While it is well-established that Treg cells play a central role in maintaining immune system balance, the roles of Treg cells in HBV disease have been largely inconclusive . Inconsistent study designs and suboptimal definition of Tregs, such as relying solely on co-expression of CD4+ and CD25hiFoxp3+ to identify this subset, are the most likely culprits. In this study, we employed the CD39/ENTPD1, a newly described ectoenzyme that mediates regulatory T cell functions, to define the discrete subsets in Foxp3+CD4+ Treg cells. We first characterized the peripheral and intrahepatic CD39+FoxP3+CD4+ T cells (CD39+ Tregs), most of which were found to be CD45RA- Treg cells and displayed stronger suppressive effects than CD39- Tregs. Then, we exploited this approach to investigate the role of Tregs in HBV infection and demonstrated, for the first time, that the CD39+ Tregs are closely correlated to serum HBV copies and ALT, but not correlated with HBeAg in CHB patients.
In this study, FACS analysis of a large number of patient samples revealed a discrete distribution of CD39 expressing CD4+ T cells in healthy controls and hepatitis B patients, which were discernable by their levels of CD39 expression. Accordingly, the CD39+CD4+T cells were categorized among three groups: CD39low, CD39int, and CD39high. Furthermore, the three groups showed distinctive association patterns with hepatitis disease status. Although we also observed an association between the frequency of total CD39+ Treg cells with disease status, the subtypes of CD39+CD4+ T cells provided more detailed information than the total CD39+CD4+ T cells alone. For instance, when analysis was performed with the CD39high/int/low groups, the HBV viral copies were found to be significantly increased in the CD39high group, as compared with the CD39low group in all HBV-infected subjects; however, the serum ALT levels were increased in the CD39low group, as compared with the CD39int/high groups in CAH patients.
Friedman and colleagues found that tag-single nucleotide polymorphisms (SNPs) of the rs10748643 genotype correlated strongly with CD39 expression, whereby GG was associated with much higher levels than AA (46-77% increase) and AG carriers displayed intermediate expression levels in cell lines from subjects with Caucasian/European, Yoruban, Chinese, and Japanese ancestry . In our study, we also observed a similar phenomenon. However, further investigations are required to determine whether the CD39 gene SNP is associated with any of the three groups (CD39low, CD39int, and CD39high) in this study. Such research is anticipated to provide further evidence for CD39 as a more suitable marker for Treg cells in combination with Foxp3.
To date, research has not been able to establish whether circulating CD4+CD25+ Treg frequency is correlated with HBV replication. By using CD39 to define the conventional Treg cells, we found in this study that an increase in the proportion of CD39+ Tregs occurred in AsCs patients; however, such an increase did not occur in the total Treg cells, which was similar to the proportions observed in healthy controls. Such a scenario demonstrated that CD39+ Tregs may suppress the HBV-specific CD8+ CTL effector function through the CD39/adenosine pathway , thereby maintaining immune tolerance to the invasive pathogen. This type of immunomodulation would be expected to further contribute to HBV replication in AsCs patients, who are characterized by high HBV load and low or normal serum levels [33, 34].
Nikolova and colleagues reported that the Treg-CD39 subset inhibits cytokine production by HIV-specific CD8+ T cells, an effect which could be partially relieved by pre-incubation of CD39+ Tregs with anti-CD39 mAb . Thus, we further analyzed the relationship between the proportion of CD39+ Tregs and clinical pathological parameters. Statistical analysis showed that the percentage of CD39+ Tregs was positively correlated with serum HBV load, especially in AsCs patients. However, we did not find any such correlation between HBV load and total Tregs. Based on the data in this study, there is no significant correlation between HBeAg expression and the frequency of total Tregs or CD39+ Tregs. It is possible that the data in our study, which was obtained from routine clinical tests, did not represent an absolute quantification of the HBeAg protein expressions. Therefore, we were limited in our ability to investigate the relationship between Treg frequency and HBeAg expression in-depth. Our follow-up studies have been designed to address this issue while evaluating the potential clinical utility of manipulating CD39+ Treg-mediated immunosuppression.
It is widely believed that the stages of CAH and ACLF in CHB infection are characterized by lower HBV copy number and higher serum ALT level, the latter of which often serves as a reliable marker of liver injury. Interestingly, in this study, we found that an increase in total Tregs was only observed in CAH and ACLF patients, as compared with AsCs patients or healthy controls. However, the significantly increased frequency of total Tregs showed no correlation with the increased serum ALT levels in these patients, thereby indicating that the increment of the total Tregs might reflect a homeostatic compensatory mechanism against inflammation. Nevertheless, the proportion of CD39+ Tregs was found to be decreased in CAH and ACLF patients, as compared to those in the AsCs patients. We further observed that the percentage of CD39+ Tregs was negatively correlated with serum ALT levels, especially in CAH patients, which might be due to subtle shifts in the subset composition of the total Tregs and which would be most clearly evidenced by the reduced frequency of CD39+ Tregs.
It has been speculated that CD39 expression may be modulated by the inflammatory milieu. In some cases, this may lead to FoxP3 dysfunction and down-regulation of functional markers, such as CD39. Such impaired CD39+ Treg cells might lose their ability to effectively hydrolyze extracellular ATP and ADP that is released by injured cells, thereby also losing their ability to inhibit the production of IL-17  or other proinflammatory cytokines, such as IL-1β [33, 36]. In a study of hepatic flares in chronic hepatitis B conducted by Tan et al., no increases in CD25+FoxP3+CD4+ T cells or CD39+CD4+ T cells were observed at any of the time points tested, suggesting that these cells may not play any role in inhibiting HBV-specific T cell responses in vivo. However, as many non-Treg CD4+ T cells also express CD39 molecules, it is possible that the role of the CD39+ Treg subset was masked. Friedman and colleagues observed that, under conditions of inflammatory bowel disease, CD39-null mice suffer more severe injury than wild-type mice and exhibit a general tendency toward exaggerated Th1/IFN-γ responses . Fletcher and colleagues reported that CD39+Foxp3+ Treg cells can suppress pathogenic Th17 cells and are impaired in cases of multiple sclerosis . Collectively, these evidences support the notion that while CD39+ Treg cells can mediate strong inhibitive effects on other immune effector cells in vivo, the CD39+ Tregs themselves can be modulated by various inflammatory factors. Loss of CD39 expression and of the related immunosuppresive functions during an inflammatory flare in CAH and ACLF patients could explain the negative correlation observed between the increased ALT levels and decreased CD39+ Tregs frequency.