key: cord-0746488-b5uuws4j
authors: Chaves, Otávio Augusto; Sacramento, Carolina Q.; Ferreira, André C.; Mattos, Mayara; Fintelman-Rodrigues, Natalia; Temerozo, Jairo R.; Vazquez, Leonardo; Pinto, Douglas Pereira; da Silveira, Gabriel P. E.; da Fonseca, Laís Bastos; Pereira, Heliana Martins; Carlos, Aluana Santana; d’Avila, Joana C.; Viola, João P. B.; Monteiro, Robson Q.; Bozza, Patrícia T.; Castro-Faria-Neto, Hugo Caire; Souza, Thiago Moreno L.
title: Atazanavir Is a Competitive Inhibitor of SARS-CoV-2 M(pro), Impairing Variants Replication In Vitro and In Vivo
date: 2021-12-24
journal: Pharmaceuticals (Basel)
DOI: 10.3390/ph15010021
sha: 002e0f685e9bb1058f366218c58a8c5e431cc92f
doc_id: 746488
cord_uid: b5uuws4j

Atazanavir (ATV) has already been considered as a potential repurposing drug to 2019 coronavirus disease (COVID-19); however, there are controversial reports on its mechanism of action and effectiveness as anti-severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Through the pre-clinical chain of experiments: enzymatic, molecular docking, cell-based and in vivo assays, it is demonstrated here that both SARS-CoV-2 B.1 lineage and variant of concern gamma are susceptible to this antiretroviral. Enzymatic assays and molecular docking calculations showed that SARS-CoV-2 main protease (M(pro)) was inhibited by ATV, with Morrison’s inhibitory constant (K(i)) 1.5-fold higher than GC376 (a positive control) dependent of the catalytic water (H(2)O(cat)) content. ATV was a competitive inhibitor, increasing the M(pro)’s Michaelis–Menten (K(m)) more than sixfold. Cell-based assays indicated that different lineages of SARS-CoV-2 is susceptible to ATV. Using oral administration of ATV in mice to reach plasmatic exposure similar to humans, transgenic mice expression in human angiotensin converting enzyme 2 (K18-hACE2) were partially protected against lethal challenge with SARS-CoV-2 gamma. Moreover, less cell death and inflammation were observed in the lung from infected and treated mice. Our studies may contribute to a better comprehension of the M(pro)/ATV interaction, which could pave the way to the development of specific inhibitors of this viral protease.

The 2019 coronavirus disease was first reported in Wuhan (China) and its etiological agent, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), spread globally into the second pandemic of the 21st century, after the influenza A(H1N1) pandemic in 2009. SARS-CoV-2 continuously circulated, even among individuals with preexisting immunity, and since COVID-19 began until the time of writing has caused about 271 million confirmed cases along with 5.2 million deaths worldwide [1] [2] [3] [4] [5] . SARS-CoV-2 variants of concern (VoC) might escape the humoral immune response to natural infection or vaccination [4, 5] , which reinforces the necessity of specific antiviral treatments. After almost a two-year effort in the repurposing of clinically approved drugs, limited benefits for COVID -19 patients have been demonstrated [5] . Thus, it is necessary to improve the pre-clinical characterization of repurposed drugs to rationalize further clinical studies, in terms of posology and susceptibility of VoC, and learn from their interactions with the viral target enzyme.

During the SARS-CoV-2 life cycle, viral surface glycoprotein spike interacts with human angiotensin converting enzyme 2 (ACE2) and with the extracellular protease transmembrane protease serine 2 (TMPRSS2) to enter the cells. Positive-sense SARS-CoV-2 genomic RNA is translated to polyproteins that are auto-cleaved by viral papain-like (PL pro ) and major (M pro ) proteases. These matured non-structural proteins (nsp), such as nsp12/8/7 (RNA polymerase complex), nsp13 (helicase), nsp14/10 (exonuclease and N7-ethyltransferase), nsp15 (endonuclease) and nsp16 (2'-O-methyltransferase), are responsible for RNA replication and transcription of structural genes. These structural genes will be translated, assembled with the newly synthetized genomic RNA and bud as a new virus progeny. These events represent key steps during viral life cycle to potentially limit SARS-CoV-2 replication [6] .

We have previously described that atazanavir (ATV, Figure 1 ), a clinically approved human immunodeficiency virus (HIV) protease inhibitor [7] , is endowed with anti-SARS-CoV-2 activity [8] . Although we described early in the pandemic outbreak the ATV's activity against SARS-CoV-2, by in silico and cell-based assays (in Vero-E6 and A549 cell lines), proper characterization of its enzymatic target, pharmacology on type II pneumocytes (the most affect cell in severe COVID-19 [9, 10] ) and antiviral activity in infected mice have not been described. Nevertheless, numerous clinical trials were initiated for out-and in-patients with COVID-19 to test ATV combined or not with other commercial drugs, such as ritonavir or dexamethasone [8, 11, 12] ; however, no definitive response has yet emerged from these clinical trials.

spread globally into the second pandemic of the 21st century, after the influ pandemic in 2009. SARS-CoV-2 continuously circulated, even among in preexisting immunity, and since COVID-19 began until the time of writi about 271 million confirmed cases along with 5.2 million deaths worldwid CoV-2 variants of concern (VoC) might escape the humoral immune respo infection or vaccination [4, 5] , which reinforces the necessity of specific ments. After almost a two-year effort in the repurposing of clinically ap limited benefits for COVID -19 patients have been demonstrated [5] . Thus, to improve the pre-clinical characterization of repurposed drugs to ratio clinical studies, in terms of posology and susceptibility of VoC, and learn fr actions with the viral target enzyme.

During the SARS-CoV-2 life cycle, viral surface glycoprotein spike int man angiotensin converting enzyme 2 (ACE2) and with the extracellular p membrane protease serine 2 (TMPRSS2) to enter the cells. Positive-sense SA nomic RNA is translated to polyproteins that are auto-cleaved by viral pap and major (M pro ) proteases. These matured non-structural proteins ( nsp12/8/7 (RNA polymerase complex), nsp13 (helicase), nsp14/10 (exonu ethyltransferase), nsp15 (endonuclease) and nsp16 (2'-O-methyltransferas sible for RNA replication and transcription of structural genes. These struct be translated, assembled with the newly synthetized genomic RNA and virus progeny. These events represent key steps during viral life cycle to p SARS-CoV-2 replication [6] .

We have previously described that atazanavir (ATV, Figure 1 ), a clinic human immunodeficiency virus (HIV) protease inhibitor [7] , is endowed w CoV-2 activity [8] . Although we described early in the pandemic outbreak tivity against SARS-CoV-2, by in silico and cell-based assays (in Vero-E6 lines), proper characterization of its enzymatic target, pharmacology on ty cytes (the most affect cell in severe COVID-19 [9, 10] ) and antiviral activity in have not been described. Nevertheless, numerous clinical trials were initiat in-patients with COVID-19 to test ATV combined or not with other com such as ritonavir or dexamethasone [8, 11, 12] ; however, no definitive res emerged from these clinical trials. Studies reconfirmed, by bioinformatics and basic enzymatic inhibitio ATV targets SARS-CoV-2 major protease (M pro ), an enzyme responsible fo of 11 sites of the viral polyprotein, a key-step in virus life cycle [13, 14] . Th by Pfizer of PAXLOVID TM (PF-07321332), combined with ritonavir, (Figu mentary Material) showed clinical efficacy in reduce hospitalization by 80% pro Figure 1 . The 2D-and 3D-chemical structure for atazanavir (ATV).

Studies reconfirmed, by bioinformatics and basic enzymatic inhibition curves, that ATV targets SARS-CoV-2 major protease (M pro ), an enzyme responsible for the cleavage of 11 sites of the viral polyprotein, a key-step in virus life cycle [13, 14] . The development by Pfizer of PAXLOVID TM (PF-07321332), combined with ritonavir, (Figure S1 , Supplementary Material) showed clinical efficacy in reduce hospitalization by 80%, reconfirming that M pro is a very important druggable target [15] . Thus, other drugs that target this enzyme deserve a more detailed characterization, to provide insight on pharmacophoric regions for next generation of anti-COVID-19 antivirals.

Anti-M pro from ATV has been disputed as controversial depending on assay conditions [16] . This apparent paradox reinforces that detailed mechanism of action and robust series of pre-clinical experiments should be conducted to shed light on the ATV inhibitory mechanism and the susceptibility of contemporaneous SARS-CoV-2 strains to this drug. Here, we address additional explanations to controversial effects of ATV on M pro , characterizing it as a competitive inhibitor of this viral enzyme that requires a catalytic water to be effective. ATV possesses anti-SARS-CoV-2 activity against B.1 and gamma strains on Calu-3 cells, a model of type II pneumocytes. ATV reached the plasma and lungs of treated Swiss-Webster mice and protected SARS-CoV-2-infected transgenic mice expression in human angiotensin converting enzyme 2 (K18-hACE2) from mortality. Moreover, ATV reduced virus-induced inflammation and cell death in bronchoalveolar lavage (BAL), and lung damage in infected and treated animals. This study compiles pre-clinical results that may allow further rationalization of clinical trials against COVID-19.

To advance on details on how ATV inhibits SARS-CoV-2 M pro , we initially performed a dose-dependent inhibition curve. ATV was slightly less potent than the positive control GC376 [17] (Figure 2A ). The Morrison's inhibitory constant (K i ) values for GC376 and ATV were 208 ± 0.15 and 703 ± 79 nM, respectively. Next, we tested ATV against various concentrations of M pro substrate. We observed that maximum velocity (V max ) values in the presence and absence of ATV were not different, while the Michaelis-Menten constant (K m ) values increased significantly in the presence of this drug ( Figure 2B ), indicating a competitive inhibition profile. Our results are specific to M pro because, differently from the positive control GRL0617, ATV did not inhibit SARS-CoV-2 papain-like protease (PL pro ) ( Figure 2C ). Since structural improvements on the M pro active site were determined more recently [18, 19] , SARS-CoV-2 molecular docking calculations were carried out to test if a catalytic water (H 2 O cat ) was required for ATV action. Since the highest docking score value was obtained in the presence of H 2 O cat , molecular docking calculations suggested a dependence of ATV potency of water content into the catalytic site of M pro - Figure 2D for M pro and ATV in the presence of H 2 O cat and Figure S2 in the Supplementary Material for the 2D-plot image corresponding to the main amino acid residues which interact with ATV in the absence and presence of H 2 O cat . The GOLD 2020.2 software ranked the ten best docking poses and at analyzing each pose, none of them showed a conformation for the interaction between the pyridyl group of ATV and the catalytic water.

The competitive inhibition implies that M pro 's substrate concentration may affect its susceptibility to ATV. Thus, we tested if ATV's potency is also affected in type II pneumocyte cell line (Calu-3) infected with different SARS-CoV-2 MOIs. In cell-based assays, the intracellular concentration of M pro 's substrate would be proportional to the virus input. Indeed, ATV's EC 50 value for the SARS-CoV-2 B.1 lineage varied in a MOIdependent way (Table 1) , similarly to remdesivir (RDV) ( Table 1) , a competitive inhibitor of the SARS-CoV-2 RNA polymerase complex under clinical use [20, 21] . The SARS-CoV-2 gamma VoC is similarly susceptible to ATV, compared to its predecessor strain B.1 (Table 1) . Since the CC 50 values of 312 ± 8 and 512 ± 30 µM for ATV and RDV, respectively, their selective index (SI) values were consistent with an adequate safety profile in vitro (Table 1) . 

We first aimed evaluate ATV's pharmacokinetics profile over time in the plasma and lungs of Swiss-Webster mice treated with 60 mg/kg of this drug, a dosage equivalent to its plasma exposure in humans under treatment against HIV. Upon treatment with 60 mg/kg, ATV concentration in the plasma was similar to the standard treatment of 300 mg in humans ( Figure 3A ) [22] . Interestingly, ATV seems to be concentrated in the lung of the treated animals ( Figure 3B ).

Next, we infected K18-hACE2-transgenic mice with SARS-CoV-2 gamma VoC and treated them with a daily oral dose of 60 mg/kg ATV, initiating 12 h after infection. Whereas the infection kills all animals within 6 days, a statistically significant increase in animal survival was observed in the infected and ATV-treated mice ( Figure 4A ). ATV 

We first aimed evaluate ATV's pharmacokinetics profile over time in the plasma and lungs of Swiss-Webster mice treated with 60 mg/kg of this drug, a dosage equivalent to its plasma exposure in humans under treatment against HIV. Upon treatment with 60 mg/kg, ATV concentration in the plasma was similar to the standard treatment of 300 mg in humans ( Figure 3A ) [22] . Interestingly, ATV seems to be concentrated in the lung of the treated animals ( Figure 3B ).

Next, we infected K18-hACE2-transgenic mice with SARS-CoV-2 gamma VoC and treated them with a daily oral dose of 60 mg/kg ATV, initiating 12 h after infection. Whereas the infection kills all animals within 6 days, a statistically significant increase in animal survival was observed in the infected and ATV-treated mice ( Figure 4A ). ATV protected the mice to continue to lose weight at the 6th day after infection ( Figure 4B ). In the bronchoalveolar lavage (BAL) of the treated animals, ATV significantly decreased SARS-CoV-2 RNA levels ( Figure 4C ), cell death-based on lactate dehydrogenase (LDH) activity ( Figure 4D ATV reduced the SARS-CoV-2-induced IL-6 levels in the BAL and in the lungs of treated animals ( Figure 5 ). In the lung, levels of TNF-α and KC were also reduced in the infected/untreated over untreated mice ( Figure 5 ). These results are in line with our previous description that ATV decreases the levels of SARS-CoV-2-induced pro-inflammatory cytokines in monocytes [8] and virus-triggered pyroptosis [23] . Moreover, SARS-CoV-2 infection provoked severe lung injury, leading to hemorrhage, and shrinking of the lobe, bronchiole and alveoli (Figure 6 ), which was reduced by ATV. This protection is the consequence of the direct antiviral and anti-inflammatory activity of ATV, since this molecule could not prevent hemorrhage as an anti-clotting agent ( Figure S3, Supplementary Material) . and mononuclear cells counts (E) and immunocytochemical staining (F) were assessed in the BAL six days after infection in the indicated experimental groups. All the analysis were performed with 6 mice/group.

ATV reduced the SARS-CoV-2-induced IL-6 levels in the BAL and in the lungs of treated animals ( Figure 5 ). In the lung, levels of TNF-α and KC were also reduced in the infected/untreated over untreated mice ( Figure 5 ). These results are in line with our previous description that ATV decreases the levels of SARS-CoV-2-induced pro-inflammatory cytokines in monocytes [8] and virus-triggered pyroptosis [23] . Moreover, SARS-CoV-2 infection provoked severe lung injury, leading to hemorrhage, and shrinking of the lobe, bronchiole and alveoli (Figure 6 ), which was reduced by ATV. This protection is the consequence of the direct antiviral and anti-inflammatory activity of ATV, since this molecule could not prevent hemorrhage as an anti-clotting agent ( Figure S3 

Repurposing of clinically approved drugs was considered an accelerated strategy to combat SARS-CoV-2 infections [24, 25] . However, limited clinical benefit has been documented for most repurposed drugs, whereas in the meantime orally available antiviral drugs against COVID-19 demonstrated clinical efficacy to reduce hospitalization, such as molnupiravir [26] and PAXLOVID TM [15] . Molnupiravir and PAXLOVID TM target the viral 

Repurposing of clinically approved drugs was considered an accelerated strategy to combat SARS-CoV-2 infections [24, 25] . However, limited clinical benefit has been documented for most repurposed drugs, whereas in the meantime orally available antiviral drugs against COVID-19 demonstrated clinical efficacy to reduce hospitalization, such as molnupiravir [26] and PAXLOVID TM [15] . Molnupiravir and PAXLOVID TM target the viral RNA synthesis and M pro , respectively. Besides detailed mechanism of action, these compounds fulfilled pre-clinical steps of investigation, from cell-based to animal models, to allow further clinical development to be conducted at target plasmatic concentration. Here, we performed pre-clinical study steps for ATV by enzymatic, in silico calculations and cell-based assays, combined with animal models ( Figure S4 in the Supplementary Material). The comprehension that ATV inhibits competitively M pro and requires a catalytic water to enhance ATV's interactions ATV onto the catalytic pocket and, more specifically, with residue His-41, which is one of the constituents of the catalytic dyad, may allow further development of anti-SARS-CoV-2 analogs.

We originally demonstrated that ATV inhibited SARS-CoV-2 M pro preparations and virus replication [8] . The controversial effects of ATV on SARS-CoV-2 M pro are documented in the literature [13, 15] . In line with studies that demonstrated SARS-CoV-2 susceptibility to ATV, we demonstrated here that M pro was inhibited by this drug, with K i 1.5-fold higher than GC376 and in a competitive fashion, increasing the enzyme's K m more than sixfold. It is unlikely that the other SARS-CoV-2 protease, PL pro , was targeted by ATV in these assays, because its K i for PL pro was above the threshold of in vitro inhibition.

The amino acid residue His-41 of M pro requires water to catalyze proteolytic cleavage [18, 19] and ATV targets this moiety. The presence of 20% glycerol in enzyme assay mixture reduces the water content and prevents ATV activity, that is why Ma and Wang [16, 27] could not identify this HIV protease inhibitor as a potential antiviral against COVID-19, whereas Li et al. [13] reached results similarly to ours. Our enzymatic assay includes bovine serum albumin (BSA) to avoid non-specific binding of small molecules to M pro . In our M pro FRET-based enzymatic assay, reactions were allowed to occur overnight to increase the sensitivity, instead of 1 h [16] . Although one might argue that a longer reaction time will increase the rate of false positive results, the following cell-based assays provide an additional layer of evidence that ATV activity against SARS-CoV-2 is credible.

Conversely, whereas ATV depends on water and multiple transient interactions, PF-07321332, which demonstrated clinical benefit [15] , was a covalent inhibitor of M pro [28] . Indeed, PF-07321332 directly target a key amino acid residue, the catalytic cysteine 145 [28] . Additionally, differently than in silico results obtained for ATV, the PF-07321332 is more buried into the protease pocket, probably due to the ligand steric area (781.51 and 493.69 Å 2 for ATV and PF-07321332, respectively) which result in a hydrogen bonding interaction between PF-07321332 and His-163 that is considered as a crucial residue for natural peptidic substrates [28, 29] .

Considering the substrate-dependent inhibitory effect of ATV, we hypothesized ATV's potency in cell-based assays would be influenced by varying the virus input, because higher numbers of virus particles should translate into increased quantities of M pro substrate. To test this hypothesis, Calu-3 cells, which is more closely resemble type II pneumocytes than A549 and Vero cells [9, 10] , were infected with a 100-fold different multiplicity of infections (MOI) of SARS-CoV-2, B.1 lineage, and treated with ATV or RDV, a competitive inhibitor of viral RNA synthesis. The MOI-dependent inhibition was consistent for ATV and RDV. The similarity in the cell-based potency of ATV to its K i for M pro further reinforces the conclusion that M pro is the target of ATV in vitro.

The emerging Brazilian SARS-CoV-2 gamma variant, initially detected in the state of Amazonas, was responsible for a public health calamity, spreading rapidly in Brazil and considered as one of variant of concern by WHO. The mutations found in gamma variant have been associated with increased transmissibility, higher viral load, propensity for immune evasion and SARS-CoV-2 reinfection [30, 31] . ATV presented adequate inhibitory At plasma exposures similar to humans [22] , ATV enhanced by 30% the survival of K18-hACE2-transgenic mice infected with SARS-CoV-2 gamma strain and decreased virusinduced cell death and inflammation. Despite there is not a statical difference in the body weight between the untreated and treated groups, the protective effect of ATV was viewed in relation to the significantly decrease of SARS-CoV-2 RNA levels in the bronchoalveolar lavage (BAL), as well as a significant reduction in the number of mononuclear leukocytes, polymorphonuclear leukocytes and LDH levels. In the BAL infected and treated animals displayed lower levels of IL-6, TNF-a, KC and PF4, compared to untreated mice. The in vivo anti-inflammatory of ATV observed here is in line with the ability of this molecule to inhibit pyroptosis in human primary monocytes for both BAL and lung samples clearly showing that ATV protect both cell death and inflammation [23, 32] .

Our study shows that ATV inhibits SARS-CoV-2 M pro with a mechanism of action different than PAXLOVID TM , showing ATV/M pro interface could give insights for further drug development. The in vitro and in vivo results reconfirm, in different magnitudes, SARS-CoV-2 susceptibility to ATV. Further ongoing clinical trials will determine if standard dose, used in HIV-treatment can prevent severe COVID-19 or if it is necessary higher doses [33] . According to ATV's monography [22] , doses three times higher could be used for shorter periods of time, when compared to the life-lasting HIV treatment.

Carboxymethyl cellulose (CMC), tris(4-(dimethylamino)phenyl)methylium chloride (crystal violet), formaldehyde, hematoxylin, eosin, phosphate buffer solution (PBS), ethanol, dimethyl sulfoxide (DMSO), ketamine, xylazine, trisodium citrate, calcium chloride (CaCl 2 ) and remdesivir (RDV) were purchased from Sigma-Aldrich/Merck (St. Louis, MO, USA). Atazanavir sulfate (ATV) was kindly donated from Instituto de Tecnologia de Fármacos, Farmanguinhos, Rio de Janeiro, Brazil.

African green monkey kidney (Vero, subtype e6) and human lung epithelial (Calu-3) cells were cultured in high-glucose Dulbecco's modified eagle medium (DMEM-HyClone, Logan, UT, USA) supplemented with 100 U/mL penicillin, 100 µg/mL streptomycin (P/S-Thermo Fisher Scientific ® , Waltham, MA, USA) and 10% fetal bovine serum (FBS-HyClone, Logan, UT, USA). The cells were incubated at 310 K in 5% of carbon dioxide (CO 2 ).

The SARS-CoV-2 B.1 lineage (GenBank #MT710714) and gamma variant (also known as P1 or B.1.1.28 lineage; #EPI_ISL_1060902) were isolated on Vero E6 cells from nasopharyngeal swabs of confirmed cases. All procedures related to virus culture were handled at biosafety level 3 (BSL3) multiuser facility at Fundação Oswaldo Cruz (FIOCRUZ), Rio de Janeiro, Brazil, according to world health organization (WHO) guidelines [34] .

The ATV capacity to inhibit enzymatic velocity of PL pro and M pro from SARS-CoV-2 was determined by the commercial kit provided by BPS Bioscience ® company (catalog number: #79995-1 and #79955-1, respectively) following the procedure and recommendations from literature and manufacturer [35] [36] [37] . Basically, 100 nM PL pro was incubated in 50 mM HEPES pH 7. KTSAVLQSGFRKME-EDANS with CAS number 730985-86-1) and ATV or GC376 (positive control), at concentrations ranging from 0 to 10 µM. Fluorescence signal was measured at an emission wavelength of 460 nm with excitation at 360 nm in a GloMax ® (Promega, Madison, WI, USA) plate reader. Morrison's inhibitory constant (K i ) was calculated by non-linear regression using GraphPad Prism 9. The Michaelis-Menten plot was conducted for 88.8 nM M pro incubated overnight in assay buffer with substrate concentrations varying from 0 to 100 µM in the presence and absence of 2.5 µM of ATV. After fluorescence quantification, the Michaelis-Menten constant (K m ) and maximum velocity (V max ) were calculated by non-linear regression using GraphPad Prism 9. The value was presented as mean ± standard deviation (SD).

The crystallographic structure of M pro was obtained from Protein Data Bank (PDB), with access code 7K40 [38] . The chemical structure for ATV was built and minimized in terms of energy by Density Functional Theory (DFT) via Spartan'18 software (Wavefunction, Inc., Irvine, CA, USA) [39] . The molecular docking calculations were performed with GOLD 2020.2 software (Cambridge Crystallographic Data Center Software Ltd., CCDC, Cambridge CB2 1EZ, UK) [40] . Hydrogen atoms were added to the protease following tautomeric states and ionization data, which are inferred by the GOLD 2020.2 software at pH 7.4. The number of genetic operations (crossing, migration, mutation) during the search procedure was set as 100,000. Redocking studies were carried out with the crystallographic ligand boceprevir (PDB code: 7K40), obtaining the lowest root mean square deviation (RMSD) value by ChemPLP function. It was defined 8 Å radius around the active binding site and the figures were generated with PyMOL Delano Scientific LLC software (DeLano Scientific LLC: San Carlos, CA, USA) [41] , while the 2D-plot was obtained by ProteinsPlus platform (Zentrum für Bioinformatik, Universität Hamburg, Germany) [42] . After 48 h, the supernatants were harvested, and infectious virus titers were quantified by plaque forming assays according to previous publications [8, 43, 44] .

To perform the virus titration, Vero cells (2.0 × 10 4 cell/well) in 96-well plates (Nalge Nunc Int, Rochester, NY, USA) were infected with log-based dilutions of the yield reduction assays' supernatants for 1 h at 37 • C in 5% of CO 2 . After the incubation, medium containing 1.8% CMC with 5% FBS was added and incubated at 37 • C with 5% CO 2 for 72 h. The cells were fixed with 10% formaldehyde in PBS and stained with a 0.04% solution of crystal violet in 70% methanol. The virus titers were calculated by scoring the plaque-forming unit-(PFU/mL) and a non-linear regression analysis of the dose-response curves was also performed to calculate the 50% effective concentration (EC 50 ). All experiments were carried out at least three independent times, including a minimum of two technical replicates in each assay, and each data was analyzed from Prism GraphPad software 8.0 (Windows GraphPad Software, San Diego, CA, USA). The value was presented as mean ± standard deviation (SD).

Vero cells (2.0 × 10 4 cell/well) were treated for 3 days with different concentrations of ATV or RDV (ranging from 1 to 600 µM) as previously described by us [8, 44] . The 50% cytotoxic concentration (CC 50 ) was calculated by a non-linear regression analysis from a dose-response curve. All experiments were carried out at least three independent times and each data was analyzed from Prism GraphPad software 8.0 (Windows GraphPad Software, San Diego, CA, USA). The results were presented as mean ± standard deviation (SD). The selectivity indexes (SI) for ATV and RDV were calculated through the ratio between CC 50 and EC 50 values.

ATV's concentration in the plasma and lungs of adult Swiss-Webster mice (8-15 weeks) was evaluated over time. Animals were treated with an oral dose of 60 mg/kg of ATV for 12-time intervals-00:05, 00:10, 00:20, 00:40, 01:00, 02:00, 03:00, 04:00, 06:00, 08:00, 10:00 and 12:00 h. Each treatment group had five animals. After these periods of times, total blood and lungs were collected. Plasma was obtained by blood centrifugation at 8000× g for 15 min. 

Experiments with transgenic mice expressing human ACE-2 receptor (K18-hACE2mice), were performed in Animal Biosafety Level 3 (ABSL-3) multiuser facility, according to the animal welfare guidelines of the Ethics Committee of Animal Experimentation (CEUA-INCa, License 005/2021) and WHO guidelines [34] . The animals were obtained from the Oswaldo Cruz Foundation breeding colony and maintained with free access to food and water at 29-30 • C under a controlled 12 h light/dark cycle. Experiments were performed during the light phase of the cycle.

For infection procedures, mice were anaesthetized with 60 mg/kg of ketamine and 4 mg/kg of xylazine and inoculated intranasally with DMEM high glucose (MOCK), or 10 5 PFU of SARS-CoV-2 gamma strain in 10 µL of DMEM high glucose. It was used six mice per experimental group: MOCK (non-infected); SARS-CoV-2-infected without treatment (NIL) and SARS-CoV-2-infected and treated with ATV. The animals were treated with a daily dose of 60 mg/kg of ATV for seven days.

The animals were monitored daily during seven days for survival and body-weight analysis. In the case of weight loss higher than 20% euthanasia was performed to alleviate animal suffering. In the last day, the bronchoalveolar lavage (BAL) from both lungs was harvested by washing the lungs once with 1 mL of cold PBS. After centrifugation of BAL (1500 rpm for 5 min), the pellet was used for total and differential leukocytes counts (diluted in Turk's 2% acetic acid fluid) using a Neubauer chamber. The lactate dehydrogenase (LDH) quantification was performed with centrifuged BAL supernatant to evaluated cell death (CytoTox96, Promega). Differential cell counts were performed by cytospin (Cytospin3; centrifugation of 350× g for 5 min at room temperature) and stained by the May-Grünwald-Giemsa method.

After BAL harvesting, lungs were perfused with 20 mL of saline solution to remove the circulating blood. Lungs were then collected, pottered and homogenized in 500 µL of a phosphatase and protease inhibitor cocktail Complete, mini EDTA-free Roche Applied Science (Mannheim, Germany) for 30 s, using an Ultra-Turrax Disperser T-10 basic IKA (Guangzhou, China).

The viral RNA from samples collected in the in vivo assays was quantified through quantitative reverse transcription polymerase chain reaction (RT-PCR). Total RNA was extracted using QIAamp Viral RNA (Qiagen, Germantown, MD, USA), according to manufacturer's instructions. Quantitative RT-PCR was performed using Quanti Tect Probe RT-PCR Kit (Qiagen) in a StepOne Plus™ Real-Time PCR System (Thermo Fisher Scientific).

Amplifications were carried out in 15 µL reaction mixtures containing 2× reaction mix buffer, 50 µM of each primer, 10 µM of the probe and 5 µL of RNA template. Primers, probes and cycling conditions recommended by the Centers for Disease Control and Prevention (CDC) protocol were used to detect the SARS-CoV-2 (CDC 2020). Amplification of the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a reference for the number of cells. The cycle threshold (CT) values for this target were compared to those obtained with different cell quantities (10 7 to 10 2 ), for calibration.

The levels of IL-6, TNF-α, KC and PF4 were quantified in BAL samples from uninfected (MOCK), infected without treatment (NIL) and infected and treated animals by ELISA, using specific kits and following the manufacturer's instructions (R&D Systems). Cell death was determined according to the activity of LDH in the BAL as previously described in the Section 4.8.

Histological features related to the injury caused by SARS-CoV-2 infection were analyzed in the lungs of K18-hACE2 mice. Inflammatory and vascular infiltrates and evidence of cell degeneration was evaluated to characterize the level of the tissue damage. The collected material was fixed with formaldehyde (4%), dehydrated and embedded in paraffin to the obtention of tissue slices through the use of a microtome. The slices were fixed and stained with hematoxylin and eosin for microphotographs analysis.

Human blood samples were collected from healthy donors in 3.8% trisodium citrate (9:1, v/v), and platelet-poor plasma was obtained by centrifugation at 3000 rpm for 10 min. Plasma (100 µL) was incubated with 1 µL of ATV at various concentrations (diluted in DMSO) for 2 min at 310 K. Plasma clotting was initiated by the addition of 100 µL of 25 mM CaCl 2 , and the time for clot formation was recorded on a KC-4 Delta coagulometer (Tcoag, Wicklow, Ireland). Time for clot formation was recorded in triplicates.

The following are available online at https://www.mdpi.com/article/ 10.3390/ph15010021/s1, Figure S1 : Chemical structure for (A) PF-07321332 and (B) ritonavir, the active principles of PAXLOVID TM from Pfizer. Figure S2 . (A) Superposition of the best docking pose for the interaction between M pro and ATV in the presence and without the catalytic water H 2 O cat (ATV in pink and brown, respectively). The 2D-representation for the interaction among the amino acid residues from the catalytic pocket of M pro with (B) ATV in the presence of H 2 O cat and (C) ATV without H 2 O cat . For better interpretation the M pro structure was represented only in the monomeric form with the domains I, II and III in light red, orange and gray, respectively. The catalytic dyad His-41 and Cys-145 are represented as sticks in cyan, while the amino acid residues which interact hydrophobically with ATV are in green in the 2D-plot image. Figure S3 

COVID-19: A brief history and treatments in development

COVID-19: The first documented coronavirus pandemic in history

WHO Coronavirus (COVID-19) Dashboard. Available online

Progress of the COVID-19 vaccine effort: Viruses, vaccines and variants versus efficacy, effectiveness and escape

Repurposing drugs for treatment of COVID-19

Coronavirus biology and replication: Implications for SARS-CoV-2

Atazanavir: Its role in HIV treatment

Atazanavir, alone or in combination with ritonavir, inhibits SARS-CoV-2 replication and proinflammatory cytokine production

Comparative replication and immune activation profiles of SARS-CoV-2 and SARS-CoV in human lungs: An ex vivo study with implications for the pathogenesis of COVID-19

Comparative tropism, replication kinetics, and cell damage profiling of SARS-CoV-2 and SARS-CoV with implications for clinical manifestations, transmissibility, and laboratory studies of COVID-19: An observational study

The Nitazoxanide Plus Atazanavir for COVID-19 Study (NACOVID)

Atazanavir and Dexamethasone Combination Therapy for the Treatment of COVID-19 Infection (NATADEX)

Identify potent SARS-CoV-2 main protease inhibitors via accelerated free energy perturbation-based virtual careening of existing drugs

Structure of Mpro from SARS-CoV-2 and discovery of its inhibitors

Pfizer's Novel COVID-19 Oral Antiviral Treatment Candidate Reduced Risk of Hospitalization or Death by 89% in Interim Analysis of Phase 2/3 Epic-HR Study

Dipyridamole, chloroquine, montelukast sodium, candesartan, oxytetracycline, and atazanavir are not SARS-CoV-2 main protease inhibitors

Repurposing the HCV NS3-4A protease drug boceprevir as COVID-19 therapeutics

Structural plasticity of SARS-CoV-2 3CL Mpro active site cavity revealed by room temperature X-ray crystallography

Crystallographic structure of wild-type SARS-CoV2 main protease acyl-enzyme intermediate with physiological C-terminal autoprocessing site

Mechanism of SARS-CoV-2 polymerase stalling by remdesivir

Remdesivir in adults with severe COVID-19: A randomised, double-blind, placebo-controlled, multicentre trial

Product Monograph Atazanavir Capsules

SARS-CoV-2 engages inflammasome and pyroptosis in human primary monocytes

An integrated drug repurposing strategy for the rapid identification of potential SARS-CoV-2 viral inhibitors

Repurposing approved drugs for guiding COVID-19 prophylaxis: A systematic review

Merck and Ridgeback's Investigational Oral Antiviral Molnupiravir Reduced the Risk of Hospitalization or Death by Approximately 50 Percent Compared to Placebo for Patients with Mild or Moderate COVID-19 in Positive Interim Analysis of Phase 3

GC-376, and calpain inhibitors II, XII inhibit SARS-CoV-2 viral replication by targeting the viral main protease

Crystal structure of SARS-CoV-2 main protease in complex with protease inhibitor PF-07321332

Supervised molecular dynamics (SuMD) insights into the mechanism of action of SARS-CoV-2 main protease inhibitor PF-07321332

SARS-CoV-2 variant of concern P.1 (Gamma) infection in young and middle-aged patients admitted to the intensive care units of a single hospital in Salvador

Genomics and epidemiology of a novel SARS-CoV-2 lineage in Manaus

Comparing the effectiveness of atazanavir/ritonavir/dolutegravir/hydroxychloroquine and lopinavir/ritonavir/hydroxychloroquine treatment regimens in COVID-19 patients

Available online

Practical guidelines for collection, manipulation and inactivation of SARS-CoV-2 and COVID-19 clinical specimens

Identifcation of ebselen and its analogues as potent covalente inhibitors of papain-like protease from SARS-CoV-2

Learning from the past: Possible urgent prevention and treatment options for severe acute respiratory infections caused by 2019-nCoV

Crystal structure of SARS-CoV-2 main protease provides a basis for design of improved α-ketoamide inhibitors

Structural similarity of SARS-CoV2 Mpro and HCV NS3/4A proteases suggests new approaches for identifying existing drugs useful as COVID-19 therapeutics

GOLD: Protein-Ligand Docking Software. Available online

Repurposing the ebola and marburg virus inhibitors tilorone, quinacrine, and pyronaridine: In vitro activity against SARS-CoV-2 and potential mechanisms

In vitro antiviral activity of the anti-HCV drugs daclatasvir and sofosbuvir against SARS-CoV-2, the etiological agent of COVID-19

The authors declare no conflict of interest and the funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.