Bictegravir Dissociation Half-life from HIV-1 G140S/Q148H Integrase-DNA Complexes Poster 497 Kirsten White, PhD , Inc. Kirsten White*, Anita Niedziela-Majka, Nikolai Novikov, Michael Miller, Haolun Jin, Scott Lazerwith, and Manuel Tsiang 333 Lakeside Drive Foster City, CA 94404 Gilead Sciences, Foster City, CA Tel: (650) 522-5162 E-mail: [email protected]

Introduction Methods (cont'd) Results Conclusions

♦♦ Bictegravir (BIC; GS-9883) (Figure 1) is a novel, potent INSTI that improves upon existing INSTIs and is part of a complete single tablet regimen with the Figure 2. SPA-Based INSTI Binding Assay Table 1. Susceptibility of INSTIs to Clinically-derived Isolates of HIV-1 Table 4. Dissociation Half-lives of INSTIs from WT and G140S/Q148H ♦♦ Bictegravir is a potent INSTI with a high barrier to 1, 2,3 F/TAF backbone 3 Containing G140S + Q148H ± other INSTI-R Substitutions HIV-1 Integrase-DNA Complexes by the Equilibrium Binding kon determined from direct binding of H-INSTI 3 resistance and improved activity compared to DTG, RAL, k determined from competition binding of H-INSTI with cold/non-radioactive INSTI a Model ♦♦ Bictegravir has an improved resistance profile compared to off Susceptibility of HIV-1 to INSTIs Bound and EVG against HIV-1 with INSTI resistance mutations (EVG), (RAL), and (DTG), particularly for patient Radiolabeled Dissociation t of INSTI from IN-DNA Complexes Wild-Type G140S/Q148H ± Other Mutants (n=16) 1/2 6 IN/DNA INSTI isolates with high-level INSTI resistance containing combinations of mutations INSTI Sensitivity (% of isolates in vitro Assembly Fold- Wild-Type IN G140S/Q148H Mutant IN 4,6 c p-value INSTI such as E92Q+N155H or G140C/S+Q148R/H/K (Table 1) EC50 (nM) Change below cut-offs) b vs BIC ♦♦ Bictegravir has the longest measured dissociation half- In biochemical studies, the apparent dissociation half-life of DTG from IntegraseD imer vs WT <2.5-fold <4-fold ♦♦ t1/2 (hr)* p-value vs BIC t1/2 (hr)* p-value vs BIC integrase-DNA complexes was shown to be longer than RAL or EVG and was BIC 1.9 3.4 ± 1.7 50% 75% -- life from wild-type HIV-1 IN-DNA complexes compared 3’-processed predicted to correlate with potent antiretroviral activity and a higher genetic donor DNA particles DTG 2.8 7.6 ± 4.3 0% 25% < 0.001 BIC 38 ± 19 -- 2.5 ± 0.07** -- to DTG, RAL, and EVG barrier to resistance5 RAL 7.1 >143 0% 0% < 0.001 3’ 3’ DTG 16 ± 9 0.017 0.65 ± 0.2*** 0.0076 ♦♦ Bictegravir also has the longest measured dissociation ♦♦ In this study, we report the dissociation kinetics of bictegravir and other INSTIs Biotin/Streptavidin EVG 2.3 >150 0% 0% < 0.001 RAL 5.2 ± 0.6 0.003 ND ND interaction half-life from HIV-1 IN-DNA from wild-type and G140S/Q148H HIV-1 integrase-DNA complexes in vitro a. PhenoSense IN assay (Monogram Biosciences) EVG 1.5 ± 0.2 0.0006 ND ND mutant G140S/Q148H b. Mean ± SD SPA Bead Surface Light c. The 2.5-fold cutoff is a standard biological cut-off in the Monogram assay; 4-fold is the lower clinical cut-off *Average ± standard deviation from 2 to 7 experiments complexes compared to DTG

for DTG in the PhenoSense IN assay **BIC t1/2 with G140S/Q148H mutant IN-DNA complexes was statistically longer than the EVG dissociation t1/2 with wild-type IN-DNA complexes. –– RAL and EVG did not bind efficiently enough for ***DTG t1/2 with G140S/Q148H mutant IN-DNA complexes was statistically shorter than the EVG dissociation t1/2 Methods with wild-type IN-DNA complexes. dissociation measurements ♦♦ Exponential Decay Analysis: The apparent association t1/2 was determined by curve fitting the binding phase using equation (1): Table 2. Apparent Association Half-life of INSTIs from HIV-1 Integrase- ♦♦ Bictegravir has the longest measured dissociation half-life compared to ♦♦ Long residence times of INSTIs on the integrase-DNA Figure 1. Structure of Bictegravir (BIC) and other INSTIs −kt −= eMy )1( (1) where M is the plateau and t 2/1 = /)2(ln k DNA Complexes DTG, RAL, and EVG complex have been correlated with potent antiretroviral –– Significantly longer from wild-type HIV-1 IN-DNA complexes activity and a high barrier to resistance in vitro5 ♦♦ The apparent dissociation t1/2 was determined by curve fitting the Apparent Association t1/2 of INSTI from IN-DNA Complexes competition binding phase using equation (2): INSTI Wild-Type IN G140S/Q148H IN –– Significantly longer from G140S/Q148H HIV-1-DNA complexes –– The long plasma half-life and high C of BIC in vivo − = min kt t 2/1 /)2(ln k Association Association 3 = eMy )( (2) where M is the starting value and p-value vs BIC p-value vs BIC should also contribute to a high resistance barrier t1/2 (min)* t1/2 (min)* k k Figure 3. Association and Dissociation Time Course of BIC and DTG ♦♦ The on and off values of the compounds were determined by curve fitting BIC 35 ± 5 -- 37 ± 4 -- ♦♦ Phase 3 clinical studies of the single tablet regimen of from WT HIV-1 Integrase-DNA Complexes and Equilibrium an equilibrium binding model (scheme A) simultaneously to both the DTG 36 ± 4 0.648 24 ± 5 0.1050 Binding Model Fit BIC/FTC/TAF are ongoing. binding and competition binding phases. The actual dissociation t1/2 was RAL 24 ± 8 0.020 nd nd

calculated as (ln 2)/koff EVG 15 ± 6 0.00003 nd nd *Average ± standard deviation from 2 to 7 experiments 120 120 k-2 koff kon k2 nd = No data available due to low binding of INSTI to the mutant IN-DNA complexes (scheme A) 100 BIC+WT IN/DNA 100 DTG+WT IN/DNA ECf EC C + E + L EL ELf d d k2 kon koff k–2 80 80 ♦♦ BIC and DTG have similar association rates to IN-DNA complexes References 60 60 Qi at time = t c

2+ 40 Metal-Chelating Core: Oxygen atoms chelate a pair of Mg ions and bind the integrase catalytic active site 3 Boundximum 40 1. Lazerwith et al., “Discovery of Bictegravir (GS-9883), a Novel, Unboosted, Halogenated Phenyl: Interacts with the integrase pocket that is normally occupied by the terminal 3’ base of viral DNA where E = integrase-donor DNA complex, L = H-INSTI, C = cold (non- Once-Daily HIV-1 Integrase Strand Transfer Inhibitor (INSTI) with Improved 20 20 % Ma % Maximum Boun radioactive) INSTI, EL = 3H- intermediate complex, EL = 3H final complex, Table 3. Dissociation Half-lives of INSTIs from WT HIV-1 Integrase-DNA % Maximum Boun f Complexes Pharmacokinetics and In Vitro Resistance Profile.” ASM Microbe, June 19, 2016, EC = cold intermediate complex, EC = cold final complex, k and k = 0 0 Boston, MA, Poster #414. Clinically-derived HIV-1 Isolate Phenotyping: A subset of a panel of 47 f on off 0 5

♦♦ 0 5 -20 10 20 30 40 50 75 on- and off-rate constants for L and C, k & k = k and k for EL and 10 20 30 40 50 75 100 125 150

2 -2 forward reverse Dissociation of INSTI from Wild-type IN-DNA Complexes* 100 125 150 2. Tsiang et al., “Antiviral Activity of Bictegravir (GS-9883), a Potent Next Generation patient-derived HIV-1 isolates with high-level INSTI resistance were profiled 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 Time (hr) EC, i = rate of addition of C and Q = dosing function for C, (0 or 1) TimeTIME (hr)(min) HIV-1 Integrase Strand Transfer Inhibitor.” ASM Microbe, June 19, 2016, Boston, for susceptibility to BIC, DTG, EVG, and RAL using the PhenoSenseIN By Exponential Decay By Equilibrium Binding Model INSTI MA, Poster #416. assay conducted by Monogram Biosciences (South San Francisco, CA). The 120 120 Equilibrium Binding Model Analysis: The equilibrium binding model Apparent t1/2 p-value vs BIC t1/2 (hr) p-value vs BIC 3. Gallant et al., “Novel Integrase Strand Transfer Inhibitor Bictegravir 10 Day panel was chosen from the available Monogram HIV-1 library and included (hr) [**] 100 BIC+140/148 IN/DNA 100 DTG+140/148 IN/DNA incorporates a correction for the gradual sedimentation of the SPA beads Monotherapy in HIV-1-Infected Patients.” ASM Microbe, June 19, 2016, Boston,

all available isolates with >2.5-fold reduced susceptibility to DTG (n=24) as BIC 135 ± 20 [na] -- 38 ± 19 -- d by replacing k and k with decreasing hyperbolic functions of time (z 80 80 well as representative isolates with EVG and/or RAL resistance mutations on off on MA, Poster #415. 79 ± 13 [71] < 0.0001 16 ± 9 0.017 Bound and z ) where tc is the time of cold competitor addition and r and n are DTG 60 60 (n=23). Here, the 16 isolates with G140S + Q148H ± other INSTI resistance- off 4. Jones et al., “Bictegravir (GS-9883), a Novel HIV-1 Integrase Strand Transfer RAL 14 ± 3 [8.8] < 0.0001 5.2 ± 0.6 0.003 correction factors. 40 Inhibitor (INSTI) with Optimized In Vitro Resistance profile.” ASM Microbe, June 19, associated substitutions are described. 40 rk − )1( k EVG 3.6 ± 0.7 [2.7] < 0.0001 1.5 ± 0.2 0.0006 2016, Boston, MA, Poster #413. on on LBound, #BIC_Set1AM 20 20 LBound, LBound, #DTG_Set1AM % Maximum Boun 3 = kz (for t < t ) z = + (for t > t ) % Maximum ♦♦ SPA Assay: The apparent association and dissociation kinetics of H-labelled onon c on c 5. Hightower et al., “Dolutegravir (S/GSK1349572) Exhibits Significantly Slower c )( +− rttn r *Average ± standard deviation from 5 to 7 experiments 0 BIC, DTG, EVG, and RAL were measured using wild-type or G140S/Q148H 0 **Published t1/2 values from Hightower et al., Antimicrobial Agents and Chemotherapy. (2011) 55(10):4552-4559. Dissociation then Raltegravir and Elvitegravir from Wild-Type and Integrase

-20 0 5 -20 0 5 rk − )1( k 10 20 30 40 50 75 HIV-1 integrase-DNA complexes and a scintillation proximity assay at 37ºC off off 10 20 30 40 50 75

0 1000 2000 300 0 4000 500 0 6000 100 7000 800125 0 9000150 Inhibitor-Resistance HIV-1 Integrase-DNA Complexes.” Antimicrobial Agents and

0 1000 2000 3000 4000 5000 6000 100 7000 8000125 9000150 5 = kz (for t < t ) z = + (for t > t ) TimeTIME (hr)(min) Time (hr) as previously described (Figure 2) . Data were analyzed using the published off off c off +− c TIME(min) Chemotherapy. (2011) 55(10):4552-4559. c )( rttn r ♦♦ The Exponential Decay Method of analysis results in an overestimation of exponential decay method. During the long time course required for these 6. Tsiang et al., “Antiviral Activity of Bictegravir (GS-9883), a Novel Potent HIV-1 Stand assays, gradual sedimentation of the SPA beads resulted in overestimates of ♦♦ This deviation correction, enabled the determination of an initial half- the dissociation half-lives due to an experimental artifact; the Equilibrium Transfer Inhibitor with an Improved Resistance Profile.” Antimicrobial Agents and Binding Method more accurately describes the kinetics dissociation half-lives of the INSTIs. Therefore, an equilibrium binding model life (i.e. actual dissociation t1/2 = (ln2)/koff ) and a terminal half-life (i.e. Chemotherapy. (2016) 60(12):7086-7097. was developed here and was used for the analysis of the G140S/Q148H a prolonged t1/2 = r(ln2)/koff due to the gradual sedimentation of the SPA mutant IN-DNA complexes. beads).

Presented at the Conference on Retroviruses and Opportunistic (CROI), February 13-16, 2017, Seattle, WA © 2017 Gilead Sciences, Inc. All rights reserved.