Mice

C57BL/6, CD45.1 (inbred C57BL/6), BALB/c, Thy1.1 (CBy.PL(B6)-Thya/ScrJ), Rag1^−/− (B6.129S7-Rag1tm1Mom/J), CD11c-iDTR (B6.FVB-1700016L2RikTg(Itgax-DTR/EGFP)57Lan/J), CAG-eGFP (C57BL/6-Tg(CAG-EGFP)1Osb/J) and CAG-DsRed (B6.Cg-Tg(CAG-DsRed*MST)1Nagy/j) mice were purchased from Charles River or The Jackson Laboratory. Clec4f-iDTR-YFP mice were generated as previously described³¹. MHC-II^−/− (B6.129S2-H2^dlAb1Ea/J) mice were obtained through the Swiss Immunological Mutant Mouse Repository (Zurich, Switzerland). MUP-core transgenic mice (lineage MUP-core 50, inbred C57BL/6, H-2K^b) have been previously described^6,29. HBV replication-competent transgenic mice (lineage 1.3.32, inbred C57BL/6, H-2K^b) have been previously described²⁵. HBV transgenic mice were used as C57BL/6 or C57BL/6 mice crossed with BALB/c H-2^bxd F1 hybrids. Cor93 TCR transgenic mice (lineage BC10.3, inbred CD45.1) have been previously described²³. Env28 TCR transgenic mice (lineage 6C2.36, inbred Thy1.1 BALB/c) have been previously described²³. Env126 TCR transgenic mice (Rag1^−/− CD45.1/2 C57BL/6 inbred) were generated in this laboratory, in collaboration with BioNTech. For intravital imaging experiments, Cor93 and Env28 TCR transgenic mice were crossbred with the CAG-DsRed lineage and Env126 TCR transgenic mice were crossbred with the CAG-eGFP lineage for homozygous fluorescent protein expression. Mice were housed under specific pathogen-free conditions and used at 8–10 weeks of age. No randomization was used in group allocation; experimental groups were predefined based on genetic background or treatment. In all experiments, mice were matched for age, sex and (for the 1.3.32 animals) serum HBeAg levels before experimental manipulations. In selected experiments, 1.3.32 mice were matched for serum HBV DNA levels before experimental manipulations. All experimental animal procedures were approved by the Institutional Animal Committee of the San Raffaele Scientific Institute and are compliant with all relevant ethical regulations. Mice were maintained on an ad libitum diet (VRF1 (P), Special Diets Services, 801900) in specific pathogen-free-controlled housing, with a 12-h light–dark cycle and stable environmental conditions.

Bone marrow chimeras

Bone marrow chimeras were generated by treatment of HBV transgenic mice with four consecutive doses of 20 mg per kg body weight of busulfan (1,4-butanediol dimethanesulfonate; Sigma-Aldrich, B2635) before receiving the new bone marrow (6 h after last dose); mice were allowed to reconstitute for at least 8 weeks before use.

Single-cell sorting

For Env126 transgenic mouse generation, WT C57BL/6 mice were immunized weekly with 100 µg Env_126–138 peptide and 50 µg polyI:C for 3 weeks by subcutaneous injection. Seven days after the third injection, 5 × 10⁶ splenocytes were cultured for 15 h in the presence of 4 µg ml⁻¹ peptide or with 2 µg ml⁻¹ concanavalin A (Sigma-Aldrich). Activated cells were labeled using IFNγ-secretion assay-APC kit and were further enriched via anti-APC-microbeads according to the manufacturer’s protocol (Miltenyi, 130-090-433, 130-090-855) using the magnetic-activated cell sorting technique. CD4⁺CD8⁻IFNγ^hi single T cells (anti-CD4, clone GK1.5, eBioscience, 48-0041-82 and anti-CD8α clone 53-6.7, BD Biosciences, 560777) were sorted (BD FACSAria) into wells of 96-well V-bottom plates (Greiner Bio-One) containing 6 μl of a mild hypotonic cell lysis buffer per well consisting of 0.2% Triton X-100, 0.2 μl RiboLock RNase inhibitor (Thermo Scientific, EO0381), 5 ng poly(A) carrier RNA (Qiagen, 1068337) and 1 μl dNTP mix (10 mM, Biozym, 331520) in RNase-free water. Plates were sealed, centrifuged and stored at −65 °C to −85 °C directly after sorting.

For Env126 CD4⁺ T_N cell and T_eff cell sorting, single-cell suspensions were prepared from mouse liver and spleen according to previously described protocols. Cells were subsequently stained with the following panel: LIVE/DEAD Fixable Far-Red dye (Invitrogen, L34973), CD11b-eFluor450 (M1/70, eBioscience, 48-5698-82), CD19-eFluor450 (1D3, eBioscience, 48-0193-82), NK1.1-Pacific Blue (PK136, BioLegend, 108722), CD8a-eFluor450 (53-6.7, eBioscience, 48-0081-82), CD45.1-Alexa Fluor 488 (A20, BioLegend, 110718), CD45.2-PE (104, BioLegend, 109808), Ter119-PE-Cy7 (TER-119, BioLegend, 116222), B220-PE-Cy7 (RA3-6B2, BD Bioscience, 552772) and CD4-Alexa Fluor 647 (GK15, BioLegend, 100426). Cells were resuspended at a concentration of approximately 15–20 × 10⁶ cells per ml in sorting buffer (PBS with 0.04% BSA Invitrogen UltraPure, AM2616). Live CD4⁺ CD45.1⁺, CD45.2⁺, CD11b⁻, CD19⁻, NK1.1⁻, CD8a⁻, Ter119⁻ and B220⁻ cells originally from Env126 CD45.1.2 mice were sorted using a BD FACSAria cell sorter equipped with a 100-µm nozzle at 20 PSI and maintained at 4 °C. Sorted target cells were collected in 1.5-ml microcentrifuge tubes pre-coated overnight with 1% UltraPure BSA at 4 °C. The final cell purity exceeded 98%.

Cloning and selection of antigen-specific TCRs

Plates with sorted CD4⁺CD8⁻IFNγ^hi T cells were thawed, and template-switch cDNA synthesis was performed as described⁶⁵ with adapted primers specific to murine TCRα and TCRβ constant genes. TRA and TRB fragments were further amplified⁶⁶, sequenced and the respective V(D)J junctions analyzed using the IMGT/V-Quest tool. DNAs of novel and productively rearranged corresponding TCR chains were digested using NotI and cloned into pST1 vectors⁶⁷ containing the appropriate constant region for in vitro transcription of complete TCRα/β chains. The reactivity of the isolated TCRα/β chains was validated in active C57BL/6 T cells transfected with respective TCR mRNA, using Env_126–138 pulsed bone marrow-derived DCs in an IFNγ ELISpot assay, as previously described⁶⁸. TCR7 (alpha chain: V7-4*2orV7D-4J15C and beta chain: V4D2J1-6C1) was used for the generation of the Env126 Tg mouse.

Viruses and viral vectors

Mice were injected intravenously (i.v.) with 10⁶ plaque-forming units of replication-competent recombinant VSV expressing HBsAg (rVSV^Env) vector 24 h after Env126 CD4⁺ T cell adoptive transfer. All infectious work was performed in designated BSL-3 workspaces, in accordance with institutional, national and international guidelines.

T cell isolation, adoptive transfer and in vivo treatments

In selected experiments, WT C57BL/6 mice were immunized subcutaneously with 100 μg of Env_126–138 peptide and 50 μg of polyI:C immune adjuvant at three time points every 7 days.

Isolation of CD8⁺ T cells from spleens of Cor93 or Env28 mice and CD4⁺ T cells from spleens of Env126 transgenic mice were performed as previously described⁶. Isolation of CD4⁺ T cells from spleens of Env126 transgenic mice was performed by negative immunomagnetic selection using the EasySep Mouse CD4⁺ T Cell Isolation Kit (StemCell Technologies). Mice were injected i.v. with 10⁶ HBV-specific (Cor93 or Env28) naive CD8⁺ T cells and/or HBV-specific effector CD4⁺ T cells (Env126 T_eff cells). In the indicated experiments, mice were splenectomized and treated i.v. with 200 μg of anti-CD62L monoclonal antibody (clone MEL-14, BioXcell) 48 h and 4 h before cell injection and 24 h after cell transfer. Splenectomy was performed according to standard procedures⁶⁹ and as previously described⁶. For FTY720 treatment, mice were injected with 1 mg per kg body weight of fingolimod (Sigma) intraperitoneally (i.p.) every 48 h throughout the experiment. In selected experiments, mice were treated i.p. with busulfan and reconstituted i.v. with bone marrow from CD11c-iDTR-GFP mice (10⁷). In selected experiments, DCs and KCs were depleted by injecting i.p. 25 ng per gram body weight of diphtheria toxin (Sigma-Aldrich) 3 days and 1 day before T cell transfer and 1 day and 3 days after cell transfer. In the indicated experiments, KCs were depleted by i.v. injection of clodronate-containing liposomes (10 μl per gram body weight; Liposoma B’V) 2 days before T cell injection^70,71. In the indicated experiments, mice were treated i.p. with 250 μg of anti-CD40L monoclonal antibody (clone MR-1) or 250 μg anti-CD40L monoclonal antibody (clone D265A), 250 μg of anti-IL-12 p40 monoclonal antibody (clone C17.8), 250 μg of anti-IL-27 p28 monoclonal antibody (clone MM27.7B1) and 2.5 μg of rIL-27 (Fc: LALA-PG-KIH heterodimer, Adipogen). All monoclonal antibodies were purchased from BioXcell, unless otherwise indicated.

In vitro cell activation

In the indicated experiments, Env126 CD4⁺ T cells were cultured in vitro under 4 h of stimulation in lymphocyte complete medium: RPMI 1640, penicillin–streptomycin (100 IU ml⁻¹ and 100 µg ml⁻¹, Corning), 2 mM l-glutamine (Corning), 50 µM 2-mercaptoehanol (Sigma-Aldrich), HEPES 10 mM (Corning), NEAA (non-essential amino acids, Corning) at 100 µM. HBsAg_126–138 was used at a concentration of 1 µg ml⁻¹. Ionomycin calcium salt (Sigma-Aldrich), was used at a concentration of 1 µg ml⁻¹ in combination with PMA (Sigma-Aldrich) at a concentration of 50 ng ml⁻¹. In the indicated experiments, Env126 CD4⁺ T cells were cultured in vitro for 48 h in complete medium with HBsAg_126–138 peptide (1 µg ml⁻¹) or with Dynabeads Mouse T-Activator CD3/CD28 (Thermo Fisher Scientific, 11456D) following the manufacturer’s instructions. For ex vivo intracellular staining to assess cytokine production before FACS analyses, cell suspensions were incubated for 4 h at 37 °C in complete medium, with HBsAg_126–138, HBsAg_28–39 or HBcAg_93–100 at 2 µg ml⁻¹ with 1 µg ml⁻¹ brefeldin A (Sigma-Aldrich) and Monensin (BioLegend) according to the manufacturer’s instructions. In selected experiments, in vitro cultured CD8⁺ T cells were stimulated for 48 h with rIL-27 at 25 ng ml⁻¹ (R&D Systems, 7430-ML).

In vitro cell labeling

In the indicated experiments, WT or Env126 CD4⁺ T cells were labeled with CellTrace Violet Cell Proliferation Kit (Invitrogen, C34571) following the manufacturer’s protocol. Briefly, WT or Env126 CD4⁺ T cells were isolated, washed and resuspended in PBS at a concentration of 10⁶ cells per ml in the working dye solution for 20 min at 37 °C. After the incubation, complete medium was added to the cell suspension that was rested at room temperature for 5 min to dilute and remove free remaining dye. Labeled cells were then resuspended in complete medium ready to be cultured.

Cell isolation and flow cytometry

Briefly, cells were plated at 2 × 10⁶ cells per well in a 96-well U-bottom tissue culture plate with cognate peptide (2 µg ml⁻¹) for 4 h in an incubator in complete RPMI medium containing 10 µg ml⁻¹ of brefeldin A and 10 U ml⁻¹ of IL-2 (EL-4 supernatant). At the end of the stimulation, cells were washed and stained as described above. Cell viability was assessed by staining with Viability 405/520 fixable dye (Miltenyi, 130-130-404) and LIVE/DEAD Fixable Far-Red dye (Invitrogen, L34973). In each experiment, Env28 CD8⁺ T cells were identified as live, CD45⁺, CD8⁺CD4⁻ Thy1.1⁺; Env126 CD4⁺ T cells were identified as live, CD45⁺, CD8⁻CD4⁺ CD45.2⁺CD45.1⁺; and Cor93 CD8⁺ T cells were identified as live, CD45⁺, CD8⁺CD4⁻ CD45.1⁺.

Staining of cell surface markers was performed with Brilliant Stain buffer (BD Biosciences, 566349) according to the manufacturer’s instructions. Antibodies are described in this order: target, clone name, catalog number and dilution to ensure precise identification of the reagent used. Antibodies from BD Biosciences: anti-CD45.2 (104, 564616; 1:200 dilution), anti-Thy1.1 (OX-7, 740917; 1:200 dilution), anti-CD8 (53-6.7, 612898; 1:300 dilution), anti-B220 (RA3-6B2, 564662; 1:300 dilution), anti-CD19 (1D3, 562291; 1:200 dilution), anti-CD3 (145-2C11, 562286; 1:100 dilution), anti-Ly6G (1A5, 562700; 1:300 dilution), anti-CD49b (DX5, 562453; 1:100 dilution), anti-CD44 (IM7, 569705, 560781; 1:200 dilution), anti-CD69 (H1.2F3, 612793, 552879; 1:100 dilution), anti-CD86 (GL1, 564199; 1:100 dilution), anti-IA-IE (M5/114.15.2, 748846, 2G9, 569244; 1:300 dilution), anti-H2-kB (AF6-88.5, 742861; 1:200 dilution), anti-CD40 (44986, 745218; 1:100 dilution), anti-TIM4 (21H12, 742774; 1:100 dilution), anti-CD28 (37.51, 740466; 1:100 dilution), anti-CD70 (FR70, 740931; 1:100 dilution), anti-Ly6A/E (D7, 756372; 1:100 dilution), anti-Ly108 (13G3, 755697; 1:100 dilution), anti-TIGIT (1G9, 744212; 1:100 dilution), anti-NKG2A/C/E (20d5, 740549; 1:100 dilution), anti-2B4 (2B4, 740671; 1:100 dilution), anti-Bcl-6 (K112-91, 562401; 1:75 dilution), anti-IFNγ (XMG1.2, 562333; 1:200 dilution), anti-T-bet (O4-46, 569089; 1:75 dilution) and anti-TNF (MP6-XT22, 563943; 1:100 dilution). From BioLegend: anti-CD45 (30-F11, 103108; 1:400 dilution), anti-CD45.1 (clone A20, 110743; 1:300 dilution), anti-CD4 (GK1.5, 100480; 1:200 dilution), anti-CD62L (MEL-14, 104438, 104453; 1:200 dilution), anti-CD80 (16-10A1, 104738; 1:100 dilution), anti-CD25 (PC61, 102075, 102020; 1:100 dilution), anti-CD11c (QA18A72, 161107; 1:100 dilution), anti-CD11b (M1/70, 101285; 1:100 dilution), anti-CD44 (IM7, 103028; 1:200 dilution), anti-F4/80 (BM8, 123110, 123130; 1:100 dilution), anti-CD64 (X54-5/7.1, 139311; 1:100 dilution), anti-CD69 (H1.2F3, 104527; 1:100 dilution), anti-ICAM-1 (YN1/1.7.4, 116116; 1:100 dilution), anti-ESAM (1G8/ESAM, 136203; 1:100 dilution), anti-CD206 (C068C2, 141719; 1:100 dilution), anti-Lag-3 (C9B7W, 125248; 1:100 dilution), anti-CD39 (Duha59, 143819; 1:100 dilution), anti-CD31 (390, 102424; 1:100 dilution), anti-IL-2 (JES6-5H4, 503837; 1:100 dilution), anti-CD107a (1D4B, 121608; 1:80 dilution), anti-ICOS (15F9, 107716; 1:100 dilution), anti-PD-1 (29F.1A12, 125253; 1:100 dilution), anti-NK1.1 (S17016D, 156529; 1:100 dilution), anti-KLRG1 (2F1/KLRG1, 138410; 1:100 dilution), anti-CXCR3 (S18001A, 155923; 1:100 dilution), anti-CTLA-4 (UC10-4B9, 106314; 1:75 dilution), anti-4-1BB (17B5, 107105; 1:100 dilution), anti-OX40 (OX-86, 119414; 1:100 dilution), anti-TNF (MP6-XT22, 506313; 1:200 dilution), anti-TIM3 (RMT3-23, 119738; 1:100 dilution), anti-CXCR5 (L138D7, 145532; 1:75 dilution), anti-CXCR6 (SA051D1, 151111; 1:100 dilution), anti-IFNγ (XMG1.2, 505830; 1:200 dilution). From eBioscience: anti-CD4 (RM4-5, 48-0042-82; 1:200 dilution), anti-CD27 (LG.7F9, 25-0271-80; 1:100 dilution), anti-Ki-67 (SolA15, 48-5698-82; 1:400 dilution). Antibodies from Invitrogen: anti-Foxp3 (FJK-16s, 53-5773-82; 1:100 dilution), anti-CD11c (N418, 25-0114-82; 1:100 dilution), anti-T-bet (4b10, 25-5825-80; 1:75 dilution), anti-TOX (TXRX10, 50-6502-82; 1:100 dilution), anti-Grzm-B (GB11, GRB04, GRB05; 1:80 dilution). Human antibodies included: anti-CD3 (clone HIT3a, BioLegend, 300306; 1:40 dilution), anti-CD4 (clone SK3, BD Bioscience, 563875; 1:20 dilution), anti-CD8 (clone RPA-T8, BD Bioscience, 557746; 1:100 dilution) and anti-IFNγ (clone B27, BioLegend, 506510; 1:20 dilution).

Recombinant dimeric H-2L^d–Ig or H-2K^b–Ig fusion proteins (BD Biosciences) complexed with peptides derived from HBsAg (Env_28–39, IPQSLDSWWTLS, Primm) or from HBcAg (Cor_93–100, MGLKFRQL, Primm), respectively, were prepared according to the manufacturer’s instructions. Dimer staining was performed as described and used to quantify HBV-specific CD8⁺ T cells (Env28 and Cor93)⁶. Peptide HBs_126–138-loaded tetramer (I-A^b Env_126–138) and human CLIP peptide-loaded negative control (I-A^b CLIP) were provided by the National Institutes of Health (NIH) Tetramer Core Facility. Tetramer staining was performed by incubating cell suspensions at 37 °C for 1 h according to the manufacturer’s instructions. Flow cytometry staining for all transcription factors and intranuclear proteins was performed using Foxp3/Transcription Factor Staining Buffer Set (eBioscience, 00-5523-00), following the manufacturer’s instructions. All flow cytometry analyses have been performed in FACS buffer comprising PBS with 2 mM EDTA and 2% FBS (Corning). Flow cytometry analyses were performed on the BD FACSCanto Clinical Flow Cytometry System, BD FACSymphony A5 Cell Analyzer or Cytek Aurora. Flow cytometry data were analyzed with FlowJo software (TreeStar, v10).

Southern blot analysis

Genomic DNA was isolated from frozen livers using the phenol–chloroform method and analyzed for intrahepatic HBV DNA contents by Southern blotting, as previously described⁷. No molecular weight markers were included, as the migration pattern of HBV DNA forms, including RC, DS, SS DNA and integrated TG, has been well characterized²⁵. Band identity was assigned based on electrophoretic mobility and comparison to these widely recognized standards. All blots are shown for qualitative purposes only.

Biochemical analysis

The extent of hepatocellular injury was monitored by the measuring of sALT activity at different and multiple time points after experimental manipulation and cell transfer, as already described⁶. Normal intervals of sALT activity are between 20 and 80 U l⁻¹.

RT–PCR

Total RNA was extracted from frozen livers using the ReliaPrep RNA Tissue Miniprep System (Promega), according to the manufacturer’s instructions, as described⁶. Genomic DNA contamination was removed using TURBO DNA-free DNase (Ambion). Around 1 μg of total RNA was reverse transcribed with Superscript IV Vilo (Life Technologies) before qPCR analysis for mouse Il27 (TaqMan Mm00461163, Life Technologies) and Il12b (TaqMan Mm 01288989, Life Technologies). All experiments were performed in triplicate and normalized to the reference gene GAPDH (TaqMan Mm99999915, Life Technologies). All nucleic acids were purified from serum using the MiniElute Virus Spin Kit (Qiagen). Viremia was determined by plotting a standard curve using HBV plasmid DNA and the Core TaqMan probe (forward TACCGCCTCAGCTCTGTATC, reverse CTTCCAAATTAACACCCACCC, probe TCACCTCACCATACTGCACTCAGGCAA). Reactions were run and analyzed on the QuantStudio 5 instrument (Life Technologies).

Confocal immunofluorescence and histochemistry

At the time of autopsy, mouse livers were perfused via the inferior vena cava with 10 ml of PBS as liver pieces were surgically removed and instantly fixed in paraformaldehyde at 4% concentration for 16–20 h. The solutions were then dehydrated in 30% sucrose before their embedding into OCT freezing media (Bio-Optica). Liver sections were obtained with a Leica CM1520 Cryostat (20 µm per section). Then, liver sections were stained as already described⁶. Antibodies used included anti-F4/80 Alexa Fluor 647 (Invitrogen, MF48021) and anti-CD38 Brilliant Violet 421 (BioLegend, 102732). All images were acquired using a Mavig RS-G4 confocal microscope.

For all the immunohistochemical staining, liver pieces were harvested in zinc formalin solution and transferred into 70% ethanol solutions 24 h later. All the tissues were then embedded in paraffin and stained as previously described⁶. Primary antibodies included monoclonal rat anti-Ki-67 (clone TEC-3, Dako, M7249), polyclonal rabbit anti-HBcAg (Dako, B0586), monoclonal rabbit anti-cytokeratin 7 (clone ERP17078, Abcam, ab181598), monoclonal rabbit anti-CD3 (clone SP7, Abcam, ab16669) and polyclonal rabbit anti-cleaved caspase-3 (Asp175, CST, 9661). Bright-field images were acquired using an Aperio Scanscope CS2 microscope and ImageScope software (Leica Biosystems), following the manufacturer’s guidelines. In each experiment, histochemical analysis was conducted for every mouse using the median-left liver lobe, following a standardized protocol established in collaboration with our institutional histopathology core facility. The entire lobe section was scanned to avoid selection bias and ensure inclusion of all tissue regions. For qualitative assessment, regions of interest were selected from these whole-lobe scans based on tissue integrity and staining quality, each covering an area of approximately 600 µm². These regions of interest were chosen to capture diverse hepatic microenvironments (for example, periportal, pericentral and mid-zonal regions), rather than isolated or extreme features. Selection was performed independently by two investigators (blinded to condition) and confirmed by a pathologist to ensure consistency and reproducibility. All HBcAg IHC stainings are presented as a qualitative assessment of antigen distribution, illustrating its presence or absence across conditions. The number of animals and sections analyzed per group is indicated in each figure legend. Immunohistochemical imaging analyses were performed with QuPath 0.5.0-x64 software.

Intravital multiphoton microscopy

Liver intravital multiphoton microscopy was performed as described⁶. Liver sinusoids were visualized by injecting nontargeted Quantum Dots 655 (Invitrogen) i.v. during image acquisition. Images were acquired with a LaVision BioTec TriMScope II coupled to a Nikon Ti-U inverted microscope enclosed in a custom-built environmental chamber (Life Imaging Services) that was maintained at 37–38 °C with heated air. Fluorescence excitation was provided by two tunable femtosecond-pulsed titanium–sapphire lasers (680–1,080 nm, 120-fs pulse width, 80-MHz repetition rate, Ultra II, Coherent), an Optical Parametric Oscillator (1,000–1,600 nm, 200-fs pulse width, 80-MHz repetition rate, Chameleon Compact OPO, Coherent). The setup includes four non-descanned photomultiplier tubes (Hamamatsu H7422-40 GaAsP High Sensitivity PMTs and Hamamatsu H7422-50 GaAsP High Sensitivity red-extended PMT from Hamamatsu Photonics K.K.) and a ×25, 1.05-NA, 2-mm working distance, water-immersion multiphoton objective (Olympus). For four-dimensional analysis of cell migration, stacks of 5–13 square xy sections (1,024 × 1,024 pixel) sampled with 4-μm z-spacing were acquired every 20 s for 20 min. Sequences of image stacks were transformed into volume-rendered, four-dimensional time-lapse movies with Imaris (Bitplane, v9.1.2). The three-dimensional (3D) positions of the cell centroids were segmented by the semiautomated cell tracking algorithm of Imaris.

The single-cell speed was calculated as the mean 3D velocity per cell track, while the instantaneous velocity represents the 3D velocities of each single step in a track. The arrest coefficient was defined as the percentage of time a track instantaneous velocity remains below a given motility threshold (here 5 μm min⁻¹). The track displacement (distance between the initial and the final position of a cell) was divided by the total track length to obtain the corrected straightness, a measure of the directionality of a track. Finally, instantaneous speed, turning angle changes, mean displacement versus time and the motility coefficient were calculated from the x, y and z coordinates of the cell centroids using custom-designed scripts in RStudio.

scRNA-seq library preparation

Env126 CD4⁺ T_eff cells and Env126 CD4⁺ T_N cells were fixed according to the Evercode Fixation (Parse Biosciences) protocol, and single-cell RNA libraries were prepared using the Evercode split-pool combinatorial barcoding technology approach using the ParseBio Evercode WT v2 kit. Sequencing was performed on the Illumina NovaSeq 6000 sequencer with the S4 flow cell 200 cycle kit according to the sequencing protocol provided by the Evercode kit (v1.5) at a desired depth of 30,000 reads per cell.

scRNA-seq data analysis

Parse Biosciences analysis pipeline ‘split-pipe’ (v.1.1.0p) was used with default parameters to perform demultiplexing, alignment of sequencing reads against the GRCm38.93 genome and generation of the cell-by-gene count matrix. A raw counts matrix of the aggregated libraries containing the counts for all the samples was imported into R (version 4.2.3) and processed with custom scripts. The ‘isOutlier’ function from the ‘scuttle’ package⁷² was used to identify outlier and low-quality cells in an unbiased way. For each of the following features, cells with values greater or lower than four median absolute deviations were discarded: library size, number of features, percentage of mitochondrial gene contamination, percentage of ribosomal gene contamination, percentage of hemoglobin gene contamination, percentage of platelet gene contamination. After quality filtering, only cells with a number of features greater than 200 and genes expressed in at least three cells were kept. Doublet removal was performed with the scDblFinder R package⁷³; only the cells automatically classified as a ‘singlet’ by the algorithm in the package were kept. Cells that were not CD4⁺ T cells were discarded from each sample individually by calling the ‘SingleR’ function using the ‘celldex MouseRNAseqData’ dataset^74,75 aided by manual curation. A Seurat (version 4.3.0)⁷⁶ object was created for each filtered sample and merged into a single dataset. The complete dataset was normalized and scaled following the standard Seurat workflow (scaling to 10,000 and log transformation log1p). The top 3,000 genes identified by Seurat’s FindVariableFeatures using the vst method were used for principal component analysis. Uniform manifold approximation and projection reduction was then applied on the first 30 principal components. Graph-based clustering was performed to cluster cells according to their gene expression profile with the FindClusters function in Seurat with default parameters, and a resolution value of 0.2 was chosen. Differential gene expression was calculated with the FindAllMarkers function in Seurat. Gene-set or signature scores were obtained with the AddModuleScore Seurat function. CD4⁺ T cells signatures were obtained from Kiner et al.⁷⁷. Heat maps were prepared with the ComplexHeatmap package; feature plots were generated with the function FeaturePlot_scCustom from the scCustomize R package^78,79. Gene-set enrichment analysis was performed using the fgsea R package (v.1.8.0)⁸⁰ on all genes ranked by log₂ fold change (from FindMarkers Seurat output) from the selected comparison. P values were calculated using 1,000,000 permutations and were corrected using the Benjamini–Hochberg method. Gene sets with a false discovery rate <0.05 were considered significant. R packages used for data analysis were scuttle (v.1.8.4), scDblFinder (v1.12.0), SingleR (v2.0.0), Seurat (v4.3.0), scCustomize (v1.1.3), celldex (v1.8.0).

Study participants and approval

A total of 12 individuals with chronic HBV infection (HBsAg⁺) were included (Supplementary Table 2), of whom 4 had HBeAg-positive and 8 had HBeAg-negative chronic HBV infection. All 12 individuals were treatment-naive, had no known history of hepatitis (normal ALT) and were non-cirrhotic with a normal FibroScan result. Supplementary Table 2 summarizes the available clinical and virological parameters. Blood donors were recruited from the viral hepatitis clinic at The Royal London Hospital. Written informed consent was obtained from all participants. The study was conducted in accordance with the Declaration of Helsinki and approved by the Barts and the London NHS Trust local ethics review board and the NRES Committee London–Research Ethics Committee (reference 10/H0715/39) and by the Singapore National Healthcare Group ethical review board (DSRB 2008/00293).

Clinical and virological parameters

On recruitment to the study, viral serology and HBV DNA levels were tested. HBsAg, HBeAg and anti-HBe levels were measured with a chemiluminescent microparticle immunoassay (CMIA; Architect Assay, Abbott Diagnostics). HBV DNA levels in serum were quantified by real-time PCR (COBAS AmpliPrep/COBAS TaqMan HBV test v2.0; Roche Molecular Diagnostics).

HBV peptide library

A set of 15-mer peptides overlapping by ten amino acids was used to identify HBV-specific T cells. The peptides covered the sequences of the HBV protein core, envelope (S) and polymerase (Pol) of genotypes A, B, C and D and were purchased from T Cell Diagnostics. Similarly, 15-mer peptides overlapping by ten amino acids that covered the sequence of hCMV protein UL55 were used to identify CMV-specific T cells and were purchased from GenScript.

PBMC isolation and T cell culture

PBMCs were isolated from peripheral blood by Ficoll gradient and cryopreserved. Cells were thawed and T cell lines were generated as follows: 20% of PBMCs were pulsed with 10 μg ml⁻¹ of the overlapping HBV or CMV peptides for 1 h at 37 °C, subsequently washed, and co-cultured with the remaining cells in AIM-V medium (Gibco; Thermo Fisher Scientific) supplemented with 2% AB human serum (Gibco; Thermo Fisher Scientific). T cell lines were cultured for 10 days in the presence of 20 U ml⁻¹ of rIL-2 (R&D Systems) with or without 50 ng ml⁻¹ of rIL-27. Half of the medium was refreshed on day 4 and day 8.

ELISpot assays

ELISpot assays for the detection of IFNγ-producing cells were performed on in vitro-expanded T cell lines using either HBV peptide pool core, envelope (S), polymerase 1 (Pol 1), polymerase 2 (Pol 2) or CMV peptide pool UL55. T cell lines were incubated overnight at 37 °C with pools of HBV/CMV peptides (2 μg ml⁻¹), where final dimethylsulfoxide concentrations did not exceed 0.2%. IFNγ ELISpot assays (Millipore) were performed as previously described⁸¹. Spot-forming units were quantified with ImmunoSpot 7.0.26.0.

Liver histology from HBV-infected chimpanzees

Liver tissue from an HBV-infected chimpanzee (A3A005) was analyzed. All relevant details regarding sex, age, and body weight before infection are available in the published dataset⁹. The animal was handled in accordance with humane care and use guidelines established by the Animal Research Committees at the NIH, The Scripps Research Institute and Bioqual Laboratories. It was housed individually at Bioqual Laboratories, an AAALAC International-accredited institution under contract with the National Institute of Allergy and Infectious Diseases. The chimpanzee was inoculated with 10⁴ genome equivalents of HBV, a dose that typically results in a self-limited infection⁹. Liver tissue samples, measuring 5 to 10 mm in length, were obtained via needle biopsy and fixed in 10% zinc formalin for histological analysis. Formalin-fixed paraffin-embedded liver sections were subjected to multiplex immunofluorescence staining using the Opal detection system (Akoya Biosciences) and visualized with fluorescence microscopy. Sections were first deparaffinized and subjected to heat-induced epitope retrieval using Bond Epitope Retrieval Solution (Leica Biosystems). Immunostaining was performed on an automated Leica Bond RX platform using ready-to-use primary antibodies for CD4 (clone 4B12, Leica PA0427), CD8 (clone C8/144B, Dako) and CD68 (clone KP1, Ventana 790-2931). Images were acquired using an HT 2.0 Akoya scanner at a magnification of ×20. Image analysis was performed using Phenochart software (Akoya Biosciences) to assess spatial interactions within the hepatic microenvironment.

Statistical analyses

All results are expressed as means ± s.e.m. All statistical analyses were performed in Prism (GraphPad Software, v10). Statistical tests, sample size and P values are indicated in each figure. No statistical methods were used to predetermine sample sizes, but our sample sizes are similar to those reported in previous publications^6,7,24. Data distribution was assumed to be normal, but this was not formally tested. Data collection and analysis were not performed blind to the experimental conditions. No data were excluded from the analyses unless classified as outliers using the ROUT method (Q = 5%), which combines robust regression and false discovery rate control. This method was used to objectively remove technical artifacts while minimizing subjective bias.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.