The cytotoxicity of lopinavir or ritonavir was determined in permissive African green monkey kidney Vero E6 cells (ATCC-1586) by CCK8 assay before the antiviral assay was carried out. Then, Vero E6 cells were treated with a series of concentrations of the two compounds, followed by infection with a clinical isolate of SARS-CoV-2 (nCoV-2019BetaCoV/Wuhan/WIV04/2019) (Zhou et al. 2020) at a multiplicity of infection (MOI) of 0.2. At 48 h p.i., viral copy numbers in the cell supernatant were quantified using quantitative real time PCR (qRT-PCR). The results showed that the half maximal inhibition concentration (IC50) and half cytotoxic concentration (CC50) values of lopinavir were ~ 12.01 μmol/L and 80.82 μmol/L, respectively, with the selectivity index (SI) of ~ 6.73 (Fig. 1A, upper left panel). For ritonavir, the IC50 and CC50 values were 19.88 μmol/L and 94.71 μmol/L (SI = 4.76), respectively (Fig. 1A, upper right panel). Meanwhile, the infected cells were fixed and subjected to immunofluorescence assay (IFA) using anti-sera against viral nucleocapsid protein (NP), and the results were in accordance with the qRT-PCR results, as evidenced by the decreased NP expression levels after the treatment of lopinavir and ritonavir at the indicated concentrations (Fig. 1B, upper panels).
Figure 1. In vitro inhibition of viral protease inhibitors against SARS-CoV-2 in Vero E6 cells. Vero E6 cells were infected with SARS-CoV-2 at an MOI of 0.2 in the treatment of different concentrations of the indicated compounds or DMSO control. At 48 h p.i., cell supernatants were collected and cells were fixed. A The viral yield in the cell supernatant was quantified by qRT-PCR. Cytotoxicity of these drugs to Vero E6 cells was measured by CCK-8 assays. The left and right Y-axis of the graphs represent mean % inhibition of virus yield and % cytotoxicity of the drugs, respectively. The experiments were done in triplicates. B Fixed cells were subjected to IFA by employing anti-NP rabbit sera. The nuclei were stained with Hoechst dye. Bars, 100 μm.
To further investigate whether lopinavir and ritonavir are effective in inhibiting SARS-CoV-2 replication in human cells, the efficacy of these two drugs in human liver cancer Huh7 cells were evaluated (MOI = 0.1). The results showed that lopinavir (IC50 = 7.79 μmol/L; CC50 = 64.43 μmol/L; SI = 8.27) and ritonavir (IC50 = 11.68 μmol/L; CC50 = 83.73 μmol/L, SI = 7.17) could also block virus infection in Huh7 cells (Fig. 2A, upper panels). Similarly, the IFA confirmed the inhibitory effect of lopinavir and ritonavir against SARS-CoV-2 (Fig. 2B, upper panels). To be noted, the lower MOI was used for infection of Huh7 cells (MOI = 0.1) compared to that for Vero E6 cells (MOI = 0.2), therefore less amounts of Huh7 cells compared to Vero E6 cells were found to be infected at 48 h p.i. (Fig. 1B and 2B, the DMSO control group). This may partially explain the decreased IC50 values for both drugs in Huh7 cells compared to the counterparts in Vero E6 cells.
Figure 2. In vitro inhibition of viral protease inhibitors against SARS-CoV-2 in Huh7 cells. Huh7 cells were infected with SARS-CoV-2 at an MOI of 0.1 in the treatment of different concentrations of the indicated compounds. At 48 h p.i., cell supernatants were collected and cells were fixed. A The viral yield in the cell supernatant was quantified by qRT-PCR. Cytotoxicity of these drugs to Vero E6 cells was measured by CCK-8 assays. The left and right Y-axis of the graphs represent mean % inhibition of virus yield and % cytotoxicity of the drugs, respectively. The experiments were done in triplicates. B Fixed cells were subjected to IFA by employing anti-NP rabbit sera. The nuclei were stained with Hoechst dye. Bars, 100 μm.
To explore at which stages of infection lopinavir and ritonavir functioned, time-of-drug-addition assay was performed as described previously (Liu et al. 2020). Compounds were added during or after virus entry, and virus production in the cell supernatants and NP expression in the infected cells were measured at 14 h p.i. As shown in Fig. 3A, both compounds inhibited SARS-CoV-2 infection after virus entry, but not during virus entry, suggesting that these drugs acted at post-entry stages of the SARS-CoV-2 infection in Vero E6 cells. This is in accordance with the presumed mode of action of lopinavir and ritonavir as viral protease inhibitors, which is supposed to inhibit virus replication by interfering viral polypeptide processing. However, this hypothesis needs further experimental evidence.
Figure 3. Time-of-drug-addition experiment of viral protease inhibitors. Time-of-drug-addition experiment of lopinavir and ritonavir (A) or rupintrivir and AG7404 (B). For "full-time" treatment, Vero E6 cells were pre-treated with indicated compounds for 1 h, and infected with SARS-CoV-2. Two hours later, the virus–drug mixture was removed, and the cells were cultured with compound-containing medium until the end of the experiment. For "Entry" treatment, Vero E6 cells were pre-treated with indicated compounds for 1 h, and infected with SARS-CoV-2. Two hours later, the virus–drug mixture was removed, and the cells were cultured with fresh culture medium until the end of the experiment. For "Post-entry" experiment, Vero E6 cells were infected with SARS-CoV-2. Two hours later, the virus-containing medium was removed, and the cells were cultured with compound-containing medium until the end of the experiment. For all treatments, an MOI of 0.2 was used for lopinavir and ritonavir groups, and an MOI of 0.05 was used for rupintrivir and AG7404 groups. At 14 h p.i., virus yield in the infected cell supernatants was quantified by qRT-PCR (left) and NP expression in infected cells was analyzed by Western blot (right).
As mentioned before, apart from lopinavir and ritonavir, we also tested rupintrivir and AG7404. In both Vero E6 cells and Huh7 cells, rupintrivir inhibited SARS-CoV-2 infection only at high drug concentrations (Vero E6: IC50 = 34.08 μmol/L; CC50 > 100 μmol/L, SI > 2.93; Huh7: IC50 = 25.38 μmol/L; CC50 > 100 μmol/L, SI > 3.94); while AG7404 inhibited SARS-CoV-2 infection at even higher concentrations (Vero E6: IC50 ~ 195.8 μmol/L CC50 > 400 μmol/L, SI > 2.04; Huh7: IC50 = 92.55 μmol/L; CC50 > 400 μmol/L, SI > 4.32) (Figs. 1A and 2A, lower panels). The antiviral activities of these compounds were also confirmed by IFA (Figs. 1B and 2B). By using high drug concentrations, both rupintrivir and AG7404 were demonstrated to inhibit SARS-CoV-2 at the stages post virus infection (Fig. 3B), which was similar to the possible mechanism of lopinavir and ritonavir in antagonizing SARS-CoV-2 infection.
To explore the molecular interaction pattern, the coordinates of X-Ray crystal structure of the 3CLpro from the SARS-CoV-2 (PDB number: 6LU7; Resolution: 2.16Å) were used as the template to dock the four inhibitors (lopinavir, ritonavir, rupintrivir and AG7404). The structure was refined by performing an energy minimization with Schrodinger software using protein preparation module. The docking results indicated that Lopinavir as ligand gave the highest score (− 9.327) (Fig. 4), suggesting it might be a potent inhibitor interacting with the SARS-CoV-2 3CLpro active site residues. Comparative active site analysis of SARS-CoV (1Q2W) and SARS-CoV-2 3CLpro showed highly conserved residues Thr23 (in model Thr25), His39 (His41), Cys42 (Cys44), Tyr48 (Tyr54), Cys139 (Cys145), His157 (His163), Glu160 (Glu166), and Gln183 (Gln189). The active site residues in SARS-CoV-2 3CLpro His41, Cys145 and Glu166 are directly involved in biding with lopinavir which was successively docked over the active site of the crystal structure. The docking scores of − 8.522, − 7.239 and − 6.414 were obtained for ritonavir, rupintrivir and AG7404, respectively (Fig. 4), implying that the binding of the catalytic site was not as ideal as lopinavir. The crystal structures of SARS-CoV-2 3CLpro in complex with inhibitors supported the antiviral efficacy of the four drugs (Figs. 1 and 2), and shed light onto the possible drug binding mechanism, as well as the development of novel 3CLpro inhibitors against novel SARS-CoV-2.
Figure 4. Docking of compounds on active site of SARS-CoV-2 3CLpro homology model. For molecular docking, the compounds (lopinavir, ritonavir, rupintrivir, AG7404) were built and subjected to the optimize geometry calculation in mechanics using MMFF94 parameter at DS2.5. The compounds were then docked at the homology model of SARS-CoV-2 3CLpro using Schrodinger Glide program. The docking scores of these four compounds were calculated and listed below.
To assess the antiviral activity against SARS-CoV-2 by lopinavir and ritonavir in vivo, and figure out whether lopinavir and ritonavir still worth further clinical trials, in vitro IC50 values toward SARS-CoV-2 and HIV and in vivo exposure were summarized and compared (Table 1). The Cmax of lopinavir and ritonavir following administration of 400 mg of lopinavir and 100 mg of ritonavir to healthy male volunteers are 13.5 μmol/L and 0.15 μmol/L, respectively, however, it is well recognized that only a free drug that is unbound to plasma proteins can move freely across cellular membranes to distribute between plasma and tissue (http://www.fda.gov/cder/foi/nda/2000/21-226_Kaletra_biopharmr_P1.pdf). Thus, it is generally thought that free concentrations in plasma should be equivalent to free concentrations in tissue at the steady state, which contributes to therapeutic effects (Smith et al. 2010). Considering the free concentration of lopinavir and ritonavir on clinic relevant level (0.2 μmol/L and 0.002 μmol/L, respectively) is well below the in vitro inhibition of SARS-CoV-2 in Vero E6 cells and Huh7 cells, the antiviral effect against SARS-CoV-2 is not likely achievable in vivo (or in clinic) with the dosing regimen in the proposed COVID-19 treatment. In contrast, the mean IC50 values of lopinavir against several clinical isolates of HIV-1 subtype B ranged from 4 to 11 nmol/L in the absence of serum (Abbott Laboratories, European Medicines Agency). Even At the end of the dosage interval, the mean trough concentrations (Ctrough) of lopinavir was as low as 5.3 μmol/L with protein adjusted unbound concentration of ~ 80 nmol/L, which is still more than 10-fold higher than its in vitro IC50 against HIV (Smith et al. 2010). Regarding to ritonavir, considering the low free plasma concentration (2–0.6 nmol/L), the in vivo activity of ritonavir could be ignored. To sum, the IVIVE of lopinavir and ritonavir suggested that these two drugs are far from capable of inhibiting SARS-CoV-2 in vivo.
Drug Virus In vitro efficiacy (µmol/L) In vivo concentration (µmol/L) IC50 (µmol/L) Cmax Unbound Cmax Cmin Unbound Cmin Lopinavir SARS-CoV-2 7.79–12.01 13.5b 0.2c 5.3d 0.08d HIV 0.004–0.011a Ritonavir SARS-CoV-2 11.68–19.88 0.15e 0.002 0.038e 0.0006 HIV 0.022–0.16e Unbound Cmax and unbound Cmin represent the free plasma concentration after protein-binding correction. Protein bindings of lopinavir and ritonavir are 98%–99%. 98.5% were used here to calculate the unbound values.
In vitro anti-HIV activity values used here obtained from the in vitro activity study absence of human albumin.
aData obtained from Abbott Laboratories and European Medicines Agency.
bData obtained from http://www.fda.gov/cder/foi/nda/2000/21-226_Kaletra_biopharmr_P1.pdf.
cData obtained from reference
Gulati et al. (2009).
dData obtained from reference
Murphy et al. (2001).
eData obtained from reference Cvetkovic and Goa (2003).
Table 1. Comparation of lopinavir and ritonavir in vitro antiviral efficacy against SARS-CoV-2 or HIV and in vivo plasma exposures.