Breakthrough SARS-CoV-2 Infections in the PROVENT Prevention Trial Were Not Associated With AZD7442 (Tixagevimab/Cilgavimab) Resistant Variants

Abstract

Background

We report spike protein-based lineage and AZD7442 (tixagevimab/cilgavimab) neutralizing activity of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants identified from breakthrough infections in the PROVENT preexposure prophylaxis trial.

Methods

Variants identified from PROVENT participants with reverse-transcription polymerase chain reaction-positive symptomatic illness were phenotypically assessed to determine neutralization susceptibility of variant-specific pseudotyped virus-like particles.

Results

At completion of 6 months' follow-up, no AZD7442-resistant variants were observed in breakthrough coronavirus disease 2019 (COVID-19) cases. SARS-CoV-2 neutralizing antibody titers were similar in breakthrough and nonbreakthrough cases.

Conclusions

Symptomatic COVID-19 breakthrough cases in PROVENT were not due to resistance-associated substitutions in AZD7442 binding sites or lack of AZD7442 exposure.

Clinical Trials Registration

NCT04625725.

Symptomatic COVID-19 breakthrough cases in the PROVENT preexposure prophylaxis trial were not due to resistance-associated substitutions in AZD7442 (tixagevimab/cilgavimab) binding sites or lack of AZD7442 exposure.

Although coronavirus disease 2019 (COVID-19) vaccines have been widely effective, a substantial population of immunocompromised individuals have an inadequate response to vaccination and remain at risk of severe disease [1]. Moreover, emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants that can escape from neutralizing antibodies (nAbs) may compromise the effectiveness of vaccines and monoclonal antibody (mAb)-based therapies [2–4].

Primary results from the phase 3 PROVENT preexposure prevention trial showed that a single 300-mg intramuscular dose of AZD7442 (tixagevimab and cilgavimab) significantly reduced symptomatic COVID-19 by 77% versus placebo and was well tolerated [5]. This result led to emergency use authorization of AZD7442 by the US Food and Drug Administration and UK Medicines and Healthcare products Regulatory Agency, and approval by the European Medicines Agency for preexposure prophylaxis of COVID-19 in immunocompromised individuals. There were few immunocompromised participants in PROVENT, although reports of the effectiveness of AZD7442 for prevention of COVID-19 in this population in clinical practice have been published [6, 7].

Monitoring for potential emergence of AZD7442-resistant variants has continued throughout clinical development and is ongoing to identify potential resistance-associated substitutions that escape one or both mAb components. The primary objective of the current analysis was to evaluate the S-protein–based lineage of SARS-CoV-2 variants identified from breakthrough infections in the phase 3 PROVENT trial, and evaluate the corresponding in vitro potency of AZD7442 and its components, to determine whether resistant variants had arisen during the course of AZD7442 administration. Pharmacodynamic exposure in breakthrough infections was additionally assessed through nAb characterization in serum at scheduled visits.

METHODS

Study Design and Procedure

PROVENT is a phase 3, double-blind, multicenter trial (ClinicalTrials.gov No. NCT04625725) in which 3460 and 1737 participants were randomized to AZD7442 and placebo, respectively [5]. Details of the PROVENT trial design, including the participant population and interventions, sample collection, and genotypic analysis of S-protein variants, have previously been reported [5].

Briefly, nasopharyngeal swabs identified as positive by SARS-CoV-2 reverse-transcription polymerase chain reaction (RT-PCR) assay at illness visit 1 (or first available illness visit) were subjected to viral nucleic acid extraction. Following extraction, the full-length S gene (amino acids 1–1274) was amplified and sequenced using a single-tube population-based RT-PCR method and validated next-generation sequencing assay (Monogram Biosciences). Amino acid polymorphisms were identified (consensus reported at ≥25% frequency; minor variant reported at ≥3% frequency) and compared with the SARS-CoV-2 Wuhan-Hu-1/2019 strain. An S-protein–only lineage identification tool, Hedgehog version 1.0.5 (https://github.com/cov-lineages/hedgehog), was utilized to classify SARS-CoV-2 S-protein sequences to PANGO lineages, version 1.2.86 [8].

nAb titers against the ancestral SARS-CoV-2 strain were assessed in serum samples of participants who had a negative or missing RT-PCR test at baseline collected at scheduled study visits (days 1, 8, 29, 58, 92, and 183) in a validated live neutralization assay (plaque reduction neutralization test [PRNT₈₀]; Viroclinics Biosciences). nAb results were censored at the time of COVID-19 illness, vaccination, or unblinding, and were stratified by breakthrough infection status. Immunogenicity analyses were grouped according to the actual treatment received. Further details of the nAb and AZD7442 pharmacokinetic assessments in PROVENT have previously been reported [5]. A 99.5% tolerance interval for log₂ nAb titers was obtained for nonbreakthrough patients at each time point.

In Vitro Microneutralization of SARS-CoV-2 S-Protein Pseudotyped Lentivirus

SARS-CoV-2 S-protein variants identified from participants with RT-PCR–positive symptomatic illness in PROVENT were incorporated into SARS-CoV-2 Wuhan-Hu-1/2019 (GenBank accession number, NC_045512) + D614G S-protein pseudotyped lentiviruses and assessed for susceptibility to AZD7442 and its component mAbs via microneutralization assay. Generation of S-protein pseudotyped lentivirus and pseudovirus microneutralization assays were performed as previously reported [9], with several modifications. Briefly, serial dilutions of mAbs were prepared in a 384-well microtiter plate and preincubated with pseudovirus for 30 minutes at 37°C. Following incubation, AD293-ACE2-ARCB cells that stably express ACE2 were added to the wells and the plates were incubated for 48 hours at 37°C.

Quantification of relative infectivity was determined by measuring the luminescence of the expressed luciferase activity as relative luminescence units on an EnVision 2105 Multimode Plate Reader (Perkin Elmer) using the Steady-Glo Luciferase Assay System (Promega), according to the manufacturer's recommendations. Percent inhibition was calculated by normalization to the virus-only control. For the research-grade assay, half-maximal inhibitory concentration (IC₅₀) values were determined by nonlinear regression using GraphPad Prism software, version 9.0.0 (GraphPad Software).

The average IC₅₀ value for each mAb was determined from at least 2 independent experiments. To support regulatory submission, the assay methodology underwent a qualification based on specificity, intermediate precision, interassay precision, and intraassay precision (further details included in Supplementary Methods and Supplementary Table 1). For assays performed using the qualified pseudovirus microneutralization assay, IC₅₀ values were subsequently determined by fitting a 4-parameter logistics model to the replicate relative luminescence signal obtained across a dilution of mAb concentrations and back-calculating the X₅₀ (the concentrations that give 50% inhibition between the virus-only and the no-virus control) using ordinary least squares in RStudio, version 4.0.2 (PBC) (Supplementary Table 2).

The corresponding fold-change IC₅₀ against each variant was determined for each mAb individually and together (tixagevimab and cilgavimab) compared with the reference SARS-CoV-2 Wuhan-Hu-1/2019 + D614G S-protein pseudovirus. Reference IC₅₀ was measured within each run of the research-grade assay, whereas an average reference IC₅₀, determined during qualification, was utilized for all qualified assay runs.

RESULTS

Variants Identified From Breakthrough Symptomatic SARS-CoV-2 Infections in the Phase 3 PROVENT Trial Remained Susceptible to AZD7442

PROVENT data were available from trial recruitment (21 November 2020 to 22 March 2021) until all participants had reached their scheduled 6-month follow-up (day 183) visit (14 October 2021) [5]. Baseline characteristics of participants with COVID-19 were largely balanced between treatment arms, except that a higher proportion of AZD7442 participants had COVID-19 comorbidities and risk of severe disease at baseline, compared with placebo (Supplementary Table 3); such imbalances were not observed in the overall PROVENT population [5]. SARS-CoV-2 isolates from participants in the AZD7442 and placebo groups with symptomatic COVID-19 were sequenced to identify S-protein–based lineage, and co-occurring substitutions were phenotypically evaluated for their susceptibility to AZD7442 by S-protein pseudotyped lentiviral-based in vitro microneutralization assay. Sequencing data were available for 20 of 45 participants at illness visit day 1 through day 183 (nasopharyngeal swabs from 18 participants with COVID-19 were not available for sequencing; 7 participants had failed sequencing results), with 14 unique full-length S-protein coding sequences obtained across 7 different SARS-CoV-2 lineages and minor variant substitutions present in 3 participants (Table 1 and Supplementary Tables 4 and 5). None of the observed lineages were preferentially identified within the AZD7442 treatment group compared with the placebo group. Susceptibility analysis depicted no reduction in AZD7442 neutralization activity against any of the lineages (Table 1). A 47-fold reduction in neutralization activity of cilgavimab against the Delta variant (B.1.617.2) was observed; however, all identified lineages were associated with <5-fold reduction in susceptibility with AZD7442 compared with the reference strain (Table 1).

Table 1.

Summary of AZD7442 In Vitro Neutralization Against SARS-CoV-2 Spike-Based Lineages Detected at Illness Visit(s) Through Day 183 in PROVENT

WHO Designation	SARS-CoV-2 Spike-Based Lineagea,b	Participants, No. (%)c			IC50, ng/mL (Fold-Reduction in Susceptibility)d,e
		Placebo (n = 33)	AZD7442 (n = 12)	Total (n = 45)	Tixagevimab	Cilgavimab	AZD7442
None, ancestral	A_22f	2 (6.0)	2 (16.7)	4 (8.9)	1.40g (1.00)	4.46g (1.00)	2.26g (1.00)
None, ancestral	A_1	0 (0.0)	1 (8.3)	1 (2.2)	0.94 (0.54)	4.67 (0.81)	1.23 (0.35)
Alpha	B.1.1.7_1	5 (15.2)	0 (0.0)	5 (11.1)	3.78 (5.64)	11.89 (3.40)	9.05 (4.23)
Beta	B.1.351	0 (0.0)	1 (8.3)	1 (2.2)	13.58 (6.95)	5.32 (0.46)	11.73 (2.24)
Delta	B.1.617.2h	5 (15.2)	1 (8.3)	6 (13.3)	3.37 (2.99)	140.26 (47.07)	6.07 (2.79)
Epsilon	B.1.429	0 (0.0)	2 (16.7)	2 (4.4)	0.83 (1.23)	8.05 (2.22)	2.05 (0.95)
Iota	B.1.526	1 (3.0)	0 (0.0)	1 (2.2)	9.41 (8.61)	3.26 (0.58)	5.24 (1.87)
Not availablei		17 (51.5)	1 (8.3)	18 (40.0)	NA
Sequencing failed		3 (9.1)	4 (33.3)	7 (15.6)	NA

SARS-CoV-2 spike-based lineage is a human readable set description, combining the most recent common ancestor of all lineages within a given set (precision) and a unique identifier for a given set assignment. Analysis of all participants through to day 183 (14 October 2021).

Abbreviations: COVID-19, coronavirus disease 2019; IC₅₀, half-maximal inhibitory concentration; mAb, monoclonal antibody; NA, not applicable; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; WHO, World Health Organization.

A spike-only version of Pangolin COVID-19 lineage assigner (Hedgehog version 1.0.5) was used to classify SARS-CoV-2 spike nucleotide sequences to current Pango lineages (version 1.2.86) or sets of lineages [8].

Results for individually evaluated pseudotyped viruses expressing the full contextual set of substitutions identified within the SARS-CoV-2 spike variant proteins of each lineage are available in Supplementary Table 6.

Distribution of SARS-CoV-2 spike variant data by treatment group based on available illness visit day 1 data through to last participant day 183; first episode illness visit day 1 sequence data available.

Reduction in in vitro potency across indicated variants; testing was performed using the recombinant spike pseudotyped virus microneutralization assay; values depict the IC₅₀ of mAb required for a 50% reduction in infection compared with wild-type reference strain (SARS-CoV-2 Wuhan-Hu-1/2019 + D614G); mean fold-change IC₅₀ reported for variants evaluated in research-grade assay n ≥ 2; the degree of reduced susceptibility (fold-change) is indicated by normal or bold text: no reduction <5 or minimal reduction ≥5 to <10 (normal text), moderate reduction ≥10 to <100 (bold text).

AZD7442 is a combination of tixagevimab and cilgavimab mAbs at a 1:1 ratio.

Includes subvariant B.1.1.315_1.

Mean IC₅₀ (n = 24) from research-grade assays depicted for reference strain (Wuhan-Hu-1/2019 + D614G).

Includes subvariants B.1.617.2_1, _3, _4, AY.3., and AY.103.

Sample not available for testing at central laboratory, quantity not sufficient, participant consent withdrawal.

Serum nAb Exposure Among Participants With Breakthrough SARS-CoV-2 Infections in PROVENT

Pharmacodynamic activity was assessed through nAb activity in serum samples and utilized as a surrogate for AZD7442 exposure. Overall, study participants receiving AZD7442 who were RT-PCR–negative or missing at baseline had nAb titers below the lower limit of quantification of the assay prior to administration of AZD7442. nAb responses peaked by the day 8 specimen collection, consistent with prior reports [5, 10], and remained elevated over baseline for at least 6 months in study participants with no symptomatic infection or nonstudy COVID-19 vaccination. nAb titers of all AZD7442 recipients with breakthrough COVID-19 infection were within 99.5% tolerance intervals of AZD7442 recipients with no reported infections (Figure 1), indicating that the nAb titers were equivalent in the 2 groups. Additionally, SARS-CoV-2 viral loads assessed at time of symptomatic onset were similar by treatment arm (Supplementary Figure 1).

Figure 1.

Serum nAb titers among participants with and without breakthrough COVID-19 in PROVENT. nAb results were censored at the time of COVID-19 illness, vaccination, or unblinding, and were stratified by breakthrough infection status. The box denotes IQR and the line inside the box denotes median. Any points >1.5 × IQR from the box were considered outliers and are not displayed. The whiskers that extend from the box indicate the minimum and maximum after removing the outliers. Boxplots were created using the log₂-transformed nAb titers. Day 1 represents the study day with the last nonmissing value taken prior to the first dose. Titer values measured as < LLOQ were imputed to a value that is half of the LLOQ. Titer values measured as > ULOQ were imputed at the ULOQ value. Abbreviations: COVID-19, coronavirus disease 2019; IQR, interquartile range; LLOQ, lower limit of quantification; nAb, neutralizing antibody; ULOQ, upper limit of quantification.

DISCUSSION

The phase 3 PROVENT trial demonstrated the efficacy of a single dose of AZD7442 for the prevention of COVID-19 [5]. The combination mAb approach of AZD7442 is advantageous as it provides redundancy in protection, in case of virus mutation or escape from a single component of the combination [10]. Indeed, in this analysis of breakthrough infections from PROVENT, there was no indication of an increased proportion of any variant in the AZD7442 group compared with the placebo group, and the susceptibility analysis depicted no reduction in neutralization activity of AZD7442 against any of the lineages identified. Taken together, these data suggest that SARS-CoV-2 breakthrough infections seen in PROVENT were not due to AZD7442 resistance, and these results support the strategy of using 2 potent SARS-CoV-2 neutralizing mAbs in combination for preexposure prophylaxis to SARS-CoV-2 [10]. However, transferability of these findings is limited by the emergence of new viral variants since the study period covered in this analysis, including some sublineages of the Omicron variant that may escape the strong neutralizing potency of AZD7442 [2, 4]. This reinforces the need for ongoing viral surveillance to monitor the effectiveness of immune therapies against newly emerging viral variants.

Following AZD7442 administration, nAb levels peaked at the earliest measured time point and remained above baseline through at least 6 months. Median nAb levels were similar between AZD7442 participants with or without symptomatic COVID-19, indicating that participants with breakthrough infections had AZD7442 exposure levels similar to those without COVID-19. By comparison, correlates of protection for nAb levels following AZD7442 administration are not established. Similarly, a study of a respiratory syncytial virus-neutralizing mAb examined correlates of protection between neutralizing activity and clinical end points and suggested that there is an upper threshold level of protection, beyond which efficacy plateaus regardless of further nAb increases [11]. These data are consistent with observations of correlates of protection with COVID-19 vaccines [12], where no titer threshold level of binding nor nAb was found to mediate complete protection against symptomatic COVID-19; although these observations may be confounded by cellular responses to vaccination, such cellular responses are not thought to prevent symptomatic illness but rather limit severity of disease [13].

Approximately 60% (7/12) of breakthrough cases from PROVENT had sequence data available and clinical samples represented 6 months of follow-up, and there was potential for AZD7442-resistant viral variants to emerge as mAb concentration levels waned over time. Therefore, surveillance for the emergence of new viral variants (beyond the 6-month surveillance period used in the current analysis) is ongoing. Furthermore, there were few immunocompromised individuals included in the PROVENT trial (n = 196; 3.8% of the total trial population), and so careful monitoring of variants in immunocompromised individuals who have received AZD7442, yet experienced breakthrough infections, is critical. AZD7442 is being evaluated as an outpatient treatment in individuals with mild-to-moderate COVID-19 in the phase 3 TACKLE trial [14], in which monitoring for treatment-emergent variants is also being conducted.

In conclusion, there was no evidence that breakthrough COVID-19 following AZD7442 as preexposure prophylaxis in the phase 3 PROVENT prevention trial was due to AZD7442-resistant variants. Pharmacodynamic activity of AZD7442, as indicated by nAb levels, was consistent across participants with and without COVID-19 breakthrough infections. These results support the importance of continued molecular surveillance to inform clinical practice and ensure the effectiveness and optimal protection of AZD7442 against COVID-19.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments . Medical writing support was provided by Shaun W. Foley, BSc (Hons) and Rob Campbell, PhD, and editorial support was provided by Jess Galbraith, BSc, all of Core Medica, London, UK, supported by AstraZeneca according to Good Publication Practice guidelines. AstraZeneca would like to thank Danny Duijsings and Viroclinics Biosciences team for performing the PRNT₈₀ microneutralization assay using serum samples from the PROVENT clinical trial, as well as Jonathan Toma and Monogram Biosciences for performing the next-generation sequencing assay using nasopharyngeal swab samples from the PROVENT clinical trial.

Financial support. This work was supported by AstraZeneca and includes data from the PROVENT trial that was funded by AstraZeneca and the United States Government. AZD7442 is being developed with support from the United States Government, including federal funds from the Department of Health and Human Services, Administration for Strategic Preparedness and Response, Biomedical Advanced Research and Development Authority in partnership with the Department of Defense, and Joint Program Executive Office for Chemical, Biological, Radiological and Nuclear Defense (contract No. W911QY-21-9-0001).

Presented in part: IDWeek 2022, 19–23 October 2022, Washington, DC, poster.