GS-9620 [8-(3-(pyrrolidin-1-ylmethyl)benzyl)-4-amino-2-butoxy-7,8-dihydropteridin-6(5H)-one] is a potent, orally bioavailable small-molecule agonist of Toll-like receptor 7 (TLR7) developed for finite treatment of chronic hepatitis B viral (HBV) infection, with the goal of inducing a liver-targeted antiviral effect without inducing the adverse effects associated with current systemic interferon-α (IFN-α) therapies. We characterized the pharmacodynamic response of GS-9620 in CD-1 mice and cynomolgus monkeys following intravenous or oral administration and showed that GS-9620 induces the production of select chemokines and cytokines, including IFN-α and interferon-stimulated genes (ISGs). It is noteworthy that we also demonstrated that, in animals and healthy human volunteers, oral administration of GS-9620 can induce a type I interferon-dependent antiviral innate immune response, as measured by whole-blood mRNA of the ISGs 2′5′-oligoadenylate synthetase 1 (OAS1) and myxovirus resistance 1 (MX1), without the induction of detectable systemic IFN-α, i.e., a presystemic response. Additionally, presystemic induction of hepatic OAS1 and MX1 mRNA was observed in CD-1 mice in the absence of detectable systemic IFN-α. We propose that the mechanism of this presystemic response is likely its high intestinal absorption, which facilitates localized activation of TLR7, probably in plasmacytoid dendritic cells at the level of gut-associated lymphoid tissue and/or the liver. This localized response is further supported by data that indicate only minimal contributions of systemic immune stimulation to the overall pharmacodynamic response to orally administered GS-9620. These data demonstrate that GS-9620 can induce an antiviral innate immune response without inducing a systemic IFN-α response and thus suggest the therapeutic potential of this approach in the treatment of chronic HBV infection.
Toll-like receptor 7 (TLR7) is a pathogen recognition receptor that plays an important role in the detection of, and the innate immune response to, pathogens. TLR7 is predominantly activated by viral single-stranded RNA and is localized within the endolysosomal compartments of plasmacytoid dendritic cells (pDCs) and B lymphocytes in humans and nonhuman primates (Jarrossay et al., 2001; Kadowaki et al., 2001; Hornung et al., 2002; Ketloy et al., 2008). Activation of pDCs plays an important role in the innate response to viral pathogens, and these cells are responsible for the majority of type I interferon (IFN) production during the acute phase of a viral infection. The induction and secretion of endogenous IFNs (e.g., IFN-α, IFN-β) also facilitate the development of an efficient adaptive immune response (Colonna et al., 2004; Feld and Hoofnagle, 2005; Kanzler et al., 2007; Barbalat et al., 2011). Interferons induce the transcription of interferon-stimulated genes (ISGs), which generate an antiviral state within cells (de Veer et al., 2001; Schoggins et al., 2011), as well as induce the production of other cytokines and chemokines which facilitate intercellular communication and cellular trafficking.
Administration of exogenous IFN-α (e.g., recombinant IFN-α or pegylated IFN-α) provides therapeutic benefit in patients with chronic hepatitis B (HBV) and C viral infections. In HBV, durable responses, normalization of alanine aminotransferase, and sustained reduction in viremia have been attained in a small percentage (<20%) of patients treated for 1 year with exogenous IFN-α (Marcellin et al., 2009; Moucari et al., 2009; Reijnders et al., 2011). A key observation is loss of HBV surface antigen (HBsAg) and seroconversion for anti-HBsAg antibody following IFN-α treatment increases for several years after discontinuation. This supports the hypothesis that viral control is due to immune modulation and slow induction of a protective antiviral immune response.
However, exogenous IFN-α treatment has several notable barriers to patient acceptance. These include the necessity of subcutaneous administration, and treatment-limiting adverse effects (Marcellin et al., 2009; Moucari et al., 2009; Reijnders et al., 2011). The low rate of HBsAg loss and seroconversion and high rate of adverse events associated with current therapies illustrate the need for new approaches to induce a protective antiviral immune response and durable cure in patients with chronic HBV. Consequently, orally administered immunomodulatory therapies are an attractive approach to developing tolerable, safer therapies that are more broadly applicable and potentially associated with higher long-term viral eradication rates in a larger percentage of treated patients.
GS-9620 [8-(3-(pyrrolidin-1-ylmethyl)benzyl)-4-amino-2-butoxy-7,8-dihydropteridin-6(5H)-one], a dihydropteridinone derivative, was designed to induce a presystemic response, defined as local stimulation of innate immune cells, probably in pDCs in gut-associated lymphoid tissue (GALT) and/or the liver without the induction of a systemic IFN-α response. GS-9620 possesses high absorption properties in the gastrointestinal tract but has moderate clearance during first-pass hepatic metabolism. These properties minimize systemic exposure of GS-9620 following oral administration while targeting exposure to, and stimulation of, innate immune cells within the gastrointestinal tract and/or liver. This presystemic response can be detected in whole blood by the presence of 2′5′-oligoadenylate synthetase 1 (OAS1) and myxovirus resistance 1 (MX1) mRNA. Both OAS1 and MX1 are specific and sensitive biomarkers of IFN-α induction (Schattner et al., 1981; von Wussow et al., 1990; Roers et al., 1994). The therapeutic benefit of TLR7 stimulation by oral GS-9620 in animal models of chronic HBV infection has been shown in both chimpanzees infected with HBV and the woodchuck hepatitis virus model. In both models, oral administration of GS-9620 reduced viremia and induced an antiviral immune response, and in woodchucks, GS-9620 induced an antibody response to viral surface antigen with subsequent clearance of viral S-antigen (Menne et al., 2011; Lanford et al., 2013).
We conducted a series of studies to explore the pharmacodynamic and pharmacokinetic properties of GS-9620, a selective TLR7 agonist, and to determine whether oral administration of low doses of GS-9620 could induce OAS1 and MX1 mRNA in peripheral blood in the absence of detectable systemic IFN-α in mice and cynomolgus monkeys. Additionally, we assessed the relative contribution of systemic immune stimulation to the pharmacodynamic response to GS-9620.
Materials and Methods
In addition to the information described in the following sections, details on experimental approaches and reagents are available in the Supplemental Materials.
Nuclear Factor κB Reporter Assay
Murine TLR activity assays were performed at InvivoGen (San Diego, CA; http://www.invivogen.com/hek-blue-tlr-cells).
Splenocyte and Peripheral Blood Mononuclear Cell In Vitro Cytokine Release Assays
Cytokine release in mice (splenocytes) and cynomolgus monkeys [peripheral blood mononuclear cells (PBMC)] was assessed by Eurofins (Bothell, WA) using standard methods. Secreted proteins in cell culture supernatants were assessed by multiplex immunoassay using species-specific, standard Luminex (Austin, TX) and/or enzyme-linked immunosorbent assay technologies.
In Vivo Studies
In vivo studies were conducted at Covance Laboratories, Inc. All protocols were approved by the respective Institutional Animal Care and Use Committees.
Blood was collected at various time points and processed for serum and/or RNA. In the mouse studies, three to five mice were euthanized at each time point via cardiac puncture under CO2 anesthesia. For cynomolgus monkeys, serial blood samples were collected from conscious animals via the femoral vein or other suitable veins.
Cytokine evaluation was performed by Eurofins. Cytokine detection in serum was accomplished by a multiplex immunoassay using species-specific, standard Luminex and/or enzyme-linked immunosorbent assay technologies.
As described earlier, whole blood was collected into tubes containing K2EDTA. Isolated whole-blood total RNA was quantified by absorbance spectroscopy, and equal quantities were used as a template for 1-step reverse-transcription polymerase chain reaction TaqMan reactions to measure the level of RNA transcripts of the interferon-stimulated genes OAS1 and MX1 and the control genes glyceraldehyde-3-phosphate dehydrogenase or 18S. Individual primer/probe sets were purchased from Life Technologies (Carlsbad, CA). Each mRNA signal was normalized to the mRNA signal of the endogenous control (glyceraldehyde-3-phosphate dehydrogenase or 18S ribosomal RNA) in the same sample using a duo-plex approach.
Unless otherwise stated, male and female animals administered the same dose were combined, and values below the lower limit of quantification (LLOQ) were excluded from analysis. Fold increases in ISG values were calculated for each individual animal relative to the predose value if available, or were calculated relative to the ISG mean from animals treated concurrently with vehicle. Mean ± S.D. fold increases in ISG for each dose level were calculated using individual fold change values. Where data are presented in the absence of systemic IFN-α, the absence of systemic IFN-α is defined as IFN-α levels below the LLOQ at all time points where IFN-α was evaluated for the individual animal.
Human Sample Analysis.
Human blood samples were collected from healthy volunteers and analyzed for IFN-α, MX1, and OAS1 as previously described (Lopatin et al., 2013).
GS-9620 Is Selective for TLR7 over Other TLRs in a Nuclear Factor κB–Dependent Cell-Based Reporter Assay.
To assess the selectivity of GS-9620 for TLR7 versus other TLRs, GS-9620 was evaluated in a nuclear factor κB (NF-κB)–dependent reporter cell–based assay with murine TLRs. In this assay, GS-9620 activated mouse TLR7 at concentrations ≥100 nM (Fig. 1A) with no detectable activation of the other TLRs, including TLR8, up to the maximum concentration tested (100 µM; Fig. 1B), whereas all TLRs were activated by their respective positive control agents. Prior studies evaluated the selectivity of GS-9620 for human TLR7 and demonstrated an approximately 30-fold selectivity of GS-9620 for activation of human TLR7 over TLR8 with no detectable activity on other human TLRs in similar reporter cell–based assays (Tumas et al., 2011). However, it should be noted that NF-κB is only one of several transcription factors activated by TLR7 signaling (He et al., 2013), and therefore the reporter assay used in this study does not necessarily permit extrapolation about GS-9620 function in different cells or species, but merely served as a tool to investigate receptor selectivity.
GS-9620–Dependent In Vitro Cytokine Secretion.
The pharmacodynamic response of GS-9620 was initially assessed in splenocytes isolated from CD-1 mice and in PBMCs isolated from healthy cynomolgus monkeys. By comparing the pattern of cytokines and chemokines induced in these primary cellular assays, we assessed the selectivity of GS-9620 for TLR7, as measured by induction of IFN-α, versus TLR8, as measured by the induction of tumor necrosis factor-α (TNF-α), in a PBMC assay with a complete complement of immune competent cells. Additionally, we assessed how species-specific differences in the expression of TLR7 and TLR8 can affect the observed GS-9620 selectivity in primary cells. It has been reported that murine TLR7 expression is notable on both conventional dendritic cells (cDCs) and pDCs, which is in contrast to cynomolgus monkeys and humans, where TLR7 expression is largely restricted to pDCs and TLR8 expression is predominantly expressed in cDCs (Edwards et al., 2003; Ketloy et al., 2008).
Not unexpectedly and consistent with the reported differences in TLR7/8 expression in the mouse, the pattern of cytokines and chemokines induced by GS-9620 was notably different in murine splenocytes when compared with the pharmacodynamic response in cynomolgus monkey PBMCs. In murine splenocytes, the minimum effective concentration (MEC) of GS-9620 to induce IFN-α was 30 µM, whereas the MEC for the induction of TNF-α was 30 nM (Table 1). The relative amount of TNF-α induced was also higher than the induced levels of IFN-α in these cultures, and dose-response relationships appeared roughly linear (Fig. 2A).
In contrast to the mouse, GS-9620 preferentially induced IFN-α over TNF-α at low concentrations in cynomolgus monkey PBMCs. The amount of TNF-α induced was markedly lower than IFN-α at GS-9620 concentrations less than 10 µM, and dose-dependent induction of TNF-α was only observed at concentrations greater than 1 µM (Fig. 2B). This is consistent with GS-9620 selectivity for TLR7 versus TLR8 at lower concentrations and combined TLR7 and TLR8 stimulation at higher concentrations of GS-9620. TLR7 agonism preferentially mediates secretion of IFN-α in PBMC cultures, whereas TLR8 stimulation in cDC and monocyte PBMC subsets is viewed as the dominant trigger for secretion of inflammatory cytokines, e.g., TNF-α (Gorden et al., 2005; Thomas et al., 2007). Although GS-9620 induced IFN-α at a mean (± S.D.) MEC of 308 ± 463 nM and induced TNF-α at a mean (± S.D.) MEC of 285 ± 478 nM (Table 1), the similar MEC for IFN-α and TNF-α is not unexpected, as TLR7-dependent activation of pDCs can result in proinflammatory cytokine secretion on a small scale, as was observed here. This observed pattern and profile of induced cytokines in the cynomolgus monkey was generally comparable to that which we have reported in human PBMCs, where we have shown that the mean MEC for induction of IFN-α and TNF-α were 66 and 3650 nM, respectively, in the same set of donors (Tumas et al., 2011).
GS-9620 Has High Intestinal Absorption, but Low to Intermediate Metabolic Stability and Bioavailability.
GS-9620 was designed to have high intestinal absorption and moderate clearance, primarily via hepatic metabolism by CYP3A4, during hepatic first-pass extraction to maximize exposure to interferon-producing cells in the GALT and/or liver while minimizing systemic exposure. In vitro experiments with hepatic microsomal fractions from CD-1 mice, cynomolgus monkeys, and humans revealed a GS-9620 half-life of 4.9, 5.9, and 30.5 minutes, respectively, indicating that GS-9620 will have moderate to high hepatic metabolic clearance in all species. Consistent with the in vitro hepatic microsomal prediction, GS-9620 has low oral bioavailability in mouse (0.2%) and monkey (1.1%) and moderate bioavailability in dog (20%) (Table 2). However, data from a study in portal vein–cannulated male beagle dogs demonstrated that the amount of GS-9620 entering the portal vein was high (82%; Supplemental Table 1), indicating that GS-9620 is readily absorbed from the gastrointestinal tract following oral administration.
IFN-α Response Is Higher following Oral versus Intravenous Administration Despite Similar Systemic Exposure.
To assess the relative contribution of systemic GS-9620 exposure to the peripheral induction of IFN-α, IFN-α response after a single intravenous versus oral administration of GS-9620 was directly compared in CD-1 mice and cynomolgus monkeys. Comparable systemic exposure [assessed by both maximum plasma/serum concentration (Cmax) and area under the concentration-time curve) following intravenous and oral administration resulted in notably higher peripheral IFN-α concentration after oral administration in both mice and cynomolgus monkeys (Table 3). In mice, the mean GS-9620 Cmax following an oral dose at 100 mg/kg and intravenous infusion at 1 mg/kg was similar (277 vs. 305 ng/ml, respectively). However, despite this similar systemic exposure, the mean peak serum IFN-α levels following oral administration was more than 100-fold higher than that from intravenous infusion (6572 vs. 60 pg/ml, respectively; Table 3).
A similar pattern of pharmacodynamic response was observed in cynomolgus monkeys. An oral dose at 2 mg/kg resulted in a Cmax that was approximately 2-fold lower than that achieved after intravenous infusion of 0.1 mg/kg GS-9620 to male cynomolgus monkeys (66 vs. 128 ng/ml, respectively; Table 3). However, the mean peak IFN-α serum level induced after oral dosing was more than 60-fold higher than that produced after intravenous infusion (119,000 vs. 1840 pg/ml, respectively; Table 3). These data suggest that systemic exposure to GS-9620 via the intravenous route generates a minimal pharmacodynamic response when compared with exposure to GS-9620 via the oral route.
Profiling of Pharmacodynamic Response to GS-9620 Following Oral Administration.
Consistent with observations from in vitro experiments described earlier, oral administration of single ascending oral doses of GS-9620 to CD-1 mice and cynomolgus monkeys resulted in dose-dependent induction of serum cytokines and chemokines. In CD-1 mice, single ascending oral doses of GS-9620 ranging from 0.1 to 50 mg/kg resulted in dose-dependent induction of IFN-α and cytokines and chemokines, generally at doses ≥1 mg/kg (Fig. 3A). Similar to the findings attained from in vitro mouse splenocyte cultures, the proinflammatory cytokines interleukin-1β and TNF-α were induced in the mouse concurrently with the induction of IFN-α, which is consistent with the different expression of TLR7 in the mouse.
In contrast to the in vivo mouse pharmacodynamic profile, but consistent with the results from cynomolgus monkey in vitro PBMC studies, in vivo administration of single oral doses of GS-9620 (0.05–10 mg/kg) to cynomolgus monkeys preferentially induced dose-dependent amounts of IFN-α at low oral doses, whereas induction of TNF-α was only observed at doses ≥1.5 mg/kg (Fig. 3B). These data indicate that GS-9620 can induce a selective, TLR7-dependent pharmacodynamic response in vivo with TLR8-dependent cytokines and chemokines only observed after administration of high oral doses.
Relationship of GS-9620 Exposure, Serum IFN-α, and Induction of ISGs in Whole Blood.
As a biomarker for an antiviral immune response, peripheral whole-blood mRNA expression of the ISGs OAS1 and MX1 was evaluated, and we characterized the kinetic relationship of ISG induction to GS-9620 serum exposure and peripheral blood IFN-α concentration in cynomolgus monkeys after a single oral dose of 0.15 mg/kg (Fig. 4). The induction of peripheral IFN-α and ISG was transient and appeared sequential following the rapid absorption and clearance of GS-9620 (Cmax = 1.6 ± 1.1 ng/ml; time to reach Cmax = 0.8 ± 0.3 hours). Peak induction of peripheral IFN-α was observed at approximately 4 hours postdose and was transient; levels rapidly decreased following 4 hours postdose and returned to baseline levels by 12 hours postdose. This peak peripheral IFN-α concentration was followed by induction of peripheral ISGs, which were maximal at approximately 8 hours postdose. ISG induction was also transient and returned to baseline by 24 hours postdose. These data are consistent with a single, transient activation of TLR7, followed by the sequential induction of IFN-α, which in turn induces the ISGs OAS1 and MX1.
GS-9620–Dependent ISG Induction Is Independent of Detectable Peripheral Interferon-α.
It is noteworthy that we also observed dose-dependent induction of ISG expression in peripheral blood without concurrent induction of systemic IFN-α in individual cynomolgus monkeys and mice after a single oral dose of GS-9620 (Fig. 5; Table 4). These data, described later, suggest that local induction of IFN-α induces ISGs in blood cells trafficking out of tissues, which then can be detected as peripheral ISG induction in whole blood.
In mice orally dosed with 0.3 mg/kg GS-9620, elevated peripheral blood and hepatic gene expression of OAS1 and MX1 was observed in a time-dependent manner in the absence of concurrently detectable systemic IFN-α (Fig. 5; Table 4). Hepatic gene expression of MX1 and OAS1 was induced in the absence of detectable systemic IFN-α (i.e., >3-fold increase from vehicle control) in individual animals at 4 (MX1 only), 8, and 12 hours postdose (Fig. 5). Importantly, no ISG induction was observed in any animal at 24 hours postdose. These data are consistent with the sequential induction of ISGs following TLR7 activation described earlier. However, the possibility that systemic IFN-α was induced at an earlier time point in these animals cannot be excluded, because serial sampling was not feasible in this experiment.
In cynomolgus monkeys, MX1 and OAS1 were induced (i.e., >3-fold increase from vehicle-treated monkeys) in the absence of systemic IFN-α following administration of a single dose of GS-9620 as low as 0.05 mg/kg, the lowest dose evaluated (Table 4). Despite the lack of detectable systemic IFN-α in these cynomolgus monkeys, induction of MX1 and OAS1 mRNA was dose-dependent. Interestingly, when the ISG response was compared between animals with and without detectable systemic IFN-α, mean MX1 and OAS1 fold increase in animals with systemic IFN-α was within 3-fold of mean ISG induction observed in animals without detectable serum IFN-α (Table 4). This is despite mean systemic IFN-α as high as 28,186 pg/ml in animals administered 1.5 mg/kg. These data suggest that systemic IFN-α is not a dominant driver of GS-9620–dependent induction of ISGs in peripheral blood. This, along with the observations following oral versus intravenous administration, further support the hypothesis that the pharmacodynamic response is driven by presystemic activation of TLR7, in either the gut-associated lymphoid tissue and/or liver.
The induction of an ISG response in whole blood, without induction of detectable levels of serum IFN-α, was also observed in a single dose study in healthy human volunteers (Lopatin et al., 2013). After single ascending oral doses of GS-9620 ranging from 0.3 to 12 mg, dose-dependent induction of MX1 and OAS1 mRNA relative to predose was observed following GS-9620 oral doses 2 mg and higher (Lopatin et al., 2013). Of note, no subject had detectable systemic IFN-α at doses ≤6 mg (LLOQ = 15.6 pg/ml). One subject at 8 mg and four subjects at 12 mg had detectable IFN-α. Similar to the cynomolgus monkeys, induction of detectable IFN-α did not appear to have a notable effect on ISG induction. In the four subjects with detectable IFN-α at 12 mg, ISG induction was not notably different than in subjects without detectable systemic IFN-α (Table 4). These data suggest that, at the level of peripheral blood gene expression, a type I IFN-independent antiviral response can be invoked after oral administration of GS-9620 in humans without concurrent induction of detectable systemic IFN-α.
Profiling of GS-9620–Induced Cytokine and Chemokine Secretion In Vitro and In Vivo.
In vitro stimulation of cynomolgus monkey PBMCs with GS-9620 at concentrations less than 10 µM results in the induction of a pattern of cytokines and chemokines that are consistent with the biology of TLR7. At concentrations of GS-9620 higher than 10 µM, additional cytokines such as TNF-α were induced in PBMCs, which is consistent with stimulation of TLR8 (Gorden et al., 2005; Thomas et al., 2007). Similarly, oral administration of GS-9620 to cynomolgus monkeys in vivo preferentially induced IFN-α and a pattern of antiviral serum cytokines and chemokines at low oral doses, whereas TNF-α was only observed following relatively high oral doses. In general, the cytokine and chemokine patterns observed across ascending doses in vivo were similar to or overlapped with the cytokine/chemokine profile induced in in vitro PBMC studies. The differences could be a result of differentials in the half-life or distribution of cytokines in vivo versus in vitro, or may reflect the response to tissues other than PBMCs which are exposed to GS-9620 in vivo.
The in vitro and in vivo pharmacodynamic profile induced by GS-9620 in the mice was notably different from that seen in cynomolgus monkeys and from that which we have previously reported in humans (Tumas et al., 2011; Lopatin et al., 2013). Oral administration of GS-9620 to CD-1 mice resulted in dose-dependent induction of IFN-α and other cytokines and chemokines; however, in contrast to the cynomolgus monkeys, induction of the proinflammatory cytokines interleukin-1β and TNF-α occurred at relatively low doses and were observed concurrent with the induction of IFN-α in mice (Figs. 2 and 3A). Unlike human and cynomolgus monkey dendritic cell subsets in which TLR7 is exclusively expressed by pDCs and TLR8 expression is predominantly found in cDCs, the pattern of murine TLR7 expression is notably different, as expression of murine TLR7 is found in both pDCs and cDCs (Edwards et al., 2003; Ketloy et al., 2008). As a consequence, the response profiles observed with a TLR7 agonist in mice or mouse splenocytes are more reminiscent of profiles observed with dual TLR7/TLR8 agonists in monkeys and humans (Ito et al., 2002; Gorden et al., 2005; Thomas et al., 2007).
The pattern of cytokine and chemokine induced in human PBMCs (Tumas et al., 2011) by GS-9620 in vitro was similar to the pattern we described for cynomolgus monkey PBMCs. Likewise, the in vivo responses to GS-9620 after oral dosing in human volunteers (Lopatin et al., 2013) were similar to the in vivo responses in cynomolgus monkeys described herein. In contrast, induction of cytokines and chemokines in vitro and in vivo in mice was generally consistent with the notion of different TLR7 expression profiles in murine DC subsets. Protein sequence comparison of mouse, human, and cynomolgus TLR7 and TLR8 showed that sequence identity between human TLR7 and mouse TLR7 or cynomolgus TLR7 is approximately 81 or 98%, respectively; sequence identity between human TLR8 and mouse TLR8 or cynomolgus TLR8 is approximately 70 or 97%, respectively (Supplemental Fig. 2). This analysis suggests that selectivity for TLR7 versus TLR8, as observed with the human TLRs, is likely maintained in the cynomolgus monkey.
GS-9620–Dependent Induction of Antiviral Gene Expression In Vivo.
Oral administration of GS-9620 induces presystemic IFN-α and induction of MX1 and OAS1 mRNA in peripheral blood. The sequential time course for induction of IFN-α and these ISGs is consistent with activation of TLR7 leading to the release of IFN-α in the tissues containing pDCs, which in turn triggers the induction of IFN-α response elements in adjacent cells in paracrine fashion. However, we also observed that single, low oral doses of GS-9620 caused dose-dependent induction of OAS1 and MX1 without concurrent detectable serum IFN-α in individual cynomolgus monkeys and mice (Fig. 5; Table 4). In addition, elevated intrahepatic expression of OAS1 and MX1 mRNA was observed in mice after a single oral dose of GS-9620 despite the absence of detectable peripheral IFN-α concurrently (Fig. 5). These data suggest that TLR7 agonism can occur in tissues during oral absorption of GS-9620, resulting in local induction of a response, although IFN-α is not detectable in serum but is clearly active in these local tissues as evidenced by ISG induction. In addition, despite notable induction of systemic IFN-α following higher doses of GS-9620 to cynomolgus monkeys, OAS1 and MX1 mRNA induction were only less than 3-fold higher when compared with induction in individuals with undetectable systemic IFN-α (Table 4). Of note, these observations in animals are further corroborated by findings in healthy human subjects, where >3-fold ISG mRNA induction over predose occurred after administration of single doses of ≥2 mg, despite undetectable peripheral IFN-α (Lopatin et al., 2013). In addition, in human volunteers with systemic IFN-α, ISG induction was not notably different from those subjects without detectable systemic IFN-α, which again is similar to our results in cynomolgus monkeys (Table 4).
These data suggest that GS-9620–dependent ISG induction may mainly be due to a localized, presystemic response rather than due to systemic immune stimulation, and further suggest that the contribution of systemic immune stimulation to the overall response is low. This local innate immune response without a systemic IFN-α response might indicate a potential for harnessing the benefit of an innate immune response in the liver without systemic adverse side effects. Although the possibility that systemic IFN-α is produced at levels below the limit of detection for the assays used in these studies cannot be excluded, these data suggest that GS-9620–dependent ISG induction can be mounted independent of appreciable concentrations of systemic IFN-α. In contrast, induction of ISGs with therapeutic doses of recombinant IFN-α is dependent on systemic IFN-α, which is obviously associated with flu-like symptoms and other adverse events. In the studies described here with GS-9620, IFN-α levels at or below the LLOQ are considerably lower than what is currently administered for approved exogenous IFN-α therapies. Therefore, it can be extrapolated that, even if low (undetectable) systemic IFN-α levels are present, systemic levels are unlikely to approach concentrations at which adverse events associated with existing interferon therapies were observed. Indeed, healthy human volunteers who were orally administered a single dose of GS-9620 had minimal adverse events that were consistent with systemic IFN-α exposure, and these were only observed at dose levels ≥8 mg (Lopatin et al., 2013). We hypothesize that the induction of an antiviral immune signature devoid of detectable systemic IFN-α is caused by presystemic stimulation of TLR7, likely in pDCs at the level of the GALT and/or in the liver, e.g., presystemic response. Figure 6 presents the postulated underlying mechanism for the presystemic response to oral GS-9620.
The high absorption of GS-9620 allows for the targeted exposure to interferon-producing cells in the GALT and/or liver following oral administration. Interferon-producing cells present in the GALT and/or liver, likely pDCs, are activated by local exposure to GS-9620 and produce IFN-α. This localized production of IFN-α stimulates ISG induction in lymphocytes and other cells as they circulate through the GALT and/or liver, which can then be detected in the peripheral blood (Fig. 6). ISG induction may occur in the liver by a similar effect—through either local IFN-α produced from stimulated pDCs residing in the liver or from a first-pass effect on the liver from portal blood containing IFN-α produced from GALT-resident cells.
The therapeutic benefit of TLR7 stimulation by oral GS-9620 in animal models of chronic HBV infection has been shown in both chimpanzees infected with HBV and the woodchuck hepatitis virus model (Menne et al., 2011; Lanford et al., 2013). Importantly, in uninfected and chronic HBV–infected chimpanzees, GS-9620 induced ISGs in the liver and/or in PBMC with no detectable increase in serum levels of IFN-α at low doses. Consequently, GS-9620 may be an effective therapy for viral hepatitis at doses that produce no detectable serum IFN-α levels and, therefore, minimize tolerability limitations associated with systemic IFN-α therapy. Currently, weekly administration of GS-9620 is being evaluated in patients infected with chronic HBV.
The authors acknowledge Paul Duatschek for in vitro pharmacodynamic analysis, Dr. Robert Lanford for ISG sample analysis, Bing Lu for PK sample analysis, Covance Laboratories for study conduct, and Eurofins Laboratories for cytokine/chemokine sample analysis.
Participated in research design: Fosdick, Zheng, Pflanz, Frey, Hesselgesser, Halcomb, Tumas, Wolfgang.
Conducted experiments: Frey, Hesselgesser, Pflanz.
Performed data analysis: Fosdick, Zheng, Pflanz, Frey, Hesselgesser, Tumas, Wolfgagng.
rote or contributed to the writing of the manuscript: Fosdick, Zheng, Pflanz.
- Received July 16, 2013.
- Accepted October 16, 2013.
This work was supported by Gilead Sciences, Inc., Foster City, CA.
This work was presented previously, in part, at two meetings: Tumas D, Zheng X, Lu B, Rhodes G, Duatschek P, Hesselgesser J, Frey C, Henne I, Fosdick A, Halcomb R, and Wolfgang G (2011) Preclinical characterization of GS-9620, a potent and selective oral TLR7 agonist (Abstract 1776). 46th Annual Meeting of the European Association for the Study of the Liver; 2011 March 30–April 3; Berlin, Germany; and Tumas D, Fosdick A, Zheng X, Lanford R, Hesselgesser J, Frey C, Wolfgang G, Halcomb R, and Lopatin U (2012) GS-9620, induces a presystemic innate immune response in the absence of serum interferon or adverse effects in both non-clinical species and humans (Abstract 18640). 47th Annual Meeting of the European Association for the Study of the Liver; 2012 April 18–22; Barcelona, Spain.
- conventional dendritic cells
- maximum plasma/serum concentration
- gut-associated lymphoid tissue
- HBV surface antigen
- hepatitis B virus
- interferon-stimulated genes
- lower limit of detection
- minimum effective concentration
- myxovirus resistance 1
- nuclear factor κB
- 2′5′-oligoadenylate synthetase 1
- peripheral blood mononuclear cell
- plasmacytoid dendritic cells
- Toll-like receptor 7
- tumor necrosis factor
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics