Advancing TCR-T cell therapy in hepatocellular carcinoma: lessons from the ADP-A2AFP trial

Hepatocellular carcinoma (HCC), the most common primary liver malignancy, typically arises in the context of chronic liver inflammation driven by metabolic dysfunction, long-term alcohol exposure, viral hepatitis, and other etiologies. Within this environment, T cells play a central role in tumor immune surveillance and elimination. Their importance in HCC is highlighted by the clinical success of immune checkpoint inhibitors (ICIs), which restore exhausted T-cell responses and have become a cornerstone of systemic therapy (1,2). However, the efficacy of ICIs largely depends on the presence of tumor-specific T cells, which are often absent or present at very low frequencies in most tumor tissues. Adoptive transfer of T cells engineered with tumor-specific T-cell receptors (TCRs) or chimeric antigen receptors (CARs) can provide a defined population of effector cells capable of directly recognizing and targeting cancer cells.

In adoptive T-cell therapy, T cells are isolated from a patient or donor, expanded and/or genetically modified ex vivo, and reinfused to enhance antitumor immunity. Among these approaches, CAR-T-cell therapy involves engineering T cells to express synthetic receptors that recognize membrane-bound tumor-associated antigens (TAAs). These receptors combine an antigen-binding domain with intracellular signaling domains, including CD3ζ and costimulatory motifs, to enhance T-cell activation, persistence, and cytotoxicity. However, the development of engineered T-cell therapies for HCC remains particularly challenging. A major limitation is the scarcity of truly tumor-specific antigens, as many TAAs are also expressed in normal or regenerating liver tissue, raising concerns about on-target, off-tumor toxicity (3). In addition, the dense tumor stroma and the highly immunosuppressive tumor microenvironment (TME) characteristic of HCC further impair T-cell trafficking and function.

TCR-engineered T cells (TCR-T) offer an alternative approach. Unlike CAR-T cells, TCR-T cells express modified T-cell receptors that recognize intracellular antigens presented in the context of specific human leukocyte antigen (HLA) molecules. This enables targeting of a broader repertoire of tumor antigens but imposes HLA restrictions that limit patient eligibility. Indeed, most TCR-T cells investigated to date have been derived from HLA-A*02:01-restricted antigen libraries, thereby limiting their applicability to only a subset of patients. To overcome this limitation, alternative strategies are being explored, such as optimizing TCR sequences to enable recognition of the same epitope across multiple HLA-A*02 variants.

Alpha-fetoprotein (AFP) is a oncofetal glycoprotein normally produced by the fetal liver that is largely absent in healthy adults but re-expressed in approximately two-thirds of cases HCC, where its serum levels correlate with tumor burden, and aggressive clinical behavior. Beyond its prognostic-marker utility (4,5), tumor-derived AFP may also contribute to tumor progression and immune escape through its immunomodulatory effects (6). These features make AFP not only a clinically relevant biomarker but also a biologically meaningful TAA, supporting its rationale as a target for engineered T-cell therapies in HCC.

Previously, AFP₁₅₈-specific TCRs been isolated (7), supporting the feasibility of targeting AFP with TCR-based immunotherapies. Moreover, studies have explored alternative immunotherapeutic strategies directed against AFP-derived epitopes, including CAR-T cells engineered to recognize the AFP_158–166 peptide in complex with HLA-A*02:01 (8), highlighting the translational interest in AFP as an immunologically actionable TAA.

In this context, Meyer et al. reported a phase I, first-in-human clinical trial (NCT03132792) evaluating ADP-A2AFP, an autologous TCR T-cell therapy targeting AFP presented by HLA-A*02:01 (9). Previously, the affinity of this therapeutic construct ADP-A2AFP (referred to as AFP^c332 in the preclinical study) was engineered to significantly enhance its binding affinity, reducing the equilibrium dissociation constant from 754 to 10.6 µM (10).

In Meyer et al. study, autologous T cells were collected via leukapheresis, transduced with a lentiviral vector encoding an AFP-specific TCR, expanded ex vivo, and reinfused into the corresponding patients with AFP-expressing liver cancer. Strict eligibility criteria, including exclusion of patients expressing other HLA-A*02 alleles to avoid off-target reactivity, further constrained enrollment. Safety was the primary endpoint, while objective response rate according to response evaluation criteria in solid tumors (RECIST) v1.1 served as the secondary endpoint.

Ultimately, 21 patients (20 with advanced HCC and one with hepatoid gastric carcinoma) received treatment following lymphodepleting chemotherapy with fludarabine and cyclophosphamide. Patients were treated with different doses ranging from 0.8×10⁸ to 10.0×10⁹ of engineered T cells, and repeat dosing was feasible in only two cases. Clinical efficacy was modest, with one complete response (CR) and one partial response (PR) (overall response rate 9.5%), both associated with reductions in serum AFP levels. Twelve patients achieved stable disease (SD), while seven experienced disease progression. Disease control rate (CR + PR + SD) was 66.7%.

The safety profile was generally consistent with other adoptive T-cell therapies. One patient experienced dose-limiting toxicity (grade 4 atrial fibrillation) and died from cholangitis deemed unrelated to treatment. Cytokine release syndrome occurred in six patients, including one grade 4 event, and was managed with a recombinant humanized anti-interleukin-6 receptor monoclonal antibody tocilizumab when required. Hematologic toxicities, including grade 3–4 leukopenia, were common and likely attributable to lymphodepleting chemotherapy.

Although the study did not demonstrate sufficient efficacy to support continued development of ADP-A2AFP in its current form, it provides important insights for the advancement of adoptive T-cell therapies in HCC. Notably, this trial evaluated the peripheral persistence of ADP-A2AFP and found that persistence was greater among participants in the dose-expansion cohorts. This is an important finding, as limited persistence continues to represent a major challenge and barrier to durable efficacy in most adoptive cell therapies.

Second, the trial highlights the substantial challenges associated with patient selection and treatment feasibility. Of 477 patients screened over 5 years, only 39 were deemed potentially eligible, underscoring the restrictive nature of HLA requirements and clinical constraints in this population. Furthermore, the time required for patient screening, TCR-T cell manufacturing, and product delivery can create a substantial interval between enrollment and treatment initiation, during which disease progression may occur. This represents a common challenge of engineered T-cell therapies compared with conventional systemic treatments. Beyond these biological and clinical barriers, engineered T-cell therapies remain highly resource-challenging. They require specialized Good Manufacturing Practice (GMP)-compliant manufacturing facilities, complex logistics, and highly trained personnel. The work by Meyer et al. illustrates both the feasibility of this approach. However, the significant challenges that may currently limit its broader implementation, raising important questions regarding scalability and integration into routine clinical practice.

In this context, alternative strategies that avoid the need for ex vivo cell engineering may offer more accessible solutions. For example, the delivery of small interfering RNA (siRNA) targeting AFP with lipid nanoparticles could reduce AFP protein levels, potentially limiting its immunosuppressive effect (6).

Another major challenge highlighted by this study is the intratumoral heterogeneity of AFP expression, which may contribute to variable responses and antigen escape. Recent research has identified additional immunogenic epitopes in HCC (e.g., targeting MAGE-A3 or LRRC46) (11), suggesting that multi-targeted approaches could improve therapeutic efficacy. Indeed, strategies incorporating dual- or multi-antigen targeting T cells may enhance specificity while reducing the risk of resistance. Furthermore, the inherently low affinity of TCRs for self-antigens, such as AFP, limits the therapeutic potential of TCR-T cells to target AFP-expressing tumor cells and represents a significant obstacle to the development of effective TCR-T cell therapies for HCC. Affinity enhancement has emerged as a key strategy to overcome this limitation, as illustrated by ADP-A2AFP TCR-T cells; however, it may lead to altered specificity and/or toxicity. In parallel, a recent preclinical study demonstrated that TCR affinity can be optimized without compromising specificity (12). Notably, AFP-specific A01-m14 TCR-T cells showed potent antitumor activity in human HCC organoids and in vivo murine models, together with a favorable safety profile, and are currently being evaluated in a phase I clinical trial in patients with AFP-positive HCC (ClinicalTrials.gov: NCT06515314).

The immunosuppressive TME in HCC represents an additional critical barrier. AFP-producing tumors are often associated with T-cell exhaustion, including increased expression of programmed cell death (PD)-1 (13-15). This provides a strong rationale for combination strategies. In particular, integrating adoptive T-cell therapies with immune checkpoint blockade (e.g., PD-1/PD-L1 inhibitors) may enhance T-cell persistence and function while counteracting exhaustion.

From a translational perspective, deeper characterization of the engineered T-cells, circulating immune cells and TME remains important for understanding determinants of response to adoptive T-cell therapies. Comprehensive profiling of T-cell phenotype, differentiation state, metabolic fitness, and persistence—both prior to infusion and during treatment—may help identify factors associated with clinical benefit. Likewise, deep characterization of the TME, when tumor tissue is available, could provide insights into mechanisms of resistance. In the study by Meyer et al., limited tissue availability restricted such analyses. Future studies may benefit from integrated longitudinal immune profiling to better distinguish responders from non-responders.

In summary, the clinical trial by Meyer et al. reflects sustained and important efforts to advance precision immunotherapy in HCC. Continued progress in cell engineering, together with a deeper understanding of tumor-immune interactions and the TME, will be critical to refining these approaches and unlocking their full therapeutic potential.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.