Effect of dipterinyl calcium pentahydrate on hepatitis B virus replication in transgenic mice

Hepatitis B virus (HBV) causes both transient and persistent infections of the liver in humans. The number of chronic HBV carriers is estimated to be 400 million worldwide; nearly 25% of which are projected to succumb to liver failure or liver cancer [1]. Additionally, HBV infection remains an important cause of acute and chronic liver disease in the United States [2]. Dipterinyl calcium pentahydrate (DCP), shown in Figure 1, has demonstrated significant antitumor activity associated with plasma IL-12 concentration increases in MDA-MB-231 (human breast cancer) xenographs in nude mice [3, 4]. This finding, along with previous work demonstrating IL-12 suppression of HBV replication in transgenic mice [5], prompted us to investigate the activities of DCP in the HBV transgenic mouse model.

DCP is a stable, sparingly soluble compound that can be solubilized in aqueous solutions to 440 μM with sonication. For this study, an orally administered suspension was used. It was hypothesized that because of the antitumor changes elicited by DCP, as well as the anti- and pro-inflammatory plasma cytokine/chemokine concentration changes reported previously, DCP would reduce liver HBV DNA levels and possibly other HBV parameters in transgenic mice carrying an infectious clone of HBV. The investigators anticipated that DCP might work via cytokine/chemokine modulatory mechanisms similar to those described by others [5‐9].

Evidence for immunomodulation by DCP was further investigated by measurement and analysis of the enzyme indoleamine 2,3-dioxygenase (IDO) serum metabolites, tryptophan (Trp) and kynurenine (Kyn). Tryptophan (Trp) is the substrate for IDO, a key immunological inhibitor of T cells, and an identified tumor escape mechanism [10]. Recent studies have demonstrated that IDO-mediated immune suppression is prevalent in hepatitis B infection [11]. The IDO enzymatic product, kynurenine (Kyn), and its downstream metabolites, kynurenine, 3-hydroxykynurenine, and 3-hydroxyanthranilic acid, are directly involved in the regulation of T cells and other lymphocytic cell types, i.e., NK cells and B cells [12]. We have shown previously that DCP inhibits IDO activity in human PBMCs (Peripheral Blood Mononuclear Cells) [3]. Certain neurotoxic end-products of the tryptophan-kynurenine pathway, such as quinolinic acid, produced under chronic inflammatory conditions (e.g., cardiovascular disease, multiple sclerosis, diabetes, cancer, and major depression) may contribute to the brain damage seen in depression and dementia [13]. For the study described here, the calculation of the serum Kyn-to-Trp ratio (Kyn/Trp) allowed us to estimate the extent of IDO activity in the serum of the HBV mice [14].

The following viral, IDO, and cytokine/chemokine measures were collected in this study: liver HBV DNA (Southern), liver HBV DNA (PCR), liver HBV RNA (PCR), HBe antigen (ELISA); Average # HBcAg Nuclei, Average # HBcAg Cytoplasms, # HBcAg Nuclei per Quarter Field; serum Tryptophan, Kynurenine, Kyn/Trp, IL-1a, IL-1b, IL-2, IL-3, IL-4, IL-6, IL-9, IL-10, IL-12, MCP-1, TNF-α, MIP-1, GM-CSF, RANTES; and liver IL-6. See additional file 1: "Summary Statistics" for the data summary and statistical results from this study.

Table 1, Table 2 gives the derivation of the significant (p = .001) DCP Effects Model (DCP-EM) variable cluster from the linear multivariate stepwise regression of DCP dosage versus all the viral, IDO, and cytokine/chemokine measures. The logarithmically transformed liver HBV DNA (Southern), liver HBV DNA (PCR), liver HBV RNA (PCR), and HBe antigen (ELISA) measures were included in this analysis as in previous studies [9, 18, 19]. The DCP Effects Model linear stepwise regression analysis identified a significant (p = .001) cluster of variables responding to the DCP treatment. Stepwise regression analyses, in general, derive a small cluster of variables from a much larger set of variables in order to model a particular variable of interest, in the present study, the DCP dose. Taken individually, the component variables of a derived cluster do not achieve the same level of significance as the cluster taken as a whole. In the present study, Log liver HBV DNA (PCR) emerged as the single strongest viral variable which significantly and inversely modeled DCP dosage. None of the other viral measures significantly further explained the DCP dosage variance. These other viral measures included: 1) liver HBV DNA (Southern), 2) liver HBV DNA (PCR), 3) liver HBV RNA (PCR), 4) the log liver HBV DNA (Southern), and 5) log liver HBV RNA (PCR). The DCP Effects Model regression, however, identified the serum measures Kyn/Trp, MCP-1, and GM-CSF as significant additional variables which further accounted for DCP dosage variance (Table 1, Table 2).

The molecular biology of HBV RNA synthesis, which was not affected by any of the treatments in the HBV transgenic mice, is significantly different from that of the natural HBV infection.

The HBV RNA in the transgenic mice is produced primarily from an HBV transgene, unlike a natural infection wherein it is expressed from HBV covalently closed circular DNA (cccDNA) (Raney et al., 2001) [20]. Unlike the natural infection, HBV produced from the transgene in mice cannot infect mouse cells for successive rounds of viral replication. In the natural infection, HBV RNA and HBe and HBs proteins are derived from successive rounds of replication to produce increasing levels of cccDNA, and consequently, increasing levels of HBV RNA and proteins. In this regard the use of the HBV transgenic mice is analogous to HBV stably transfected, cells such as Hep G 2.2.15 cells (Iyer et al., 2004a) [21]. It was not unexpected, therefore, that a selective reduction of HBV DNA would not necessarily affect HBV RNA or protein levels, since these levels can be independently derived from the HBV RNA transcribed from the transgene. [9]

The Figure 2 profile of log rel. liver HBV DNA (PCR) by treatment shows significant efficacy for the female mice at 23 mg/(kg d) DCP (p < .05) as compared to female controls. The figure also shows that log rel. liver HBV DNA (PCR) is significantly lower than controls for 10 mg/(kg d) adefovir dipivoxil (p < .05) in both sexes. Moreover, Figure 2 shows that there are significant gender differences in viral DNA levels for all DCP dosage levels tested. A similar treatment response pattern was found for the log liver HBV DNA (Southern) measure (see additional file 1: "Summary Statistics").

In order to explain the Figure 2 DCP/gender differences, an ADJUSTED log rel. liver HBV DNA (PCR) was calculated for each mouse based upon the DCP-EM regression from Table 1, Table 2, rearranged algebraically as shown in the legend of Figure 3. The resultant ADJUSTED log rel. liver HBV DNA (PCR) values demonstrated no significant gender differences among the mice for any treatment group. Therefore, the log rel. liver HBV DNA (PCR) gender differences were accounted for by adjusting for the Kyn/Trp, MCP-1, and GM-CSF serum levels for each mouse, as shown in the re-plotted ADJUSTED log rel. liver HBV DNA (PCR) scores in Figure 3.

The significant Table 1, Table 2 linear stepwise regression quantifies a DCP Effects Model (DCP-EM) for the HBV mice showing a significant relationship between the inhibition of hepatitis B DNA replication and DCP dosage in the context of three serum immunological changes. The three immunological changes found to be significantly associated with DCP's reduction of viral DNA levels are: 1) decreased MCP-1; 2) decreased Kyn/Trp (an estimation of IDO activity); and 3) increased GM-CSF. Among female HBV mice, 23 mg/(kg d) DCP was found to reduce control values of log rel. HBV DNA (PCR) to levels comparable to those achieved by the antiviral agent adefovir dipivoxil (Figure 2). When serum Kyn/Trp (IDO activity), MCP-1, and GM-CSF levels were taken into account and ADJUSTED log rel. liver HBV DNA (PCR) values calculated for each mouse (Figure 3), gender differences in log rel. HBV DNA levels vanished, implicating the involvement of these immunological variables in the gender differences.

Table 1, Table 2 shows that DCP decreases MCP-1 serum levels. MCP-1 (Monocyte chemotactic protein-1) is a small cytokine belonging to the CC chemokine family, also known as Chemokine (C-C motif) ligand 2 (CCL2), which recruits monocytes, memory T cells, and dendritic cells to sites of tissue injury and infection [22, 23]. A number of genetic loci have been proposed to be associated with persistent hepatitis B virus (HBV) infection [24], including the MCP-1 -2518 G/G genotype and G allele, more frequent in HBV patients than controls (P < 0.01, after multiple corrections Pc < 0.05). In other studies with HBV-associated hepatoma cells, MCP-1 is found to be overexpressed [25]. Table 1, Table 2 also shows that DCP increases GM-CSF serum levels. GM-CSF induces the differentiation of granulocyte, macrophage, and eosinophil precursor cells, as well as the proliferation of monocyte-macrophages, T lymphocytes, keratinocytes, and endothelial cells. GM-CSF is produced by T lymphocytes, macrophages, and several cell types in extramedullary sites, where it may act in a paracrine manner to regulate the local response to antigenic challenge [26]. Clinical trials evaluating the efficacy and safety of GM-CSF as an adjuvant to hepatitis B vaccine in patients with end-stage renal disease showed improved seroprotection rates with HBV vaccine after GM-CSF administration [27]. Therefore, serum MCP-1 is a HBV risk factor decreased by DCP, while GM-CSF, whose serum level is increased by DCP, has been shown to be a effective HBV vaccine adjuvant.

The inhibition of the enzyme IDO by DCP is significantly correlated to the decrease of HBV DNA in the DCP Effects Model (Table 1, Table 2). Notably, the mean control value for female HBV mouse serum Kyn/Trp (22.2 ± 2.2 uM/mM) is greater than the mean control serum Kyn/Trp for male HBV mice (12.8 ± 1.1 uM/mM). Furthermore, the mean female HBV mouse Kyn/Trp value is closer in magnitude to normal human serum Kyn/Trp (26.5 - 45.0 uM/mM) [14, 28, 29]. These findings, taken together with the finding that the serum IDO gender difference for the HBV mice accounts, in part, for the gender differences seen in the Figure 2 DCP dose-response diagram, and strongly suggests that DCP has greater anti-HBV efficacy in female HBV mice partly because to their elevated serum IDO.

One could argue that the lower IDO activity in males would render this pathway more susceptible to the effects of DCP, and that a lower dose of this compound would be effective in targeting the more limiting IDO metabolites in male mice. However, since indoleamine 2, 3-dioxygenase (IDO) metabolizes the essential amino acid tryptophan in mammals, catalyzing the initial and rate-limiting step in the de novo biosynthesis of nicotinamide adenine dinucleotide (NAD) [30], excessively low levels of IDO, and consequently NAD, would be debilitating to the male mice. Males likely have metabolic mechanisms to prevent the IDO/NAD pathway from dropping below basal levels.

The observation that female transgenic mice have significantly higher serum IDO activity than the male mice is biologically relevant to HBV expression in these transgenic mice in the following ways. IDO immuno-inhibition has been found to take two forms, 1) as a depletor of the nutrient tryptophan, and 2) through the direct action of its enzymatic products, kynurenine, 3-hydroxykynurenine, and 3-hydroxyanthranilic acid on immune cells [10, 12, 14, 31, 32]. In previous studies, female HBV mice were identified to have slightly lower levels of liver and serum HBV DNA than male mice [33, 34]. Thus, these two gender-differentiated measures, liver HBV DNA levels and IDO activity, might be connected mechanistically. For example, HBV surface antigens can be regulated in HBV transgenic mice by sex steroids and glucocorticoids [35]. Similarly, both IDO, and its related hepatic enzyme TDO (hepatic tryptophan pyrrolase; tryptophan 2,3-dioxygenase) can be upregulated by steroids like estrogen [36‐38]. Therefore, there are links between the lower levels of liver HBV DNA in female mice, and hormone-regulated tryptophan metabolism. Further evidence for a gender-HBV linkage is that hepatocellular carcinoma (HCC), one of the human cancers etiologically related to HBV, affects males in a significantly higher proportion than females [39]. A similar gender effect is seen in HBV-transgenic C57BL/6 mice [40]. Also, male gender is a predictor of poor response to IFN-α treatment of chronic hepatitis B (CHB) [41].

The involvement of IDO activity in determining the levels of liver HBV DNA can explain the mechanism of the anti-HBV activity of DCP. Since DCP is known to inhibit IDO in human PBMCs [3], the higher IDO activity found in the sera of the female HBV mice allows for a proportionally greater degree of IDO inhibition by DCP, and consequently relatively greater immuno-enhancement, since IDO is a Th1-inhibitory enzyme. A higher IDO activity in the female HBV mice provides more metabolic targets for inhibition by DCP, and a greater de-repression of Th1 immune responses. In humans, altered IDO-related serum metabolite concentrations, i.e., lower tryptophan and kynurenine levels, have been found in young adult females as compared to young adult males [42].

The serum IDO differences between humans and mice might also be related to iNOS differences because 1) human macrophages are deficient in high output NO production [43, 44], and 2) NO interferes with IDO expression and function [45]. Thus, the lower serum Kyn/Trp (i.e., lower IDO activity) in mice as compared with humans might be due to the higher NO levels in mice. Further examination of these possible factors influencing serum IDO is of interest since endogenous IDO levels are found to play a significant role in the antiviral activity of the immunomodulator DCP.

DCP inhibition of serum IDO as part of its anti-HBV effect reduces circulating kynurenine levels, and increases tryptophan availability, both of which enhance T cell-, B cell- and NK-mediated immunity [10, 12, 14, 31, 32]. Studies linking the collapse of HBV-specific CD8 T-cells, and impaired innate immunity (NK cells) in persons with chronic hepatitis B (CHB) have been reviewed [46]. HBV-specific CD8⁺ Tcells are associated with viral control (HBV DNA load) [47, 48], and reducing HBV replication to <10⁷ copies/ml is suggested to maximize the chances of vaccines to expand a broad repertoire of HBV-specific CD8⁺ cells [48]. Our data shows that DCP significantly inhibits IDO in the HBV transgenic mice. Moreover, a number of studies show a significant positive linkage between IDO and the PD-1/PD-L1 pathway [49‐54]. PD-1 receptors are found on activated T cells, B cells, macrophages, and DCs [55], and expansion of the PD-1 pathway by T cells was associated with viral replication in a model of chronic HBV infection [56]. Therefore, a DCP-driven decrease of IDO activity should, in turn, downgrade the PD-1 pathway, expand HBV-specific T cells, and lead to a decrease in HBV DNA. Further study into these connections can more firmly establish this model.

A recent review of the woodchuck hepatitis virus model [57] discusses immunotherapeutic approaches based upon the reconstitution of Th1 immune responses. IFN-γ antiviral immune responses stimulate IDO activity in a self-limiting manner due to the T cell inhibitory effects of the IDO-produced kynurenines. By inhibiting IDO with inhibitors such as DCP, the Th1 type immune response can be enhanced while the negative effects of kynurenine are decreased. However, while IDO inhibition probably improves immune functioning, it may also serve to enhance the rate of viral reproduction due to the inhibition of tryptophan deprivation and consequent enhancement of growth conditions. On the other hand, it is also possible that some of the immunological effects of DCP are due to a general growth inhibition that diminishes virus multiplication. Also, the present study did not measure the plasma levels of HBV DNA, and the HBs and HBe antigens, which would be critical for the evaluation of the effects of DCP upon circulating HBV titers, viral assembly, and infectivity.

A linear stepwise regression model, the DCP Effects Model (DCP-EM), was generated from viral and immunological measures collected from hepatitis B virus (HBV) transgenic mice treated with dipterinyl calcium pentahydrate (DCP). Three significant serum changes were found to be associated with the inhibition of HBV DNA (PCR) replication by DCP. These immunological changes were: 1) decreased MCP-1; 2) decreased Kyn/Trp (an estimation of IDO activity); and 3) increased GM-CSF. Among the female HBV mice, 23 mg/(kg d) DCP was found to reduce log HBV DNA to levels comparable to those achieved by the antiviral agent adefovir dipivoxil.