Hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC) is a serious health problem in Asia, since the efficacy of current systemic therapies for advanced HCC is still suboptimal with a median overall survival often less than 2 years [1]. New therapy modalities are needed and in this issue of Clinical and Molecular Hepatology, Wan et al. [2] described the experimental efforts to develop and test both in animal model and then in a patient with advanced HBV-related HCC (HBV-HCC), a cellular therapy based on adoptive transfer of autologous T cells engineered to be specific for an HBV epitope (the protein fragment presented by human leukocyte antigen [HLA]-class I and recognized by a specific T cell receptor) presented on the surface of HBVHCC. The rationale to target HBV antigen as a tumor-specific antigen in HBV-HCC is due to the ability of HBV to integrate into the DNA of hepatocytes and therefore be present in the great majority of transformed HCC cells [3].

Wan et al. [2] utilized a lentiviral vector to introduce the genetic information of V alpha and V beta regions of a T cell receptor specific for the complex of HLA-A2 molecule and S domain region 20–28 into T cells and elegantly demonstrated in vitro that this lentiviral transduction of T cells generated T cells stably expressing the selected T cell receptor (TCR) (specific for the HLA-A2/S_20-28 region). The engineered HBV-specific TCR T cells were first demonstrated to recognize HBV+ hepatoma cells both in vitro and in an animal model prior to their use in treating a patient with HBV-HCC. Using the developed technology, a large quantity of HBV-specific TCR T cells was produced under Good Manufacturing Practice conditions. The expression of the HBV-TCR on the engineered T cells was tested in vitro and then the autologous HBV-specific TCR-expressing T cells were infused into a lymphodepleted (fludarabine/cyclophosphamide) patient with multifocal HBV-HCC and chronic HBV infection in a single dose of 7.9x10⁷xkg (total 5.9x109 T cells).

Even if the critics might point out that this cellular treatment didn’t achieve a complete eradication of the HCC lesion and negativity of hepatitis B surface antigen (HBsAg), the work confirms the potential of this immunotherapeutic strategy for HBV-HCC and chronic HBV treatment. In the treated patient, engineered HBV-TCR T cells efficiently reached the liver, were functional in vivo and targeted both HBV-infected and HBV-expressing HCC cells, showing a shrinkage of the HCC lesion (70% reduction of the initial volume), a reduction of the quantity of HBsAg (a reduction of almost 4 log) and persistence of the HBV-TCR T cells after infusion. Importantly, though the infusion of a single large quantity of engineered TCR T cells immediately caused the triggering of a cytokine release syndrome (with high values of IL-6 detected in serum, onset of fever and hypotension) followed by alanine aminotransferase (ALT) elevation (up to 1,440 IU/L at day 3 after infusion), all symptoms were reversible.

In order to fully appreciate the significance of this new data in terms of safety and of clinical effectiveness against HBV-HCC and HBV infection, we need to consider these findings in the context of the various efforts by different laboratories (mainly from the Protzer lab and our laboratory) to develop T cell therapy for HBV infection and its complications that started more than 15 years ago (Table 1).

The possibility of modifying the specificity of T cells through expression of a chimeric antigen receptor (CAR) or a classical TCR started back in 2008. Ulrike Protzer's group was the first to engineer HBV-specific T cells utilizing a chimeric antigen receptor specific for HBsAg [4]. The work was a logical extension, in the HBV field, of the progressive development of molecular methods that utilize viral vectors to introduce CAR or TCR into T cells that endow them with novel specificity [5]. For the readers not familiar with the difference between a CAR and a TCR, we would direct them to the many specific and detailed reviews on the arguments that have been written in the last 15 years (see for example [6]). Nevertheless, to briefly summarize, a CAR consists of an extracellular antibody recognition site coupled with a constant region of a TCR (in its cognate form or modified with the insertion of intracellular co-stimulatory domains) that delivers the activation signal. The main difference is that the CAR does not recognize the HLA-class I/viral or tumor epitope complex like a conventional TCR; it instead recognizes the whole conformation of an antigen expressed on the surface of the target cells.

The CAR-T cell strategy has practical advantages in terms of translation to clinic. A single CAR can be used to engineer the T cells of different patients, while a TCR can only be used in patients who possess the HLA-class I molecules that present the correct targeted epitope [6]. This means that developing a TCR-based strategy is more complex; you need to have a large TCR library, and the number of patients that you can treat with TCR-redirected T cells is limited by their HLA-class I profile.

However, despite the practical advantages of HBV-specific CAR-T cells, their clinical utilization in HBV chronic patients or in HBV-HCC has not been expanded for two fundamental reasons. The first one is that HBsAg is present in high concentrations in the serum of the patients with CHB infection and also in some of the patients with HBV-HCC. Such circulating HBsAg can directly bind to the HBsAgspecific CAR present on the surface of the engineered T cells and potentially interfere with their activation and target recognition [7]. The second and perhaps even more important reason, particularly in patients with HBV-HCC, is that most of the HBV-HCC do present HBV-DNA integration, but such integration is not complete and thus rarely codes for the whole HBsAg [8]. Instead, HBV-HCC tumors with partial HBV-DNA integration can still present fragments of HBV proteins associated with their HLA-class I molecules derived from the processing of chimeric/fragmented HBV proteins (often translated from HBV envelope or X regions integrated into the host genome) and thus can be recognized by HBV-specific TCR T cells [8] but not by HBV-specific CAR-T cells.

Indeed, after the development of HBsAg-specific CAR T cells [4,9], our group [10,11] and then Protzer’s group [12] isolated different HBV-specific TCRs with different specificities and HLA-restrictions that recognize both HBV-HCC and HBVinfected hepatocytes in vitro and in animal models [13,14].

The demonstration that T cells modified with an HBVspecific TCR can efficiently recognize HBV-HCC did not resolve all the problems of this therapy approach, particularly those related to safety. Utilizing HBV antigen as a tumor antigen for HBV-HCC was not only driven by the frequent HBV-DNA integration found in 80–90% of all HBVHCC, but also because it was relatively easy to isolate and clone HBV-specific TCRs from patients who have resolved acute HBV infection. While HBV-HCC can express other classical tumor antigens (i.e., alpha-fetoprotein, NY-Eso-1, Glypican-3, Mage-3) and T cells specific for these classical tumor-associated antigens are present in HBV-HCC patients [15], TCRs specific for tumor associated antigens are often of low affinity and require modifications in order to efficiently recognize tumor cells. In contrast, T cells recognizing HBV antigens present in patients that resolved acute HBV infection express high-affinity TCR and recognize with high efficiency the low quantity of HBV epitope present on the surface of the target cells that can be HBV-infected hepatocytes and/or HCC cells with HBV-DNA integration [10,12]. However, this inability to distinguish between normal HBVinfected hepatocytes and HCC cells that are expressing HBV antigens from DNA integration is a drawback; the risk of inducing severe liver damage after HBV-TCR T cell treatment in patients with chronic HBV infection who developed HCC cannot be ignored. This is why initial tests of HBVTCR engineered T cell therapy in patients were done in individuals with HBV metastasis after liver transplantations [8,16] (after showing that the transplanted liver was not HBV-infected or mismatched in relation to the HLA restriction of the adoptively transferred HBV-specific TCR-T cells) or by using T cells that express the HBV-TCR transiently through mRNA electroporation [17,18].

It is therefore important to see in a patient with advanced HBV-HCC and HBV chronic infection that the infusion of a large number of T cells that stably express an HBV-specific TCR causes a marked elevation of ALT at day 3 that was, however, reversible and associated with a significant drop of HBsAg [2]. Even though the quantity of HBV-infected hepatocytes can be highly variable in patients with chronic HBV infection [19] and HBV-HCC, it is reassuring to see how a liver chronically infected with HBV can tolerate and recover after the induction of robust inflammation.

We would argue that the observation of a marked expansion of the adoptively transferred HBV-TCR T cells in vivo shown by these new data might also indicate that the number of infused HBV-TCR T cells could be substantially reduced and perhaps this could also avoid or lessen the cytokine-release syndrome (high level of IL-6 and hypotension controlled with Tocilizumab, corticosteroids and noradrenaline) that was observed in the patient immediately after infusion (2 hours). Infusion of lower quantities of HBV-TCR T cells (1x10⁵ to 5 x10⁶xkg) [17,18] produced with mRNA technology in HBV-HCC patients has not been associated with rapid cytokine-release syndrome symptoms, even though in some patients ALT elevation (with increased bilirubin level) was observed.

Furthermore, in addition to dosage and safety, other questions regarding the effectiveness of the treatment arise from the data. For instance, it's uncertain whether HBV-specific TCR T cells that expanded in vivo maintain their functionality over time. The initial expansion and subsequent persistence of the adoptively transferred HBV-TCR T cells in this patient are important observations. They show that the HBV-TCR T cells, after recognizing their "hepatic" target, were not immediately tolerized, but instead proliferated and caused a clear reduction of HBsAg. This reduction serves as indirect evidence of their in vivo functionality. However, after the initial HCC shrinkage, the patient did not completely eliminate the tumor and also did not completely suppress HBsAg. This might indicate that after the initial recognition, the HBV-TCR T cells have gradually lost their effector function.

There may be other reasons for the persistence of HBVHCC lesions and HBV infection, such as the inability of HBV-TCR T cells to reach their targets or low expression or mutation of the HBV epitope targeted by the HBV-TCR T cells. Whatever is the cause, it will be important to understand whether stably engineered HBV-TCR T cells not only proliferate in vivo but also maintain their functionality.

This understanding is vital to better design adoptive T cell therapy strategies in HBV-HCC patients, which at the moment are split between the use of T cells that stably express their TCR (like the one of Wan et al.) [2] or the use of T cells that express their TCRs transiently [20] (a strategy used mainly by us).

Wan and colleagues argue in their discussion that the persistence of the adoptively transferred TCR-T cells in treated patients is an indispensable requirement for clinical efficacy. We would argue here that it is certainly correct to note that the persistence of CAR-T cells in patients with liquid tumors is associated with therapeutic success [21]. Nevertheless, the clinical success of adoptive T cell therapy against solid tumors has been extremely poor so far, and thus, potential biomarkers of clinical success may not depend on T cell persistence.

Unlike the group of Protzer, we favour a strategy of HBVTCR T cell treatment utilizing T cells that are not genetically modified but express the HBV-TCR transiently through mRNA electroporation. This strategy was selected not only for safety (it allows infusion of escalating doses of HBVTCR T cells in patients and a controlled analysis of the side effects) but also because the repetitive infusion of newly engineered HBV-TCR T cells might escape functional exhaustion induced by the immunosuppressive environment of HBV-HCC [22]. In addition, the repetitive infusions of highly activated HBV-TCR T cells can, despite their transient persistence alter this immunosuppressive environment. As Wan et al. also argue in their discussion, the clinical efficacy of HBV-TCR T cell therapy against HBV-HCC might not be only associated with the direct HBV-HCC cell lysis by the adoptive transferred T cells. It might be also dependent on their ability to modify the HCC microenvironment that can facilitate the induction of novel endogenous tumorspecific T cells. This hypothesis was indeed suggested by initial observations from a clinical trial using mRNA-based HBV-TCR T cells where the response to therapy (fully and partially observed in 1 and 3 out of 8 primary HBV-HCC patients treated, respectively) was not directly proportional to the number of infused TCR-T cells, but was instead associated with the induction of inflammatory liver events [18]. Demonstrating that the clinical effectiveness of TCR-T cell therapy against a solid tumor is linked to its capacity to boost an endogenous immune response, rather than to the quantity and persistence of adoptively transferred T cells, might change the approach to adoptive T cell therapy against solid tumors.

Therefore, we need to understand whether the success of T cell therapy against HBV-related HCC is associated with the preservation of the function of adoptively transferred T cells or whether a hit-and-run strategy using T cells with limited functional lifespan can be less or more efficacious.

What is certain is that an answer to these questions can only be derived from high-quality data obtained from the detailed study of treated patients, like the study published in this issue of Clinical and Molecular Hepatology. Only by analysing the mechanisms of action of the adoptively transferred T cells, whether they can proliferate and maintain functionality or modify the HCC microenvironment, will we be able to use this therapy more effectively and improve the clinical needs of patients with HBV related HCC. Such studies might also clarify whether we need to use different T cell populations for adoptive T cell therapy. In the future, we might use Mucosal Associated Invariant T cells (MAIT cells) [23] that are known to have a tropism for the liver, instead of normal T cells, or utilize mixtures of T cells engineered with HLA-class I and HLA-class II TCRs that have also been recently cloned [24]. The indication of successful HBV-TCR T cell therapy could then be extended to other HBV-related pathologies and, for example, be used to achieve functional HBV cure7 or as a preventive therapy for HBV-HCC relapses after liver transplantations [25].

The work ahead of us might be long and arduous, but the reward for patients and science will be substantial.