An in silico disease model for the development of FXR ...
1
An in silico disease model for the development of FXR agonist EYP001 as a therapy for HBV infection Claudio Monteiro (1), Lara Bruezière (1), Samuel Laheux (1), Marc Hommel (1), Pauline Radreau (2), Elise Roy (2), Raphaël Darteil (2), Patrice André (3), Jean-Pierre Boissel (1), Jacky Vonderscher (2) and Pietro Scalfaro (2) (1) Novadiscovery, Lyon, France, (2) ENYO pharma, Lyon, France, (3) CIRI - INSERM U1111 – CNRS UMR5308 - Université Lyon 1- ENS de Lyon BACKGROUND Chronic infection with hepatitis B virus (HBV) increases the risk of death from cirrhosis or liver cancer. The farnesoid X receptor (FXR) is an investigational target for HBV infection therapies in view of its putative role in modulating HBV replication and in decreasing the pool of intracellular HBV DNA. The FXR agonist EYP001 was well tolerated by healthy and HBV-infected subjects in a phase 1 study. In hepatocytes, EYP001 inhibited the ex vivo secretion of HBV DNA and the HBV antigens HBsAg and HBeAg, whereas the antiviral entecavir (ETV) reduced HBV DNA secretion only. We suggest that therapies combining the FXR agonist EYP001 with standard treatments such as nucleoside analogs or interferons will increase chances of curing chronic hepatitis B through their highly probable synergistic effects in terms of immunomodulation and in decreasing HBV replication. We used computational modeling to improve the design, dosage, timing, and patient selection for combination therapies based on EYP001 treatment. METHODS An in silico disease model of chronic HBV patients has been built based on public and expert knowledge, non-clinical, and clinical data. The Computational Model is a system of differential equations that integrates 300+ biological variables and 600+ parameters. With 7 mechanistic submodels (including the effect of FXR agonist on HBV replication, HBV excretion, bile acid physiology and EYP001, ETV, and pegylated interferon [PEG-IFNα2a] drug models), the model has been used to predict quantitative efficacy of treatments on disease-related endpoints (eg. plasma HBV DNA and HBsAg concentrations) in a virtual population. All the submodels are ultimately combined into a multi-scale Computational Model simulating the dynamics of biological entities at the molecular, cellular and organ levels. Figure 1 illustrates the structure of the integrated Computational Model. The Computational Model was written and implemented through Novadiscovery’s proprietary simulation framework and its various tools (Jinkō). The virtual population and exploration tools were used to calibrate the model: 1,000 virtual patients were generated by randomly sampling from a set distribution for each of descriptors (representing the n model parameters). Virtual patients were ranked and selected on the basis of a score translating physiological and biological constraints that the model should comply with, as well as data from Phase I studies. This results in a n-dimension space domain where the parameter values meet the constraints. These simulated results will be quantitatively validated with data from upcoming Phase II studies. The model has been used to explore the effects of multiple combinations of EYP001 with ETV and/or PEG-IFNα2a therapies via the simulation of an in silico Phase II trial. We compared different combinations of treatments according to the following factors: / Drug combination (no treatment, EYP001a monotherapy, EYP001a + ETV & EYP001a + PEG-IFNα2a bitherapies, EYP001a + ETV + PEG-IFNα2a tritherapy). / Treatment duration (24 & 48 weeks). We also simulated the effect of run-in periods (ETV 1 month, PEG-IFNα2a 1 month, ETV 4 months, PEG-IFNα2a 4 months, tritherapy 1 month) before the administration of treatment combinations. RESULTS Phase 1 results were well reproduced in silico, including the effects of EYP001 on the FXR response markers 7α-hydroxy-4-cholesten-3-one (C4) and fibroblast growth factor 19 (FGF19) (Figures 2 and 3), and on HBV DNA plasma concentrations. The model accurately predicts the short-term impact of FXR agonism on FGF19 concentrations and its impact in reducing HBV DNA and HBsAg concentrations (Figures 4 and 5). In silico Phase II trial simulations allowed us to identify the 24-week bi-therapy of EYP001 and PEG-IFNα2a as the best combination in reducing plasma HBV DNA and HBsAg levels, and in minimizing relapse after a 24-week follow-up period (Figures 4 and 5). The percentage of relapsing patients 2 months after the end of 24-week treatment was 26.5% for the bitherapy with EYP001 QD 400 mg + PEG-IFNα2a, whereas it was of 94.2% for patients receiving PEG-IFNα2a only, and 100% for the no treatment group. DISCUSSION & CONCLUSIONS / To date, and despite decades of research, there is still no curative treatment for chronic hepatitis B. The last few years have seen an increased interest for combination of antiviral therapies with the common objective of increasing the rate of HBV eradication. / We used computational modeling to improve the design, dosage, timing, and patient selection for combination therapies based on EYP001 treatment. / The in silico model reproduced well all drugs plasma concentration profiles as well as their independent characteristics and effects. / These preliminary results suggest EYP001 combined with PEG-IFNα2a is an optimal regimen and support selection of EYP001 regimens in Phase 2 trials. / Next step will include best virtual responders characterization for patients selection of the next clinical trial. ACKNOWLEDGEMENTS This study was supported by Enyo Pharma SA. Figure 2: Kinetics of EYP001a in plasma for the administered doses of 250 mg (left panel) and 500 mg (right panel). The black dots and the brown bars represent the median and the minimum-maximum range of experimental data, respectively. Red solid lines correspond to the EYP001a simulated concentrations. The kinetics of EYP001 in plasma for administered dose of 250 mg [EYP001] in plasma (mol/L) Time after EYP001 administration (h) The kinetics of EYP001 in plasma for administered dose of 500 mg Figure 3: Kinetics of FGF19 (left panel) and C4 (right panel) biomarkers in plasma for the administered dose of EYP001a of 500 mg. Black dots separated by black lines and brown bars represent the median data and the minimum-maximum range of experimental data, respectively. Red solid lines represent the simulated concentration of biomarkers. The kinetics of C4 in plasma for administe- red dose of 500 mg. EYP001 is administe- red every day from day 5 to day 10. Day 8 is displayed. [Biomarkers] in plasma (mol/L) Time after EYP001 administration (h) The kinetics of FGF19 in plasma for administered dose of 500 mg. EYP001 is administered every day from day 5 to day 10. Day 8 is displayed. Figure 1: Computational Model structure: submodels are shown as blue circles and entities used as connectors for the submodels are shown as green circles. Cholesterol & bile acids HBV submodel Immune system Fibrosis DNA HBV HBsAg PDGF MMPs TIMPs 29 BIOREACTIONS 11 ENTITIES 43 PARAMETERS 158 BIOREACTIONS 103 ENTITIES 253 PARAMETERS 55 BIOREACTIONS 28 ENTITIES 63 PARAMETERS 55 BIOREACTIONS 43 ENTITIES 119 PARAMETERS 55 BIOREACTIONS 43 ENTITIES 59 PARAMETERS NTCP LRH1 SHP HNF4α EYP001a drug model 6 BIOREACTIONS 3 ENTITIES 43 PARAMETERS PEG-IFNα2a drug model ETV drug model 11 BIOREACTIONS 9 ENTITIES 20 PARAMETERS FXR Acetyl COA pgRNA cccDNA NK cells pgRNA, HBV polymerase, HBcAg, rcDNA Capsid Figure 5: HBsAg levels in plasma (log10 (copies/mL)) after 24 weeks of treatment and 24 weeks of follow-up. Comparison of control (EYP001a QD 400 mg) with ETV + EYP and IFN + EYP bitherapies and EYP + IFN + ETV tritherapy. (ETV: entecavir; IFN: pegylated interferon alfa-2a; EYP: EYP001a). Control: EYP001a QD400mg Bitherapy: ETV + EYP001a Control: EYP001a QD400mg Bitherapy: IFN + EYP001a Control: EYP001a QD400mg Tritherapy: ETV+IFN+EYP001a Concentration of HBsAg in plasma (log10(cp/mL)) Time (weeks) Figure 4: HBV DNA levels in plasma (log10 (copies/mL)) after 24 weeks of treatment and 24 weeks of follow-up. Comparison of control (EYP001a QD 400 mg) with ETV + EYP and IFN + EYP bitherapies and EYP + IFN + ETV tritherapy. (ETV: entecavir; IFN: pegylated interferon alfa- 2a; EYP: EYP001a). Control: EYP001a QD400mg Bitherapy: ETV + EYP001a Control: EYP001a QD400mg Bitherapy: IFN + EYP001a Control: EYP001a QD400mg Tritherapy: ETV+IFN+EYP001a Concentration of HBV DNA in plasma (log10(cp/mL)) Time (weeks) www.novadiscovery.com NOVADISCOVERY, 1 Place Giovanni da Verrazzano, 69009 Lyon, France, [email protected] | ENYO PHARMA, Bioserra 1, 60 Avenue Rockefeller Bâtiment B, 69008 Lyon, France, [email protected]
Transcript of An in silico disease model for the development of FXR ...
An in silico disease model for the development of FXR agonist EYP001 as a therapy for HBV infection
Claudio Monteiro (1), Lara Bruezière (1), Samuel Laheux (1), Marc Hommel (1), Pauline Radreau (2), Elise Roy (2), Raphaël Darteil (2), Patrice André (3), Jean-Pierre Boissel (1), Jacky Vonderscher (2) and Pietro Scalfaro (2) (1) Novadiscovery, Lyon, France, (2) ENYO pharma, Lyon, France, (3) CIRI - INSERM U1111 – CNRS UMR5308 - Université Lyon 1- ENS de Lyon
BACKGROUNDChronic infection with hepatitis B virus (HBV) increases the risk of death from cirrhosis or liver cancer. The farnesoid X receptor (FXR) is an investigational target for HBV infection therapies in view of its putative role in modulating HBV replication and in decreasing the pool of intracellular HBV DNA. The FXR agonist EYP001 was well tolerated by healthy and HBV-infected subjects in a phase 1 study. In hepatocytes, EYP001 inhibited the ex vivo secretion of HBV DNA and the HBV antigens HBsAg and HBeAg, whereas the antiviral entecavir (ETV) reduced HBV DNA secretion only. We suggest that therapies combining the FXR agonist EYP001 with standard treatments such as nucleoside analogs or interferons will increase chances of curing chronic hepatitis B through their highly probable synergistic effects in terms of immunomodulation and in decreasing HBV replication. We used computational modeling to improve the design, dosage, timing, and patient selection for combination therapies based on EYP001 treatment.
METHODSAn in silico disease model of chronic HBV patients has been built based on public and expert knowledge, non-clinical, and clinical data. The Computational Model is a system of differential equations that integrates 300+ biological variables and 600+ parameters. With 7 mechanistic submodels (including the effect of FXR agonist on HBV replication, HBV excretion, bile acid physiology and EYP001, ETV, and pegylated interferon [PEG-IFNα2a] drug models), the model has been used to predict quantitative efficacy of treatments on disease-related endpoints (eg. plasma HBV DNA and HBsAg concentrations) in a virtual population. All the submodels are ultimately combined into a multi-scale Computational Model simulating the dynamics of biological entities at the molecular, cellular and organ levels. Figure 1 illustrates the structure of the integrated Computational Model.The Computational Model was written and implemented through Novadiscovery’s proprietary simulation framework and its various tools (Jinkō). The virtual population and exploration tools were used to calibrate the model: 1,000 virtual patients were generated by randomly sampling from a set distribution for each of descriptors (representing the n model parameters). Virtual patients were ranked and selected on the basis of a score translating physiological and biological constraints that the model should comply with, as well as data from Phase I studies. This results in a n-dimension space domain where the parameter values meet the constraints.
These simulated results will be quantitatively validated with data from upcoming Phase II studies.The model has been used to explore the effects of multiple combinations of EYP001 with ETV and/or PEG-IFNα2a therapies via the simulation of an in silico Phase II trial. We compared different combinations of treatments according to the following factors: / Drug combination (no treatment, EYP001a monotherapy, EYP001a + ETV & EYP001a + PEG-IFNα2a bitherapies, EYP001a + ETV + PEG-IFNα2a tritherapy). / Treatment duration (24 & 48 weeks).
We also simulated the effect of run-in periods (ETV 1 month, PEG-IFNα2a 1 month, ETV 4 months, PEG-IFNα2a 4 months, tritherapy 1 month) before the administration of treatment combinations.
RESULTSPhase 1 results were well reproduced in silico, including the effects of EYP001 on the FXR response markers 7α-hydroxy-4-cholesten-3-one (C4) and fibroblast growth factor 19 (FGF19) (Figures 2 and 3), and on HBV DNA plasma concentrations.
The model accurately predicts the short-term impact of FXR agonism on FGF19 concentrations and its impact in reducing HBV DNA and HBsAg concentrations (Figures 4 and 5). In silico Phase II trial simulations allowed us to identify the 24-week bi-therapy of EYP001 and PEG-IFNα2a as the best combination in reducing plasma HBV DNA and
HBsAg levels, and in minimizing relapse after a 24-week follow-up period (Figures 4 and 5). The percentage of relapsing patients 2 months after the end of 24-week treatment was 26.5% for the bitherapy with EYP001 QD 400 mg + PEG-IFNα2a, whereas it was of 94.2% for patients receiving PEG-IFNα2a only, and 100% for the no treatment group.
DISCUSSION & CONCLUSIONS / To date, and despite decades of research, there is still no curative treatment for chronic hepatitis B. The last few years have seen an increased interest for combination of antiviral therapies with the common objective of increasing the rate of HBV eradication.
/ We used computational modeling to improve the design, dosage, timing, and patient selection for combination therapies based on EYP001 treatment. / The in silico model reproduced well all drugs plasma concentration profiles as well as their independent characteristics and effects.
/ These preliminary results suggest EYP001 combined with PEG-IFNα2a is an optimal regimen and support selection of EYP001 regimens in Phase 2 trials. / Next step will include best virtual responders characterization for patients selection of the next clinical trial.
ACKNOWLEDGEMENTSThis study was supported by Enyo Pharma SA.
Figure 2: Kinetics of EYP001a in plasma for the administered doses of 250 mg (left panel) and 500 mg (right panel). The black dots and the brown bars represent the median and the minimum-maximum range of experimental data, respectively. Red solid lines correspond to the EYP001a simulated concentrations.
The kinetics of EYP001 in plasma for administered dose of 250 mg
[EYP
001]
in p
lasm
a (m
ol/L
)
Time after EYP001 administration (h)
The kinetics of EYP001 in plasma for administered dose of 500 mg
Figure 3: Kinetics of FGF19 (left panel) and C4 (right panel) biomarkers in plasma for the administered dose of EYP001a of 500 mg. Black dots separated by black lines and brown bars represent the median data and the minimum-maximum range of experimental data, respectively. Red solid lines represent the simulated concentration of biomarkers.
The kinetics of C4 in plasma for administe-red dose of 500 mg. EYP001 is administe-red every day from day 5 to day 10. Day 8 is displayed.
[Bio
mar
kers
] in
plas
ma
(mol
/L)
Time after EYP001 administration (h)
The kinetics of FGF19 in plasma for administered dose of 500 mg. EYP001 is administered every day from day 5 to day 10. Day 8 is displayed.
Figure 1: Computational Model structure: submodels are shown as blue circles and entities used as connectors for the submodels are shown as green circles.
Cholesterol & bile acids
HBV submodel
Immune system
Fibrosis
DNA HBV
HBsAg
PDGFMMPsTIMPs
29 BIOREACTIONS11 ENTITIES
43 PARAMETERS
158 BIOREACTIONS103 ENTITIES
253 PARAMETERS
55 BIOREACTIONS28 ENTITIES63 PARAMETERS
55 BIOREACTIONS43 ENTITIES
119 PARAMETERS
55 BIOREACTIONS43 ENTITIES59 PARAMETERS
NTCPLRH1SHP
HNF4α
EYP001a drug model
6 BIOREACTIONS3 ENTITIES
43 PARAMETERS
PEG-IFNα2a drug model
ETV drug model
11 BIOREACTIONS9 ENTITIES20 PARAMETERS
FXRAcetyl COA
pgRNAcccDNANK cells
pgRNA, HBV polymerase,
HBcAg, rcDNA Capsid
Figure 5: HBsAg levels in plasma (log10 (copies/mL)) after 24 weeks of treatment and 24 weeks of follow-up. Comparison of control (EYP001a QD 400 mg) with ETV + EYP and IFN + EYP bitherapies and EYP + IFN + ETV tritherapy. (ETV: entecavir; IFN: pegylated interferon alfa-2a; EYP: EYP001a).
Control: EYP001a QD400mgBitherapy: ETV + EYP001a
Control: EYP001a QD400mgBitherapy: IFN + EYP001a
Control: EYP001a QD400mgTritherapy: ETV+IFN+EYP001a
Conc
entra
tion
of H
BsAg
in p
lasm
a (lo
g10(
cp/m
L))
Time (weeks)
Figure 4: HBV DNA levels in plasma (log10 (copies/mL)) after 24 weeks of treatment and 24 weeks of follow-up. Comparison of control (EYP001a QD 400 mg) with ETV + EYP and IFN + EYP bitherapies and EYP + IFN + ETV tritherapy. (ETV: entecavir; IFN: pegylated interferon alfa-2a; EYP: EYP001a).
Control: EYP001a QD400mgBitherapy: ETV + EYP001a
Control: EYP001a QD400mgBitherapy: IFN + EYP001a
Control: EYP001a QD400mgTritherapy: ETV+IFN+EYP001a
Conc
entra
tion
of H
BV D
NA
in p
lasm
a (lo
g10(
cp/m
L))
Time (weeks)
www.novadiscovery.comNOVADISCOVERY, 1 Place Giovanni da Verrazzano, 69009 Lyon, France, [email protected] | ENYO PHARMA, Bioserra 1, 60 Avenue Rockefeller Bâtiment B, 69008 Lyon, France, [email protected]
