The Concept of Immune Tolerance in Chronic Hepatitis B Virus Infection Is Alive and Well

Chronic infection with the hepatitis B virus (HBV) is typically established neonatally and progresses through different phases. Historically, the phases have been designated as immune tolerant (IT); HBeAg⁺, immune active (IA⁺); HBeAg⁻, immune active (IA⁻); and immune clearance (IC). In this issue of Gastroenterology, Mason et al¹ compared HBV integration, clonal hepatocyte expansion, and HBV-specific T-cell responses in the various phases of chronic HBV infection. Based on these findings and previous reports by Bertoletti and colleagues argue that the IT phase is neither disease free “nor is it associated with an immune profile of T-cell tolerance” and that “the data do not support the hypothesis that CHB in children and young adults is characterized by a state of immunological tolerance”.^1–4 As a consequence, they suggest changing the nomenclature from IT to “high replication, low inflammation (HRLI) phase.”

The authors’ skepticism is based primarily on the observation that HBV-specific T cells can be detected during the IT phase, although “these HBV-specific T cells were weak and functionally impaired as one would expect to see in patients with CHB,”² which raises the question of the fate of the strong and functional HBV-specific T cells. Immune tolerance as it applies to an infectious agent like HBV needs to be precisely defined. T-cell tolerance is clonal in nature and occurs 1 clone at a time either by deletion or by nondeletional mechanisms. Because T-cell tolerance is clonal, it is heterogeneous whether specific for different viral proteins or for different epitopes on the same protein. For example, in an HBV transgenic mouse model, the structural proteins (HBsAg, HBc/HBeAgs) elicit T-cell tolerance, whereas the polymerase does not.⁵ Similarly, in HBeAg transgenic mice 1 T-cell epitope is tolerogenic and another T-cell epitope on the same HBeAg is not.⁶ Evidence for immune tolerance to HBV is best defined operationally as the deletion or silencing of a sufficient number of T-cell clones to preclude significant liver injury and subsequently clearance or modulation of the infection (ie, the IT phase) but does not require a complete absence of all HBV-specific T cells. In contrast, the authors suggest that an IT patient should be defined as “immunologically inert.” Not even self-tolerance is defined by a total absence of self-reactive T cells. Deletion of 1 to several critical T-cell clones (eg, through exposure to the HBeAg in utero) may “tip the balance” toward persistence in neonatal infection.^7,8 Once persistence is established, tolerance can be widened to other epitopes and maintained by constant exposure to secreted HBeAg and HBsAg.⁹ The mechanisms of T-cell tolerance can include clonal deletion, clonal anergy, clonal ignorance, T regulatory activity, and regulatory pathways such as “immune checkpoint” inhibitory receptors.^10–12 Clonal deletion preferentially occurs in high-affinity T cells; therefore, low-affinity T cells are more likely to persist and more likely to be susceptible to all the various regulatory mechanisms operative in the HBV-specific T-cell repertoire during chronic infection.¹³ The liver also has unique immune regulatory functions that promote tolerance to antigens encountered locally.¹⁴ These factors may explain the low frequency of T cells, characteristic regulatory markers (ie, PD1, CTLA4, and TIM3) and poor function of T cells in chronic infection^{10–12,15–17} (Figure 1). Based on this discussion, it is not surprising that the authors detected functionally impaired, HBV-specific T cells during the IT phase of chronic infection.

Figure 1.

T-cell tolerance is clonal in hepatitis B virus (HBV) infection. A variety of T-cell clones of different specificities and affinities are available to respond to HBV infection in adults, who clear >90% of infections. In contrast, in the setting of neonatal infection, the HBV antigens behave as neo–self-antigens and elicit tolerance as opposed to activation of HBV-specific T cells, resulting in chronicity in >90% of infections. As depicted, T-cell tolerance is mediated by clonal deletion or by nondeletional mechanisms based on clonal affinities. The deletion or down-regulation of high-affinity T-cell clones results in lower numbers of low-affinity T cells remaining to represent the HBV-specific repertoire in chronic infection. These functionally impaired T cells are not directly detectable ex vivo and require in vitro clonal expansion of peripheral blood mononuclear cell to detect. In utero infection permits all structural and nonstructural HBV antigens (HBs, HBe/HBc, Pol, and X) to gain access to the developing thymus of the fetus potentially eliciting T-cell tolerance to the entire virion. In postnatal infection only the HBeAg may have access to the fetus in utero. Lesser rates of chronicity occur in perinatal infections depending on age at the time of infection. Chronic neonatal infection can progress from the immune tolerant (IT) phase characterized by high serum viral loads and no-to-minimum liver injury to the HBeAg⁺, immune active (IA⁺) phase in early adulthood characterized by fluctuating and decreased serum virus and increased liver injury yet failure to clear the infection. The IA⁻ phase is characterized by anti-HBe seroconversion and greater reductions in serum virus and more apparent liver injury, which can result in spontaneous clearance in a minority of infections. The timing of the transition from the IT to the IA phases correlates with the age-dependent involution of the thymus and may be analogous to the waning of self-tolerance. Neonates are not born tolerant to HBV antigens or infection and can respond to HBs vaccination and clear an acute infection if infected with the HBeAg⁻ mutant, respectively. Therefore, T-cell tolerance to HBV is an active process requiring exposure to HBV neo–self-antigens before or early in infection. Vaccination after chronic infection is established is ineffective in adults and children, and demonstrates the existence and relevance of immune tolerance in all phases of chronic infection.

Mason et al¹ also argue that “there is no quantifiable difference in antiviral immunity between the HRLI phase and IA⁺ patients.” First, it is important to note that HBV-specific T cells in all phases of chronic HBV infection are very difficult to detect ex vivo, especially in the blood, as compared with acute infection,^11,15,16 which is an a priori argument for immune tolerance. To detect HBV-specific T cells, most studies expand peripheral blood mononuclear cells in culture for 10 days in vitro.^1,2 In this study, the IT and IA⁺ patients were similar in age (mean age of the IT patients [23.4 years] was actually older than the IA⁺ patients [21 years]), mean serum HBV DNA levels, mean serum HBsAg titer; all patients in both groups were HBeAg⁺ and there were no significant histologic differences. Given the age, virologic, serologic, and histologic similarities between these patients, it is not surprising that no differences in HBV-specific T-cell responses, HBV integration events, or hepatocyte clonal expansion were observed in IT and IA⁺ patients. Perhaps, the selection of younger (ie, 0–14 years) IT patients would have yielded differences. In contrast, it was surprising that IT and IA⁻ patients demonstrated similar HBV-specific T-cell, levels indicating a lack of correlation with viral load, liver injury (alanine aminotransferase [ALT]), HBeAg status, and hepatocyte clonal size, which all differed between the IT and IA⁻ patients. Therefore, what is the nature and relevance of the detected T cells to HBV clearance? The in vitro expanded T cells may represent the maximum HBV-specific T-cell repertoires in these patients, but in vitro clonal expansion also obscures any differences in in vivo clonal frequencies and the function of these T cells in vivo is unknown. How is the detection of such T cells evidence for the absence of T-cell tolerance (eg, T cells may be quiescent in vivo) or evidence for an ongoing antiviral immune response in IT patients?

The authors cite a report from Vanwolleghem et al¹⁸ that demonstrates increased interferon-stimulated gene transcripts in IT patients as compared with IA patients. However, innate immune cells are not subject to antigen-specific, clonal tolerance and the up-regulation of interferon-stimulated gene transcripts most likely represents sensing of the higher HBV loads in IT patients. More important, B-cell, T-cell, and natural killer cell–related genes are all transcriptionally more active in the IA phase as compared with the IT phase.¹⁸ The authors also cite Wang et al¹⁹ for “virological data showing sequence evolution of HBV with increasing age in a cohort of IT patients.” According to the same Wang et al paper, “During the IT phase, when host immunity was feeble (ALT < 20 U/liter), viral nucleotide diversity decreased while copy numbers increased. Rates of evolutionary change derived for different patients were in a very narrow range. As the disease progressed toward the immunoclearance phase (ALT > 20 U/liter), viral diversity increased but copy number decreased. Evolutionary rates varied among patients in accordance with their levels of ALT.” These 2 reports do not support the view that “there is no quantifiable difference in antiviral immunity between the HRLI phase and IA⁺ patients”.

Finally Bertoletti et al suggest that HBV exposure in utero triggers a state of “trained immunity” that “challenges the role of immune tolerance in viral persistence after neonatal infection.”^1,4 Trained immunity in neonates may provide cross-protection against bacterial infections as well as against the primary pathogen. However, neonates of HBV-infected, HBeAg⁺ mothers are not protected against the primary HBV pathogen and become chronically infected in approximately 90% of cases. Further, there is little evidence for the HBV specificity of the induction or the HBV specificity of the effects of trained immunity. Therefore, it is not clear how this concept challenges the role of HBV-specific immune tolerance in neonates.

Implications for Therapeutic HBV Vaccine Design and Immunotherapy

Debating the role of immune tolerance in chronic HBV infection is not merely academic. I believe that the IT phase as well as all phases of chronic HBV infection are influenced directly to different degrees by immune tolerance mechanisms; therefore, the design of therapeutic vaccines, immunotherapies, and the decision of when to start treatment in chronic patients need to address the issue of a compromised HBV-specific T-cell repertoire. Although experimental (HBV transgenic), virologic, clinical, and pathologic findings are consistent with a role for immune tolerance in all phases of chronic HBV infection,^5–7,16,17 it is difficult to directly prove in humans. HBV has provided a rare opportunity to directly test the existence of human immune tolerance by immunization. Vaccination with a variety of mostly envelope HBV antigens, including recombinant proteins, DNA, peptides, and immune complexes often in combination with antiviral therapies has failed to elicit immune responses sufficient to modulate chronic infections whether in adults or in the young during the IT phase.^20,21 Poor to nonresponse to immunization, which defines immunologic tolerance in animals, provides definitive evidence for immune tolerance during chronic HBV infection. The important question is not whether T-cell tolerance exists in chronic HBV, but how to reverse it? The goal of therapeutic vaccination and immunotherapies is to activate non-deleted but low-affinity and/or regulated (ie, PD1⁺, CTLA4⁺, TIM3) HBV-specific T cells. Recently, significant progress has been made using innovative and diverse immunization strategies to overcome T-cell tolerance in animal models of chronic HBV infection²² and to restore HBV-specific T-cell function by blocking immune checkpoint inhibitory receptors.^11,12

The basic data in these studies are not being challenged. What is being challenged is the interpretation that minimal liver injury, HBV integration, clonal expansion of hepatocytes, and the labored detection of low levels of HBV-specific T cells in IT patients is inconsistent with HBV-specific T-cell tolerance mechanisms, which can maintain viral persistence for decades. Therefore, I do not agree that there is sufficient evidence to change the nomenclature from the IT phase to the HRLI phase of chronic HBV infection. The HRLI term is descriptive but the IT designation provides a mechanistic explanation for the high replicative and low inflammatory state. Further, it is appropriate and necessary to consider immune parameters in defining the various phases of chronic HBV infection inasmuch as the disease is primarily immune mediated.