https://orcid.org/0000-0002-6143-3104
Worldwide, approximately, 400 million people have chronic hepatitis B virus (HBV) infection (chronic hepatitis B [CHB]), a leading cause of liver-related death [1]. A highly effective vaccine has prevented millions of infections but belies the challenge to treat and cure those with CHB. The complexity of current HBV treatment guidelines is partially attributed to the suboptimal, oral antiviral treatments for CHB that are available; nucleos(t)ide analogues (NUCs) only suppress viral replication but do not eliminate viral genomes from hepatocytes in the liver where they reside. Treatment is reserved only for persons with signs of liver damage, especially when there is abundant viral replication. Furthermore, when treatment is initiated, it is lifelong for the majority of people. Therefore, better medicines are needed for CHB.
Management of CHB requires monitoring the blood for hepatitis B surface antigen (HBsAg). People who have the lowest levels of HBsAg (historically called “HBsAg loss”) are considered functionally cured. Because a small proportion of people treated with NUCs achieve HBsAg loss after NUC discontinuation, there is hope for this type of cure. Functional cure is most similar to spontaneous HBV recovery in that HBsAg is not detectable in blood, but some hepatocytes retain HBV viral genomes [2, 3]. People who achieve HBsAg loss are considered only functionally cured because these genomes can reactivate during immunosuppression. Drug development has been focused on expanding the proportion achieving functional cure, but there has only been incremental progress.
DIFFICULTY OF ACHIEVING HBsAg LOSS WITH NUC THERAPY
The intricacies of HBsAg loss are best appreciated by understanding the basics of the HBV life cycle (Figure 1). The HBV genomic template for viral replication in the hepatocyte is the covalently closed circular DNA (cccDNA), a stable minichromosome residing in the nucleus. This cccDNA assembles with histone proteins and persists in every infected hepatocyte for its lifespan, comprising a reservoir that resists elimination. Host RNA polymerases transcribe cccDNA into all the viral messenger RNAs (mRNAs) required for replication, including a pregenomic RNA (pgRNA) and 2 mRNAs that are translated to HBsAg. pgRNA is packaged along with the HBV polymerase into a viral capsid that is enveloped and secreted to infect new hepatocytes. The polymerase protein reverse-transcribes pgRNA to yield a relaxed circular HBV DNA. However, in 10% of reverse transcriptions, pgRNA is instead converted into a double-stranded linear HBV DNA [4]. Rather than yielding cccDNA in the newly infected cell, a double-stranded linear HBV DNA integrates into the human genome (iDNA); however, this iDNA is not a full-length genome and cannot produce new full virions. In many cases, it is sufficient to produce HBsAg. Thus, HBsAg derives from both cccDNA and iDNA.
Schematic illustrations of hepatitis B virus (HBV)–infected hepatocytes. Covalently closed circular DNA (cccDNA) is the genomic template for viral RNAs, including pregenomic (pgRNA), which is encapsidated and reverse-transcribed in infectious virions that are secreted into blood and measured as serum HBV DNA. cccDNA is also transcribed into HBV surface (S) messenger RNAs (mRNAs) encoding hepatitis B surface antigen (HBsAg), which is secreted into blood. Integrated HBV DNA (iDNA) can also be transcribed into HBV S mRNAs encoding HBsAg. On the left is a hepatocyte from someone who is hepatitis B e antigen (HBeAg) positive or untreated; in these hepatocytes, cccDNA produces an abundance of viral mRNAs that contribute to high levels of serum HBV DNA and HBsAg. iDNA-derived S mRNAs also contribute to HBsAg production, but to a lesser extent than cccDNA. On the right is a hepatocyte from someone who is HBeAg negative or treated with nucleos(t)ide analogues (NUCs); in these hepatocytes, cccDNA-derived transcripts are dramatically reduced and reverse-transcription is lessened, resulting in a major decrease in serum HBV DNA levels. iDNA-derived S mRNAs do not appear to be reduced in these people, leading to sustained levels of HBsAg in blood. Abbreviations: dslDNA, double-stranded linear DNA; rcDNA, relaxed circular DNA.
NUCs inhibit reverse transcription by interfering with the viral polymerase: thus, it may appear obvious that HBsAg loss does not occur with NUC therapy since NUCs do not affect cccDNA transcription by host RNA polymerases. Intriguingly, we and other groups have discovered that NUC therapy is associated with substantially decreased cccDNA transcription (and even silencing) through yet undetermined mechanisms [5–9] (Figure 1). So, if cccDNA is the primary source of HBsAg, then NUC therapy should lead to substantial declines in HBsAg. However, the conundrum is that HBsAg loss rarely occurs, and quantitative levels barely change during NUC therapy. This apparent paradox may be best explained by focusing on HBsAg production from iDNA.
EVIDENCE FOR iDNA AS A MAJOR SOURCE OF HBsAg
CHB goes through phases that are unique among chronic viral infections [10]. Early after infection, HBV DNA levels are high, corresponding to active transcription from cccDNA, hepatitis B e antigen (HBeAg) is detectable in blood, and there is negligible hepatic inflammation (termed the HBeAg-positive chronic infection phase); there is subsequent increase in hepatic inflammation (HBeAg-positive chronic hepatitis phase). Later, HBeAg often disappears and is accompanied by the emergence of anti-HBe antibodies and lower HBV DNA levels that correspond to reduced or inactive transcription from cccDNA (HBeAg-negative phases). In most adults, CHB is diagnosed at this stage.
The first demonstration of the importance of iDNA as a source of HBsAg is a tour-de-force study by Wooddell et al [11] who showed that a small interfering RNA (siRNA) administered to persons with HBeAg-positive CHB led to nearly 100-fold decreases in HBsAg levels. In contrast, persons with HBeAg-negative CHB had minimal HBsAg declines. Using RNA sequences from archived liver samples from chimpanzees with HBeAg-negative and HBeAg-positive CHB, the team deduced that in HBeAg-negative CHB, mRNA encoding HBsAg largely derived from iDNA. However, sequences from HBeAg-positive CHB revealed that HBsAg derives chiefly from cccDNA. With administration of a redesigned siRNA accounting for iDNA-derived and cccDNA-derived transcripts, persons with HBeAg-negative CHB had declines in HBsAg levels comparable to those in persons with HBeAg-positive CHB. Thus, they concluded that iDNA, which itself is not replication competent, was nevertheless sufficient to produce transcripts that could be translated into HBsAg. Several groups have since confirmed that iDNA-derived transcription is a major contributor to HBsAg production in HBeAg-negative CHB [12, 13].
Our team provided additional evidence for iDNA as an important source of HBsAg. We hypothesized that people who lacked meaningful declines in HBsAg levels during NUC therapy were likely to have iDNA-derived HBsAg. We developed a multiplex droplet digital PCR assay that distinguished cccDNA or iDNA as the predominant source of HBsAg and applied it to liver tissue. Persons with largely cccDNA-derived mRNAs had substantial declines in HBsAg levels, whereas persons with largely iDNA-derived mRNAs had minimal declines [7]. Our data using single-cell methods from paired liver tissue samples suggests that while NUC therapy is associated with reductions in cells with cccDNA-derived HBsAg, cells with iDNA-derived HBsAg persist for years [14]. Collectively, these studies strongly indicate that a major impediment to a functional HBV cure is the persistence of cells with iDNA-derived HBsAg production.
TRANSLATING KNOWLEDGE ABOUT HBsAg PRODUCTION INTO AN HBV CURE
This clinical question then arises: If there are some individuals in whom NUC therapy leads to HBsAg production only from iDNA that is maintained off treatment, should HBsAg loss be required for functional cure? One study demonstrated that, among those who achieved HBV DNA suppression with NUC therapy, the addition of HBsAg loss did not affect the risk of progressive liver disease but it did decrease the risk of hepatocellular carcinoma. In addition, there may be immunologic effects of HBsAg, irrespective of its source, that warrant treatment [15]. Thus, it seems prudent to try and achieve HBsAg loss.
Further advances in achieving durable HBsAg loss after NUC discontinuation will require a better understanding of iDNA. However, the leading emerging treatments—siRNAs and antisense RNAs—have all been redesigned to address iDNA- and cccDNA-derived transcripts simultaneously. There is therefore a risk of a nihilistic view if all HBsAg is grouped together, irrespective of source. Nonetheless, even with the redesigned siRNAs, the best combination therapies have yielded functional cures in only 10%–30% of trial participants. This may be because the turnover of hepatocytes with iDNA-derived HBsAg differs markedly from that of hepatocytes with cccDNA-derived HBsAg. Such information will require further detailed prospective studies of liver tissue samples from people with CHB, including both HBeAg-positive and HBeAg-negative patients, on long-term NUC therapy to build models of decay rates. In addition, clinical trial durations must be adjusted to account for the dynamics of the cell population relevant to the treatment given.
Other emerging agents, such as the capsid assembly modulators, may address only cccDNA-based replication and not iDNA. Furthermore, we need to develop tools that can be applied to blood samples that can differentiate HBsAg deriving primarily from iDNA or cccDNA; such tools could then be used in clinical trials to learn whether the effectiveness of novel therapeutics depends on the HBsAg source. Along those lines, if there are combinations of treatments that are more efficacious in either people with largely cccDNA-derived HBsAg or those with iDNA-derived HBsAg, it would be critical to stratify clinical trials by the source of HBsAg in participants.
Translational studies have brought us closer to developing a functional HBV cure, but there are critical gaps that need to be bridged. Doing so will require continued linking between outcomes in clinical trials and molecular features of the virus in liver tissue and blood samples from people with CHB. This approach will require intensive basic-clinical collaborations that could transform lives for 400 million people.
Notes
Acknowledgments. Figure 1 was created using Biorender.com.
Financial support. This work was supported by National Institute of Allergy and Infectious Diseases__ grants R01 AI116868, R56 AI138810, R21 AI157760, and R21 AI165166.
References
Author notes
Potential conflicts of interest. Both authors: No reported conflicts. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
