Abstract # AASLD Washington DC

Transcription

1 Modeling Hepa--s Delta Virus (HDV) Dynamics During Ritonavir Boosted Lonafarnib Treatment The LOWR HDV-3 Study Pree$ Dubey 1*, Christopher Koh 2*, Pallavi Surana 2, Susan L. Uprichard 1, Ma Ai Thanda Han 2, Nancy Fryzek 2, Gitanjali Subramanya 1, Devika Kapuria 2, Ohad Etzion 2,3, Varun Takyar 2, Yaron Rotman 2, Cihan Yurdaydin 4, Jeffrey S. Glenn 5, ScoU J. Cotler 1, Theo Heller 2* and Harel Dahari 1* 1 The Program for Experimental and Theore$cal Modeling, Division of Hepatology, Loyola University Medical Center, Maywaood, IL, USA, 2 Liver Diseases Branch, NIDDK, NIH, 3 Soroka Univeristy Medical Center, Beer Sheva, Israel; 4 Ankara University, Ankara, Turkey, 5 Division of Gastroenterology and Hepatology, Stanford University School of Medicine *These authors contributed equally. Abstract # AASLD Washington DC

2 Disclosures My presenta+on includes discussion of off-label/inves+ga+onal use of lonafarnib and ritonavir Harel Dahari: Research Grant: Eiger BioPharmaceu+cals, Inc. Jeffrey Glenn: Board membership: Eiger BioPharmaceu+cals, Inc., Riboscience, LLC Consul+ng: Gilead,Janssen, Sundise, Genentech, Merck, Roche, Romark Laboratories, StemCells Inc. Cihan Yurdaydin: Stock/Shareholder: Eiger Group Interna+onal Inc., Eiger BioPharmaceu+cals, Inc., Riboscience, LLC Consul+ng: AbbVie, BMS, Gilead, Roche, Merck, Boehringer Ingelheim, Janssen, Eiger Biopharma Research Grant: BMS, Roche, Eiger Biopharmaceu+cals, Inc. All remaining authors have no financial rela+onships to disclose

3 HDV Infec-on: the least common form of viral hepa--s has the worst outcomes Ø 15-2 million people are infected worldwide with HDV Ø Up to 8% of pa+ents with HDV may develop cirrhosis within 5-1 years Ø Higher risk for hepa+c decompensa+on leading to death & development of HCC compared to HBV mono-infected pa+ents Hughes SA, et al. Lancet 211;378; Colombo M et al. Gastroenterol 1983; 85 Manesis EK et al. J Hep 213;59; Fa^ovich G et al. Gut 2; 46 Romeo R, et al. Gastroenterol 29; 136

4 The quest for beler therapies against HDV Ø Interferon-alpha therapy is unsa+sfactory <3% achieve HBsAg loss and become HDV RNA nega+ve Extended dura+on of therapy (>1 yr) does not improve HDV response rates Ø Nucleos(t)ide analogues are ineffec+ve Ø Daily IV infusion (for 4 weeks) of silibinin is ineffec+ve (case study) Ø Inves+ga+onal Therapies (in clinical trials) Human sodium-taurocholate cotranspor+ng polypep+de Myrcludex B Nucleic acid polymers REP 2139 Prenyla+on inhibi+on Lonafarnib Peg-Interferon-lambda Wedemeyer H, et al. N Eng J Med 211;27 Bogomolov P, et al. Hepatology 216;65 Hamid et al. Hepatology abstract # 927 (AASLD 217) Wedemeyer H, et al. Hepatology 213;58 Bazinet M, et al. Lancet Gastroenterol Hepatol 217 Heller T, et al. Aliment Pharmacol Ther 214;4 Koh C, et al. Lancet Infect Dis 215;15 Reungoat E, et al. Hepatology 216; 64:1(Suppl): 928A

5 HDV life cycle and prenyla-on inhibitor lonafarnib (LNF) Ø HDV has a nega+ve sense, circular RNA genome 17 bases in length Ø Encodes only one protein of its own, HDAg, in two forms, large (LHDAg) and small (SHDAg) Ø Uses hepa++s B surface an+gen (HBsAg) to create its envelope and to achieve secre+on Prenyla+on (lipid modifica+on) plays a vital part in the life cycle of HDV. Lonafarnib (LNF) disrupts prenyla+on of the LHDAg and prevents its ability to interact with HBsAg Bordier et al. J Clin Invest 23;112:47-414; Taylor JM. Virology 26;344:71-76

6 The beginning of an era of HDV dynamics Ø Understanding HDV dynamics is in its infancy due to limited studies with frequent kine+c data and the availability of only one medica+on (peg-ifn-α) with ac+vity against HDV Ø In a previous study modeling HDV and HBsAg kine+cs, we provided ini+al insights into HDV-HBsAg-host dynamics and IFN-α s mode of ac+on (MOA) and its efficacy against HDV Guedj J, et al Hepatology 214; 6(6):192-1

7 Lonafarnib (LNF) phase 2 program provides important data (N>1) for viral kine-c analysis and mathema-cal modeling Proof of Concept Monotherapy N = 14 LOWR HDV 1 ± RTV or PEG IFN-α N = 21 LOWR HDV 2 Dose Finding ± PEG IFN-α N = 58 LOWR HDV 3 QD Dose N = 21 LOWR HDV 4 Dose-Escala+on N = 15 MHH Hannover Medical School Koh et al. Lancet Infect Dis 215;15: ; Yurdaydin, C et al. Journal Hepatology, 216; 64(Suppl no 2): S587 Modeling HDV kinetics during LNFbased therapy provides a novel opportunity to further characterize: Ø HDV-host dynamics Ø the an+viral effect of LNF and IFN-α and their MOA against HDV Ø the dynamics between HDV and HBV replica+on Koh et al., EASL 217 Abstract #LBP-519 Wedemeyer et al., EASL 217 Abstract #PS-39 Yurdaydin, C. et al, AASLD 216 Abstract #1845

8 Modeling provides a strong ra-onale for ritonavir boos-ng of LNF PK/PD modeling predicts that a LNF monotherapy dose of 61 mg bid would achieve 99% efficacy, however, LNF maximum tolerated dose was 2 mg bid. LNF 1 mg bid with ritonavir (RTV) boos+ng exceeded the predicted 99% efficacy concentra+on and was associated with drama+c HDV viral load declines and be^er tolerability than higher doses of LNF monotherapy Canini L & Koh C et al. Hepatology Communica+ons 217; 1: ; Yurdaydin C et al. Journal Hepatology, 216; 64(Suppl no 2): S587

9 How can modeling be used to op-mize therapy? Ø Real-+me modeling of viral kine+cs can be used to individualize dura+on of therapy Ø Modeling empowers a pa+ent to par+cipate in shared decision making regarding length of treatment Dahari et al 215; Liver Interna+onal ; 35(2):

10 LOWR-3 study design Phase 2a, Double Blinded, Randomized, Placebo-Controlled Study Lonafarnib (LNF) 5, 75 or 1 mg with ritonavir (RTV) 1 mg daily LNF 5 mg + RTV 1 mg daily x 24 wk (n=4) LNF 75 mg + RTV 1 mg daily x 24 wk (n=4) Eligibility Screening LNF 1 mg + RTV 1 mg daily x 24 wk (n=4) PBO x 12 wks -> LNF 5 mg + RTV 1 mg daily x 12 wk Post-Therapy Follow-up x 24 wk PBO x 12 wks -> LNF 75 mg + RTV 1 mg daily x 12 wk PBO x 12 wks -> LNF 1 mg + RTV 1 mg daily x 12 wk * All eligible subjects were placed/maintained on HBV nucleos(t)ide therapy for the dura+on of the study

11 Pa-ent characteris-cs * All subjects were placed/maintained on HBV nucleos(t)ide therapy for the dura+on of the study * No difference in baseline parameters between placebo and treatment groups CharacterisNc Result (n=21) Male Sex (%) 13 (62) Median Age (IQR) 4 (32, 49) BMI 25.7 (22.5, 22.8) Race (%) Caucasian 1 (48) Asian 1 (48) Black 1 (5) Median Laboratory Results (IQR) ALT U/L 7 (45,99) AST U/L 4 (35, 57) HBV DNA IU/mL <21 (<21, 23) HDV RNA log IU/mL 6.82 (6.18, 7.25) Fibroscan kpa 7.9 (5.4, 12.)

12 Goals To characterize viral kine+cs and provide insights into HDV-HBsAg-host dynamics during LNF+RTV treatment using mathema+cal modeling Analysis based on 12 pa+ents dosed for 24 weeks with frequent sampling LNF 5 mg + RTV 1 mg daily x 24 wk (n=4) LNF 75 mg + RTV 1 mg daily x 24 wk (n=4) Eligibility Screening LNF 1 mg + RTV 1 mg daily x 24 wk (n=4) PBO x 12 wks -> LNF 5 mg + RTV 1 mg daily x 12 wk Post-Therapy Follow-up x 24 wk PBO x 12 wks -> LNF 75 mg + RTV 1 mg daily x 12 wk PBO x 12 wks -> LNF 1 mg + RTV 1 mg daily x 12 wk

13 Viral kine-c analysis iden-fied 4 dis-nct palerns of response Median HDV Change (log IU/ml) N=3 N=3 N=3 N=3 Non-responders Triphasic Rebounders Flat ParNal Responders Time (week)

14 Four palerns of response seen in each LNF dosing group Log HDV RNA IU/ml Triphasic Response LNF 5 mg QD / RTV 1 mg QD Time (wk) Log HDV RNA IU/ml Flat ParNal Response HDV RNA (IU/mL) LNF 5 mg QD / RTV 1 mg QD Time (wk) Log HDV RNA IU/ml Rebound LNF 5 mg QD / RTV 1 mg QD Time (wk) Log HDV RNA IU/ml Non-Response LNF 5 mg QD / RTV 1 mg QD Time (wk)

15 HBsAg and ALT kine-c characteriza-on Ø HBsAg levels did not change on therapy Ø ALT Normaliza+on Ø 1% of triphasic responders Ø 67% of flat par+al responders Ø ALT levels fluctuated in rebounders

16 LNF and RTV kine-cs Ø There was no associa+on between LNF conc. and viral response pa^erns Ø Peak RTV levels > 12 ng/ml were associated with a triphasic response HDV RNA (log cp/ml) HDV RNA (IU/mL) Conc of RTV RTV (ng/ml) Time (wk)

17 HDV rebound associated with decline in LNF conc. There is no evidence of virological resistance LNF 5 mg QD + RTV 1 mg QD Log HDV RNA cp/ml Time (wk) HDV RNA (IU/mL) LNF ConcentraNon (ng/ml) LNF Conc (ng/ml) LNF 1 mg QD + RTV 1 mg QD Log HDV RNA cp/ml Log HDV RNA cp/ml LNF 75 mg QD + RTV 1 mg QD Time (wk) LNF ConcentraNon (ng/ml) LNF ConcentraNon (ng/ml)

18 Modeling HDV RNA, HBsAg and ALT kine-cs Ø (T) is the number of HBsAgproduc+ve cells that could be infected with HDV (V) to become (I) Ø T+I can proliferate with maximum prolifera+on rate r, according to a blind homeostasis process. Ø Pre-treatment steady state was assumed Ø LNF efficacy (%) in blocking HDV produc+on is represented by ε

19 Triphasic cases model fits well - HBsAg - HDV RNA - ALT Modeling curves: HDV RNA (solid line), HBsAg (dashed line) and ALT (do^ed line)

20 Flat par-al response model fits well - HBsAg - HDV RNA - ALT Modeling curves: HDV RNA (solid line), HBsAg (dashed line) and ALT (do^ed line)

21 Rebounders model fits well - HBsAg - HDV RNA - ALT Modeling curves: HDV RNA (solid line), HBsAg (dashed line) and ALT (do^ed line)

22 Triphasic response is explained by drug efficacy (ε) higher than pa-ent s cri-cal drug efficacy (ε c ) Pt Viral kinenc pa_ern (LNF dose) EsNmated LFN +RTV efficacy in blocking HDV producnon (ε) EsNmated crincal drug efficacy in each panent (ε c ) 1 Triphasic (5) Triphasic (75) Triphasic (1) Flat par+al (5) Flat par+al (75) Flat par+al (1) Rebound (5) Rebound (75) Rebound (1) When ε > ε c viral load is predicted to decline in a triphasic manner When ε < ε c viral load is predicted to decline in a flat par+al response manner. es+mated ε before viral rebound

23 Predic-ng -me to < 1 HDV par-cle in the en-re extracellular-body fluid HDV RNA (log IU/ml) LNF/RTV 5/1 mg PaNent 1 Model curve HDV RNA assay limit (14 IU/ml) <1 HDV RNA in extracellular fluid Projected Nme to HDV RNA < 1 parncle aher 49 wk of LNF/RTV Time (weeks) HDV RNA (log IU/ml) LNF/RTV 1/1 mg PaNent 3 Model curve HDV RNA assay limit (14 IU/ml) <1 HDV RNA in extracellular fluid Projected Nme to HDV RNA < 1 parncle aher 51 wk of LNF/RTV Time 25 (weeks)

24 Summary Ø LNF+RTV was safe and generally well tolerated Ø The kine+c analysis indicates that there is no associa+on between ritonavir boosted lonafarnib dosing (5, 75 or 1 mg daily) and viral response pa^ern (i.e., triphasic response, flat-par+al response and rebound) Ø The model was able to reproduce the observed viral, HBsAg and ALT kine+cs in each pa+ent and provided insights into viral-host-drug dynamics Ø Modeling results show high an+viral efficacy (95%) with ritonavir boosted lonafarnib

25 Discussion Ø The results suggest that on treatment (real +me) modeling may help individualize the length of lonafarnib-based therapy that is needed to achieve HDV clearance

26 Acknowledgements Christopher Koh Theo Heller Jay H. Hoofnagle T. Jake Liang Yaron Rotman Pallavi Surana Varun Takyar Nancy Fryzek Ohad Etzion Peter Pinto Ma Ai Thanda Han Erin Wolff Rachel Bishop Raissa Canales PreeN Dubey Gitanjali Subramanya Susan L Uprichard Sco^ J Cotler Cihan Yurdaydin Jeffrey S Glenn This research was supported by NIH grant R1-AI78881

27 Adverse Events Nausea Diarrhea Anorexia Abdominal BloaNng VomiNng Headaches FaNgue Significant Weight Change 8 (38%) 12 (57%) 12 (57%) 13 (62%) 2 (9%) 3 (14%) 1 (48%) (%) Symptom Group Significance Score Difference (1 pt scale) P value Nausea 5 vs 1 mg LNF Abdominal Bloa+ng 75 vs 1 mg LNF No significant changes in weight in all subjects No Grade 3 or 4 adverse events No discon+nua+ons due to adverse events