significantly enhanced the production of cytokines IL-12p40 and IL-18 when compared with other TLR agonists. LPS stimulation induced cytokines IL-6, IL-1β and IL-23p19. Relative frequencies of CD14⁺HLA-DR⁺ monocytes producing these proinflammatory cytokines in response to various TLRs engagement are shown ( Figures 1E–J ). The frequencies of CD14⁺HLA-DR⁺ monocytes or the expression levels of the TLR8 receptor in CD14⁺HLA-DR⁺ monocytes did not change with any of the TLR agonist treatments ( Supplementary Figures 1A–C ). TGF-β1 was induced only by TL8-506 and no significant induction of IL-27p28 was observed ( Supplementary Figures 1D, E ). These data demonstrate that cytokines IL-12 and IL-18 are induced in monocytes from CHB patients when stimulated with TLR8 agonists.

Having shown potent induction of T_FH polarizing cytokine IL-12p40 (12) in monocytes from CHB samples, we next examined the impact of TLR8 agonism on T_FH phenotypes. Circulating CD4⁺CXCR5⁺ cells in peripheral blood share characteristics and functional phenotypes with GC-like T_FH cells (20). The expression of signature cytokine IL-21, BCL-6 (a master regulator of T_FH cells programing and differentiation) and CD40 ligand (CD40L) in T_FH cells provide signals to B cells that are required for Ig class-switching recombination and high affinity antibody productions (21). The expression of inducible co-stimulatory molecule (ICOS) indicates T_FH development, migration and their active state (22). We asked whether TLR8-induced cytokines modulate these immunophenotypic profiles of circulating T_FH cells (cT_FH). CHB patient samples were stimulated with TLR8-specific agonists ssRNA40 or TL8-506 in the presence or absence of mitogenic stimulation using PMA/Ion (required for induction of BCL-6 and ICOS). The sequential gating strategy for T_FH cell panel is presented in Supplementary Figure 2A . Flow plots ( Figure 2A ) show that the frequency of IL-21 producing cT_FH cells (IL-21⁺CXCR5⁺) gated on CD4⁺ T cells was significantly increased in conditions with ssRNA40 or TL8-506. The data represented for ssRNA40 stim or TL8-506 stim are response found in TLR8 agoniost+PMA/Ion re-stimulated response subtracted with PMA/Ion alone response ( Figures 2A–F ). The co-expression of GC-like T_FH markers IL-21⁺BCL-6⁺ and ICOS⁺BCL-6⁺ were significantly increased in ssRNA40 stim or TL8-506 stim conditions ( Figures 2B–F ). Since proinflammatory cytokines IL-12p40 and IL-18 were induced in monocytes in response to TLR8 agonism ( Figure 1 ), we measured levels of these cytokines in the PBMC culture supernatants. As expected, cultures incubated with TLR8-specific agonists ssRNA40 or TL8-506 markedly induced the secretion of both these cytokines ( Figures 2G, H ).

Next, we asked whether TLR8-specific signaling modulates subsets of circulating T_FH cells in CHB patients. When compared with untreated negative control, PBMCs in conditions with ssRNA40 or TL8-506 had significantly increased frequencies of cT_FH subsets cT_FH1 (CXCR3⁺CCR6^-) and T_FH1/17 (CXCR3⁺CCR6⁺) ( Supplementary Figures 2B, D, G ), while no differences were observed for global cT_FH (CXCR5⁺CD4⁺ gated on CD3⁺ T lymphocytes) cells, subsets cT_FH2 (CXCR3^-CCR6^-) and cT_FH17 (CXCR3^-CCR6⁺) ( Supplementary Figures 2C, E, F ). It has been previously reported that cT_FH1 and cT_FH1/17 are the source for IL-21 and IFNγ, whereas cT_FH2 and cT_FH17 are source for anti-inflammatory cytokines IL-4, IL-10 and IL-17A. These results collectively suggest that TLR8-specific signaling induces IL-21 producing T_FH subsets in the setting of CHB.

After in vitro demonstration of T_FH differentiation with TLR8 agonism, we tested samples from CHB patients given a single oral dose of Selgantolimod. Surprisingly, in post-Selgantolimod treated samples that had been obtained only 8 hours after dosing, a significant induction of GC-like signature T_FH factors IL-21⁺BCL-6⁺, ICOS⁺BCL-6⁺ and ICOS⁺CD40L⁺ cells were observed ( Supplementary Figures 3A–H ).

Next, we employed a cytokine-independent method for detection of HBsAg-specific T_FH cell activation by measuring the upregulation of activation induced markers (AIM) CD69, CD25, OX40, PD-L1 and CD40L (16, 17). Both frequencies of CD3⁺CD4⁺CXCR5⁺CXCR3^-PD-1⁺ T_FH and their activation, indicated by surface expression of OX40⁺CD25⁺, PD-L1⁺CD25⁺ and CD69⁺CD40L⁺, was increased in vaccinated individuals upon in vitro stimulation with HBV antigens or SEB compared to the negative control ( Figures 3A–F , Supplementary Figure 4 ). Expectedly, baseline samples from CHB patients displayed low HBV specific activation of T_FH indicating defective response. Importantly, a significant increase in AIM response was evident in 8 hour samples, which was comparable to HBV vaccinated samples ( Figures 3A–F ). Additionally, the frequencies of PD-1⁺CXCR3^- cT_FH cells, but not total CXCR5⁺ cT_FH cells, were significantly higher in samples collected at post-TLR8 time-point ( Supplementary Figures 4B, C ). These data demonstrate that TLR8-specific signaling can normalize defective antigen specific T_FH response.

To determine whether the activation of TLR8 pathway and resulting T_FH differentiation impacts B cell populations in CHB infection, frequencies of various B cell subsets were examined after incubation of PBMC with ssRNA40, TL8-506 or R848 in presence or absence of HBV antigens for 5 days. Frequencies of CD19⁺ B cells did not change with these treatments ( Supplementary Figures 5A–C ). Activation of TLR8 pathway with ssRNA40 or of TLR7/8 with R848 significantly expanded the frequencies of memory B cells (CD19⁺CD27⁺), plasma B cells (CD19⁺CD27⁺⁺CD38⁺), plasmablasts (CD19⁺CD27⁺⁺CD38⁺CD138^-) and plasma cells (CD19^-CD138⁺) ( Figures 4A–E and Supplementary Figures 5D, E ). While TL8-506 treatment did not result in significant changes in frequencies of these populations, it increased naïve B (CD19⁺CD27^-CD21⁺) and long-lived plasma (CD19^-CD138⁺) cells ( Figure 4F and Supplementary Figure 5D ). Addition of HBV antigens in addition to TLR8 agonism didn’t alter naïve B (CD19⁺CD27^-CD21⁺), resting memory B (CD19⁺CD27⁺CD21⁺) or activated memory B cells CD19⁺CD27⁺CD21^-) ( Figures 4A–I and Supplementary Figures 5D, E ).

Previously we reported that atypical memory B cells that expressed high levels of FcRL family receptors were expanded in CHB patients and this sub-population lacks HBsAg-specific B cell response (6). Here we observed that TLR8 stimulation expanded memory B cells with an atypical phenotypic signature (CD19⁺CD21^-CD27^-) ( Figure 4C, I ). We believe that this subset lacking CD21 and CD27 that is expanded in response to TLR7/8 stimulation does not represent an elevation of dysfunctional cells. R848 is typically used to increase memory B cells for measuring B cell responses with ELISPOTs for chronic HBV infection (9, 23). It is also known that during malaria, B cells with similar ‘atypical’ phenotype are in fact functional and produce neutralizing antibodies (24).

In order to discern the mechanism behind TLR8-mediated differentiation and proliferation of T_FH cells, we performed co-culture experiments of isolated monocytes and CD4⁺ T cells. CD14⁺ monocytes were treated with ssRNA40 or TL8-506, in presence of isotype control IgG or anti-human IL-12 or anti-human IL-18 neutralizing (blocking) antibodies followed by co-culture with autologous naïve CD4⁺ T cells for 6 days. Monocytes stimulated with TLR8-specific agonists induced T_FH cell markers on CD4⁺ T cells (CD4⁺CXCR5⁺IL-21⁺, IL-21⁺ICOS⁺ and ICOS⁺PD-1⁺) in samples from healthy individuals and CHB patients. Untreated monocytes co-cultured with autologous naïve CD4⁺ T cells failed to induce IL-21 and other differentiation factors ( Figures 5A–C and Supplementary Figures 6A–C ).

In humans, TLR8 pathway-specific secretion of proinflammatory cytokine IL-12 sends the signal responsible for IL-21 production and drives the T_FH cell differentiation (12, 25). In CHB patients, T_FH cell frequencies are high due to hyperactivation and these cells lack IL-21 cytokine production. We therefore explored whether TLR8-specific cytokine induction will promote circulating T_FH cell differentiation and recover defective IL-21 production in PBMCs of CHB patients. IL-12p40 and IL-18 secretion in monocytes and IL-21 production from newly generated T_FH cells were quantitatively analyzed in supernatants of co-culture experiments and naïve CD4⁺ T cell culture by Luminex multiplex immunoassay. As expected, IL-12 neutralization (ssRNA40+α-IL-12 and more clearly TL8-506+α-IL-12) in CD14⁺ monocytes-naïve CD4⁺ T cells co-culture significantly decreased IL-21 production as well as the co-expression of GC-like T_FH markers IL-21⁺ICOS⁺ and ICOS⁺PD-1⁺, on cT_FH cells (CD3⁺CD4⁺CXCR5⁺); IL-18 neutralization (ssRNA40+α-IL-18 or TL8-506+α-IL-18) did not have similar effect ( Figures 5A–C and Supplementary Figures 6A–C ). Correspondingly, the secretion of cytokine IL-21 was significantly reduced in co-cultures pre-treated with IL-12 neutralizing antibody (ssRNA40+α-IL-12 and more notably with TL8-506+α-IL-12). IL-18 blocking again did not have an effect on IL-21 production in either ssRNA40+α-IL-18 or TL8-506+α-IL-18 antibody treatment conditions ( Figures 5D–F ). IL-18 production remained intact with α-IL-12 treatments. To further justify that TLR8-specific IL-12 cytokine-dependent signal is responsible for T_FH cell differentiation and IL-21 cytokine production, we tested naïve CD4+ T cells isolated from PBMCs of HC and CHB patients that were left untreated or directly incubated with recombinant human IL-12 or IL-18 and cultured for 5 days in the absence of CD14⁺ monocytes. The frequency of IL-21 producing T_FH cells was significantly increased when incubated with rhIL-12, but not with rhIL-18 ( Figures 5A–F and Supplementary Figure 6F ). It is important to note here that effect of TLR8 was not specific to CXCR5⁺ T_FH and the agonist activated CD4⁺ T cells including CXCR5^- cells in an IL-12 dependent manner ( Supplementary Figures 6D, E ). Taken together, these data demonstrate that TLR8-specific IL-12-induction in monocytes plays a critical role in the T_FH cells differentiation, which can restore the deficient IL-21 production in CHB.

To evaluate the utility of newly generated T_FH cells as true helpers of B cells, we sorted CD4⁺CXCR5⁺CXCR3^-PD-1⁺ICOS⁺ T_FH generated with UT, ssRNA40 or TL8-506 pre-exposed CD14⁺ monocytes and assessed their ability to promote plasma B cell differentiation after co-culture with autologous naïve B cells for 7 days. T_FH generated from TLR8 primed monocytes enhanced the differentiation of global memory B (CD27⁺), plasma B (CD27⁺⁺CD38⁺), long-lived antibody-producing plasma cells (CD27⁺⁺CD38^-CD138⁺) and plasmablasts (PBs) (CD27⁺⁺CD38⁺CD138^-), whereas T_FH derived from UT or naïve CD4⁺ T cells co-cultured with B cells failed to induce memory, plasma B and long-live plasma cell differentiation ( Figures 6A–J ).

Additionally, significantly higher frequencies of IgG⁺ (CD19⁺CD27⁺) memory B cells and robust IgG production in TL8-506 primed T_FH-naïve B co-culture condition compared to naïve B without or with naïve CD4⁺ T or UT_T_FH co-culture conditions was present ( Figures 6I, J ). Flow plot in Figure 6D shows ssRNA40 and TL8-506-primed T_FH-naïve B co-culture conditions resulted in a reduction in naïve B (CD19⁺CD21⁺CD27^-) and increase in resting memory B (CD19⁺CD21⁺CD27⁺), activated memory B (CD19⁺CD21^-CD27⁺) and atypical memory B (CD19⁺CD21^-CD27^-) cell populations, relative to non-TLR8 conditions. Thus T_FH differentiated with TLR8-treatment support the generation of memory B cells and their subsets in CHB samples.

To assess whether this mechanism of enhancing T_FH response impacts B cell response in vivo, we next examined the effect of Selgantolimod oral treatment in paired samples from before and 8 hours after treatment. For this experiment, in addition to phase 1b clinical samples, clinical samples from phase 1a trial in CHB negative healthy subjects with known positive HBV vaccination status were tested for HBsAg-specific IgG and total IgG with ELISPOT assays. We defined responders as those who exhibited an increase of ≥1.5-fold SFU/ml after Selgantolimod administration. Using this as criteria, there were 4/11 responders for HBs specific and 6/13 responders for total IgG response among phase 1a samples ( Figures 7A–C ). In samples from CHB patients (phase 1b), HBs specific IgG spots increased in 6/17 subjects and total IgG in 5/18 subjects ( Figures 7D–F ). Most subjects that showed improvement in spots had lower baseline IgG response, though there was no clear cutoff for determining responders. These outcomes suggest that TLR8 treatment has potential to augment memory B cell response in a subset of CHB subjects. It will be important to identify characteristics of responders or non-responders; however, here we did not see positive correlation between T_FH response and ELISPOT response in the small number of responders.