Viral target- and metabolism-based rationale for combined use of recently authorized small molecule COVID-19 medicines: molnupiravir, nirmatrelvir and remdesivir

Abstract

The COVID-19 pandemic remains a major health concern worldwide, and SARS-CoV-2 is continuously evolving. There is an urgent need to identify new antiviral drugs and develop novel therapeutic strategies. Combined use of newly authorized COVID-19 medicines including molnupiravir, nirmatrelvir and remdesivir, has been actively pursued. Mechanistically, nirmatrelvir inhibits SARS-CoV-2 replication by targeting the viral main protease (M^pro), a critical enzyme in the processing of the immediately translated coronavirus polyproteins for viral replication. Molnupiravir and remdesivir, on the other hand, inhibit SARS-CoV-2 replication by targeting RNA-dependent RNA-polymerase (RdRp), which is directly responsible for genome replication and production of subgenomic RNAs. Molnupiravir targets RdRp and induces severe viral RNA mutations (genome), commonly referred to as error catastrophe. Remdesivir, in contrast, targets RdRp and causes chain termination and arrests RNA synthesis of the viral genome. In addition, all three medicines undergo extensive metabolism with strong therapeutic significance. Molnupiravir is hydrolytically activated by carboxylesterase-2 (CES2); nirmatrelvir is inactivated by cytochrome P450-based oxidation (e.g., CYP3A4); and remdesivir is hydrolytically activated by CES1 but covalently inhibits CES2. Additionally, remdesivir and nirmatrelvir are oxidized by the same CYP enzymes. The distinct mechanisms of action provide strong rationale for their combined use. On the other hand, these drugs undergo extensive metabolism that determines their therapeutic potential. This review discusses how metabolism pathways and enzymes involved should be carefully considered during their combined use for therapeutic synergy.

1. INTRODUCTION

Pandemics cause global disruptions to public health, financial system and even social stability tremendously. Throughout the human history, quite a few pandemics have been well documented [1–6] such as the smallpox pandemic of 1870–1874 and the ongoing HIV/AIDS pandemic starting in 1981 [7, 8]. It is estimated that the smallpox pandemic has killed between 300–500 million people [9–11], representing approximately 20% of the then world population. The ongoing HIV/AIDS pandemic is the longest in record and has lasted for over four decades with approximately 40 million people killed [7, 12, 13]. The current SARS-CoV-2 pandemic of 2019 (severe acute respiratory syndrome-associated coronavirus-2), commonly referred to as COVID-19, was reported in December of 2019 [1]. As of today, the confirmed COVID-19 cases have reached the number of over 755 million with a total mortality of over 6.8 million worldwide [14].

COVID-19 symptoms manifest in a phase-specific manner: mild, pulmonary and hyperinflammatory [15–17]. These phases are interconnected in terms of viral replication and clinical manifestations. The mild phase is characterized by general malaise but lacks apparent respiratory changes [16]. The managing priority for patients in this phase is focused on prevention of spread (i.e., isolation), symptomatic management (e.g., fever reduction), monitoring of disease progression and antiviral therapy if needed. The pulmonary phase is characterized by sustained viral replication and pneumonia associated symptoms such as respiratory dysfunction (e.g., dyspnea) [18, 19]. Management of the pulmonary phase includes the use of antiviral therapy, anti-inflammatory agents, anticoagulants, or in combination [19]. The hyperinflammatory phase, directly associated with hospitalization and mortality, is manifested by acute respiratory distress syndrome and multi-organ failure [20–22]. Calming the cytokine storm with supplemental oxygen is a priority [23].

Immuno-prevention and -therapy are effective approaches to prevent and treat COVID-19 infection and progression. So far, all of the approved vaccines target the Spike protein, a surface protein of SARS-CoV-2 [24–26]. This viral protein initiates the cellular infection by interacting with the host receptor ACE2 (angiotensin-converting enzyme 2) and is an excellent target for vaccine development and immunotherapy [27–31]. Likewise, many immunotherapeutic COVID-19 agents target the Spike protein [29–31]. Etesevimab is a fully human monoclonal neutralizing antibody, binds specifically to SARS-CoV-2 and prevents its infection [32]. This monoclonal antibody is commonly used together with bamlanivimab, another monoclonal antibody [33]. The combination was authorized in December of 2021 for the treatment of mild and moderate COVID-19 in adults and children 12 or older. The Spike protein, on the other hand, is extensively glycosylated, undergoes conformational changes and exhibits great potential for rapid mutations [34–36]. These characteristics are linked to waning immunity against COVID-19, as shown by breakthrough infection and multiple surges [37–39].

The success in developing COVID-19 vaccines and antibody-based therapy, in such a short period of time, represents a significant stride in the scientific community and for the public health [40, 41]. The efforts of developing COVID-19 therapeutics, like those for vaccine development, are unprecedented as well [42, 43]. In a short span of two years, more than dozens of therapeutic targets have been identified [42–47], and hundreds of clinical trials have been completed or are ongoing [48]. These targets represent a comprehensive list from viral targets such as RNA-dependent RNA-polymerase (RdRp) to host proteins such as TMPRSS2 (transmembrane protease, serine 2). TMPRSS2 is a facilitator of SARS-CoV-2 infection [42, 43]. Many existing antiviral drugs have been shown to exert inhibitory activities against SARS-CoV-2 [49]. Lopinavir, a popular antiviral agent against HIV, has been reported to inhibit SARS-CoV-2 at an IC₅₀ of 12 μM [50]. Favipiravir, an anti-influenza viral agent, inhibits SARS-CoV-2 at an IC₅₀ of 9.6 μM [51]. The IC₅₀ values (half-maximal inhibitory concentration) for both drugs are within their peak plasma concentrations [51]. On the other hand, the Food and Drug Administration (FDA) has authorized the use of several small molecule drugs specific to COVID-19 therapy including: molnupiravir, nirmatrelvir and remdesivir.

2. THERAPEUTIC CHARACTERISTICS OF MOLNUPIRAVIR, NIRMATRELVIR AND REMDESIVIR

Remdesivir received emergency use authorization from the FDA in May of 2020, and the full approval in October of the same year [52]. Molnupiravir and nirmatrelvir both received emergency use authorization a year later [53]. Nirmatrelvir is functionally derived from and developed based on lufotrelvir, a phosphate prodrug of a hydroxyketone that was designed to target SARS-CoV-2 [54]. However, nirmatrelvir but not lufotrelvir is orally active [55]. In contrast, both molnupiravir and remdesivir were developed originally to treat other viral infections. Remdesivir was intended initially for hepatitis C and later for Ebola [56]. Molnupiravir, along with its parent compound EIDD-1931, was intended initially to treat Venezuelan equine encephalitis virus and later human influenza virus [57]. While all three drugs have been approved for the treatment of COVID-19, they differ markedly in various aspects including viral targets, metabolism and pharmacokinetic determinations.

2–1. Mechanism of action

These small molecule COVID-19 drugs were approved in a timeframe of less than two years. This is unprecedented as new medicines typically take as many as ten years or even longer to get final approvals [58]. Excitingly these COVID-19 drugs differ in mechanism of action, suggesting combined use for additive or even synergistic activity. Mechanistically, nirmatrelvir inhibits SARS-CoV-2 replication by targeting the viral main protease (M^pro) [59]. This protease hydrolytically processes the immediately translated coronavirus polyproteins for assembling the viral replication complex [59]. The inhibition of nirmatrelvir toward M^pro is achieved by covalently interacting with the thiol group of catalytic cysteine-145 (Cys145) to form a thioimidate adduct [59, 60]. Crystallographic study has suggested that the covalent link for the thioimidate adduct is reversible [59]. In support of this possibility, dilution of M^pro - nirmatrelvir complexes by 100-fold led to a recovery of enzymatic activity [60]. Covalent-reversible inhibitors, compared with their covalent-irreversible counterparts, have several advantages, notably avoidance of producing new antigenic epitopes.

In contrast to nirmatrelvir, molnupiravir and remdesivir inhibit SARS-CoV-2 replication by targeting viral RdRp (Table 1) [61, 62]. However, molnupiravir and remdesivir lead to different outcomes. Remdesivir causes RdRp to pause or induces chain termination, whereas molnupiravir causes RdRp to introduce widespread errors of the viral genome, leading to lethal mutagenesis (Table 1). The incorporation efficiency of natural nucleotides over that of NHC-TP (tri-phosphorylated N-hydroxycytidine), the therapeutic metabolite of molnupiravir, into model RNA substrates follows the order GTP (12,841) > ATP (424) > UTP (171) > CTP (30), indicating that NHC-TP competes predominantly with CTP for incorporation (Table 1) [63]. Importantly, 3CL^pro and RdRp, compared with the Spike protein, are more conserved [64–67]. For example, the Omicron isolates have only two missense mutations across the replicase-transcriptase complex [64]. RdRp, the core protein of the complex, has only a single mutation (i.e., P323L) [64]. This mutation is not in the RNA binding site and may not cause significant changes catalytically [64]. Likewise, M^pro exhibits high sequence and structural conservation [65], although some mutations have been identified [66].

Table 1.

Pharmacological characteristics

Item	Molnupiravir	Nirmatrelvir	Remdesivir
Target	RdRp	Mpro	RdRp
Mode of action	Reversible	Covalent	Reversible
Consequence	Errors introduced	Assembling↓↓	Pausing/termination
Ki or IC50	GTP/CTP (428)a	0.933–3.110 nMb	0.032 μMc

^a[133];

^b[60, 140];

^c[63].

2–2. Clinical characteristics

In addition to mechanism of action, these COVID-19 drugs differ in several major clinical characteristics (Table 2). Molnupiravir is authorized for use in patients at 18 years of age or older; nirmatrelvir in patients at 12 years of age or older; and remdesivir in patients as young as 28 days old [68]. Both molnupiravir and nirmatrelvir are administered orally [69], whereas remdesivir is given through intravenous infusion [70]. Molnupiravir and nirmatrelvir are given twice a day but molnupiravir must be given 12 h apart. Molnupiravir and remdesivir are packaged singly, whereas nirmatrelvir is packaged along with the boosting agent ritonavir. The co-package is commonly referred as Paxlovid [71, 72]. With an exception of remdesivir, neither molnupiravir nor nirmatrelvir is authorized for use longer than 5 consecutive days or for initiation of treatment in patients requiring hospitalization due to severe COVID-19 [71, 72].

Table 2.

Efficacy and safety of molnupiravir, nirmatrelvir and remdesivir

Characteristics	Molnupiravir	Nirmatrelvir	Remdesivir
Combined recipe	No	Yes	No
Dosing regimens	800 mg/twice/12 h*	300 mg/twice/day*	100 mg/day*
Administrative route	Oral	Oral	Intravenous
Adults; children ages 12 years and older		Yes
Adults and children (28 days of age)			Yes
Adults	Yes
Over placebo in hospitalization/death	6.77 over 9.72%a	0.77 over 6.41%b
Days to recovery (Treatment/control)			0.73/1.10c
Over placebo in serious adverse events	7.18 over 11.41%a	6.68 over 16.14%b	0.69–0.78c

^*Fact sheet for healthcare providers: Emergency use of authorization for lagevrio^™ (molnupiravir) capsules. Revised EUA Authorized Date: 08/2022; Fact sheet for healthcare providers: Emergency use of authorization for Paxlovid^™. Original EUA Authorized Date: 12/2021. Veklury FDA Approval History (https://www.drugs.com/history/veklury.html).

^a[73];

^b[74];

^c[36, 88].

Clinical trials have shown the superiority of all three medicines over placebo controls. As summarized in Table 2, the risk of hospitalization for all-cause mortality is 6.77% in the molnupiravir group, whereas the risk increases to 9.72% in the corresponding placebo group [73]. The relative risk of nirmatrelvir over placebo groups is more drastic: 0.77 versus 6.41% [74]. Similar trends are observed in regard to serious adverse events with a ratio of 0.63 (7.18/11.41%) for molnupiravir (treatment over placebo), 0.41 (6.68/16.14%) for nirmatrelvir and 0.69–0.78 for remdesivir, respectively (Table 2) [73, 75–77]. The serious adverse events vary depending on a medication. Compared to the placebo, the molnupiravir group has a higher incidence of bacterial pneumonia, nausea and dizziness [73], and the nirmatrelvir group has a higher incidence of dysgeusia, diarrhea, elevated alanine aminotransferase, headache, creatinine clearance, nausea and vomiting [74]. The remdesivir groups from multiple clinical trials have a wide range of serious adverse events from cardiovascular events, to pulmonary disorders, and to hepatic dysfunction [36, 75–79].

2–3. Broad spectrum of activities against SARS-CoV-2 variants

Molnupiravir, nirmatrelvir and remdesivir target M^pro or RdRp, which are sequence- and/or even structurally conserved compared to the Spike protein [65]. It is therefore anticipated that these medicines are effective against SARS-CoV-2 variants. Since the start of the pandemic, the SARS-CoV-2 virus has mutated over time, producing variants (e.g., Delta) and even subvariants (e.g., Omicron B1 and B2) [80]. Currently, Omicron subvariants are the dominant circulating SARS-CoV-2 [81, 82]. As summarized in Table 3, all three drugs remain effective against variants/subvariants based on both cytopathogenic reduction and viral replication assays [83–86]. Newly emerged variants are generally more sensitive than earlier variants to these drugs. For example, Omicron B1 is almost twice as sensitive as alpha variant based on the cytopathogenic reduction with the former having an EC₅₀ value of 1.9 (half maximal effective concentration), whereas the latter is 3.6 μM. A similar ratio of sensitivity between these two variants occurs with nirmatrelvir and remdesivir (Table 3). It should be noted that these COVID-19 drugs, molnupiravir and remdesivir in particular, have been shown to exert inhibitory effects towards other viruses. For example, molnupiravir has been shown to potently inhibit the replication of influenza virus, a member of the family Orthomyxoviridae [87]. Likewise, members of the family Filoviridae are highly sensitive towards remdesivir with an EC₅₀ value of as low as 3 nM, 10 or more times of the sensitivity of SARS-CoV-2, a member of the family Coronaviridae [36, 86].

Table 3.

Activities toward variants based on cytopathogenic reduction (^aEC₅₀) or viral replication (^bIC50)

Drug/metabolite	Alpha	Beta	Gamma	Delta	Omicron B1	Omicron B2	GHB
Molnupiravir	3.6a				1.9a		3.9a
NHC	2.3a, 0.34b	1.5a	2.0a, 0.24b	3.3a, 0.29b	0.34b	0.46b	2.2a
Nirmatrelvir	0.28a, 1.78 b	0.14a	0.28a, 1.65b	0.21a, 1.80b	0.14a	1.9b	0.11a
Remdesivir	0.077a	0.063a	0.074a		0.048a		0.05a
GS-441524	0.76a	0.77a	0.90a	0.87a	0.50a		0.81a

NHC: N-hydroxycytidine, the hydrolytic metabolite of molnupiravir; GS-441524: major metabolite of remdesivir in the blood.

^a[86];

^b[135].

3. METABOLISM AND ACTIVE TRANSPORT

Molnupiravir, nirmatrelvir and remdesivir, undergo extensive metabolism [36, 89–96], primarily through hydrolysis, oxidation or both (Table 4). All three drugs undergo hydrolysis, whereas nirmatrelvir and remdesivir also undergo oxidation. Hydrolysis of molnupiravir and remdesivir represents therapeutic activation, whereas both oxidation and hydrolysis of nirmatrelvir represents inactivation. On the other hand, metabolism is inherently linked to active transporters across cell membranes [97, 98]. Molnupiravir and remdesivir’s hydrolytic metabolites are negatively charged, thus requiring drug transporters for membrane translocation.

Table 4.

Metabolism pathways

Pathway	Molnupiravir	Nirmatrelvir	Remdesivir
Hydrolysis	Yes	Yes*	Yes
Oxidation		Yes	Yes
Significance (hydrolysis)	Activation		Activation
Significance (oxidation)		Inactivation	TBD

^*Via bacteria in the gastrointestinal tract [96]. TBD: to be determined.

3–1. Metabolism of molnupiravir

Molnupiravir is an isopropylester and structurally belongs to a superfamily of nucleoside analogs (Fig. 1). Members of this superfamily are widely used for anticancer and antiviral therapies [99–101]. Like many nucleoside drugs, molnupiravir is initially hydrolyzed to NHC and subsequently undergoes phosphorylation to form triphosphate. It is the triphosphate metabolite that targets RdRp and exerts potent antiviral activity [102]. Carboxylesterases constitute a large class of serine hydrolases with high catalytic efficiency. In humans, CES1 and CES2 are two carboxylesterases with known pharmacological and toxicological significance [103–106]. These two carboxylesterases, on the other hand, differ in their substrate specificity and tissue distribution [107–109]. CES1 preferably hydrolyzes esters with an acyl moiety relatively larger than its alkoxy moiety [93] and opposite is true with CES2 [104, 105].

Fig 1. Hydrolysis of molnupiravir by CES2.

Molnupiravir is a substrate of CES2 and the hydrolysis leads to the formation of N- hydroxycytidine. Boxed is the acyl moiety.

The isopropylester of molnupiravir has an alkoxy moiety relatively larger than its acyl moiety (boxed in Fig. 1). It is assumed to be hydrolyzed by CES2. Consistent with the assumption, hydrolysis of molnupiravir is correlated significantly with the expression of CES2 but not CES1, and the hydrolytic activity toward molnupiravir shows similarly to the tissue distribution of CES2 [95]. For example, kidney and intestine have abundant expression of CES2 and highly hydrolyze molnupiravir. The involvement of CES2 but not CES1 in the hydrolysis of molnupiravir is confirmed by recombinant enzymes [95]. Finally, several CES2 genetic polymorphic variants exhibit significantly altered activity toward molnupiravir. For example, the variants A178V and F485V have significantly decreased hydrolysis but the opposite is true with the variant R180H [95].

NHC, the hydrolytic metabolite of molnupiravir, undergoes three-step phosphorylation to produce monophosphate, diphosphate, and ultimately triphosphate, respectively. Potent antiviral activity is mainly driven by the triphosphate metabolite [36, 102]. The identities of kinases for the phosphorylation remain to be determined. Several kinases, such as adenylate kinase 2, pyruvate kinases and thymidine kinase 1, have nevertheless been implicated in such catalytic actions toward nucleoside analog drugs such as tenofovir [110]. These kinases have a broad tissue distribution, but a majority of them are expressed in cell- and organ-specific manner. The metabolic fate of these phosphorylated metabolites remains largely unknown. A recent study reports that NUDT18, a nucleoside diphosphate linked to moiety-X (NUDIX) hydrolase, has been shown to hydrolyze NHC triphosphate [111]. NUDIX hydrolases constitute a superfamily of hydrolases that known to play vital roles in energy metabolism and nutrition homeostasis [112].

3–2. Metabolism of nirmatrelvir

Nirmatrelvir is a fluorine-containing compound with a core-structure of azabicyclo hexane (Fig. 2). It has an oxopyrrolidin group, a cyano moiety and several amide bonds (Fig. 2). In vitro incubation of nirmatrelvir with liver microsomes, in the presence of NADPH, leads to degradation but the magnitude of the degradation is species-dependent [71]. Rat microsomes from both male and female produce a single major metabolite (M4). The peak corresponding to this metabolite is split into two, presumably eluted from interconverting diastereomers. In contrast, microsomes from human or monkey liver produce additional metabolites. In addition to M4, incubation with monkey microsomes results in two more metabolite peaks that have a retention time of 7.1 and 7.2 min, respectively [71]. Even with M4, the corresponding peak produced with monkey microsomes is wider and split into four. Finally, the peak area of the parent drug, upon incubation with monkey microsomes, is much smaller than that with rat microsomes, suggesting greater intrinsic clearance of nirmatrelvir in monkey liver.

Fig 2. Hydroxylation of nirmatrelvir.

The hydroxylation is catalyzed by several CYP enzymes, but primarily by CYP3A4. The hydroxylation (in Red) occurs at different atoms, resulting in the formation of M1, M2, M3 and M4, respectively. The red arrows indicate amide bonds for hydrolysis.

Human microsomes produce the most number of metabolites among these species, designated as M1, M2, M3 and M4, respectively. Metabolite M4 is shared by the incubation with both rat and monkey microsomes. M2 has a retention time of 7.2 min and is also present in the incubation with monkey microsomes [71]. M1, on the other hand, has a shortest retention time among all metabolites generated with microsomes from all three species. Human liver microsomes produce an additional metabolite M3, but this unique metabolite is minimal based on the peak area. Overall, the relative abundance among these metabolites is in the order of: M4>M2>M1>M3. Critically, the metabolite profile generated by human liver microsomes is almost identical to that generated by CYP3A4 [71]. Indeed, H-NMR spectra establish that all of these metabolites are produced by monohydroxylation in various positions such as the dimethyl 1–3 azabicyclo [3.1.0] hexane fused ring (Right of Fig. 2).

The dominant involvement of CYP3A4 in the metabolism of nirmatrelvir serves the foundation for this antiviral agent to be co-packaged with ritonavir (as Paxlovid), a strong CYP3A4 inhibitor and pharmacokinetic boosting agent [74]. Indeed, none of the metabolites generated by human liver microsomes are detected in participants who receive the co-package formulation, and the parent drug is the only drug entity in the plasma [96]. On the other hand, there are several notable metabolites present in urine and/or feces, designated as M5, M7, M8 and M9, respectively. M9 is present in the feces only, whereas M7 in the urine only. M5 and M8 are present in both urine and fecal samples [96]. M8 is generated through hydrolysis of the amide bond indicated by an arrow in red (Left of Fig. 2). M5 is also a hydrolytic metabolite through the cleavage of another amide bond as indicated by a red arrow. M5 is the precursor of M7, a glucuronide.

3–3. Metabolism of remdesivir

Remdesivir, like molnupiravir, is a carboxylic ester and structurally belongs to the superfamily of nucleoside analogs [99, 101]. Unlike molnupiravir, remdesivir has a core structure of hydroxyphenoxyphosphine (Top of Fig. 3) that is critical for therapeutic activation. Incubation with human liver microsomes, in the presence of NADPH, leads to a half-life of 1.1 min [92]. However, the absence of NADPH prolongs the half-life by almost 30 fold, to 30.6 min. It has been reported that remdesivir is a substrate of CYP2C8, 2D6 and 3A4. Cobicistat, a CYP3A4 inhibitor, completely inhibits oxidative metabolism of remdesivir, pointing to a major role of CYP3A4 in the oxidation. Nevertheless, both mono- and di-oxidation metabolites are detected, although the mono-oxidation metabolite reaches the peak much faster (5 min) [92]. Mass-spectrometric analysis of fragmentation pathways suggests that the initial oxidation occurs at the para position of the phenyl ring (Right of Fig. 3 in Red).

Fig 3. Metabolism of remdesivir.

This virial agent is a substrate of hydrolases (e.g., CES1) and CYP enzymes (e.g., CYP3A4). The hydrolytic metabolite is further hydrolyzed by HINT1 (Histidine triad nucleotide-binding protein 1, an amidase. It is likely that oxidized metabolite undergoes hydrolysis by CES1, Cat A or both.

It is well established that therapeutic activation of remdesivir is achieved by initial hydrolysis followed by phosphorylation [113]. The hydrolysis is catalyzed by CES1 and cathepsin A [91, 94, 114]. The relative contribution of these two enzymes to the hydrolysis varies depending on the tissue or cell type. While cathepsin A is kinetically more favorable toward remdesivir [91], CES1 is generally more abundant [94, 114]. It remains to be determined whether the oxidative metabolites of remdesivir are substrates of CES1 and/or cathepsin A (Fig.3). Nevertheless, the hydrolytic metabolite (alanine intermediate) generated by CES1 or cathepsin A undergoes further hydrolysis by histidine triad nucleotide-binding protein 1 to monophosphate (Fig. 3). The monophosphate metabolite undergoes two-step phosphorylation to form the final antiviral metabolite. As discussed with molnupiravir, the kinases for the phosphorylation remain to be determined. Likewise, NUDT18 has been shown to hydrolyze the therapeutically active metabolite as seen with NHC-TP triphosphate [111].

3–4. Drug transporters for molnupiravir, nirmatrelvir and remdesivir

Metabolism and active transport are highly coordinated events for drugs and their metabolites [115]. In contrast to metabolizing enzymes, the identities of transporters remain largely unknown for the membrane crossing of molnupiravir, nirmatrelvir and remdesivir or their metabolites. Nevertheless, it has been reported that equilibrative nucleoside transporters (ENTs) modestly support the uptake of molnupiravir but its hydrolytic metabolite NHC is a much better substrate of these transporters [116]. In VeroE6 P-glycoprotein knockout cells, nirmatrelvir has been shown to potently inhibit the replication of SARS-CoV-2 viruses, suggesting that nirmatrelvir is a substrate of P-glycoprotein [117], although the involvement of P-glycoprotein has not been fully confirmed [118]. In contrast to molnupiravir and nirmatrelvir, remdesivir has been identified to be a substrate of multiple transporters including BSEP, MATE1, MRP4, NTCP, OATP1B1/1B3 and OCT1 [119, 120]. These transporters are abundantly expressed in the liver. Interestingly, remdesivir is also a substrate of ENTs with a potency of 11–17 times of that of molnupiravir and 6 times of NHC, the hydrolytic metabolite of molnupiravir [116].

The higher efficiency of ENTs on the active transport of NHC, compared with that of its parent drug molnupiravir likely signifies a higher efficacy potency of NHC against SARS-CoV-2. As shown in Table 3, NHC is 1.6–5.6 times as potent as molnupiravir. Interestingly, GS-441524, the hydrolytic metabolite of remdesivir, is consistently less potent than its parent drug remdesivir. The differences in the relative lipophilicity between the parent drug and its metabolite likely accounts for such discrepancy. Remdesivir and its major metabolite GS-441524 have an XLogP3 value (an indicator for lipophilicity) of −1.4 and +1.9, respectively. These values constitute a relatively large net difference (i.e., 3.3), suggesting passive diffusion favoring remdesivir over GS-441524 for membrane-crossing. Remdesivir undergoes hydrolysis by CES1 and is eventually converted to GS-441524 [93]. The significantly higher efficacy of remdesivir than GS-441524 points to rapid diffusion and efficient hydrolysis in the SARS-CoV-2 infected cells. In contrast, molnupiravir and its hydrolytic metabolite NHC have an XLogP3 value of −0.8 and −2.2, resulting in a relatively small net difference (i.e., 1.4) [95]. Such difference points to an involvement of both passive and active transport to a similar extent for both molnupiravir and its hydrolytic metabolite NHC. As a result, NHC is moderately more potent than its parent drug molnupiravir.

4. POTENTIAL OF COMBINED USE OF SMALL MOLECULE COVID-19 MEDICINES

A combination medication has become an attractive strategy for various types of diseases, even for preventative purposes. This is particularly true with infectious diseases. The use of two or more antiretroviral medicines, commonly referred to as “cocktail therapy”, is currently the standard treatment for HIV infection and prophylactic purpose (human immunodeficiency virus) [121]. Combinations are usually made based on several important considerations: mechanism of action, host-metabolism and both. M^pro is the target of nirmatrelvir and conserved among coronaviruses [Lu et al., 2018]. RdRp, the target of molnupiravir and remdesivir, is highly conserved across all families of RNA virus [122]. As such, confirmation of properly combined use of these medicines for therapeutic synergy will likely have a broader clinical significance even beyond COVID-19 pandemic.

4–1. Metabolism-based rationale for combined use

As discussed above, molnupiravir, nirmatrelvir and remdesivir all undergo extensive metabolism. The metabolism may represent therapeutic activation or inactivation depending on a drug or the type of metabolism. For example, nirmatrelvir is metabolized primarily by oxidation and oxidized metabolites of nirmatrelvir no longer have therapeutic activities [71]. The oxidation of nirmatrelvir is primarily catalyzed by CYP3A4. Interestingly, this very P450 also oxidizes remdesivir, although the therapeutic potential of oxidized remdesivir remains to be determined [92]. Nevertheless, co-presence of remdesivir would competitively inhibit the oxidation of nirmatrelvir, thus enhancing the therapeutic activity of nirmatrelvir (Fig. 4). Nirmatrelvir is co-packaged along with the boosting agent ritonavir [123]. Ritonavir is a potent CYP3A4 inactivator and reduces the metabolism of nirmatrelvir [72]. Clearly remdesivir-nirmatrelvir co-package for COVID-19 has advantages over ritonavir-nirmatrelvir combination as both remdesivir and nirmatrelvir have anti-SARS-CoV-2 activity.

Fig 4. Potential interactions of remdesivir with nirmatrelvir or molnupiravir.

Remdesivir likely competitively inhibits nirmatrelvir, primarily through CYP3A4 (Top) or with molnupiravir through ENT transporters (Bottom). In addition, remdesivir covalently inhibits CES2, leading to decreased hydrolysis of molnupiravir. However, the covalent inhibition is reduced by CES1-hydrolysis of remdesivir.

Remdesivir, on the other hand, is hydrolytically activated, and the hydrolysis is catalyzed by CES1 [93]. At the same time, remdesivir is a covalent inhibitor of CES2 [94], a hydrolase that therapeutically activates molnupiravir [95, 103]. It has been shown that the expression of CES1 is inversely correlated with the inhibition of CES2, suggesting that remdesivir but not its hydrolytic metabolite is a CES2 inhibitor. Nevertheless, co-presence of remdesivir would likely impact the hydrolytic activation of molnupiravir. CES2 is abundantly expressed in the intestine and liver [124, 125]. Inhibition of CES2 by remdesivir likely increases the absorption and bioavailability of molnupiravir. In addition to hydrolytic metabolism, remdesivir and molnupiravir share uptake transporters (ENTs) for membrane crossing [116], representing another potential interaction for altered therapeutic activity of remdesivir and molnupiravir (Fig. 4). To take the advantage of remdesivir-molnupiravir interactions, the timing of drug administration is very important.

4–2. Viral target-based rationale for combined use

The mechanism of action is another critical factor to be considered for combined use of antiviral agents. It is generally accepted that combined use of drugs with distinct mechanisms of action has a reasonable chance of success in improving efficacy and safety. The anti-hepatitis C viral pill Harvoni, for example, contains a combination of ledipasvir/sofosbuvir and is more effective with a broader spectrum than sofosbuvir alone [126]. Ledipasvir is an inhibitor of NS5A, a hepatitis C virus protein [127], whereas sofosbuvir inhibits the RdRp of hepatitis C virus [128]. Likewise, many HIV medicines are fixed combinations of two or more drugs. Some of the combined drugs even share the metabolism of action. For example, Truvada contains emtricitabine and tenofovir disoproxil fumarate, and both HIV drugs are nucleoside reverse transcriptase inhibitor [129]. Truvada and Decovy are the only FDA approved medications for PrEP (pre-exposure prophylaxis) [129].

M^pro is the target of nirmatrelvir and RdRp is the target of molnupiravir and remdesivir (Table 2). The viral genomic RNA directs immediate translation of two large open reading frames, resulting in the production of polyproteins [130, 131] (Fig. 5). These polyproteins undergo processing to produce functional proteins including RdRp. M^pro plays a major role in protein processing [132]. RdRp, on the other hand, is responsible for the genome replication and production of subgenomic mRNAs. These subgenomic mRNAs are translated for production of accessary proteins, critical for viral assembling [131]. Clearly, co-inhibition of M^pro (nirmatrelvir) and RdRp (molnupiravir or remdesivir) represents a one-two pouch and likely produces synergistic activities against SARS-CoV-2 (Fig. 5). Even molnupiravir and remdesivir, although they share the target of RdRp, likely deliver additive or even synergistic activities as well. Molnupiravir targets RdRp and induces severe viral RNA mutations, commonly referred to as error catastrophe. Remdesivir, in contrast, targets RdRp and causes chain termination and arrests RNA synthesis of viral genome [61, 63, 133].

Fig 5. Mechanistic interplay among molnupiravir, nirmatrelvir and remdesivir.

Molnupiravir and remdesivir target RdRp, whereas nirmatrelvir covalently inhibits M^pro. likely competitively inhibits nirmatrelvir, primarily through CYP3A4 (Top) or with molnupiravir through ENT transporters (Bottom). In addition, remdesivir covalently inhibits CES2, leading to decreased hydrolysis of molnupiravir. However, the covalent inhibition is reduced by CES1-hydrolysis of remdesivir.

4–3. Experimental confirmation on additive/synergistic activities of combined use

The COVID-19 pandemic remains a major health concern worldwide, and SARS-CoV-2 is continuously evolving to new variants [134]. There is an urgent need to identify new antiviral drugs and develop therapeutic strategies. Combined use of recently authorized COVID-19 medicines has been actively pursued in animal and in vitro studies. Some exciting and promising results have been recently reported on the combined use of two or all three drugs: molnupiravir, nirmatrelvir and remdesivir. Li et al [135] has reported that molnupiravir and nirmatrelvir synergistically inhibit the infection of SARS-CoV-2 but the overall inhibitory patterns vary depending on a variant (e.g., Delta and Omicron) and a cell model (e.g., Calu-1 and organoids). Similar synergy is reported with BA.1 and 2 variants in Vero E6 cells by Gidari et al [136]. The synergy of the molnupiravir and nirmatrelvir combination is also confirmed in mice [137]. Interestingly, this mouse model shows less antiviral efficacy of the nirmatrelvir-remdesivir combination. It should be noted that all three drugs are used at 20 mg/kg twice daily for 5 days with the first two given orally whereas remdesivir intraperitoneally [137]. Clearly, the dosing regimen of remdesivir is much higher than those of molnupiravir and nirmatrelvir based on the recommended clinical regimens (Table 2). Nevertheless, several case reports have shown that a molnupiravir and remdesivir combination exerts a synergistic effect on reducing the severity of symptoms as well as the duration of hospitalization of COVID-19 patients [138]. Interestingly, favipiravir, another drug that targets RdRp, synergizes with molnupiravir in reducing infectious virus titers of SARS-CoV-2 in a hamster model [139].

5. CONCLUSION

The COVID-19 pandemic remains a major health concern worldwide, and SARS-CoV-2 is continuously evolving. There is an urgent need to identify new antiviral drugs and develop novel therapeutic strategies. Molnupiravir, nirmatrelvir and remdesivir are newly authorized COVID-19 medicines that have demonstrated high efficacy with reasonable safety profiles. They differ in the mechanism of action, presenting a strong foundation for combined use. Indeed, some in vitro and animal studies have shown promise of their combined use for enhanced activities. On the other hand, these drugs undergo extensive metabolism that determines their therapeutic potential. The metabolism pathways and the corresponding enzymes involved should be carefully considered during their combined use for therapeutic synergy.

ABBREVIATIONS

CES1

carboxyleslerase-1

CES2

carboxylesterase-2

COVID-19

coronavirus disease 2019

CYP

cytochrome P450

M^pro

main protease

NHC

N-hydroxycytidine

NHC-TP

NHC-triphosphate-triphosphate

RdRp

RNA dependent RNA polymerase

SARS-CoV

severe acute respiratory syndrome coronavirus

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2