Genetically-modified, redirected T cells target hepatitis B surface antigen-positive hepatocytes and hepatocellular carcinoma lesions in a clinical setting

Article information

Clin Mol Hepatol. 2024;30(4):735-755

Publication date (electronic) : 2024 May 29

doi : https://doi.org/10.3350/cmh.2024.0058

Xueshuai Wan¹, Karin Wisskirchen², Tao Jin², Lu Yang², Xiaorui Wang², Xiang’an Wu¹, Fang Liu¹, Yu Wu¹, Christy Ma², Yong Pang², Qi Li², Ke Zhang², Ulrike Protzer^,3, Shunda Du^,1

¹Department of Liver Surgery, Peking Union Medical College Hospital, PUMC, and Chinese Academy of Medical Sciences, Beijing, P. R. China

²SCG Cell Therapy Pte. Ltd., Singapore, Singapore

³Institute of Virology, School of Medicine, Technical University of Munich/Helmholtz Munich, Munich, Germany

Corresponding author : Shunda Du Department of Liver Surgery, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, 1 Shuaifuyuan, Wangfujing, Beijing 100730, P. R. China Tel: +86-10-69152836, Fax: +86-10-69156043, E-mail: dushd@pumch.cn

Ulrike Protzer Institute of Virology, Technical University of Munich/Helmholtz Munich, Trogerstrasse 30, 81675 Munich, Germany Tel: +49-89-4140-6821, Fax: +49-89-4140-6823, E-mail: protzer@tum.de

Editor: Su Hyung Park, Korea Advanced Institute of Science and Technology (KAIST), Korea

Received 2024 January 20; Revised 2024 May 28; Accepted 2024 May 28.

Abstract

INTRODUCTION

Hepatocellular carcinoma (HCC) is the sixth most common malignancy and the third leading cause of cancer-related mortality worldwide [1]. Systemic HCC therapies encompass immune checkpoint, angiogenesis, or tyrosine kinase inhibitors, but response rates remain limited [2]. The risk of HCC recurrence is high because the therapies do not target the underlying cause of HCC development.

Hepatitis B virus (HBV) infection is the primary risk factor for HCC development, accounting for ~50% of HCC worldwide and ~85% in China [3]. Despite the implementation of vaccination programs, the projected number of liver cancer cases in China remains high, with 300,000 total cases predicted for 2030 [3]. HBV-DNA integration is detected in 86% of HBV-related HCC tumor cells [4]. HBV-DNA integration contributes to tumorigenesis and can result in persistent expression of either complete hepatitis B surface antigen (HBsAg) or fragments thereof depending on the integration sites. This renders HBsAg an interesting therapeutic target as HBsAg is clearly distinct from proteins expressed in healthy liver tissue. As such, it represents an ideal target for adoptive cell therapy (ACT) of HBV-related HCC, as it is also able to address the underlying cause of HCC development. Clonal expansion of hepatocytes with HBV-DNA integration occurs early after infection and increases with disease progression [5,6]. Consequently, targeting these potentially premalignant cells seems crucial to prevent further HCC progression and the formation of new lesions.

HBV breakpoints resulting in HBV-DNA integration most commonly localize within a 1,800 bp region of the HBV genome, preserving the structural integrity of the envelope, X and polymerase open reading frames [4]. In advanced stages of chronic hepatitis B (CHB) and HCC, a substantial proportion of HBsAg appears to emanate from integrated DNA rather than from active viral infection [7], and particularly ground glass hepatocytes contain abundant HBsAg [8]. The number of HBsAg⁺ cells determined by histological staining is variable, ranging from 8–42% within tumor tissue and 71–86% in adjacent non-tumor tissue [9-12]. Even in histologically negative HCC samples, HBsAg RNA can be detected via qPCR and Nanostring technology [13].

Immunopeptidomics analysis of HBV-HCC samples confirmed that most of the HBV peptides presented on tumor cells are derived from envelope and polymerase proteins [14,15] including the HBV peptide S₂₀ (also referred to as Env₁₈₃) presented on human leukocyte antigen (HLA)-A*02 [14]. T cells genetically modified to express HBV-specific TCRs can be redirected to recognize HBV-positive target cells [16,17]. Out of a library of HBV-specific TCRs, the most sensitive and highly specific, HLA-A*02-restricted TCR recognizing the HBV S₂₀ peptide was selected for preclinical and clinical development [17]. T cells expressing this HBV-specific TCR secrete pro-inflammatory and antiviral cytokines, and selectively eliminate HBV-positive hepatoma cells and HBV-infected target cells, thereby clearing viral infection [18].

In earlier clinical studies, T cells transiently expressing an HBV-specific TCR after RNA-electroporation were used to target HBV-HCC metastases [13]. In that setting, HLA-A2-restricted, HBV-specific T cells were repeatedly applied to patients that had been transplanted with an HLA-mismatched liver, thereby only targeting metastases of the original tumor but not the transplanted liver [13]. Given the mere mass of malignant cells within a solid tumor such as HCC, we hypothesized that tumor targeting and efficient killing of tumor cells would require a stable expression of an HBV-specific TCR to allow T cells to expand while maintaining their specificity and effector function. Therefore, a Good Manufacturing Practice (GMP)-compliant, semi-automated protocol for generating and upscaling an autologous T-cell product by lentiviral transduction and expansion was developed. Functionality and safety of the HBV-specific TCR T cell product (product code SCG101) were evaluated in preclinical models. Furthermore, for the first time, these stably TCR-expressing autologous cells were studied in a setting of primary HBV-related HCC and concomitant chronic hepatitis B. We here report the application of a single dose of SCG101 to an HLA-A*02:01-positive patient within an investigator-initiated clinical trial (NCT05339321).

MATERIALS AND METHODS

RESULTS

Specificity of the clinical T-cell product SCG101

SCG101 is a human TCR T-cell product expressing a natural, high-affinity TCR isolated from an HLA-A*02:01⁺donor with resolved HBV infection [17]. The TCR construct comprises V alpha 34 and V beta 5.1 variable domains fused to minimally-murinized cysteine-modified constant domains to enhance correct pairing of the TCR chains while keeping immunogenicity at a minimum (Fig. 1A) [19]. The TCR was cloned into a lentiviral vector for transduction of T cells with a GMP-compliant protocol. As part of the cell product assessment before clinical application, SCG101 T cells were characterized in vitro and in preclinical models.

Figure 1.

Properties of the clinical HBV-specific T cell product SCG101. (A) Schematic structure of the HLA-A*02-restricted, HBV S₂₀-specific T cell receptor (TCR) used for transduction of T cells to generate SCG101. To increase TCR expression and correct pairing, TCR chains were codon-optimized, and an additional cysteine bond (orange, dashed line) was introduced to the constant domains. The human constants harbored nine amino acids from the murine TCR constant domains as indicated (red lines). pMHC, complex of peptide and MHC. (B) Binding strengths of TCR SCG101 expressed in Jurkat cells. Jurkat-SCG101 were stained with decreasing amounts of HLA-A*02:01-S₂₀ (FLLTRILTI)-PE-labeled multimers (blue circles) or with an HLA-A*02:01-C₁₈ (FLPSDFFPSV) multimer (red circles) as negative control. The mean fluorescence intensity (MFI) was quantified by flow cytometry. (C) TCR expression in CD8⁺ and CD4⁺ T cells after lentiviral transduction was detected by co-staining S₂₀ dextramer and anti-human CD4/CD8. The binding of a PE-labeled S₂₀-multimer indicates correctly paired TCR chains and Vβ5.1-FITC staining indicates the proportion of transduced T cells. Quantification of the potential TCR mispairing rate in % as a quotient of total S₂₀ dextramer+ cells divided by (total Vβ5.1⁺ cells – endogenously expressing Vβ5.1⁺ cells). Mean±SEM for mispairing rates of six SCG101 batches are shown. Untransduced cells (UT) served as a control to quantify T cells endogenously expressing Vβ5.1. (D) Estimation of the SCG101 recognition motif by alanine scanning. Each native residue of peptide S₂₀ (FLLTRILTI) was substituted at each position for an alanine. T2 cells were pulsed with 1 µM of each modified peptide indicated on the x-axis and co-cultured with SCG101 TCR-T, the IFN-γ concentration was measured by CBA. The prototype S₂₀-gtA/D and the C₁₈ peptide were set as positive or negative controls, respectively. (E) Cross-reactivity screening of SCG101 against human peptides containing the core binding motif L3 / T4 / R5 / I6 at concentrations of 1 (blue circles) and 0.1 (red circles) µg/mL. (F) Cross-reactivity of SCG101 against human peptides containing ≥ six amino acid sites consistent with the S₂₀ peptide. S₂₀-B/C peptide FLLTKILTI represents a variant often found in HBV genotypes B and C. Mean values of duplicates are shown. The labeling indicates the name of the human gene containing the respective peptide sequence. HBV, hepatitis B virus; HLA, human leukocyte antigen; SEM, standard error of the mean; IFN-γ, interferon-gamma; CBA, cytometric bead array.

Specific binding of the TCR to its peptide-MHC complex (S₂₀ FLLTRILTI_HLA-A*02:01) was dose-dependent with a half-maximum binding activity of 2.1 nM (Fig. 1B). Within primary SCG101 T cells, CD8⁺ as well as CD4⁺ T cells were able to bind the S₂₀ multimer, indicating a transduction rate of almost 90% (Fig. 1C). Staining using an antibody specific to the human TCR variable beta chain 5.1 (TCRVβ5.1), and an HLA-A*02-S₂₀ multimer allowed to estimate TCR expression and pairing efficiency, respectively. Accordingly, TCR mispairing with endogenous TCR sequences could be deduced from positive Vβ staining in the absence of multimer-binding. The maximum likelihood of such mispairing was low, with a mean value of 7% among six batches of SCG101 T cells (Fig. 1C).

To investigate the specific binding of SCG101 further, we aimed to assess critical amino acids within the TCR epitope sequence that are either essential for recognition by SCG101 T cells or for peptide binding to the HLA-A2 molecule. The exchange of L-T-R-I at position S22-S25 led to a 90–99% reduced recognition of the peptide, indicating the potential binding motif (Fig. 1D). Next, we assessed the cross-reactivity of SCG101 with human peptides and potential off-target activity. Neither the nine human peptides identified to contain the potential binding motif L-T-R-I (Fig. 1E, Supplementary Fig. 1A) nor the 14 peptides that shared six identical amino acids with peptide S₂₀ (Fig. 1F, Supplementary Fig. 1B) led to activation of SCG101 T cells. Hence, the results demonstrated the specificity of SCG101’s TCR and indicated the absence of cross-reactivity against tested very similar human peptides. Taken together, the SCG101 T-cell product exhibited clearly recognizable levels of TCR expression and correct TCR chain pairing while retaining specificity to its cognate HBV peptide.

Potency of SCG101 T cells

We next asked whether the clinical-grade SCG101 T cell product produced using a new upscaling protocol under GMP conditions would maintain antiviral and antitumor potency [18]. The functional avidity of SCG101 was measured by secretion of pro-inflammatory cytokines towards a dose-titration of peptide S₂₀ loaded onto T2 cells compared to peptide C₁₈, which should not be recognized (Fig. 2A). IFN-γ secretion proved most sensitive with an EC₅₀ of 1.55×10^-8 M. When co-cultured with HBsAg⁺ HepG2 hepatoma cells, SCG101 T cells were able to kill rapidly and specifically up to 100% of target cells within 48 hours (Fig. 2B). Consistent with the cytotoxic capability of SCG101, the antiviral cytokine IFN-γ was produced at high levels of 20 ng/mL upon antigen stimulation at an E:T of 1:1 (Fig. 2C). HBsAg in supernatants was also reduced by approximately 50% within two days of co-culture (Fig. 2D).

Figure 2.

In vitro potency of the HBV-specific T-cell product SCG101. (A) The functional avidity of SCG101 T cells was assessed by titration of the S₂₀ peptide (blue circles) loaded on T2 cells. The concentration of IFN-γ, TNF-α, and IL-2 in supernatants after 24 hours of coculture was measured by CBA. C₁₈ peptide (red circles) loading served as a negative control. (B) Cytolysis of tumor cell lines determined by the impedance of adherent target cells via xCELLigence real-time measurement. HBsAg negative (left) or positive (right) HepG2 target cells were co-cultured with SCG101 (blue lines) or untransduced (UT, red lines) at an effector-to-target (E:T) ratio of 1:1 (solid lines) or 1:5 (dashed lines). Mean values are shown. (C) Supernatants of the co-cultures were analyzed for IFN-γ after 48 hours by CBA. (D) HBsAg in supernatants after 48 hours of co-culture was measured by a quantitative ELISA. Data from two donors (C, D mean±SEM) are shown. HBV, hepatitis B virus; IFN-γ, interferon-gamma; TNF-α, tumor necrosis factor-α; IL, interleukin; HBsAg, hepatitis B surface antigen; SEM, standard error of the mean; CBA, cytometric bead array.

In vivo efficacy of SCG101 T cells

To analyze the anti-tumor activity and tissue distribution of SCG101 in an in vivo model, we used a xenograft model with transplanted HBsAg⁺ hepatoma cells. When tumors had formed, four different SCG101 dose levels ranging from 2 to 20 million TCR⁺ T cells were applied (Fig. 3A). Animals did not lose body weight during treatment (Fig. 3B), and by day 21, tumor growth inhibition in all SCG101 TCR-T groups was >95% (Fig. 3C).

Figure 3.

In vivo anti-tumor potency of HBV-specific SCG101 T cells. (A) NOD-Cg-PrkdcSCIDIL-2Rgcnull/vst (NPG) mice (female:male=1:1) were subcutaneously inoculated with 1×10⁷ HBsAg⁺ hepatoma cells/animal in the right armpit. After six days, 60 animals were divided into six groups. Mice received either increasing numbers of SCG101, multimer positive T cells of 0.2 to 2×10⁷ TCR-T cells/animal, or a mixture of CS10, HSA, and sodium chloride (vehicle, black lines), or untransduced T cells (UT, 4.3×10⁷ cells/animal, brown lines) according to the total number of T cells in the highest SCG101 group. Injections were done intravenously and (B) body weight and (C) tumor volume were analyzed twice weekly. Mean±SEM of n=10 per group, 5 males (circles) and 5 females (squares) are shown. HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; TCR-T, T cell receptor T cells; SEM, standard error of the mean. ^***P≤0.001.

The same preclinical model was also used to study the pharmacokinetics of the highest dose of TCR⁺ T cells (Supplementary Fig. 2A). The distribution of SCG101 was analyzed by qPCR of 14 different tissues at six consecutive time points over three weeks. After a single dose of 1×10⁹ TCR⁺ T cells/kg, the cells had rapidly distributed throughout the body by day one. After a gradual decrease of SCG101 in all tissues by day seven, copy numbers increased again in the following two weeks, peaking on day 21 (Supplementary Fig. 2B), indicating T-cell expansion upon antigen recognition. SCG101 accumulated mainly in the lung, spleen, blood, and liver (Supplementary Fig. 2C). Taken together, SCG101 TCR-T mediated a high target-specific anti-tumor activity and initiated T-cell persistence in vivo.

Application of HBV-specific T cells in a clinical setting

Encouraged by the preclinical data showing the anti-tumor activity of SCG101, an investigator-initiated trial, “Clinical study of SCG101 in the Treatment of Subjects with Hepatitis B Virus-Related Hepatocellular Carcinoma (SCG101-CI-101)”, was set up. It was approved by the local ethics board. A 54-year-old HLA-A*02:01⁺ man diagnosed with CHB and HCC in 2019 was enrolled in the study and infused with SCG101 in July 2022. The patient had a history of tumor and liver segment VIII resection in 2019, followed by transarterial chemoembolization in 2019 and 2021 and sorafenib treatment for one year. The patient started to take entecavir for HBV infection management in March 2022 (Fig. 4A). The status of his liver disease was Child-Pugh class A, Eastern Cooperative Oncology Group (ECOG) performance status 0, BCLC stage B and China liver cancer (CNLC) classification stage IIb without extrahepatic metastasis, portal vein thrombosis, or other comorbidities (Fig. 4B).

Figure 4.

Patient characteristics. (A) Scheme of the medical history of HCC diagnosis and treatment of patient ST1206. Study enrollment was initiated after the failure of three prior HCC treatments (resection, transarterial chemoembolization [TACE] with Raltitrexed and Lobaplatin, and Sorafenib). (B) HLA profile and status of liver disease of patient ST1206 including Child-Pugh-Score, Eastern Co-operative Oncology Group (ECOG) performance status, Barcelona Clinic Liver Cancer (BCLC) staging, and China liver cancer (CNLC) classification. Blue lines indicate duration of systemic treatments. ETV=Entecavir. (C–F) Immunohistological analysis of a liver biopsy taken three months prior to treatment. Four pieces of non-tumor liver tissue were obtained, and representative sections are shown. Scale bars: 200 µm and 40 µm (inlay), respectively. The circle indicates the orientation of the inlay. (C) Morphological analysis using hematoxylin and eosin (H&E) staining. Immunostainings for (D) HLA-A, (E) HBcAg, and (F) HBsAg (three different sections are shown because of a diverse expression profile). HCC, hepatocellular carcinoma; HLA, human leukocyte antigen; HBcAg, hepatitis B core antigen; HBsAg, hepatitis B surface antigen.

An archived liver biopsy sample taken three months before adoptive T-cell transfer was analyzed (Fig. 4C). Despite having multiple nodules and two target lesions, obtaining tumor biopsy was unsuccessful. Histological analysis revealed an intense HLA-A staining along sinusoids and a weaker HLA-A staining on hepatocytes (Fig. 4D). The tissue stained negative for HBcAg (Fig. 4E) and an average of around 10% of hepatocytes across three different sections stained positive for HBsAg (Fig. 4F).

After signing informed consent, the patient underwent leukapheresis to obtain peripheral blood mononuclear cells to produce SCG101 with the standardized protocol under GMP conditions. The cell product contained 30% TCR⁺ T cells, with CD8⁺ and CD4⁺ T-cell frequencies of 62.12% and 36.78%, respectively (Supplementary Fig. 3A), and passed all the quality controls and release criteria (Supplementary Fig. 3B). The data cut-off was February 7, 2023.

Treatment-related events

One week before treatment, lymphodepletion was performed on three consecutive days using cyclophosphamide and fludarabine. Finally, a single dose of 7.9×10⁷ per kg corresponding to 5.9×10⁹ total TCR-T cells was infused. Upon adoptive T-cell transfer, the patient was closely monitored for treatment-related adverse events (TRAE) and blood was analyzed regularly (Fig. 5A). The patient tolerated SCG101 therapy well; no treatment-related serious adverse events (SAEs), neurotoxicity, or infusion reactions occurred and all TRAEs were manageable and reversible (Table 1, Supplementary Fig. 4). TRAEs of grade 3 or 4 included cytokine release syndrome (CRS), hypotension, cytopenia, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) increase (Table 1). Cytopenia, which was anticipated due to the Cy/Flu preconditioning, comprised a decrease in lymphocytes, monocytes, neutrophils, and total white blood cells (Table 1, Supplementary Fig. 4A–D) and, in general, recovered to grade 2 within 30 days. However, a prolonged decrease and fluctuating platelet counts constituting grade 2 to grade 4 events were observed and finally resolved on day 80 (Supplementary Fig. 4E).

Figure 5.

Liver and cytokine serum markers after transfer of HBV-specific SCG101 T cells. (A) Scheme of the drug administration (black circles) and tumor assessments (white circles). Following Cy/Flu lymphodepletion on days -6 to -4, a single dose of 7.9×10⁷/kg (5.9×10⁹ total) TCR-T cells was infused intravenously. Red lines indicate blood sample collections. (B–H) Serum markers of liver function, alanine transaminase (ALT), aspartate transaminase (AST), lactate dehydrogenase (LDH), TBIL (total bilirubin), ferritin, albumin, and the INR were measured on indicated days. A grey area indicates the normal range. (I) C-reactive protein (CRP) indicating ongoing inflammation. (J–M) Serum concentrations of IL-6, IFN-γ, IL-2, and IL-10 were determined by CBA. HBV, hepatitis B virus; Cy/Flu, cyclophosphamide plus fludarabine; TCR-T, T cell receptor T cells; LDH, lactate dehydrogenase; INR, international normalized ratio; IL, interleukin; IFN-γ, interferon-gamma; CBA, cytometric bead array; LOD, limit of detection; LD, lymphodepletion; ACT, adoptive cell transfer.

Table 1.

Treatment-related adverse events ≥grade 3

Since SCG101 can not only target the tumor but also infected hepatocytes, as observed in preclinical models [18], particular attention was given to the liver function. Pretreatment ALT was 18 U/L, the pretreatment international normalized ratio (INR) for blood clotting was 1, and the pretreatment bilirubin level was 14.0 μmol/L. ALT elevation after SCG101 infusion reached a peak of 1,404 U/L on day 3 (Fig. 5B), that of AST 1,140 U/L on day 2 (Fig. 5C), respectively, both constituting grade 4 events (Table 1). After liver protection therapy, including glutathione, diammonium glycyrrhizinate, ademetionine, and polyene phosphatidylcholine, ALT, AST, and lactate dehydrogenase (Fig. 5D) levels decreased gradually.

Elevation of liver enzymes was transient, returned to baseline on day 17, and was accompanied by a slight, transient increase in serum total bilirubin below two times the upper limit of normal (ULN) (Fig. 5E, Supplementary Fig. 4G), ferritin (Fig. 5F), creatinine and urea (Supplementary Fig. 4H, I), as well as a reduction in serum albumin below 35 g/L on a single day (Fig. 5G), while the INR remained normal (Fig. 5H).

After infusion of SCG101, a strong immune reaction followed, characterized by C-reactive protein (CRP), IL-6, IFN-γ, IL-2, and IL-10 increasing rapidly one day after infusion (Fig. 5I–M, Supplementary Fig. 4J), temperature peaking at 39.8°C (Supplementary Fig. 4L) and hypotension with a minimum blood pressure of 70/41 mmHg (Supplementary Fig. 4M). The symptoms stabilized after antipyretic and symptomatic treatment with norepinephrine and rehydration. Together, these symptoms indicated a CRS and were contained with Tocilizumab and dexamethasone (10 mg; 2x d1, 4x d2, 1x d3) and the subject started to recover by day three post-infusion. In summary, infusion of SCG101 induced non-serious side effects that were expected due to SCG101’s mode of action and were manageable with standard medication for immune-related events.

Expansion of HBV-specific T cells in peripheral blood

The efficacy of adoptive T-cell therapy in both liquid and solid tumors using gene-modified T cells is associated with the persistence of transferred cells [20,21]. We, therefore, analyzed this by flow cytometry and qPCR. Ex vivo, transferred T cells at peak expansion constituted around one-third of CD3+ T cells in blood detected by flow cytometry (Fig. 6A, Supplementary Fig. 5A) with a maximum of 4.37×10⁴ cells/mL on day 21, suggesting a strong expansion (Fig. 6B). Both CD8⁺ and CD4⁺ HBV-specific TCR⁺ T cells circulated equally indicating a slight increase in CD4⁺ SCG101 T cells in blood compared to their proportion in the infusion product (Fig. 6B, Supplementary Fig. 3A, Supplementary Fig. 5B). Interestingly, while on day seven effector memory T cells (CCR7^– CD45RA^–) were still detected at equal numbers, mainly effector T cells (CCR7^– CD45RA⁺) and T memory stem cells (T_SCM, CCR7⁺ CD45RA⁺) circulated in peripheral blood thereafter within the first four weeks (Fig. 6C, Supplementary Fig. 5C). HBV-specific T cells were able to persist at least until four months post-transfer and by then almost exclusively consisted of T_SCM cells (Fig. 6C, Supplementary Fig. 5C). Long-term survival of transferred transduced T cells was confirmed by quantification of the copy number of the lentiviral vector integrates (Fig. 6D). This qPCR analysis indicated a peak in viral-vector copies seven days post-transfer, stabilizing until day 21 and slowly dropping thereafter. Collectively, these data demonstrated an enduring persistence of engineered T cells circulating in the blood associated with the development and maintenance of a predominantly stem cell memory phenotype after antigen exposure.

Figure 6.

Ex vivo analysis of transferred T cells. (A) Flow cytometry gating strategy exemplified with a blood sample taken 21 days posttransfer. Single living lymphocytes were gated first, followed by identification of transferred cells via anti-CD3 and HBV S₂₀-dextramer staining. In two separate staining panels, dextramer-binding TCR+ T cells were stained either for the T cell subsets with anti-CD8 and anti-CD4, or for memory differentiation status with anti-CCR7 and anti-CD45RA. Given the unspecific stimulation during T cell transduction and the antigen encounter after infusion, CCR7⁺ CD45RA⁺ cells were considered T_SCM and not naïve cells. The FMO (fluorescence minus one) control shows a staining without dextramer. (B, C) Flow cytometry analysis of blood samples taken after the cell transfer, gated on CD3⁺ and HBV-S₂₀-dextramer+ cells (blue circles). TCR+ T cells stained either for the T-cell subsets with anti-CD8 (pink triangle) or antiCD4 (purple triangle), or for memory differentiation status with anti-CCR7 and anti-CD45RA. T_CM :CCR7⁺CD45RA^- (red squares), T_SCM :CCR7⁺CD45RA⁺ (green triangle), T_EM :CCR7^-CD45RA^- (black circles), T_EFF :CCR7^-CD45RA⁺ (orange diamond). Different T-cell phenotypes were analyzed on indicated days and quantified using counting beads. (D) Genomic DNA was extracted from blood samples and the viral copy number (VCN) was quantified via qPCR. HBV, hepatitis B virus; TCR, T cell receptor; CCR, C-C chemokine receptor; qPCR, quantitative polymerase chain reaction.

Antiviral and antitumor activity of T cells redirected against HBsAg

Coinciding with the expansion of HBV-specific T cells and the increase in serum ALT, serum HBsAg substantially and rapidly decreased within one week after cell transfer from 557.96 to 1.3 IU/mL. The maximum reduction from baseline was 3.84 log₁₀ to 0.08 IU/mL. It was maintained for more than six months (Fig. 7A). Serum HBV-DNA levels were low at baseline (29 IU/mL) and remained undetectable after cell infusion (Fig. 7B). Since no tumor tissue was contained in the biopsies obtained before treatment, the presence of the target on tumor cells remained unknown, and changes in HBsAg expression could only be analyzed in hepatocytes. Compared to the HBsAg expression in 10% of hepatocytes at screening (Fig. 4F), we barely observed HBsAg staining in hepatocytes by 73 days after cell infusion (Fig. 7C). Alpha-fetoprotein (AFP) levels fluctuated and remained below 90 ng/mL until increasing again by day 133 (Supplementary Fig. 4N).

Figure 7.

Antiviral and antitumor response after treatment with HBV-specific SCG101 T cells. (A) HBsAg (left axis, blue squares) was measured in serum at indicated time points. The diagnostic ELISA’s lower limit of detection (LoD) was 0.05 IU/mL. ALT values (right axis, orange line) from Figure 6B were plotted again to better visually correlate both markers. The adoptive cell transfer (ACT) time point is indicated with a dashed line. (B) HBV-DNA was measured in sera via qPCR. (C) Immunostaining of HBsAg of a liver biopsy taken 73 days after treatment. One piece of non-tumor liver tissue was obtained. Scale bars: 200 µm and 50 µm (inlay), respectively. (D) The tumor burden was analyzed via multiphasic CT scan on days -7 prior to treatment and on days 28, 63, 133, and 181 after ACT. The size of the target (T) lesion no.1 and no.2 in mRECIST (m) or iRECIST (i) and the position of non-target (NT) lesions are indicated in yellow. HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; ALT, alanine aminotransferase; qPCR, quantitative polymerase chain reaction; CT, computer-assisted tomography.

At baseline, the subject presented with multiple nodules and two target lesions in the right lobe with a 38- and 21-mm diameter, respectively (Fig. 7D, upper row). By day 28 post SCG101 infusion, a large area of tumor necrosis was observed (Fig. 7D, 2nd row), followed by a partial response with the target tumor lesions decreasing by 74.5% per mRECIST and by 47.5% per iRECIST, respectively. The patient maintained a stable disease for at least 6.9 months, suggesting clinical antiviral as well as antitumor activity of HBV-specific T cells.

DISCUSSION

HBV infection is a main driver for the development of HCC, and the available treatment options are limited. We here show that autologous, HBV-specific T cells can be manufactured using lentiviral transduction of a high-affinity TCR under GMP conditions, persist over several months, are safe and functioning in the clinical setting of an HBV-HCC, and lead to a profound reduction in viral markers and tumor mass.

SCG101 carries a natural TCR isolated from an HLA-A*02:01 donor with resolved HBV infection [17]. Hence, it had undergone negative selection against self-antigens within the thymus of the donor, and, not surprisingly, we did not observe any off-target activity in the patient. Although an alanine-substitution scan identified position five to be part of the TCR recognition motif, SCG101 was capable of recognizing the two most common HBV peptide variants FLLTRILTI (gtA,C,D) and FLLTKILTI (gtB). This cross-recognition could be explained by Arginine (R) and Lysine (K) as positively charged amino acids sharing more similar properties than Arginine and Alanine. Furthermore, preclinical toxicology and distribution studies were performed according to regulatory standards. However, these xenograft mouse models can only provide limited information since cytokines, chemokines, and receptors do not match between the human T-cell product and the murine recipient. Therefore, it is reassuring to know that T-cell transfer was also safe in a syngeneic mouse model when murine SCG101 T cells were injected into HBV⁺ HLA-A2⁺ mice [22].

Overall, the treatment was well-tolerated, all TRAEs were reversible through symptomatic interventions, and neither SAE nor dose-limiting toxicity (DLT) occurred. Grade 3 CRS became evident by an elevation of IL-6, fever, and hypotension, which likely led to increased levels of creatinine and urea in the blood due to reduced renal blood perfusion. CRS was effectively managed with tocilizumab, corticoids, and noradrenaline. Notably, in the treatment of CD19-malignancies, CRS typically occurs more than 24 hours after CAR-T-cell infusion [23]. However, in the case presented here, the onset of CRS was unexpectedly rapid, with elevated body temperature and high IL-6 levels observed as early as two hours after the infusion. Generally, it is assumed that IIFN-γ released by activated T cells leads to IL-6 production in bystander cells such as macrophages [24]. Our in vitro experiments using real-time cytotoxicity measurement revealed that T-cell activation and target cell killing occurred immediately after the co-culture began. Potentially, in the patient, effector function of the T cells started immediately after infiltration into the tissue, and the significant number of liver-resident Kupffer cell macrophages may have contributed to the rapid onset of CRS.

All grade 4 events (cytopenia and increased ALT/AST levels) were expected to be a positive sign of response to treatment. Cytopenia was intended and most likely induced by the pre-conditioning Cy/Flu regimen, which has been associated with better T-cell engraftment and outcome in CD19 CAR-T treatment [25]. The reasons why patients benefit from lymphodepletion prior to ACT are not fully understood, and co-depletion of regulatory cells [26], an increase of serum cytokines [27,28], and tolerization towards xenogeneic sequences [29] have been discussed. Cytopenia also comprised a low platelet count, which started already at screening before treatment. It might have been prolonged due to multiple factors, including lymphodepletion, tocilizumab treatment, and the study drug’s proliferation. We suppose that a more proactive use of recombinant human thrombopoietin or thrombopoietin receptor agonists can be considered to help the platelet count recover under such circumstances.

The elevation in ALT/AST levels occurred concomitantly with a substantial reduction in serum HBsAg and is in line with SCG101’s anti-tumor and anti-viral cytotoxicity demonstrated before [18]. Notably, in addition to targeting tumor cells, HBV-specific T cells also recognize and attack hepatocytes that express HBsAg or fragments thereof, either due to HBV infection or HBV-DNA integration [18], which likely led to the transient increase in serum transaminases. The ALT levels in our patient reached 35-fold the ULN. This, however, did not result in hepatic dysfunction or severe hepatic damage as evidenced by the absence of impaired liver synthetic function (bilirubin, albumin, or INR), the absence of specific symptoms such as jaundice or bleeding, and the transient nature of the flare [30]. Cytotoxicity directed towards non-tumor HBV⁺ hepatocytes is inherently linked to an effective T-cell response, either naturally generated during acute hepatitis and viral clearance or artificially generated via transfer of immune cells. [31] Indeed, transient ALT flares that are host-induced, i.e., comprising an effective immune response, but not virus-induced, can be associated with favorable outcomes [30] and viral clearance [32,33].

Nonetheless, the mitigation of excessive liver damage following adoptive T-cell transfer remains a priority. To address this, inclusion required a Child-Pugh score ≤7 and ECOG performance status of 0 or 1 as well, and pre-treatment biopsies were analyzed for HBsAg⁺ hepatocytes. Several studies have examined CHB patient biopsies to determine the frequency of HBsAg⁺ hepatocytes with high variation from 0% to 100% [34-39], with particularly high levels when ground glass hepatocytes are present [37]. The most recent and comprehensive study by Aggarwal et al. analyzed biopsies from 114 patients, finding the average proportion of HBsAg⁺ hepatocytes to be below 10%. Interestingly, they observed a lack of correlation between quantification results from two biopsies collected from the same individual and time point [36]. This aligns with our findings of heterogenic HBsAg staining for subject ST1206, suggesting that multiple biopsies may be necessary to yield reliable results representing the entire liver. This information could be complemented by serum HBsAg levels, which generally, though not always [34], correlate with intrahepatic HBsAg levels [40]. Additionally, more sophisticated probe- [13] or sequencing-based [41] assays might be necessary to assess the presence of the target peptide expressed from chimeras of HBV and host-cell nucleic acids that might not cover the full HBsAg open reading frame and might not be detectable by antibody staining [13]. We anticipate insights from the ongoing phase I/II study of SCG101 (NCT05417932) will contribute to identifying a reliable biomarker for safe patient stratification.

The patient described here was negative for HBcAg liver staining, HBeAg, and HBV-DNA, leading to the assumption that most of the serum HBsAg originated from tumor cells and/or hepatocytes with HBV-DNA integration rather than from active infection. Studies have indicated that HBsAg and HBcAg rarely colocalize in biopsy samples [9,36] and that in late-stage CHB, integrated HBV-DNA is the main source for HBsAg production [42-44]. HBsAg decline is the hallmark of effective anti-HBV treatment [45] and has not been achieved by RNA-interference-based therapies [46], likely due to integrated HBV-DNA being the source of HBsAg [47]. Therefore, targeting integrated HBV-DNA has been suggested to be a goal in drug development [48,49]. With SCG101 treatment, we observed a considerable reduction in HBsAg, approaching the detection limit of the diagnostic assay, which would have constituted a functional cure. Over six months, HBsAg levels did not rebound but were also not eliminated despite SCG101 persistence, and several reasons for this observation can be discussed. First, it is possible that HBsAg⁺ cells developed evasion strategies, such as HLA-loss or mutations in the peptide presentation pathway, similar to what has been observed in HPV-cancer patients refractory to treatment with HPV-specific T cells [50]. Second, SCG101 cells might have undergone similar exhaustion mechanisms as endogenous HBV-specific T cells in the liver and tumor [51,52]. Due to the limited amount of blood available for analysis, we were unfortunately unable to address this hypothesis. Although SCG101 T cells were still circulating in the blood after several months, we have no proof that they retained their functionality. However, in mice transduced with AAV-HBV and AAV-HLA-A2, we observed the same kinetics of ALT increase, HBsAg reduction, and persistence of HBV-specific T cells. Also, a residual low amount of HBsAg remained detectable in this syngeneic mouse model. When T cells were analyzed ex vivo, the high-avidity TCR-T cells could no longer be specifically activated by their cognate peptide [22].

Most importantly, following activation of SCG101 in subject ST1206, a sizable and sustained antitumor effect was observed 28 days after cell infusion, lasting at least until the cut-off date. While tumor shrinkage occurred directly after SCG101 infusion, without a pre-treatment tumor biopsy, we can only speculate but not prove its direct cytotoxicity on tumor cells. Alternatively, an indirect effect can be envisioned, where SCG101 cells create an inflammatory environment, helping endogenous immune cells attack the tumor. This idea was already proposed when two patients treated with RNA-electroporated TCR-T cells experienced clear shrinkage of the primary tumor [53,54] or metastases [13]. Using mRNA, anti-tumor responses occurred only after several infusions and during a time frame when the transferred cells most likely had lost their designed antigen-specificity due to the transient HBV-TCR expression after mRNA electroporation, while still other immunological alterations were detected [13,54].

Notably, four months after T-cell infusion, a moderate rise in AFP levels was observed in our patient, exceeding preinfusion levels. This AFP increase could indicate liver regeneration, as similar AFP rises have been observed in acute hepatitis B and correlated positively with the extent of transaminase elevation [55,56]. However, in this study, peak transaminase levels were followed by the AFP increase only four weeks later. Another potential hypothesis for the increase in AFP is that eliminating HBsAg⁺ tumor and premalignant cells created space for HCC cells negative for the S₂₀ target peptide of SCG101 to proliferate. Mason et al. [6] proposed the “loss of productive HBV infection, providing at least partial escape from the antiviral immune response as a major facilitator of clonal expansion”. Given the early emergence of clonal growth [5], we would expect this postulated negative effect of an antiviral immune response to also manifest in an increased number of HCC cases after spontaneous clearance in late-stage CHB. However, further investigations in larger cohorts are required to gather more data on this phenomenon.

The on-target efficacy observed with our selected dose of SCG101 and treatment scheme exceeded similar - albeit different - regimens using HBV-specific T cells transiently expressing a TCR after electroporation [13,53]. In patients with the indication of HBV-HCC-derived metastases and an HLA-A2 negative and/or HBV-negative liver after transplantation [13,57]. HBsAg was reduced by almost 50% in 2/8 patients three to six months after infusing several doses of transient, RNA-electroporated TCR T cells [53] or by 90% in a single patient one month after infusion of stably transduced TCR-T cells [57]. Although it is difficult to quantitatively compare these approaches due to different clinical settings, cell doses and duration of cellular functionality, it was remarkable that in the pioneering study of Qasim et al. [57], the stably transduced TCR-T seemed to act faster and more efficiently despite a low transduction rate and a low dose of cells infused.

This encouraged us to set up a highly efficient GMP-compliant transduction process generating higher TCR-T numbers that - as demonstrated in this study - after infusion quickly reduced HBsAg by 99.99%. Along with the stable expression of the TCR, other factors might have contributed to the clinical efficacy of SCG101. Its dual functionality in CD8⁺ and CD4⁺ T cells, the balance of both T-cell subtypes, and the shift to a memory stem cell phenotype might have supported the persistence and extended functionality of transferred T cells. A well-balanced CD8:CD4 T-cell ratio and the occurrence of the T_SCM phenotype have positively impacted the outcome of CAR-T therapy [58-60]. Potentially, also the encounter of HBsAg on hepatocytes may have facilitated the stimulation and expansion of transferred, stably TCR-expressing T cells before encountering the immunosuppressive tumor microenvironment. Others currently follow the same principle of using an RNA-vaccine to boost transferred claudin-specific CAR T cells to treat solid tumors [61].

Overall, SCG101 was shown to be a safe product with its stable TCR expression, leading to both antitumor and antiviral effector functions and sustained persistence in the patient. Further studies will determine whether adoptive T-cell therapy can maintain a favorable risk-benefit ratio and has the potential to be applied to target the cause of malignant transformation to prevent additional tumor development.

SUPPLEMENTAL MATERIAL

Article information Continued

Notes