As drug resistance becomes a major issue in the coming CVID-19 onslaughts, a strategy of using a combination of antivirals like that in HIV treatment protocols might be a better solution.
Since the beginning of the COVID-19 pandemic, scientists and researchers around the world have been working tirelessly to understand the virus responsible for the disease, SARS-CoV-2, and develop effective treatments and vaccines to combat it.
One of the key strategies in this effort has been the use of protease inhibitors, which are drugs that can block the activity of enzymes that the virus uses to replicate itself.
What are protease inhibitors?
Protease inhibitors are a class of drugs that are used to treat a variety of viral infections, including HIV, hepatitis C, and influenza. These drugs work by blocking the activity of protease enzymes, which are essential for viral replication. By inhibiting protease activity, the drugs prevent the virus from producing new copies of itself, which can help to slow or stop the spread of the infection.
Several protease inhibitors have been approved for use in treating COVID-19, including lopinavir/ritonavir and darunavir/cobicistat. These drugs have shown promise in early clinical trials, and are currently being used to treat COVID-19 patients in some countries.
What are transmissible SARS-CoV-2 variants with resistance to protease inhibitors?
However, recent studies have identified SARS-CoV-2 variants that are resistant to protease inhibitors, which could limit the effectiveness of these drugs in treating COVID-19. One of the most concerning of these variants is the B.1.427/B.1.429 variant, which was first identified in California in December 2020.
The B.1.427/B.1.429 variant has been found to be more transmissible than earlier strains of the virus, and is now the dominant strain in California. It has also been detected in other parts of the United States and around the world, raising concerns about its potential to spread rapidly and undermine efforts to control the pandemic.
Why are transmissible SARS-CoV-2 variants with resistance to protease inhibitors a cause for concern?
The emergence of transmissible SARS-CoV-2 variants with resistance to protease inhibitors is a cause for concern for several reasons. First, it could limit the effectiveness of current treatments for COVID-19, which could lead to more severe disease and higher mortality rates.
Second, it could make it more difficult to control the spread of the virus, as the variants could become dominant strains and be more easily transmitted between individuals. This could lead to more widespread outbreaks and longer-lasting pandemic.
To address the emergence of transmissible SARS-CoV-2 variants with resistance to protease inhibitors, several strategies are being pursued. These include:
- Continued monitoring and sequencing of the virus to identify new variants and track their spread: One of the most important strategies for addressing the emergence of transmissible SARS-CoV-2 variants with resistance to protease inhibitors is to continue monitoring the virus and sequencing its genome. This will allow scientists and researchers to identify new variants as they emerge and track their spread, which can help to inform public health measures and treatment strategies.
- Developing new treatments: Another strategy is to develop new treatments that are effective against the new variants. This could involve developing new protease inhibitors that are able to target the mutated spike protein, or identifying other drugs that can block viral replication through different mechanisms.
- Enhancing public health measures: In addition to developing new treatments, it is also important to continue implementing and improving public health measures to slow the spread of the virus. This could include measures such as social distancing, mask-wearing, and increased testing and contact tracing.
- Accelerating vaccination efforts: Finally, accelerating vaccination efforts is crucial for addressing the emergence of transmissible SARS-CoV-2 variants with resistance to protease inhibitors. Vaccines are a key tool for preventing severe illness and death from COVID-19, and widespread vaccination can help to reduce the spread of the virus and limit opportunities for new variants to emerge. However, it is also important to monitor the effectiveness of current vaccines against new variants and develop new vaccines or boosters as necessary.
A new research has found that ….
Major efforts continue for developing antiviral drugs to complement vaccination-based strategies for treating patients infected by SARS2 with the ultimate hopes of ending the COVID-19 pandemic and fortifying against future CoV outbreaks. Mpro inhibitors are at the forefront of CoV antiviral drug development, with Paxlovid (nirmatrelvir) already authorized for emergency clinical use in more than 65 countries and several other compounds, including ensitrelvir, in various stages of development (27).
However, drug resistance mutations have the potential to rapidly undermine these therapies. Here, we show that several naturally occurring Mpro variants already exhibit resistance to nirmatrelvir and ensitrelvir (results summarized in Tables 1 and 2).
The highest levels of resistance resulting from a single amino acid substitution identified here is A173V for nirmatrelvir (11.6-fold) and M49L for ensitrelvir (25.4-fold). Phylogenetic analyses show that these (and other) variants have arisen multiple times independently in different parts of the globe, with regional clusters and genetic linkage providing compelling evidence for transmission.
Nirmatrelvir is a substrate-mimicking covalent drug, and ensitrelvir is a nonpeptide/noncovalent inhibitor (12, 14). Consistent with distinct mechanisms of action, our studies indicate that these inhibitors are subject to at least partly nonoverlapping resistance profiles. For instance, A173V confers selective resistance to nirmatrelvir, whereas M49I and M49L confer increased resistance to ensitrelvir.
In comparison, ∆P168 appears to have a more broad-spectrum resistance phenotype. Encouragingly, antiviral assays with recombinant SARS2 corroborate our findings with two different cell-based assays and indicate that the ∆P168/A173V virus causes strong resistance to nirmatrelvir with little change in ensitrelvir susceptibility.
Variation at additional residues may also produce distinct resistance patterns when present in isolation versus in combination. For instance, T45I and D48Y exhibit a mild preferential resistance to ensitrelvir as single changes; however, when combined with ∆P168, D48Y shows strong resistance to both inhibitors and T45I cripples enzyme activity. Although our structural modeling and MD simulations provide plausible explanations for the A173V (± ∆P168) resistance phenotype, additional dedicated studies will be needed to establish other mechanisms of action.
During revision of this manuscript, multiple preprints reported Mpro mutants with resistance to nirmatrelvir (28–33). A173V was selected during serial passage experiments in the presence of boceprevir and independently in the presence of nirmatrelvir, which coupled with our results suggest that A173 may be a resistance hotspot (30, 31).
Our biochemical data using the purified A173V mutant demonstrate a lower affinity for nirmatrelvir as evidenced by a 50-fold increase in Ki with little change in catalytic efficiency against the canonical Nsp4-5 cleavage site. While these results support the selection of these mutants in serial passage, there appears to be some discrepancy in the magnitude of resistance between changes in in vitro Ki compared to changes in antiviral EC50.
This is observed for other Mpro variants, such as S144A, which is selected in serial passage and has been determined by Pfizer as causing a 90-fold increase in niramtrelvir Ki, while antiviral studies show a more modest 2-fold increase in antiviral EC50 (31). The smaller changes in antiviral EC50 may result from the fact that Mpro has a wide range of affinities [dissociation constant (Kd) measurements ranging from 28 μM to 2.7 mM] for its different polyprotein cleavage sites (34).
Therefore, while inhibitor affinity may be reduced by these mutants, it could still be sufficient to bind the enzyme before cleavage of the lower-affinity viral substrates. Therefore, resistance mutations may need to confer Ki increases of multiple orders of magnitude to cause large shifts in antiviral EC50. This interpretation is supported by our results with the ∆P168 mutant, which alone does not confer a shift in antiviral EC50; however, ∆P168/A173V has a 7.6-fold increase in EC50 compared to A173V alone (62.5-fold compared to WT).
Our data with these select mutants are concordant across four orthologous assays (two live cell assays, in vitro biochemistry, and in cellulo with replication-competent virus), suggesting that multiple mutations may be necessary to decrease drug binding affinity and cause resistance.
Consistent with our gain-of-signal assay showing only a twofold increase in background luminescence, our kinetic analyses of the purified A173V mutant indicate similar catalytic efficiency to WT (less than twofold change in kcat/KM). However, the multicycle growth kinetic assays with recombinant SARS2 show decreased replication. Independent studies have reported no change in replication kinetics for A173V, whereas another saw a decrease similar to ours that could be rescued by an additional L50F mutation (30, 31).
The reason for this discrepancy between replicative fitness of A173V from different laboratories is currently unclear, but different measurements of quantifying viral replication could be a contributing factor. Although activity is retained on the canonical Nsp4-5 substrate that is a standard for in vitro experiments, other cleavage sites along the polyprotein may be disproportionately affected by this change. For instance, our MD simulations show an increase in the dynamics of amino acids 40 to 65, which form the top of the S2 subsite that accommodates the hydrophobic P2 position of the peptide substrate.
As most cleavage sites along the viral polyprotein have a leucine at P2, phenylalanine and valine are also found at Nsp5-6 and Nsp6-7 junctions, respectively, which may be less efficiently processed by the A173V mutant. However, this variant has sufficient activity for virus replication and is observed recurrently in patient sequences and, therefore, it has the potential to contribute to clinical resistance phenotypes.
Although many groups have focused appropriately on resistance to nirmatrelvir given its early emergency use authorization by the U.S. Food and Drug Administration, ensitrelvir resistance is now equally important to understand given that emergency use authorization was granted in Japan on 22 November 2022. Along with this approval, documentation was released on serial passage experiments selecting for D48G, M49L, P52S, and S144A as resistance mutations (35).
These results support our finding of M49L showing the largest resistance phenotype using the gain-of-signal assay and the VSV-based assay. Furthermore, the selection of D48G also substantiates our hypothesis that disrupting hydrogen bonds between T45 and D48 to destabilize the structure of the helix above the S2 subsite can contribute to ensitrelvir resistance (indicated by our data for T45I and D48Y).
Another recent report has also identified M49I as conferring selective resistance to ensitrelvir and elegantly demonstrates the structural basis of this phenotype being due to the bulky isoleucine reorienting H41 and disrupting a base stacking interaction with the inhibitor (36). This is consistent with our finding that M49L causes greater resistance than M49I due to branching of the leucine side chain at the gamma carbon, which is closer to H41. Together, these findings indicate that the 45–49 region of Mpro has the potential to become a hotspot for the development of ensitrelvir resistance mutations.
By using our facile live-cell gain-of-signal assay coupled with sequence- and structure-informed mutation identification, we have been able to identify multiple changes in Mpro that confer varying degrees of resistance to nirmatrelvir and/or ensitrelvir. The resistance phenotypes described here are consistent between the four different assays we have implemented, and they are also consistent with reports by other groups through serial passage of virus or in vitro biochemical assays.
Our cell-based gain-of-signal assay has the advantage of only requiring the transfection of a single plasmid, which increases the throughput of variant testing compared to generation of recombinant virus or purification of mutant enzymes (especially those that are difficult to purify, such as the ∆P168 mutant). Using variants found within patient sequences at residues that are not strictly conserved across CoV species has helped identify changes more likely to be compatible with productive viral infection.
However, it is important to take great care when classifying variants within the GISAID database as many annotated variants are likely to be sequence artifacts. For example, we found >4000 sequences with an M165Y change, and manual inspection revealed that this is due to a single guanosine deletion in a poly-U stretch, which causes a frameshift after F160 and leads simultaneously to “detection” of H163W, E166Q, and a downstream stop codon (fig. S12, A and B).
Most of these sequences were produced using long-read nanopore technology, which has a 76-fold higher rate of indel errors compared to short-read technology (37). These sequencing and mis-annotation mistakes can lead to incorrect conclusions regarding the presence of variants in the population, and therefore, manual inspection of viral genomes is encouraged for putative changes at strictly conserved residues or those requiring multiple base changes (28). Determining the phylogenetic relationships of different variants using publicly available tools such as UShER also provides additional validation of the emergence of these variants by identifying common ancestors and dynastic relationships within distinct geographic regions (26).
Although our gain-of-signal system provides robust metrics for evaluating mutants and Mpro inhibitors in live cells, there are also a few limitations. For instance, although the relative background luminescence can serve as a proxy for catalytic activity, it is currently unclear which cellular substrates are being cleaved to cause low reporter expression, thereby limiting the correlations that may be made to activity of Mpro against individual viral polyprotein cleavage sites.
Moreover, our current approach focuses on amino acid residues that are variable (and not completely conserved) to avoid changes that would be severely deleterious, which limits the potential number of resistance pathways tested. For instance, it is difficult to predict secondary suppressor mutations that could restore fitness of a resistant but deleterious mutation. An example of this is the identification of E166V and E166A selected during serial passage to confer a high level of resistance against nirmatrelvir and other peptidomimetic inhibitors (30–32). As E166 is a highly conserved residue, it was not tested here.
Severe replication defects are evident with single substitutions at E166, but secondary mutations, such as L50F and T21I, appear to restore fitness (30–32). Thus, the gain-of-signal assay should be considered a valuable tool to study resistance mutations and complement traditional approaches such as serial passaging and competition studies to triage variants of interest that may be selected in cell culture or emerge in vivo during patient treatment before proceeding with more experimentally demanding approaches such as protein biochemistry and/or BSL3 testing with infectious viruses. Furthermore, we anticipate that the panel of mutants described here will be able to serve as an asset during the development of future generation Mpro inhibitors for rapid resistance profiling in parallel with structure activity relationship studies.
It is presently unclear what magnitude of resistance will be necessary for treatment failure in a clinical setting. Precedents with HCV NS3/4A show that single amino acid changes can elicit selective resistance of multiple orders of magnitude toward different inhibitors with minimal impact on viral fitness (38). However, resistance to HIV protease inhibitors often requires two or more mutations, with single amino acid changes typically showing modest changes in inhibitor susceptibility (39, 40).
The naturally occurring SARS2 Mpro variants described here may serve as evolutionary stepping stones for intermediate levels of resistance and provide a permissive environment enabling selection of secondary mutations that confer full drug resistance. Genetic surveillance of several of the variants identified here may be advantageous, and strategies should be taken to minimize the widespread development of resistance including the careful design of Mpro inhibitor drugs with different resistance profiles, which, encouragingly, is likely to be the case for nirmatrelvir and ensitrelvir.
The study findings were published in the peer reviewed journal: Science Advances.
