Pyrimidine inhibitors synergize with nucleoside analogues to block SARS-CoV-2

Abstract

The SARS-CoV-2 virus has infected more than 261 million people and has led to more than 5 million deaths in the past year and a half¹ ( https://www.who.org/ ). Individuals with SARS-CoV-2 infection typically develop mild-to-severe flu-like symptoms, whereas infection of a subset of individuals leads to severe-to-fatal clinical outcomes². Although vaccines have been rapidly developed to combat SARS-CoV-2, there has been a dearth of antiviral therapeutics. There is an urgent need for therapeutics, which has been amplified by the emerging threats of variants that may evade vaccines. Large-scale efforts are underway to identify antiviral drugs. Here we screened approximately 18,000 drugs for antiviral activity using live virus infection in human respiratory cells and validated 122 drugs with antiviral activity and selectivity against SARS-CoV-2. Among these candidates are 16 nucleoside analogues, the largest category of clinically used antivirals. This included the antivirals remdesivir and molnupiravir, which have been approved for use in COVID-19. RNA viruses rely on a high supply of nucleoside triphosphates from the host to efficiently replicate, and we identified a panel of host nucleoside biosynthesis inhibitors as antiviral. Moreover, we found that combining pyrimidine biosynthesis inhibitors with antiviral nucleoside analogues synergistically inhibits SARS-CoV-2 infection in vitro and in vivo against emerging strains of SARS-CoV-2, suggesting a clinical path forward.

Extended Data Fig. 1 |. High-throughput screening in Calu-3 cells to identify antivirals against SARS-CoV-2.

a–c, High throughput screening of ~18,000 drugs including the (a) UPENN library, (b) NCATS library and (c) ReFrame library in Calu-3 cells infected with SARS-CoV-2. Percent of Control (POC) for % infection is plotted versus rank of the drug across the primary screens. Yellow circles, DMSO controls; Green circles, remdesivir controls; Blue Circles, sample wells. Solid horizontal line represents 100% Infection of DMSO control wells. d, Huh7.5 or Vero (CCL81) cells were treated with the indicated nucleosides in dose response showing infection (blue) and toxicity (green). Data are presented as mean ± SD (n = 3 independent biological replicates). e, The indicated cell line was treated with increasing doses of 6-mercaptopurine and dose responses for infection (blue) and toxicity (green) are shown. Data are presented as mean ± SD (n = 3 independent biological replicates). f, p-values from Fig. 1d.

Extended Data Fig. 2 |. Structures of antiviral nucleoside analogs shown.

Nucleoside-related analog structures shown.

Extended Data Fig. 3 |. Combination of remdesivir and molnupiravir is additive.

a, The indicated cell line was treated with increasing doses of EIDD-1931 and dose responses for infection (blue) and toxicity (green) are shown. Data are presented as mean ± SD (n = 3 independent biological replicates). b, anti-SARS2 activity of remdesivir and molnupiravir as single agents. Data are presented as mean ± SD. c, anti-SARS2 activity of remdesivir in combination with a single concentration of molnupiravir or molnupiravir in combination with a single concentration of remdesivir. d, Two-dimensional representation of dose response interaction matrix for percent of control of infection mean ± SD, n = 9 independent biological replicates.

Extended Data Fig. 4 |. Inhibitors of host nucleoside metabolism are antiviral.

a, Calu-3 cells were treated with the indicated IMPDH inhibitors in dose response showing infection (blue) and toxicity (green). Data are presented as mean ± SD (n = 4 independent biological replicates). b, A549-Ace2 cells were treated with the indicated inhibitors in dose response showing infection (blue) and toxicity (green). Data are presented as mean ± SD (n = 4 independent biological replicates). c, Table of the EC50, CC50 and SI for the indicated cell line and drug (uM).

Extended Data Fig. 5 |. Combination of molnupiravir or remdesivir with DHODH inhibitors are synergistically antiviral in vitro.

a, Anti-SARS2 activity of molnupiravir, remdesivir, and Brequinar, as single agents. Data are presented as mean ± SD. b, Anti-SARS2 activity of molnupiravir or remdesivir in combination with a single concentration of Brequinar. c, Two-dimensional representation of dose response interaction matrix for percent of control of infection. Data are presented as mean ± SD (n = 6 independent biological replicates). d, Anti-SARS2 activity of molnupiravir, remdesivir, and BAY-2402234, as single agents. Data are presented as mean ± SD. e, Anti-SARS2 activity of molnupiravir or remdesivir in combination with a single concentration of BAY-2402234. f, Two-dimensional representation of dose response interaction matrix for percent of control of infection. Data are presented as mean ± SD (n = 6 for EIDD-2801, n = 8 for remdesivir for independent biological replicates).

Extended Data Fig. 6 |. Combination of molnupiravir or remdesivir with Pyrazofurin but not AVN944 is synergistically antiviral in vitro.

a, BLISS analysis in Calu3 cells with molnupiravir (EIDD-2801) or remdesivir in combination with pyrazofurin. Data are presented as mean values of Excess over BLISS. (n = 4 independent biological replicates). The statistical significance was determined by a one-sample t-test (* p < 5×10⁻²; **p < 10⁻³). b, Anti-SARS2 activity of molnupiravir, remdesivir, and AVN944 as single agents. Data are presented as mean ± SD. c, Anti-SARS2 activity of molnupiravir or remdesivir in combination with a single concentration of AVN944. d, Two-dimensional representation of dose response interaction matrix for percent of control of infection. Data are presented as mean ± SD (n = 10 for EIDD-2801, n = 6 for remdesivir for independent biological replicates).

Extended Data Fig. 7 |. Combination of molnupiravir with DHODH inhibitors is synergistic against SARS-CoV-2 variants.

a, Table of EC50, CC50 and SI for the indicated variant and the indicated drugs (uM) in Calu-3 cells. b, BLISS analysis in Calu3 cells with molnupiravir (EIDD-2801) in combination with DHODH inhibitors Brequinar or BAY-2402234 infected with the indicated variant. Data are presented as mean values of Excess over BLISS. The following number of independent replicates were performed: SARS-CoV2 Alpha; n = 8 for BAY2402234, n = 6 Brequinar; SARS-CoV2 Beta; n = 7, SARS-CoV2 Gamma; n = 6 BAY2402234, n = 8; SARS-CoV2; Delta: n = 8. The statistical significance was determined by a one-sample t-test (* p < 5×10⁻²; **p < 10⁻³, ***p < 10⁻⁴).

Extended Data Fig. 8 |. Nucleoside-related drugs are well-tolerated in air-liquid interface cultures.

a, Calu-3 cells were treated with the indicated concentrations of Molnupiravir (EIDD-2801) and infected with SARS-CoV-2. 48hpi viral replication was quantified by RT-qPCR and expression (viral RNA/18S) was normalized to vehicle treated cells. Data are presented as mean±SEM for n = 2–3 biological replicates. b, P-values for Fig. 3d. c, Nasal air-liquid interface cells were treated daily on the basolateral side with the indicated concentrations of drug, and at 72 h TEER, Cilia beating frequency (CBF) and toxicity (LDH) were measured. d, P-values for Fig. 4.

Extended Data Fig. 9 |. Combination of molnupiravir and Brequinar reduces SARS-CoV-2 infection and inflammation in vivo during therapeutic dosing.

Wild type Balb/C mice were treated with Brequinar (IP) and/or molnupiravir (EIDD-2801) (PO) daily at the indicated concentrations starting 24 h after infection. Mice (N = 5 mice per group) were intranasally inoculated with 1×10⁵ PFU/mouse of SARS-CoV-2/B.1.351. a, 3 dpi, lungs analyzed for viral titer by plaque assay, Mean±SD shown. * p < 0.05, ** p < 0.01, non-parametric one-way ANOVA with Dunnett’s multiple comparison test. (p-values: B1.135 only vs EIDD-2801 (25 mg/kg) = 0.045, vs EIDD-2801 (50 mg/kg = 0.016, vs EIDD-2801 (25 mg/kg)+ Breq (50 mg/kg) = 0.0017. b, c, 3 dpi, lungs fixed in 4% PFA for H&E staining (Scale bar = 200 μM) and quantified for interstitial inflammation. Mean±SD shown. ** p < 0.01, non-parametric one-way ANOVA with Dunnett’s multiple comparison test. (p-value = 0.0015).

Extended Data Fig. 10 |. Combination of molnupiravir or remdesivir with the protease inhibitor PF-07321332 is additive against SARS-CoV-2 (Beta, Delta).

BLISS analysis in Calu-3 cells with molnupiravir (EIDD-2801) or remdesivir in combination with PF-07321332 infected with the indicated SARS-CoV-2 variant. Data are presented as mean Excess over BLISS for n = 2 independent biological replicates (SARS-CoV-2 Beta); n = 6 independent biological replicates (SARS-CoV-2 Delta). The statistical significance was determined by a one-sample t-test (* p < 5×10⁻²; **p < 10⁻³).

Fig. 1 |. Antiviral nucleosides are highly active in respiratory cells and show cell-type-specific activity.

a, Pie chart of antivirals validated in Calu-3 cells with selective index (EC₅₀/CC₅₀) > 3. From approximately 18,000 compounds, 122 show activity. b, Sixteen nucleoside analogues validated in Calu-3 cells with the nucleoside type listed along with the EC₅₀, CC₅₀ and SI. c, Calu-3, A549-ACE2 or Caco-2 cells were treated with the indicated nucleosides in dose–response showing infection (blue) and toxicity (green). Data are presented as mean ± s.d. (n = 3 independent biological replicates). BCNA, bicyclic nucleoside analogue. d, Calu-3 cells pretreated with vehicle or 10 μM of the indicated drugs and infected with SARS-CoV-2 for 48 h and total RNA was subject to RT–qPCR analysis of viral infection. Data are presented as mean ± s.e.m. for reduction compared to vehicle control (n = 4 independent biological replicates) (***P < 0.001, one-way ANOVA; values are listed in Extended Data Fig. 1f). vRNA, viral RNA. e, RT–qPCR analysis of viral infection for the indicated drugs (remdesivir (10 μM), molnupiravir (10 μM), azathioprine (30 μM) and mercaptopurine (30 μM)) in nasal ALI cultures. Data are presented as mean ± s.e.m. (n = 2 independent biological replicates). f, BLISS analysis of the 2 × 2 combination of remdesivir and EIDD-2801 in Calu-3 cells showing additivity. Data are presented as mean values of excess over BLISS for n = 9 independent biological replicates. The statistical significance was determined by a one-sample Student’s t-test (*P < 5 × 10⁻²).

Fig. 2 |. Inhibitors of host nucleoside metabolism are antiviral.

a, Simplified schematic of nucleoside metabolism. Number of arrows indicates single versus multiple steps. CAD, carbamoyl-phosphate synthetase 2, aspartate transcarbamylase and dihydroorotase; GMPS, guanosine monophosphate synthetase; PRPP, phosphoribosyl diphosphate synthase. b, Dose–response analysis of the pyrimidine biosynthesis or purine biosynthesis inhibitors in Calu-3 cells. Infection (blue) and toxicity (green) are shown. Data are presented as mean ± s.d. (n = 3 independent biological replicates). c, Analysis of Calu-3 cells treated with the indicated inhibitor in the presence of increasing concentrations of the indicated nucleosides. Data are presented as mean ± s.d. (n = 3 independent biological replicates) for per cent infection. d, Calu-3 cells treated with the indicated inhibitor for 24 h and cell pellets subjected to liquid chromatography–mass spectrometry to quantify the nucleoside triphosphate levels. Mean ± s.e.m. is shown (n = 6 independent biological replicates). *P < 0.05, **P < 0.01, ***P < 0.001, one-way analysis of variance (ANOVA) with Dunnett’s multiple comparisons test (P = 0.0177 (ATP) P = 0.0045 (GTP) P < 0.0001 (CTP) and P = 0.0009 (UTP)).

Fig. 3 |. Combination of molnupiravir or remdesivir with DHODH inhibitors is synergistically antiviral in vitro.

a–c, BLISS analysis in Calu-3 cells with molnupiravir or remdesivir in combination with the DHODH inhibitors Brequinar (a) or BAY-2402234 (b) and the IMPDH inhibitor AVN944 (c). Data are presented as mean values of excess over BLISS. The number of independent biological replicates (n) is indicated for the combination tested. Statistical significance was determined by a one-sample Student’s t-test (*P < 5 × 10⁻², **P < 10⁻³ and ***P < 10⁻⁴). d, Calu-3 cells were treated with the indicated drugs and infected with SARS-CoV-2. Total RNA was subject to RT–qPCR 48 h after infection for analysis of infection. Data are presented as mean ± s.e.m. for reduction compared to vehicle control (n = 3 independent biological replicates). The black asterisks show the P value relative to vehicle: **P < 0.01, using one-way ANOVA. The green asterisks show P values relative to single treatment with EIDD-2801: *P < 0.05, **P < 0.01 and ***P < 0.001, using one-way ANOVA. P values are listed in Extended Data Fig. 8b. e, f, ALI-bronchial (e) or ALI-nasal (f) cells were treated with the indicated drugs and infected with SARS-CoV-2. Viral replication was quantified by RT–qPCR 72 h after infection and expression (vRNA/18S) was normalized to vehicle-treated cells. Data are presented as mean ± s.e.m. (n = 2 independent biological replicates). In d–f, the indicated drug concentration is in μM.

Fig. 4 |. Combination of molnupiravir and Brequinar reduces SARS-CoV-2 infection and inflammation in vivo.

Wild-type BALB/C mice were treated with Brequinar (intraperitoneal administration) and/or molnupiravir (oral administration) daily at the indicated concentrations starting 12 h before infection. Mice (n = 5 per group over 2 independent experiments) were intranasally inoculated with 1 × 10⁵ p.f.u. per mouse of SARS-CoV-2 (B.1.351). a–f, Lungs were analysed for viral titre 2 days after infection by plaque assay (a, d) or fixed in 4% paraformaldehyde for haematoxylin and eosin staining and quantified for interstitial inflammation (b, c, e, f). n = 5 mice per group. Mean ± s.d. is shown. *P < 0.05, **P < 0.01 and ***P < 0.001, using non-parametric one-way ANOVA with Dunnett’s multiple comparison test. The red asterisks are compared to vehicle; the blue asterisks are compared to molnupiravir. P values are listed in Extended Data Fig. 8d. Scale bars, 200 μM.