Impact of Baloxavir Resistance-Associated Substitutions on Influenza Virus Growth and Drug Susceptibility

ABSTRACT

Baloxavir marboxil (baloxavir) is a recently FDA-approved influenza virus polymerase acidic (PA) endonuclease inhibitor. Several PA substitutions have been demonstrated to confer reduced susceptibility to baloxavir; however, their impacts on measurements of antiviral drug susceptibility and replication capacity when present as a fraction of the viral population have not been established. We generated recombinant A/California/04/09 (H1N1)-like viruses (IAV) with PA I38L, I38T, or E199D substitutions and B/Victoria/504/2000-like virus (IBV) with PA I38T. These substitutions reduced baloxavir susceptibility by 15.3-, 72.3-, 5.4-, and 54.5-fold, respectively, when tested in normal human bronchial epithelial (NHBE) cells. We then assessed the replication kinetics, polymerase activity, and baloxavir susceptibility of the wild-type:mutant (WT:MUT) virus mixtures in NHBE cells. The percentage of MUT relative to WT virus necessary to detect reduced baloxavir susceptibility in phenotypic assays ranged from 10% (IBV I38T) to 92% (IAV E199D). While I38T did not alter IAV replication kinetics or polymerase activity, IAV PA I38L and E199D MUTs and the IBV PA I38T MUT exhibited reduced replication levels and significantly altered polymerase activity. Differences in replication were detectable when the MUTs comprised ≥90%, ≥90%, or ≥75% of the population, respectively. Droplet digital PCR (ddPCR) and next-generation sequencing (NGS) analyses showed that WT viruses generally outcompeted the respective MUTs after multiple replication cycles and serial passaging in NHBE cells when initial mixtures contained ≥50% of the WT viruses; however, we also identified potential compensatory substitutions (IAV PA D394N and IBV PA E329G) that emerged and appeared to improve the replication capacity of baloxavir-resistant virus in cell culture.

IMPORTANCE Baloxavir marboxil, an influenza virus polymerase acidic endonuclease inhibitor, represents a recently approved new class of influenza antivirals. Treatment-emergent resistance to baloxavir has been observed in clinical trials, and the potential spread of resistant variants could diminish baloxavir effectiveness. Here, we report the impact of the proportion of drug-resistant subpopulations on the ability to detect resistance in clinical isolates and the impact of substitutions on viral replication of mixtures containing both drug-sensitive and drug-resistant variants. We also show that ddPCR and NGS methods can be successfully used for detection of resistant subpopulations in clinical isolates and to quantify their relative abundance. Taken together, our data shed light on the potential impact of baloxavir-resistant I38T/L and E199D substitutions on baloxavir susceptibility and other biological properties of influenza virus and the ability to detect resistance in phenotypic and genotypic assays.

INTRODUCTION

Influenza is an acute, sometimes fatal, respiratory disease in humans caused by influenza type A and B viruses (IAV and IBV, respectively). Seasonal influenza epidemics have resulted in 3 to 5 million cases of severe disease and upwards of 500,000 deaths worldwide (1, 2). Moreover, novel influenza viruses have periodically emerged to cause global pandemics throughout history, with the most recent influenza pandemic occurring in 2009 (3, 4). Influenza antivirals include 3 classes of drugs: the M2 ion channel inhibitors (amantadine and rimantadine), the neuraminidase (NA) inhibitors (laninamivir, oseltamivir, peramivir, and zanamivir), and most recently, the polymerase acidic (PA) endonuclease inhibitor, baloxavir marboxil (5, 6).

The ability of influenza virus to mutate to acquire resistance to antiviral inhibition threatens the effectiveness of available countermeasures (7). Most currently circulating influenza strains are resistant to the M2 ion channel inhibitors, and, therefore, the use of these inhibitors is not currently recommended (8). Moreover, NA inhibitor-resistant seasonal IAV H1N1 viruses emerged and became widespread as recently as 2009, and while subsequently replaced by a susceptible H1N1 pandemic strain, the potential exists for NA inhibitor resistance to become widespread again (7, 9, 10).

Treatment-emergent resistance to baloxavir has been observed in up to 11% of adult/adolescent baloxavir marboxil clinical trial subjects, with increased frequencies observed in younger subjects and in IAV infections, particularly in the H3N2 subtype; baloxavir resistance has occurred more rarely in IBV infections (11–13). Baloxavir resistance is most frequently mediated by substitutions of the conserved isoleucine at PA position 38 (i.e., I38F/L/M/N/S/T), and IAVs and IBVs containing these substitutions have been reported to confer 3- to 116-fold reductions in baloxavir susceptibility in cell culture (14–19). While the most common baloxavir resistance-associated amino acid substitutions have occurred at PA I38 (11, 20), other PA substitutions associated with >3-fold-reduced susceptibility in cell culture have been observed as treatment emergent (e.g., E23G/K/R, A37T, and E199G in IAV), identified in circulating influenza virus (e.g., I38L and E199D in IAV), or selected in cell culture (e.g., E198K in IAV) (11, 18–21).

Phenotypic and genotypic evaluations of clinical isolates are key approaches to surveillance for antiviral resistance. Traditionally, methods such as quantitative reverse transcription-PCR (RT-PCR), pyrosequencing, and Sanger sequencing have been used to detect drug resistance-associated mutations. However, these methods may not always detect mutations at very low frequencies: i.e., when drug-resistant viruses are present as a small proportion of the total virus population in a clinical isolate (22, 23). Technologies such as droplet digital PCR (ddPCR) and next-generation sequencing (NGS) may provide a more sensitive means for detection of viral mutations associated with drug resistance in clinical specimens and isolates (22, 24, 25). NGS has been deployed to detect nucleotide mutations encoding PA I38F/M/S/T substitutions that may exist as minor variants in clinical isolates (26–28). ddPCR has been used to evaluate changes in the ratio of wild-type (WT) to mutant (MUT) viruses carrying PA I38T after passages in either ST6-GalI-MDCK cells or C57BL/6 mice (14). In these studies, PA I38T decreased in frequency over multiple passages in cell culture, but its proportion remained unchanged or increased in mice, indicating that the impact of the PA I38T substitution may differ depending on the target cell population or animal host. NGS and ddPCR methods were found to enable detection of low-frequency single nucleotide mutations with high accuracy and precision.

Most of our knowledge regarding the impact of baloxavir-resistant viruses is based on the analysis of homogenous influenza virus populations carrying the target mutation (11, 14, 15, 19). However, the ability to measure the impact of potential baloxavir resistance-associated substitutions on baloxavir susceptibility and viral replication kinetics may depend on their prevalence in the viral population. Previously, coculture of different mixtures of WT and PA I38T MUTs of A/Quebec/144147/2009 (H1N1)pdm09 and A/Switzerland/9715293/2013 (H3N2) viruses in the absence of baloxavir resulted in a significant reduction in the proportion of MUTs after a single passage in ST6GalI-MDCK cells, indicating that these MUTs had reduced replication capacity relative to their respective WT counterparts in this cell culture system (14). In addition, clinical isolates of IAV and IBV carrying PA I38T showed replicative deficits compared to WT IAV and IBV in MucilAir human nasal epithelial and MDCK cells as well as in a ferret animal model (26).

Understanding the impact of mutant proportions on baloxavir susceptibility of a mixed virus population containing WT and drug-resistant MUT virus will aid in identifying the best approaches for monitoring drug resistance in a clinical setting. Here, we generated recombinant IAV and IBV MUTs containing baloxavir resistance-associated PA substitutions (i.e., IAV I38L, IAV I38T, IAV E199D, and IBV I38T) and assessed their impact on viral replication and drug susceptibility when they are present at different ratios within the virus population (i.e., WT:MUT ratios [percentages] of 100:0, 90:10, 75:25, 50:50, 25:75, 10:90, and 0:100). In addition, to better understand the impact of resistance-associated substitutions on virus replication capacity and potential compensatory pathways that might increase the replication capacity of resistant viruses, we used ddPCR and NGS methods to measure changes in the WT:MUT ratios as well as identify and track potential compensatory mutations that emerged during passages in normal human bronchial epithelial (NHBE) cells.

RESULTS

Growth characteristics and polymerase activity of the WT:MUT virus and plasmid mixtures.

We first mixed WT and MUT viruses or PA plasmids at ratios (percentages) of 100:0, 90:10, 75:25, 50:50, 25:75, 10:90, and 0:100 based on the concentration of the parental viruses or plasmid stocks to achieve 10^4.5 PFU or 10⁸ plasmid copies per mixture, respectively. The actual WT:MUT virus or plasmid ratios were then verified by ddPCR and were adjusted accordingly when needed (Table 1). We next quantified virus growth of the WT:MUT virus mixtures by plaque assay in MDCK cells. As shown in Table 1, titers of all IAV and IBV mixtures were within an ≈2.5-fold difference of the parental WT viruses. No changes in plaque sizes were observed in any of the WT:MUT virus mixtures.

TABLE 1.

WT:MUT ratios evaluated in the study, growth characteristics, and polymerase activity levels

Variant	WT:MUT ratio (%)			Virus yield (log10 PFU/mL)b,c	Plaque size (mm)d	Polymerase activity (%)c,e
	Targeted	Measured by ddPCRa
		Virus	Plasmid
IAV I38L	100:0	100:0	100:0	5.3 ± 0.1	0.6 ± 0.2	100.0 ± 7.3
	90:10	91:9	91:9	5.7 ± 0.2*	0.5 ± 0.2	127.6 ± 14.5°
	75:25	78:22	77:23	5.5 ± 0.1	0.5 ± 0.2	133.5 ± 2.4°
	50:50	49:51	50:50	5.6 ± 0.1	0.4 ± 0.2	143.9 ± 2.6°
	25:75	26:74	25:75	5.7 ± 0.1*	0.4 ± 0.2	140.5 ± 4.3°
	10:90	10:90	10:90	5.2 ± 0.2	0.4 ± 0.1	138.9 ± 3.5°
	0:100	0:100	0:100	5.5 ± 0.1	0.4 ± 0.1	147.4 ± 17.1°
IAV I38T	100:0	100:0	100:0	5.3 ± 0.1	0.6 ± 0.2	100.0 ± 7.3
	90:10	91:9	88:12	5.2 ± 0.1	0.5 ± 0.2	98.4 ± 5.2
	75:25	75:25	73:27	5.1 ± 0.1	0.5 ± 0.2	111.4 ± 0.3
	50:50	51:49	51:49	5.2 ± 0.2	0.4 ± 0.2	89.6 ± 2.6
	25:75	26:74	24:76	5.0 ± 0.1*	0.4 ± 0.2	102.7 ± 5.7
	10:90	10:90	11:89	5.3 ± 0.2	0.4 ± 0.1	114.8 ± 8.0
	0:100	0:100	0:100	5.4 ± 0.1	0.3 ± 0.1	90.8 ± 7.0
IAV E199D	100:0	100:0	100:0	5.3 ± 0.1	0.6 ± 0.2	100.0 ± 7.3
	90:10	91:9	89:11	5.4 ± 0.1	0.5 ± 0.2	100.9 ± 2.3
	75:25	73:27	75:25	5.0 ± 0.1*	0.5 ± 0.2	108.5 ± 2.5
	50:50	50:50	49:51	5.0 ± 0.2	0.5 ± 0.2	102.1 ± 17.1
	25:75	24:76	25:75	5.0 ± 0.2	0.5 ± 0.1	120.0 ± 1.9*
	10:90	9:91	12:88	5.1 ± 0.1	0.5 ± 0.1	156.2 ± 7.9°
	0:100	0:100	0:100	5.5 ± 0.1	0.4 ± 0.1	145.0 ± 9.7°
IBV I38T	100:0	100:0	100:0	6.0 ± 0.3	0.5 ± 0.2	100.0 ± 3.1
	90:10	89:11	91:9	6.1 ± 0.1	0.5 ± 0.2	103.8 ± 6.3
	75:25	76:24	75:25	5.8 ± 0.1	0.6 ± 0.2	95.2 ± 7.2
	50:50	50:50	49:51	6.0 ± 0.1	0.6 ± 0.3	90.4 ± 5.8
	25:75	26:74	26:74	6.0 ± 0.2	0.7 ± 0.3	87.9 ± 3.8*
	10:90	11:89	10:90	6.0 ± 0.2	0.7 ± 0.2	87.3 ± 4.2*
	0:100	0:100	0:100	6.1 ± 0.2	0.7 ± 0.2	86.6 ± 1.0*

^aValues represent the WT:MUT ratios of each mixture as measured by ddPCR. Approximately 10^4.5 RNA or 10⁸ plasmid copies were used in each ddPCR mixture.

^bValues represent the mean log₁₀ PFU per milliliter ± standard deviation from three independent determinations. The number of PFU was measured by plaque assay in MDCK cells after incubation at 37°C for 2 days (IAV) or 33°C for 3 days (IBV) (46).

^cValues in boldface are significant (*, P < 0.05; °, P < 0.01) by one-way analysis of variance (ANOVA) compared to the value for WT virus or RNP.

^dValues represent the plaque diameter (mean ± standard deviation) as measured by Finescale comparator.

^eValues represent the RNP polymerase activity (mean ± standard deviation) compared to WT RNP activity.

We measured the RNP activity of the WT:MUT plasmid mixtures in 293T cells using the luciferase-based minigenome assay. As shown in Table 1, all IAV RNPs with I38L exhibited significantly increased polymerase activity compared to WT RNP (1.4-fold; P < 0.01). No changes were observed in the polymerase activity of the IAV RNPs containing I38T. The presence of E199D in IAV RNP at ≥75% significantly increased polymerase activity compared to WT RNP (1.4-fold; P < 0.05). In contrast, the presence of I38T in IBV RNPs at ≥75% resulted in a measurable decrease in polymerase activity (1.1-fold; P < 0.05).

Replication kinetics of the WT:MUT virus mixtures in NHBE cells.

We measured the replication kinetics of the WT:MUT virus mixtures in differentiated NHBE cells (Fig. 1). The presence of the I38L or E199D MUTs in an IAV population at ≥90% resulted in reduced viral titers at 72 h postinfection (hpi) compared to WT (log₁₀ 1.4-fold; P < 0.05). In contrast, the presence of the IAV I38T MUT did not affect WT growth kinetics at any ratio. The IBV virus mixtures that contained the I38T MUT at ≥75% grew to significantly lower titers than WT at 48 and 72 hpi (log₁₀ 1.4-fold; P < 0.05).

FIG 1.

Replication kinetics of the IAV (A and B) WT:MUT I38L, (C and D) WT:MUT I38T, (E and F) WT:MUT E199D, and (G and H) IBV WT:MUT I38T virus mixtures in NHBE cells. Cells were infected with the virus mixtures at an MOI of 0.01 for 1 h before being washed, overlaid with fresh medium, and incubated at 37°C (IAV) or 33°C (IBV). At 24, 48, and 72 hpi, supernatants were collected and titrated by plaque assay in MDCK cells. Fold changes in the area under the curve for the WT:MUT mixtures compared to the WT virus are shown in parentheses. *, P < 0.05, and °, P < 0.01, compared to the values for the respective WT virus by one-way ANOVA.

To gain insight into changes of the WT:MUT ratios after multiple replication cycles in NHBE cells, the culture supernatants were analyzed by ddPCR and NGS (Fig. 2). A strong correlation (r_s values ranged from 0.97 to 0.99) was observed between the results obtained by both methods (Table 2; see Fig. S1 in the supplemental material). Collectively, WT virus increased in abundance in all IAV and IBV WT:MUT mixtures by 72 hpi when the initial mixtures contained the WT virus at ≥25%, except one IAV mixture that contained the I38T MUT at 75%, which remained relatively steady throughout the experiment (Fig. 2). Increased MUT abundance was observed by 72 hpi in all virus mixtures containing the MUT virus at ≥90%.

FIG 2.

Changes in the WT:MUT ratios after multiple replication cycles in NHBE cells. RNA was extracted from supernatants and analyzed either by ddPCR (A, C, E, and G) or NGS (B, D, F, and H). The dashed line represents the percentage of the MUT virus needed to increase the EC₅₀ value by 3-fold compared to the respective WT virus as determined by virus yield reduction assay in NHBE cells (Table 4). Red shading indicates abundance of the MUT virus resulting in EC₅₀ values higher than 3-fold the EC₅₀ of the respective WT virus (3EC₅₀); gray shading indicates abundance of the MUT virus resulting in EC₅₀ values lower than 3EC₅₀ of the respective WT virus as determined by virus yield reduction assay.

TABLE 2.

WT:MUT ratios after 72 hpi and after the 3rd passage in NHBE cells

Variant	WT:MUT ratio (%)
	Initial	After 72 hpi		After 3rd passagea
		ddPCR	NGS	ddPCR	NGS
IAV I38L	100:0	100:0	100:0	100:0	100:0
	90:10	100:0	100:0	100:0	100:0
	75:25	90:10	86:14	91:9	92:8
	50:50	83:17	75:25	82:18	81:19
	25:75	68:32	50:50	59:41	60:40
	10:90	1:99	1:99	0:100	1:97
	0:100	0:100	0:100	0:100	0:100
IAV I38T	100:0	100:0	100:0	100:0	100:0
	90:10	100:0	100:0	100:0	100:0
	75:25	86:14	92:8	81:19	79:29
	50:50	76:24	75:25	72:28	67:33
	25:75	29:71	24:76	1:99	6:94
	10:90	0:100	1:99	0:100	2:98
	0:100	0:100	0:100	0:100	0:100
IAV E199D	100:0	100:0	100:0	100:0	100:0
	90:10	99:1	99:1	98:2	100:0
	75:25	99:1	96:4	95:5	88:12
	50:50	96:4	85:15	88:12	72:28
	25:75	53:47	57:43	35:65	39:61
	10:90	0:100	1:99	0:100	5:95
	0:100	0:100	0:100	0:100	0:100
IBV I38T	100:0	100:0	100:0	100:0	100:0
	90:10	97:3	99:1	97:3	100:0
	75:25	93:7	98:2	88:12	87:13
	50:50	80:20	89:11	80:20	85:15
	25:75	48:52	50:50	36:64	58:42
	10:90	7:93	9:91	8:92	14:86
	0:100	0:100	0:100	0:100	0:100

^aThreefold differences between values measured by ddPCR and NGS are shown in boldface.

Subsequently, we employed NGS to determine if additional PA substitutions arose during viral replication in NHBE cells (Table 3). No additional substitutions were observed in any 100% WT virus cultures, IAV WT:MUT I38T, or WT:MUT E199D mixtures. However, the PA D394N substitution was observed at 36% in 100% of the IAV I38L MUT population after 72 hpi, and the PA E329G substitution was observed at 27% in 100% of the IBV I38T MUT population after 72 hpi in NHBE cells (Table 3). We next determined if IAV PA D394N and IBV PA E329G substitutions occurred in influenza viruses circulating from 2009 to 2022 by analyzing human influenza virus sequences available in the National Center for Biotechnology Information (NCBI [accessed through www.fludb.org]). As shown in Table S1 in the supplemental material, both PA substitutions were observed at very low frequencies (<1%), indicating that these substitutions are currently uncommon.

TABLE 3.

PA substitutions that emerged during replication of the WT:MUT mixtures in NHBE cells

Variant	WT:MUT ratio (%)	hpi or passage no.	Additional PA substitution	Frequency (%) by NGS
IAV I38L	0:100	48 hpi	D394N	31
	0:100	72 hpi	D394N	36
	0:100	P1	D394N	27
	0:100	P2	D394N	28
	0:100	P3	D394N	33
	10:90	72 hpi	D394N	20
	10:90	P1	D394N	26
	10:90	P2	D394N	27
	10:90	P3	D394N	30
	25:75	72 hpi	D394N	9
	25:75	P1	D394N	22
	25:75	P2	D394N	22
	25:75	P3	D394N	24
	50:50	P1	D394N	11
	50:50	P2	D394N	11
	50:50	P3	D394N	12
	75:25	P2	D394N	10
	75:25	P3	D394N	10
IBV I38T	0:100	72 hpi	E329G	27
	0:100	P1	E329G	26
	0:100	P2	E329G	37
	0:100	P3	E329G	41
	10:90	72	E329G	21
	10:90	P1	E329G	24
	10:90	P2	E329G	31
	10:90	P3	E329G	38
	25:75	72 hpi	E329G	17

Serial passaging of the WT:MUT virus mixtures in NHBE cells.

To gain further insight into changes in the WT:MUT ratios during replication in NHBE cells, we serially passaged each mixture 3 times, and the WT:MUT ratios were determined at the end of each passage by ddPCR and NGS (Fig. 3). The WT:MUT mixtures that contained the WT virus at ≥50% exhibited increased WT abundance during passaging. In contrast, MUT virus abundance increased during passaging when it was present at ≥90% in the initial mixture. Interestingly, when the initial WT:MUT mixtures contained 75% MUT virus, the abundance of the MUT virus varied after 3 passages in NHBE cells: the abundance of the IAV I38T MUT increased; however, the abundance of other MUTs either decreased (IAV I38L) or remained steady (IAV E199D and IBV I38T) (Fig. 3). We observed that results obtained with ddPCR and NGS methods were highly correlated (r_s values ranged from 0.97 to 0.99) (Table 2; Fig. S2). Moreover, the changes in the WT:MUT ratios seen after 72 hpi (Fig. 2) and after 3 sequential passages in NHBE cells (Fig. 3) were consistent and showed that WT dominated over MUT when WT was present at ≥50% in the initial virus mixture.

FIG 3.

Changes in the WT:MUT ratios after serial passaging in NHBE cells. RNA was extracted from supernatants and analyzed by either ddPCR (A, C, E, and G) or NGS (B, D, F, and H). The dashed line represents the percentage of the MUT virus needed to increase the EC₅₀ value by 3-fold compared to the respective WT virus as determined by virus yield reduction assay in NHBE cells (Table 4). Red shading indicates abundance of the MUT virus resulting in EC₅₀ values higher than 3-fold the EC₅₀ of the respective WT virus (3EC₅₀); gray shading indicates abundance of the MUT virus resulting in EC₅₀ values lower than 3EC₅₀ of the respective WT virus as determined by virus yield reduction assay.

We also monitored for additional PA substitutions that occurred in the IAV and IBV WT:MUT mixtures after serial passaging in NHBE cells by NGS (Table 3). PA D394N and E329G substitutions emerged to frequencies of 27% in the 100% IAV I38L population and 26% in the 100% IBV I38T population after the first passage and increased slightly through passage 3. The frequencies of these substitutions increased in relative proportion to the MUT viruses after passaging.

Susceptibility of the WT:MUT mixtures to baloxavir as measured by plaque reduction and minigenome assays.

We first evaluated the susceptibility of the WT:MUT virus mixtures to baloxavir by plaque reduction assay in MDCK-SIAT1 cells (Fig. 4). We observed that IAV mixtures containing the I38L MUT at ≥25% significantly increased baloxavir 50% effective concentration (EC₅₀) values by 5.7-fold compared to those of the WT virus (P < 0.05). A similar trend was observed for the IAV WT:MUT I38T virus mixtures, when the presence of the I38T MUT at ≥25% resulted in significantly increased EC₅₀ values (25.8-fold; P < 0.05). The presence of E199D MUT in the virus mixture at ratios of ≥50% significantly increased baloxavir EC₅₀ values compared to the WT (2.3-fold; P < 0.05). Additionally, IBV mixtures containing ≥25% I38T MUT exhibited significantly increased EC₅₀ values compared to the IBV WT virus (2.7-fold; P < 0.05) (Fig. 4D).

FIG 4.

Susceptibility measurements of IAV (A) WT:MUT I38L, (B) WT:MUT I38T, (C) WT:MUT E199D, and (D) IBV WT:MUT I38T virus mixtures to baloxavir by plaque reduction assay in MDCK-SIAT1 cells. Cells were infected with influenza virus for 1 h at 37°C (IAV) or 33°C (IBV). The cells were then washed, overlaid with minimal essential medium containing 0.25% agarose and baloxavir, and incubated at 37°C (IAV) or 33°C (IBV) for 2 or 3 days, respectively. Cells were stained with 0.1% crystal violet in 10% formaldehyde solution, and the plaques were counted. The concentration of baloxavir that caused a 50% decrease in the PFU titer compared to control wells without drug was defined as the EC₅₀. The results of two independent experiments were averaged. IAV I38L, I38T, and E199D, and IBV I38T MUTs exhibited EC₅₀ value fold changes of 10.3-, 32.4-, 2.7-, and 5.0-fold when present at 100%, respectively, compared to the respective WTs. The dashed line represents the percentage of the MUT virus needed to increase the EC₅₀ value by 3-fold compared to the respective WT virus (Table 4). Red shading indicates abundance of the MUT virus resulting in EC₅₀ values higher than 3-fold the EC₅₀ of the respective WT virus (3EC₅₀); gray shading indicates abundance of the MUT virus resulting in EC₅₀ values lower than 3EC₅₀ of the respective WT virus. *, P < 0.05, and °, P < 0.01, compared to the values for the respective WT virus by one-way ANOVA.

We verified our results by using the luciferase-based minigenome assay in 293T cells, which eliminates variance caused by different replication kinetics of tested virus mixtures. IAV RNPs containing the PA I38L or PA E199D substitutions at ratios of ≥25% had a significantly increased baloxavir 50% inhibitory concentration (IC₅₀) values compared to WT RNP (2.3-fold; P < 0.05) (Fig. 5A and C). IAV RNP mixtures with ≥10% PA I38T substitution exhibited significantly increased IC₅₀ values (51.6-fold; P < 0.01) (Fig. 5B). The presence of the PA I38T substitution in IBV RNP mixtures at ≥25% resulted in significantly increased IC₅₀ values compared to WT RNP (16.3-fold; P < 0.05) (Fig. 5D). Overall, our results from the plaque reduction and minigenome assays were in good agreement (Fig. S3) and demonstrated that viral mixtures that contained IAV I38L, IAV I38T, and IBV I38T MUTs at abundances of 35%, 16%, and 82%, respectively, showed 3-fold increases in EC₅₀ values compared to the respective WT viruses (Table 4).

FIG 5.

Susceptibility of IAV RNPs containing PA (A) I38L, (B) I38T, (C) E199D, and (D) IBV RNPs containing PA I38T to baloxavir. HEK 293T cells were treated with baloxavir for 1 h before being transfected with a mixture of PB1, PB2, PA (WT:MUT ratio [percentage]), and NP plasmids. The cells were the incubated at 37°C (IAV) or 33°C (IBV) for 24 h. Cell extracts were lysed, and luciferase levels were assayed with a dual-luciferase-based assay. The concentration of baloxavir that caused a 50% reduction in RNP activity compared to untreated cells was defined as the IC₅₀. Experiments were performed at least in triplicate. IAV I38L, I38T, and E199D, and IBV I38T MUT RNPs exhibited IC₅₀ value fold changes of 5.3-, 139.5-, 2.7-, and 25.7-fold when present at 100%, respectively, compared to the respective WT RNPs. The dashed line represents the percentage of the MUT RNP needed to increase the IC₅₀ value by 3-fold compared to the respective WT RNP (3IC₅₀) (Table 4). Red shading indicates abundance of the MUT RNP resulting in IC₅₀ values higher than IC₅₀ of the respective WT RNP; gray shading indicates abundance of the MUT RNP resulting in EC₅₀ values lower than 3IC₅₀ of the respective WT virus. *, P < 0.05, and °, P < 0.01, compared to the values for the respective WT RNP by one-way ANOVA.

TABLE 4.

Percentage of MUT needed to detect a 3-fold increase in the EC₅₀ or IC₅₀ values compared to the respective WT virus

Mixture	% of MUT needed to detect 3-fold increase vs WT by:
	Plaque reduction assay in MDCK-SIAT1 cells	Minigenome assay	Virus yield reduction assay in NHBE cells
IAV WT:MUT I38L	35	66	49
IAV WT:MUT I38T	16	31	13
IAV WT:MUT E199D	-a	-	92
IBV WT:MUT I38T	82	20	10

^a- Not applicable.

Susceptibility of the WT:MUT mixtures to baloxavir as measured by virus yield reduction assay and changes in WT:MUT ratios in the presence of the drug.

We further assessed the impact of the MUT viruses on susceptibility of the WT:MUT virus mixtures to baloxavir by virus yield reduction assay in NHBE cells (Fig. 6). Virus mixtures containing the IAV I38L MUT at ≥50% demonstrated significantly reduced susceptibility to baloxavir compared to the WT virus (5.5-fold; P < 0.05). The presence of the IAV I38T MUT at ≥10% in the virus population also reduced susceptibility to baloxavir by 16.8-fold (P < 0.05). The IAV E199D MUT present at ≥75% reduced susceptibility to baloxavir (1.8-fold; P < 0.05), and IBV WT:MUT mixtures containing I38T MUT at ≥10% exhibited reduced susceptibility compared to the IBV WT virus (27.4-fold; P < 0.05) (Fig. 6D). Overall, the EC₅₀ values determined by virus yield reduction assay in NHBE cells correlated well with our results from the plaque reduction and minigenome assays (Fig. S4 and S5) and indicated that a relatively small proportion of the IAV I38T MUT (i.e., 13%) was needed to significantly reduce susceptibility of the IAV population to baloxavir in human epithelial cells compared to the proportions of I38L and E199D MUTs needed to detect reduced susceptibility (49% and 92%, respectively) (Table 4).

FIG 6.

Susceptibility of IAV (A) WT:MUT I38L, (B) WT:MUT I38T, (C) WT:MUT E199D, and (D) IBV WT:MUT I38T virus mixtures to baloxavir by virus yield reduction assay in NHBE cells. Cells were pretreated with baloxavir for 2 h before being infected with WT:MUT virus mixtures (MOI of 0.01 PFU/cell). After 1 h, the cells were washed, baloxavir-containing basal medium was replaced, and cells were incubated at 37°C (IAV) or 33°C (IBV) for 2 days. Viral titers in the collected supernatants were determined by plaque assay in MDCK cells. The concentration of baloxavir that caused a 50% decrease in the PFU titer compared to control wells without drug was defined as the EC₅₀. The results of two independent experiments were averaged. IAV I38L, I38T, and E199D, and IBV I38T MUTs exhibited EC₅₀ value fold changes of 15.3-, 72.3-, 5.4-, and 54.5-fold when present at 100%, respectively, compared to the respective WTs. The dashed line represents the percentage of the MUT virus needed to increase the EC₅₀ value by 3-fold compared to the respective WT virus (3EC₅₀) (Table 4). Red shading indicates abundance of the MUT virus resulting in EC₅₀ values higher than 3EC₅₀ of the respective WT virus; gray shading indicates abundance of the MUT virus resulting in EC₅₀ values lower than 3EC₅₀ of the respective WT virus. *, P < 0.05, and °, P < 0.01, compared to the values for the respective WT virus by one-way ANOVA.

To determine the changes in the IAV and IBV WT:MUT virus ratios due to baloxavir selective pressure, virus supernatants collected after the virus yield reduction assay were assessed by ddPCR (Fig. 7). Although 5 nM or 10 nM baloxavir completely eliminated detectable WT virus in the IAV WT:MUT I38L/T mixtures, the WT virus was still detectable in the IAV WT:MUT E199D virus mixtures at these concentrations. Decreased IBV WT abundance was also observed in the WT:MUT I38T virus mixtures at 5 nM and 10 nM baloxavir; however, at 50 nM the drug completely suppressed replication of the IBV WT virus.

FIG 7.

Abundance of the MUT virus in the supernatants collected after virus yield reduction assay in NHBE cells as measured by ddPCR. The dashed line represents the percentage of the MUT virus needed to increase the EC₅₀ value by 3-fold compared to the respective WT virus as determined by virus yield reduction assay in NHBE cells (3EC₅₀) (Table 4). Red shading indicates abundance of the MUT virus resulting in EC₅₀ values higher than 3EC₅₀ of the respective WT virus; gray shading indicates abundance of the MUT virus resulting in EC₅₀ values lower than 3EC₅₀ of the respective WT virus as determined by virus yield reduction assay.

DISCUSSION

The recent approval of baloxavir marboxil as a therapeutic agent against influenza bolstered our limited countermeasures against this virus. However, treatment-emergent resistance to baloxavir has been observed frequently in clinical trials (up to 11% in adult/adolescent subjects and up to 26% in pediatric subjects [13]) and is associated with prolonged virus shedding (11, 12, 19). Therefore, understanding the impact of baloxavir resistant-associated substitutions on virus fitness and drug sensitivity and the best approaches to measuring these phenotypes are key concerns. To date, most of our knowledge on the impact of these substitutions comes from studies of pure viral populations. However, drug-resistant viruses often exist as a mixture with WT virus within the whole viral population. In this study, we assessed the effect of baloxavir-resistant subpopulations on viral replication capacity and sensitivity to baloxavir.

Substitutions at amino acid position 38 of the PA protein have previously been shown to decrease influenza virus susceptibility to baloxavir (11, 18–20, 29–32), and PA I38T exhibits the greatest impact on drug susceptibility, with increases in EC₅₀ values of 27- to 116-fold in cell culture (18, 29–31). Indeed, our data demonstrated that only a small proportion of virus with this substitution (i.e., 13%) was needed to induce significantly reduced baloxavir susceptibility of the virus population in human respiratory epithelial cells. In contrast, the PA I38L and E199D substitutions, which have more limited impacts on baloxavir susceptibility, need to be present in higher proportions (49% to 92%) to detect a decrease in susceptibility of a virus population to baloxavir.

The PA I38T substitution had a negligible effect on viral polymerase activity and replication kinetics. In contrast, IAV PA I38L and E199D, and IBV PA I38T significantly changed (i.e., increased or decreased) polymerase activity, which was detectable when these three MUTs were present at ratios as low as 90:10, 25:75, and 25:75, respectively. Decreased viral replication levels at 72 hpi in NHBE cells were also observed for virus populations that contained these MUTs at ≥90% in the mixture. Our results on the impact of the IAV PA I38T substitution on viral replication and polymerase activity correlate well with previous reports that showed that this substitution did not alter polymerase activity or replication kinetics of A(H1N1)pdm09-like viruses in HEK 293T and ST6-GalI-MDCK cells, respectively (14, 33). We also confirmed our previous findings that the PA I38L and E199D substitutions significantly increased polymerase activity and reduced viral growth of the A/California/04/09 (H1N1) strain (19). However, these data are in contrast to those of Chesnokov et al. (15), who showed that the A/Illinois/37/2018 (H1N1) virus that carried the PA I38L substitution had viral growth characteristics similar to its WT counterpart in MDCK and MDCK-SIAT1 cells. These differences might be associated with the different virus genetic backgrounds used for evaluation of the impact of the PA I38L substitution in vitro and highlight the fact that the impact of the baloxavir resistance-associated substitutions on viral fitness may differ dramatically depending on which cell line is used.

Our results also aligned well with a previously published report in which the B/Brisbane/60/2008-like and B/Phuket/3073/2013-like viruses carrying the PA I38T substitution exhibited decreased replication kinetics and polymerase activity in MDCK and HEK 293T cells, respectively (16). Additionally, the B/Victoria/2/1987-like virus with the PA I38T substitution showed significantly decreased growth at 24 hpi in ST6GalI-MDCK and at all time points tested in human airway epithelial cells. In contrast, B/Phuket/2073/2013 with the PA I38T substitution had replication kinetics similar to that of the drug-sensitive counterpart (32). These findings indicate that the impact of the PA I38T substitution may also depend on the virus background used for cell culture evaluations. In summary, all of the baloxavir resistance-associated substitutions that we evaluated, with the exception of the IAV PA I38T substitution, significantly altered viral replication capacity in vitro, which may indicate reduced fitness of these variants in the human population.

The ability of drug-resistant viruses to spread in the human population is partially determined by their competitive fitness compared to drug-sensitive variants (14, 15, 26). It was reported previously that A/Quebec/144147/2009 (H1N1)pdm09 and rgA/WSN/33 (H1N1) viruses with PA I38T were outcompeted by the respective WT viruses in ST6GalI-MDCK and MucilAir human nasal epithelial cells, respectively (14, 26). A similar trend was observed in our study. We found that WT outcompeted most MUT viruses either after multiple replication cycles or after serial passage in NHBE cells when present at ≥50% in WT:MUT mixtures. The IAV I38T MUT was the only virus that was able to maintain the same ratio or even increase in proportion after multiple replication cycles or serial passage, respectively, when the initial WT:MUT mixture contained at least 75% MUT virus. In contrast, the rest of our baloxavir-resistant MUTs had to be present at ≥90% in the mixture with their respective WT virus in order to detect their impact on baloxavir susceptibility. Our findings are consistent with the PA I38T substitution being the most common resistance-associated substitution observed in patients who are treated with baloxavir.

The effective clinical use of baloxavir relies on rapid identification and monitoring of drug-resistant MUTs present in viral populations. Recently, NGS was implemented as a primary sequencing tool for influenza virus surveillance (24). The ability of NGS to perform whole-genome sequencing with increased sensitivity for minor variants in the virus population compared to conventional methods makes NGS a more precise tool for detection of drug resistance. ddPCR is another highly sensitive method for detection of known resistance markers, and results can be obtained rapidly (<1 day) (34, 35). In this study, we confirmed that both NGS and ddPCR can successfully detect baloxavir-resistant mutations in influenza virus populations when present at low frequencies. We also found that the ratio of the WT to MUT viruses can be measured with a high correlation between these two methods.

Substitutions that confer antiviral resistance are often accompanied by additional substitutions that compensate for the decreased viral fitness (36, 37). Substitutions in the influenza NA protein that confer resistance to NA inhibitors and negatively impact viral replication kinetics are often observed with concomitant hemagglutinin substitutions that ameliorate this effect (36). Our NGS data analysis showed that viral mixtures containing the IAV I38L or IBV I38T MUTs at ratios of ≥90% developed additional PA D394N and E329G substitutions, respectively. We previously found that the PA D394N substitution was present in A/California/04/09 (H1N1)-like virus carrying PA I38L after serial passaging in Calu-3 cells (19). Our previous study also showed that PA D394N did not affect baloxavir sensitivity of the H1N1 virus. Since the frequencies of two additional PA substitutions, IAV PA D394N and IBV PA E329G, increased with increased abundance of the respective MUTs and since they were only detected at later time points during passaging of resistant virus, one can speculate that they may have compensated, at least partially, for the loss in replication capacity conferred by the baloxavir-resistance-associated PA substitutions observed in our cell culture model. Further work on their potential impact on polymerase activity and/or viral replication is needed, as well as monitoring their frequency in surveillance sequence data, which may provide insight regarding the evolution of baloxavir-resistant viruses in the human population. It is unknown if compensatory substitutions selected in cell culture would necessarily enhance viral fitness in humans.

In conclusion, our findings shed new light on the potential impact of influenza baloxavir resistance-associated PA substitutions on viral fitness and drug susceptibility. Our data may provide some insight regarding which baloxavir resistance-associated substitutions pose the greatest risk of potentially becoming widespread. Further assessments of the impact of other PA substitutions, including E23G/K and A37T, on baloxavir susceptibility, viral replication, and polymerase activity in different virus backgrounds (i.e., H1N1 and H3N2) and in animal models are needed. Our experimental findings also provide additional support for the use of ddPCR and NGS for the detection of baloxavir resistance. These newer methods may be very useful for detecting and measuring the prevalence of resistant viruses in mixed viral populations and for identifying additional potentially compensatory mutations.

MATERIALS AND METHODS

Cells and compound.

Madin-Darby canine kidney (MDCK) and human embryonic kidney (HEK 293T) cells were obtained from American Tissue Culture Collection (Manassas, VA) and were maintained as previously described (38). MDCK cells transfected with cDNA encoding human 2,6-sialyltransferase (MDCK-SIAT1 cells) were kindly provided by Mikhail N. Matrosovich and were maintained as described previously (39). Primary NHBE cells were obtained from Lonza (Walkersville, MD) and were grown on 6.5-mm-diameter Transwell membrane supports (Corning, Inc., Corning, NY) at the air-liquid interface in serum-free and hormone- and growth factor-supplemented medium as described previously (40). Once confluent, the NHBE cells were allowed to fully differentiate for at least 3 weeks before use, and only differentiated NHBE cells were used in all the experiments. Baloxavir acid (baloxavir) was obtained from MedChemExpress (Monmouth Junction, NJ). A commonly used >3-fold change in baloxavir EC₅₀ value was chosen as the arbitrary threshold for determining reduced susceptibility in cell culture systems in this study; however, the obtained results do not necessarily reflect WHO criteria for reduced susceptibility determined in other cell culture systems carried out in WHO collaborating centers (20, 41, 42).

Viruses.

Recombinant influenza A/California/04/09 (H1N1)-like (IAV) and B/Victoria/504/2000-like (IBV) viruses were generated by DNA transfection of 293T cells (43). Point mutations to encode amino acid substitutions in the IAV PA protein (I38L, I38T, or E199D) or IBV PA protein (I38T) of the respective WT viruses were introduced using the Quick-Change site-directed mutagenesis kit (Stratagene, La Jolla, CA). Stocks of WT (i.e., IAV WT and IBV WT) and MUT (i.e., IAV I38L, IAV I38T, IAV E199D, and IBV I38T) viruses were prepared by incubation of the viruses in 10-day-old embryonated chicken eggs for 72 h at 37°C for IAV or 33°C for IBV, respectively. The entire PA genes of rescued viruses were sequenced to verify the presence of the desired mutations and absence of additional mutations. NGS sequencing of the starting MUT stocks did not detect IAV PA D394N or IBV PA E329G substitutions. All experimental work was performed in a biosafety level 2 (BSL2) laboratory approved for use of the IAV and IBV strains by the U.S. Department of Agriculture and the U.S. Centers for Disease Control and Prevention.

Virus sequence analysis.

Viral RNA was extracted from virus stocks using the Qiagen RNeasy minikit (Germantown, MD). Samples were then reverse transcribed and analyzed by PCR using universal primers specific for the PA gene as described previously (44). Sequencing was performed by the Research Central Facility for Biotechnology Resources at the U.S. Food and Drug Administration (Silver Spring, MD). DNA sequences were completed, edited, and analyzed using the DNASTAR Laser gene sequence analysis software package (Madison, WI, USA).

Primers and probes.

The primers and probes targeting IAV PA I38 (WT), IAV PA L38 (MUT), IAV PA T38 (MUT), IAV PA D199 (MUT), IBV PA I38 (WT), and IBV PA T38 (MUT) were designed with the DNASTAR Lasergene sequence analysis software package, purchased from Integrated DNA Technologies (Coralville, IA), and are available upon request.

ddPCR.

ddPCR was used to determine the WT:MUT ratios in samples collected before and after virus yield reduction, minigenome assays, multiple replication cycles, and serial passaging in NHBE cells. RNA was extracted from mixed viral stocks and cell culture supernatants using the Qiagen RNeasy minikit. ddPCR was performed using the harvested RNA samples or mixed plasmid stocks using the Bio-Rad one-step ddPCR advanced kit for probes (Hercules, CA) according to the manufacturer’s instructions. Droplets were generated using the Bio-Rad automated droplet generator, and ddPCR was performed using the Bio-Rad C1000 thermal cycler with the following protocol: 50°C for 30 min, 95°C for 10 min, and then 32 cycles at 95°C for 30 s and 55°C (IAV) or 62°C (IBV) for 1 min, followed by a postcycle step at 98°C for 10 min and an infinite hold at 4°C. The droplets were then transferred to the Bio-Rad QX200 reader and analyzed with the FAM (6-carboxyfluorescein) (PA WT) and VIC (PA MUT) channels, and the data were visualized using the Bio-Rad QuantaSoft program. No-template control reactions were used as references to set thresholds.

NGS.

NGS was used to determine the WT:MUT ratios in samples collected after multiple replication cycles and serial passaging in NHBE cells. RNA was extracted from cell culture supernatants using the Qiagen RNeasy minikit (Qiagen). The RNA library was prepared using the NEBNext Ultra II RNA library prep kit for Illumina (New England BioLabs, Ipswich, MA). Fragmentation and priming of the samples were performed in one reaction with the fragmentation buffer and random primer mix provided in the kit, with an average RNA input of 5.6 ng/sample. The double-stranded cDNA synthesis and the library construction were done according to the manufacturer’s protocol. The libraries were analyzed for size distribution and concentration with TapeStation 4200 (Agilent Technologies, Santa Clara, CA) and Qubit 4 (Invitrogen, Waltham, MA), respectively. NGS was performed using NextSeq 500 (Illumina, San Diego, CA) to produce 2 × 75-nt paired-end reads. The raw sequencing reads were then analyzed by using the FDA in-house-developed High-performance Integrated Virtual Environment (HIVE) computing environment (45). Only mutations that were present at ≥5% at nucleotide positions with a read depth of ≥1,500 were considered in order to rule out any false mutations that may be due to sequencing artifacts or background noise.

Infectivity of the recombinant viruses and optimization of the WT:MUT virus mixtures.

The infectivity of the viruses was determined by plaque assay (46). Briefly, confluent cultures of MDCK cells were incubated for 1 h with 10-fold serial dilutions of virus mixtures or samples at 37°C (IAV) or 33°C (IBV). The cells were then washed and overlaid with minimal essential medium containing 0.3% bovine serum albumin (BSA), 0.25% agarose, and 1 μg/mL l-(tosylamido-2-phenyl)methyl chloromethyl ketone (TPCK)-treated trypsin. After 3 days of incubation at 37°C (IAV) or 33°C (IBV), the cells were stained with 0.1% crystal violet in 10% formaldehyde solution and the number of PFU per milliliter and the plaque size of 10 randomly chosen plaques were determined using a Finescale magnifying comparator.

WT and MUT viruses were mixed at the following ratios (percentages) to achieve 10^4.5 PFU: 100:0, 90:10, 75:25, 50:50, 25:75, 10:90, and 0:100. ddPCR was used to quantify the actual ratio of the WT and MUT viral populations, and the input amounts of the WT and MUT viruses were adjusted accordingly. The optimized WT:MUT ratios were used in all subsequent experiments.

Viral replication kinetics.

Growth curves of the WT, MUT, and WT:MUT virus mixtures were determined by infection of NHBE cells for 1 h at a multiplicity of infection (MOI) of 0.01 PFU/cell. After incubation, the cells were washed and then incubated at 37°C (IAV) or 33°C (IBV). Samples were collected at the indicated time points and kept at −80°C until use. RNA was extracted from the samples after each time point, and WT and MUT viral populations were quantified by ddPCR and NGS.

Plaque reduction assay.

Baloxavir sensitivities of the WT, MUT, and WT:MUT virus mixtures were determined by plaque reduction assay as previously described (46). Briefly, confluent MDCK-SIAT1 cells were inoculated with 50 to 100 PFU of influenza virus for 1 h and incubated at 37°C (IAV) or 33°C (IBV). The cells were then washed and overlaid with minimal essential medium containing 0.3% BSA, 0.25% agarose, 1 μg/mL of TPCK-treated trypsin, and baloxavir (ranging from 0.05 nM to 100 nM). After 3 days of incubation at 37°C (IAV) or 33°C (IBV), the overlay was removed, the cells were stained with 0.1% crystal violet in 10% formaldehyde solution, and then the plaques were counted to determine the PFU per milliliter. The baloxavir concentration that caused a 50% decrease in the PFU titer compared to control wells without baloxavir was defined as the EC₅₀. The results of two independent experiments were averaged.

Virus yield reduction assay.

The virus yield reduction assay was performed as previously described (47). Briefly, 24-well plates of confluent NHBE cells were pretreated with baloxavir ranging from 0.05 nM to 500 nM for 2 h and then infected with the WT, MUT, and WT:MUT virus mixtures at an MOI of 0.01 PFU/cell for 1 h at 37°C (IAV) or 33°C (IBV). The cells were then washed, baloxavir-containing basal medium was replaced, and then the cells were cultured at the air-liquid interface at 37°C (IAV) or 33°C (IBV). After 2 days, aliquots of supernatant were taken from the cells and stored at −80°C until use. Supernatant viral yields were determined by plaque assay in MDCK cells. The baloxavir concentration that caused a 50% decrease in the PFU titer in comparison to control wells without baloxavir was defined as the EC₅₀. The results of two independent experiments, each consisting of two replicates, performed on different days were averaged. RNA was extracted from the selected cell culture supernatants, and WT and MUT viral populations were quantified by ddPCR.

Minigenome assay for polymerase activity.

293T cells were pretreated with concentrations of baloxavir that ranged from 0.05 nM to 1,000 nM for 1 h before being transfected with PB1, PB2, PA (WT, MUT, or WT:MUT mixtures), and NP plasmids (at a ratio [micrograms] of 1:1:1:2), a luciferase reporter plasmid (enhanced green fluorescent protein [EGFP] open reading frame in pHW72-EGFP substituted for the firefly luciferase gene), and pGL4.75[hRluc/CMV] vector, which expresses Renilla luciferase (Promega, Madison, WI). WT:MUT plasmid mixtures were mixed at ratios (percentages) of 100:0, 90:10, 75:25, 50:50, 25:75, 10:90, and 0:100 based on the concentration of plasmid stocks to achieve 10⁸ plasmid copies per mixture After 24 h, cell extracts were harvested and lysed, and the luciferase levels were assayed with a dual-luciferase assay system (Promega, Madison, WI). The concentration of baloxavir that caused a 50% reduction in RNP activity compared to control cells without baloxavir was defined as the IC₅₀. The experiments were performed in triplicate. The luciferase levels of the cells transfected with RNPs containing WT:MUT PA mixtures were compared with those transfected with respective WT RNPs. To verify the target WT:MUT PA ratio, ddPCR was performed using the initial WT:MUT PA plasmid mixtures and RNA samples extracted from mock-treated 293T cells 24 h posttransfection.

Serial passaging.

WT, MUT, and WT:MUT virus mixtures were serially passaged 3 times in NHBE cells. Confluent and differentiated NHBE cells were infected at an MOI of 0.01 PFU/cell and incubated at 37°C (IAV) or 33°C (IBV) for 1 h. The virus mixture was then removed, and cells were washed and incubated at 37°C (IAV) or 33°C (IBV). After 48 h, virus-containing supernatant samples were collected from the cells and either passaged onto fresh NHBE cells or stored at −80°C until use. RNA was extracted from the cell culture supernatants, and WT and MUT viral populations were quantified by ddPCR and NGS.

Statistical analysis.

Virus yield, plaque size and number, IC₅₀ and EC₅₀ values, and polymerase activity values were compared by analysis of variance (ANOVA). The degrees of association between plaque reduction, virus yield reduction, and minigenome assays were determined by Spearman correlation test. Probability values of ≤0.05 indicated statistically significant differences.