An orally bioavailable SARS-CoV-2 main protease inhibitor exhibits improved affinity and reduced sensitivity to mutations

Abstract

Inhibitors of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) main protease (M^pro) such as nirmatrelvir (NTV) and ensitrelvir (ETV) have proven effective in reducing the severity of COVID-19, but the presence of resistance-conferring mutations in sequenced viral genomes raises concerns about future drug resistance. Second-generation oral drugs that retain function against these mutants are thus urgently needed. We hypothesized that the covalent hepatitis C virus protease inhibitor boceprevir (BPV) could serve as the basis for orally bioavailable drugs that inhibit SARS-CoV-2 M^pro more efficiently than existing drugs. Performing structure-guided modifications of BPV, we developed a picomolar-affinity inhibitor, ML2006a4, with antiviral activity, oral pharmacokinetics, and therapeutic efficacy similar or superior to those of NTV. A crucial feature of ML2006a4 is a derivatization of the ketoamide reactive group that improves cell permeability and oral bioavailability. Last, ML2006a4 was found to be less sensitive to several mutations that cause resistance to NTV or ETV and occur in the natural SARS-CoV-2 population. Thus, anticipatory design can preemptively address potential resistance mechanisms to expand future treatment options against coronavirus variants.

Fig. 1.. Mpro inhibitor leads were derived from HCV protease inhibitors.

(A) Manual rigid docking of BPV into a 13b-${\text{M}}^{\text{pro}}$ co-crystal structure (PDB 6Y2G, protomer B). Using PyMOL, BPV was manually placed into the SARS-CoV-2 ${\text{M}}^{\text{pro}}$ active site and unconstrained bonds were rotated for optimal complementarity. (B) Structures of BPV and TPV and the derived ML1000 and ML1100. ML1000 is essentially BPV with a P1 $\text{γ}$-lactam. ML1100 replaces the bicyclic P2 proline analog of BPV with that of TPV. (C) Inhibition of ${\text{M}}^{\text{pro}}$-coil activity by GC376, ML1000, and ML1100 at 25 °C after 1 h of preincubation. Data presented as mean ± SD. (D) Crystal structure of the ML1000-${\text{M}}^{\text{pro}}$ adduct with H-bond interactions highlighted with dashed lines. In (A) and (D), as well as representations of crystal structures in all following figures, oxygen, nitrogen, sulfur, and fluorine atoms are colored by a typical red, blue, yellow, and light blue color scheme, whereas carbon atoms are colored according to the chosen custom color scheme, for example teal for ML1000 in (D).

Fig. 2.. Removal of H-bond donors improved permeability and potency.

(A) ${\text{M}}^{\text{pro}}$ inhibitors based on ML1000. Additionally, select non-constrained inhibitors (P2 = Leu) were synthesized to assess the effects of P2 rigidification. (B) Progressing from ML1000, stepwise optimization of P4 (R), P1’ (R’), and P1 was performed. (C) Inhibitor activity against purified ${\text{M}}^{\text{pro}}$-coil (${\text{IC}}_{50}$), intracellular ${\text{M}}^{\text{pro}}$ (${\text{IC}}_{50,\text{cell}}$), and SARS-CoV-2 replication (${\text{EC}}_{50}$). 95% confidence intervals (${\text{IC}}_{95}$) are reported in parentheses and data, fits, and statistical details are found in fig. S4 to S7. ${\text{IC}}_{50}$ data were measured at 25 °C after 1 h of preincubation. In contrast to NTV, the ketoamides have not reached equilibrium. n/a, not applicable. –, not available. (D) Co-crystals of ${\text{M}}^{\text{pro}}$ with ML1000 or ML1001. The arrow indicates contraction of the S4 pocket for ML1001. (E) A ${\text{M}}^{\text{pro}}$-ML1006m co-crystal. The circle highlights the position for N,N-substitution and extension of the ketoamide into S1’. (F and G) Inhibition curves for (F) ${\text{M}}^{\text{pro}}$-coil in solution and (G) SARS-CoV-2 replication in Huh7.5.1++ cells. Data are presented as mean ± SD. (H) Co-crystal of ${\text{M}}^{\text{pro}}$-ML1006a.

Fig. 3.. SARS-CoV-2 Mpro inhibitors with higher in vitro potency than NTV were discovered.

(A) P1 lactamyls and azetidine substitutions were tested to improve permeability. (B) ${\text{IC}}_{50}$ curves are shown for ML1006a, ML2006a, ML3006a, and ML4006a. (C) Tabulation of inhibitor activity against purified ${\text{M}}^{\text{pro}}$-coil (${\text{IC}}_{50}$ and ${\text{K}}_{\text{i}}$), intracellular ${\text{M}}^{\text{pro}}$ $\left({\text{IC}}_{50,\text{cell}}\right)$, and SARS-CoV-2 replication (${\text{EC}}_{50}$). ${\text{IC}}_{50}$ are reported in parentheses and data, fits, and statistical details are found in fig. S5, S6, S7, S9, and S13. ${\text{IC}}_{50}$ and ${\text{K}}_{\text{i}}$ data were measured at 37 °C after 3 h of preincubation to allow the equilibrium to settle for the ketoamides. –, not available (D) Crystal structure of the covalent adduct of ML2006a4 and SARS-CoV-2 ${\text{M}}^{\text{pro}}$. (E) Intracellular inhibition of ${\text{M}}^{\text{pro}}$ $\left({\text{IC}}_{50,\text{cell}}\right)$ was measured using a luciferase proteolysis reporter. (F) Inhibition of SARS-CoV-2 replication in Huh7.5.1++ cell cultures by ML2006a derivatives. Data in (B), (E), and (F) are presented as mean ± SD.

Fig. 4.. ML2006a4 is orally bioavailable in mice.

(A) Key parameters from non-compartmental analysis of the mean inhibitor plasma concentration from pharmacokinetic studies performed in mice $(n=3)$ with oral (po) dosing at 20 mg/kg and intravenous (iv) dosing at 2 mg/kg in a saline solution with 40% PEG300, 10% DMSO, and 5% Tween-80 (PDT). In parentheses, uncertain estimates obtained as described in Materials and Methods. For direct experimental observations, mean ± SD are reported. ^aThe unbound drug peak concentration $\left({\text{C}}_{\max ,\text{u}}\right)$ and area under the curve extrapolated to infinity $\left({\text{AUC}}_{\infty ,\text{u}}\right)$ were calculated using ${\text{f}}_{\text{u}}$ reported in table S1. ^bOral bioavailability (F). ^cPlasma clearance rate $\left({\text{CL}}_{\text{p}}\right)$. ^dVolume of distribution at steady state $\left({\text{V}}_{\text{ss}}\right)$. ^eRatio of inhibitor concentrations in lungs and plasma measured 8 h post oral dosing. ^fRTV dosed po at 20 mg/kg 30 min prior to test compound. ^{g $(n=2)$} (B) Inhibitor plasma concentrations were monitored for 24 h after oral administration of a 20 mg/kg RTV pre-dose followed by single-doses of 20 mg/kg ML2006a4, NTV, or PTV solutions in PDT. (C) Inhibitor plasma concentrations were monitored for 12 h after single-agent oral administration of 300 mg/kg ML2006a4, NTV, or PTV suspensions in 0.5% methylcellulose and 5% Tween-80 (MCT-5). Data in (B) and (C) are presented as mean ± SD.

Fig. 5.. ML2006a4 protects against SARS-CoV-2 in mice.

(A) Pharmacokinetics of the dosing regimen used for safety and efficacy testing in mice (n = 3). Plasma concentrations were monitored for 24 h after oral administration of co-suspensions of 40+20 mg/kg ML2006a4+RTV or NTV+RTV in 0.5% methylcellulose and 2% Tween-80 (MCT-2). (B) Protocol for efficacy testing of ML2006a4 in 16-week-old female BALB/c mice. After intranasal infection with SARS-CoV-2 MA10 (10⁵ TCID₅₀), the mice received inhibitors po bid for 4 days. ML2006a4+RTV and NTV+RTV were dosed as in (A), and the RTV-only control was dosed at 20 mg/kg. All infected groups contained 10 mice with 5 mice euthanized at 2 dpi and 6 dpi, respectively. Uninfected control groups contained 5 mice that were euthanized at 6 dpi. (C and D) Daily survival curves (C) and body weights (D) are shown for all five groups. No death or clinical decline was observed for any of the ML2006a4+RTV or NTV+RTV treated groups. (E and F) Infectious viral titers (E) and viral RNA loads (F) were determined in lungs harvested at 2 and 6 dpi, respectively. For the RTV-only group, data plotted at 6 dpi are from lungs collected at death (actual time 3 to 5 dpi). (G) Representative images of H&E-stained lungs collected at 2 or 6 dpi. Scale bar indicates 80 $\text{μm}$. Histopathological analysis identified peribronchiolar and perivascular infiltrations with mononuclear cells and respiratory epithelial cell injury in the three infected groups. RTV and NTV+RTV groups exhibited a high extent of infiltrations and severe injury in respiratory epithelia (black arrows, epithelial cell debris in the lumen) at 2 dpi followed by incomplete epithelial regeneration (orange arrows, missing epithelium) at 5 or 6 dpi. Data are presented as mean ± SD.

Fig. 6.. ML2006a4 exhibited improved mutation tolerance.

(A) Comparison of NTV and ML2006a4 in vitro inhibition curves against ${\text{M}}^{\text{pro}}$-coil WT and S144A. Data are presented as mean ± SD. (B) Inhibition constants, ${\text{K}}_{\text{i}}$, at 37 °C for ${\text{M}}^{\text{pro}}$-coil WT and six mutants. Colored numbers below represent fold change in ${\text{K}}_{\text{i}}$ relative to WT ${\text{M}}^{\text{pro}}$. (C) ${\text{EC}}_{50}$ values were determined for ML2026a4 that contains a smaller P3 group relative to ML2006a4. In (B) and (C), ${\text{IC}}_{95}$ are reported in parentheses and data, fits, and statistical details are found in fig. S5, S6, S13, and S19.