Cell line, viruses, and receptor
Vero E6 cells were obtained from ATCC, CRL-1586 and maintained in Minimal Eagles Medium (MEM) supplemented with 10% fetal bovine serum (FBS), 1% glutamine, and 1% penicillin and streptomycin (P/S). THP-1 cells (ATCC TIB-202) maintained in Roswell Park Memorial Institute (RPMI) medium, supplemented with 10% FBS, 1% glutamine, 1% P/S, 1% HEPES, and 50µM β-ME. HEK 293T/ACE2 cells were obtained from Drs. Michael Farzan and Huihui Mu at the Scripps Research Institute (Jupiter, FL, USA). For the challenge phase of the study, the SARS-CoV-2 (USA-WA-1/2020) passage 4 (P4) isolate was obtained from Biodefense and Emerging Infections Research Resource Repository (catalog number NR-52281, BEI Resources, GenBank accession number MN985325.1). For the in vitro neutralization assay, the SARS-CoV-2 (USA-WA-1/2020) isolate was obtained from the Center for Disease Control and provided by Dr. Matthew Frieman, University of Maryland. Histidine-tagged human ACE2 receptor was purchased from Sino Biologics (Beijing, CHN). Matrix-M™ adjuvant was provided by Novavax, AB (lot number M1-111, Uppsala, SWE)52.
NVX-CoV2373 spike glycoprotein
The SARS-CoV-2 S vaccine was constructed from the full-length, wild-type SARS-CoV-2 S glycoprotein based on the GenBank gene sequence MN908947 nucleotides 21563–25384. The native S protein was modified by mutating the putative furin cleavage site (682-RRAR-685 to 682-QQAQ-685) in the S1/S2 cleavage domain to confer protease resistance. Two additional proline amino acid substitutions were inserted at positions K986P and V987P (2P) within the heptad repeat 1 (HR1) domain to stabilize SARS-CoV-2 S in a prefusion conformation53. The synthetic transgene was codon optimized and engineered into the baculovirus vector (BV2373) for expression in Spodoptera frugiperda (Sf9) insect cells (GenScript, Piscataway, NJ, USA). NVX-CoV2373 spike trimers were detergent extracted from the plasma membrane with phosphate buffer containing TERGITOL NP-9, clarified by centrifugation, and purified by TMAE anion exchange and lentil lectin affinity chromatography. Purified NVX-CoV2373 (547 µg mL− 1, lot number BV2373-16APR20) was formulated in 25 mM sodium phosphate (pH 7.2), 300 mM NaCl, and 0.02% (v/v) polysorbate and supplied frozen at -80ºC ± 10ºC54.
Animal ethics statement
The immunization and challenge phases of the study complied with all applicable sections of the Final Rules of the Animal Welfare Act regulations (9 CFR Parts 1, 2, and 3) and Guide for the Care and Use of Laboratory Animals - National Academy Press, Washington D. C. 8th Edition, 2011 (The Guide). The study was conducted at the Texas Biomedical Research Institute (Texas Biomed, San Antonio, TX, USA), an AAALAC (Association for the Assessment and Accreditation of Laboratory Animal Care) accredited facility. The work was conducted in accordance with a protocol approved by Texas Biomed’s Institutional Animal Care and Use Committee.
Human ethics statement
The Phase 1 vaccine study was previously described19. Healthy 18-59-year-old men and non-pregnant women were included in the study. Previously infected individuals were excluded. With the exception of 6 sentinel participants vaccinated in an open-label manner, the remaining 125 participants were randomly assigned to vaccine and placebo groups in a blinded fashion. All subjects signed informed consent and safety oversight was monitored by a data monitoring board.
Animal husbandry
Animals were housed individually in stainless steel cages with wire mesh bottoms. Animals were fed commercially available certified primate diet from Purina Mills 5048 (LabDiet) and provided water ad libitum from an institutional watering system that was analyzed monthly for impurities. Environmental conditions included 12 hour light and 12 hour dark cycle with controlled temperature (74ºF ± 10ºF) and humidity (30–70% RH). Cages were cleaned daily.
Twenty-four experimentally naïve rhesus macaques (Macaca mulatta) of Chinese origin were sourced from Envigo (Alice, TX, USA). Animals were screened and determined to be negative for Simian Immunodeficiency Virus (SIV), Simian T-Lymphotropic Virus-1 (STLV-1), Simian Varicella Virus (SVV) and Macacine herpesvirus 1 (Herpes B virus), and Simian Retrovirus (SRV1 and SRV2) by polymerase chain reaction (PCR), and negative for Trypanosoma cruzi. Rectal swabs were collected and tested for Shigella, Campylobacter, Salmonella, and Yersinia. Pharyngeal swabs were used to test for Bordetella bronchiseptica. All animals were tested and verified to be negative for tuberculosis.
The vaccination phase of the study was performed in the Texas Biomed Animal Biosafety Level 2 (ABSL-2) facility. Following the immunization phase of the study, animals were transferred and acclimated for 7 days in the Texas Biomed ABSL-3 facility prior to challenge. Animals were monitored a minimum of twice daily for the duration of the study.
Study blinding
This study was blinded (assignment to vaccinated/immunized versus placebo group) to avoid bias in evaluation, euthanasia, gross pathology assessment, and qRT-PCR assay outcome. All staff performing in vitro assays were blinded to the animal vaccine dosage and to whether the animal received vaccine or placebo while performing assays and analysis.
Study design
Animals were randomly assigned to groups, with stratification across age and gender, using a computerized randomization procedure. Twenty-four (12 male and 12 female) rhesus macaques, within the age range of > 3 to < 8-year-olds and weight range ≥ 3.67 kg to ≤ 10 kg, were randomized into four immunization groups and two placebo groups. NVX-CoV2373 was formulated with 50µg Matrix-M on the day of immunization. The placebo groups received formulation buffer. Groups 1 (1 male and 1 female) received placebo in two doses spaced 21 days apart (study day 0 and 21) and group 4 (1 male and 1 female) received placebo in one dose (study day 0). Group 2 (2 females and 3 males) received 5µg NVX-CoV2373 + 50µg Matrix-M and group 3 (2 females and 3 males) received 25µg NVX-CoV2373 + 50µg Matrix-M in two doses spaced 21 days apart (study day 0 and 21). Group 5 (3 females and 2 males) received 5µg NVX-CoV2373 + 50µg Matrix-M and group 6 (3 females and 2 males) received 25µg NVX-CoV2373 + 50µg Matrix-M in one dose (study day 0). Injections (0.5 mL) were administered in the thigh muscle.
Animals were sedated by intramuscular (IM) administration of Telazol (2–8 mg kg− 1, IM) prior to vaccination, collection of blood samples, virus challenge, collection of nasal swabs, nasal washes, and bronchoalveolar lavage (BAL). For serologic assessments, serum was collected on study day 0 prior to immunization and day 21, and day 31 or 32 after the first immunization and stored at -80ºC until assayed. Nasal washes, nasal pharyngeal swabs, and BAL were collected on study day 31/32, prior to challenge.
Anti-spike IgG and IgA ELISA
Serum, nasal wash, and BAL anti-SARS-CoV-2 spike (S) protein IgG titers were determined by ELISA. Briefly, 96-well plates (Thermo Fisher Scientific, Rochester, NY, USA) were coated with 1.0 µg mL− 1 of SARS-CoV-2 S protein (BV2373, Lot# 16Apr20, Novavax, Inc. Gaithersburg, MD, USA). Plates were washed with phosphate buffered Tween (PBS-T) and non-specific binding was blocked with TBS Startblock blocking buffer (Thermo Fisher Scientific, Rochester, NY, USA). Serum samples were serially diluted 3-fold starting with a 1:100 dilution and BAL and nasal wash samples were serially diluted 2-fold starting with a 1:2 dilution, then added to the coated plates and incubated at room temperature for 2 hours. For IgG ELISA, plates were washed with PBS-T, then incubated with horseradish peroxidase (HRP)-conjugated mouse anti-monkey IgG (catalog number 4700-05, Southern Biotech, Birmingham, AL, USA) for 1 hour. For IgA ELISA, plates were washed with PBS-T and mouse anti-monkey IgA (catalog number MCA2553, Bio-Rad, Hercules, CA, USA) was added for 1 hour followed by washing with PBS-T, then incubation with HRP-conjugated goat anti-mouse IgG (catalog number 1030-05, Southern Biotech). Plates were then developed with 3,3’,5,5’-tetramethylbenzidine (TMB) peroxidase substrate (Sigma, St. Louis, MO, USA). Reactions were stopped with TMB stop solution (ScyTek Laboratories, Inc. Logan, UT, USA). Plates were read at OD 450 nm with a SpectraMax Plus plate reader (Molecular Devices, Sunnyvale, CA, USA). EC50 values and endpoint titer values were calculated by 4-parameter fitting using SoftMax Pro 6.5.1 GxP software. Individual animal anti-S IgG or IgA titers, and group geometric mean titer (GMT) and 95% confidence interval (95% CI) were plotted using GraphPad Prism 9.0 software. For serum titers below the assay lower limit of detection (LOD), a titer of < 1000 (starting dilution) was reported and a value of “50” assigned to the sample to calculate the group mean titer. For BAL and nasal wash titers below the assay LOD, a titer of < 2 (starting dilution) was reported and a value of “1” assigned to the sample to calculate the group mean titer.
Human angiotensin converting enzyme 2 (hACE2) receptor blocking antibody
Human ACE2 receptor blocking antibody titer was determined by ELISA. Ninety-six well plates were coated with 1.0 µg mL− 1 SARS-CoV-2 rS protein (BV2373, lot no. 16Apr20, Novavax, Inc., Gaithersburg, MD, USA) overnight at 4°C. Sera were serially diluted 2-fold starting with a 1:20 dilution and were added to coated wells for 1 hour at room temperature. After washing, 30 ng mL− 1 histidine-tagged hACE2 (Sino Biologics, Beijing, CHN) was added to wells for 1 hour at room temperature. HRP-conjugated mouse anti-histidine-tag IgG (1:4000) (catalog number 4603-05, Southern Biotech, Birmingham, AL, USA) was added for 1 hour followed by addition of TMB substrate. Plates were read at OD 450 nm with a SpectraMax Plus plate reader (Molecular Devices, Sunnyvale, CA, USA) and data analyzed with SoftMax Pro 6.5.1 GxP software. The % Inhibition for each dilution for each sample was calculated using the following equation in the SoftMax Pro program: 100-[(MeanResults/ControlValue@PositiveControl)*100].
Serum dilution versus % Inhibition plot was generated, and curve fitting was performed by 4-parameter logistic (4PL) curve fitting to data. Serum antibody titer at 50% inhibition (IC50) of hACE2 to SARS-CoV-2 S protein was determined in the SoftMax Pro program. The group GMT and 95% CI and individual animal titers were plotted using GraphPad Prism 9.0 software. For a titer below the assay lower limit of detection (LOD), a titer of < 20 (starting dilution) was reported and a value of “10” assigned to the sample to calculate the group mean titer.
SARS-CoV-2 neutralizing antibody assay
The SARS-CoV-2 neutralizing antibody assay was conducted in a select agent ABSL-3 containment facility at the University of Maryland, School of Medicine. Sera were diluted 1:20 in Vero E6 cell growth media and further serially diluted 1:2 to 1:40,960. SARS-CoV-2 (multiplicity of infection (MOI) of 0.01 pfu per cell) was added to the wells for 60 min at 37ºC. Vero E6 media was used as a negative control. Each serum dilution was assessed microscopically for inhibition of virus cytopathic effect (CPE) on Vero E6 cells. The endpoint titer was reported as the reciprocal of the dilution at which 99% CPE was observed at 3 days post infection19,54.
Preparation of the SARS-CoV-2 challenge stock
A fourth cell-culture passage (P4) of SARS-CoV-2 isolate USA-WA1/2020 was obtained from Biodefense and Emerging Infections Research Resources Repository (catalog number NR-52281, BEI Resources, GenBank accession number MN985325.1). Live virus stock was prepared in the Texas Biomed ABSL-3 containment facility. The stock virus was passaged for a fifth time (P5) in Vero E6 cells at a MOI of 0.001 to produce the master virus stock. The master stock was again passaged in Vero E6 cells at a MOI of 0.02 (P6) to produce the challenge stock. The P6 challenge stock had a titer of 2.10 × 106 pfu mL− 1 (Lot No. 20200320) and was stored 500 µL aliquots at -65ºC in Dulbecco’s modified essential media (DMEM) and 10% fetal bovine serum. The identity of the challenge stock was confirmed to be SARS-CoV-2 by deep sequencing and was confirmed to be identical to the published sequence (GenBank: MN985325).
SARS-CoV-2 challenge
Vaccinated and placebo animals were transferred from the ABSL-2 facility on study day 31/32 to the ABSL-3 facility and acclimated for 7 days. On the day of challenge (study day 38), animals were sedated and challenged with a total target dose of 1.05 × 106 pfu in 500 µL. The challenge dose was equally administered by the intranasal (IN) route 5.25 × 105 pfu in 250 µL and intra-tracheal (IT) route 5.25 × 105 pfu in 250 µL. IN administration was performed with an atomization device (Teleflex Intranasal Mucosal Atomization Device LMA MAD Nasal Device, Morrisville, NC, USA) and IT delivery was performed with Tracheal Mucosal Atomization Device (Teleflex Laryngo-Tracheal Mucosal Atomization Device LMA MADGIC, Morrisville, NC, USA). To confirm the challenge dose, aliquots of the challenge samples were collected prior to challenging the first animal and last animal and stored at ≤ -65ºC. A neutral red agarose overlay and conventional plaque assay were used to confirm the titer of the challenge dose. The actual pre- and post-challenge titers were 1.80 × 106 pfu mL− 1 and 7.83 × 105 pfu, respectively.
Sample collection for SARS-CoV-2 RNA quantification
Nasal pharyngeal swab collection. Animals were sedated and nasal pharyngeal swabs were collected prior to challenge (study day 31/32) and on 2, 3, 4, 6, and 7–8 days post infection (dpi). After collection, swabs were placed in a tube containing viral transport medium (VTM), then stored at ≤-60ºC until processing.
Bronchoalveolar lavage (BAL) collection. BAL aspirates were collected prior to challenge (study day 31/32) and on 2, 4 and 7–8 dpi. Animals were sedated and the trachea visualized with a laryngoscope. A sterile rubber feeding tube with stylet was inserted into the trachea and into the airway until it met slight resistance. Up to 80 mL of warm (< 40ºC) sterile saline, divided into multiple aliquots, was instilled through the tube. Aspirated fluid was dispensed into sterile vials with VTM and stored at ≤-60ºC until batch processed.
Nasal wash collection. Nasal washes were collected prior to infection (study day 31/32) and 2, 4, and 7–8 dpi. Animals were sedated and a syringe with a flexible tipped 20-22-gauge intravenous (IV) catheter was inserted into the nostril passage and a volume of 2.5-5mL of sterile saline instilled. Samples were collected in sterile conical tubes containing VTM and stored at ≤-60ºC until batch processed.
Tissue collection. Tissues were collected 7–8 dpi (study days 45–46) at the scheduled necropsy from the upper, middle and lower lobes of the lung; nasal cavity; and trachea. Tissues were weighed and stored at 80ºC ± 10ºC until batch processed. RNA was extracted analyzed for the presence of SARS-CoV-2 RNA via qRT-PCR targeting the N1 gene.
Quantification of virus load in nasal swabs/washes, BAL, and tissues
Genomic (g)RNA virus. Samples were assessed for viral load by qRT-PCR. A 250 µL aliquot of VTM inactivated in TRIzol LS reagent (catalog number 10296010, ThermoFisher Scientific) was used for isolation of total RNA. For total viral RNA, qRT-PCR targeting the nucleocapsid gene (N1) was run on duplicate samples and results reported as genome equivalents (GE) mL− 1 for nasal washes/swabs and BAL. For tissue samples, results are reported as GE µg− 1 for tissue homogenates.
Subgenomic (sg)RNA virus. Replicating virus load by qRT-PCR targeting the subgenomic envelope (E) gene RNA in 250 µL aliquot of nasal swabs, nasal washes, and BAL aspirates. The forward and reverse primers, probe, cycling conditions, and Master Mix included:
SUBGEN-FORWARD: CGATCTCTTGTAGATCTGTTCTC
E_Sarbeco_R2 Reverse Primer: ATATTGCAGCAGTACGCACACA
Probe (Thermo): FAM-MGB: ACACTAGCCATCCTTACTGCGCTTCG
TaqPath™ 1-Step RT-qPCR Master Mix, CG (catalog number A15299, ThermoFisher Scientific). Cycling parameters were 25°C 2 minutes, 50°C 15 minutes, 95°C 2 minutes; Amplification 40 × 95°C 3 seconds, 60°C 30 seconds.
Pseudovirus neutralizing antibody assay.
SARS-CoV-2 neutralization was assessed with spike-pseudotyped virus infection of HEK 293T/ACE2 cells as a function of reduction in luciferase (Luc) reporter activity. HEK 293T/ACE2 cells were maintained in DMEM containing 10% fetal bovine serum, 25 mM HEPES, 50 µg mL− 1 gentamycin and 3 µg mL− 1 puromycin. An expression plasmid encoding codon-optimized full-length spike of the Wuhan-1 strain (VRC7480), was provided by Drs. Barney Graham and Kizzmekia Corbett at the Vaccine Research Center, National Institutes of Health (USA). The D614G amino acid change was introduced into VRC7480 by site-directed mutagenesis using the QuikChange Lightning Site-Directed Mutagenesis Kit (catalog number 210518, Agilent Technologies). The mutation was confirmed by full-length spike gene sequencing. Pseudovirions were produced in HEK 293T/17 cells (ATCC cat. no. CRL-11268, Manassas, VA, USA) by transfection using Fugene 6 (catalog number E2692, Promega, Madison, WI, USA) and a combination of spike plasmid, lentiviral backbone plasmid (pCMV ΔR8.2) and firefly Luc reporter gene plasmid (pHR' CMV Luc) in a 1:17:17 ratio55. Transfections were allowed to proceed for 16–20 hours at 37ºC. Medium was removed, monolayers rinsed with growth medium, and 15 mL of fresh growth medium added. Pseudovirus-containing culture medium was collected after an additional 2 days of incubation and was clarified of cells by low-speed centrifugation and 0.45µm micron filtration and stored in aliquots at -80ºC. TCID50 assays were performed on thawed aliquots to determine the infectious dose for neutralization assays (RLU 500-1000x background, background 50–100 RLU).
For neutralization, a pre-titrated dose of virus was incubated with 8 serial 5-fold dilutions of serum samples in duplicate in a total volume of 150 µL for 1 h at 37ºC in 96-well flat-bottom poly-L-lysine-coated Biocoat plates (catalog number 354461, Corning, NY, USA). Cells were suspended using TrypLE Select Enzyme solution (Thermo Fisher Scientific) and immediately added to all wells (10,000 cells in 100 µL of growth medium per well). One set of 8 control wells received cells + virus (virus control) and another set of 8 wells received cells only (background control). After 66–72 h of incubation, medium was removed by gentle aspiration and 30 µL of Promega 1X lysis buffer was added to all wells. After a 10 minute incubation at room temperature, 100 µL of Bright-Glo luciferase reagent was added to all wells. After 1–2 minutes, 110 µL of the cell lysate was transferred to a black/white plate (Perkin-Elmer). Luminescence was measured using a PerkinElmer Life Sciences, Model Victor2 luminometer. Neutralization titers are the serum dilution at which relative luminescence units (RLU) were reduced by either 50% (ID50) compared to virus control wells after subtraction of background RLUs. Serum samples were heat-inactivated for 30 min at 56ºC prior to assay.
Antibody-dependent cellular phagocytosis and neutrophil phagocytosis
ADCP and ADNP were conducted as previously described56,57. Briefly, NVX-CoV2373 Spike protein was biotinylated using EDC (Thermo Fisher) and Sulfo-NHS (Thermo Fisher), and then coupled to yellow/green Neutravidin-conjugated beads (Thermo Fisher). Immune complexes were formed by incubating the bead + protein conjugates with diluted serum for 2 hours at 37°C, and then washed to remove unbound antibody. The immune complexes were then incubated overnight with THP-1 cells (ADCP), or for 1 hour with RBC-lyzed whole blood (ADNP). THP-1 cells were then washed and fixed in 4% PFA, while the RBC-lyzed whole blood was washed, stained for CD66b+ (Biolegend) to identify neutrophils, and then fixed in 4% PFA. Flow cytometry was performed to identify the percentage of quantity of beads phagocytosed by THP-1 cells or neutrophils, and a phagocytosis score was calculated (% cells positive × Median Fluorescent Intensity of positive cells). Flow cytometry was performed with an IQue (Intellicyt) or LSRII(BD) and analysis was performed using IntelliCyt ForeCyt (v8.1) or FlowJo V10.7.1.
Antibody-dependent complement deposition
ADCD was conducted as previously described58. Briefly, NVX-CoV2373 Spike protein was biotinylated using EDC (Thermo Fisher) and Sulfo-NHS (Thermo Fisher), and then coupled to red Neutravidin-conjugated microspheres (Thermo Fisher) or directly coupled to Carboxylate-Modified microspheres (Thermo Fisher). Immune complexes were formed by incubating the bead + protein conjugates with diluted serum for 2 hours at 37°C, and then washed to remove unbound antibody. The immune complexes were then incubated with lyophilized guinea pig complement (Cedarlane) and diluted in gelatin veronal buffer with calcium and magnesium (Boston Bioproducts) for 30 minutes. C3 bound to immune complexes was detected by fluorescein-conjugated goat IgG fraction to guinea pig Complement Ce (MP Biomedicals). Flow cytometry was performed to identify the percentage of beads with bound C3. Flow cytometry was performed with an IQue (Intellicyt) and analysis was performed on IntelliCyt ForeCyt (v8.1).
Antibody-dependent NK cell degranulation
Antibody-dependent NK cell degranulation was conducted as previously described59. NVX-CoV2373 Spike protein was coated on Maxisorp ELISA plate (Thermo Fisher), and then blocked with 5% BSA. Antibodies were then added and incubated for 2 hours at 37°C. Human NK cells were isolated from peripheral blood by negative selection using the RosetteSep Human NK cell enrichment cocktail following the manufacturer’s instructions. Human NK cells were then added to the bound antibody and incubated for 5 hours at 37°C in the presence of RMPI + 10%FBS, GolgiStop (BD), Brefeldin A (Sigma), and anti-human CD107a antibody (BD Bioscience). After incubation, cells were washed, stained with CD16, CD56, and CD3 (BD Bioscience), and fixed in 4% PFA for 15 minutes. Intracellular staining was performed using the FIX/PERM Cell fixation and permeabilization kit (Thermo), and cells were stained for interferon-γ and macrophage inflammatory protein-1β (BD bioscience). Flow cytometry was performed with an IQue (Intellicyt) and analysis was performed on IntelliCyt ForeCyt (v8.1).
Isotype and FcR-binding Luminex profiling
Isotyping and FcR profiling was conducted as previously described60,61. Briefly, antigens (NVX-CoV2373 Spike, SARS-CoV-2 Spike, S1, RBD, S2, HKU-1 RBD, or OC43 RBD) were carboxyl coupled to magnetic Luminex microplex carboxylated beads (Luminex Corporation) using NHS-ester linkages with Sulfo-NHS and EDC (Thermo Fisher), and then incubated with serum for 2 hours at 37°C. Isotyping was performed by incubating the immune complexes with secondary mouse-anti-rhesus antibody detectors for each isotype (IgG1, IgG2, IgG3, IgG4, IgA), then detected with tertiary anti-mouse-IgG antibodies conjugated to PE. FcR binding was quantified by incubating immune complexes with biotinylated FcRs (FcγR2A-1, FcγR2A-2, FcγR3A, courtesy of Duke Protein Production Facility) conjugated to Steptavidin-PE (Prozyme). Flow cytometry was performed with an IQue (Intellicyt) and analysis was performed on IntelliCyt ForeCyt (v8.1).
Multivariate analysis
A principal component analysis (PCA) was performed based on serological features using the R package ‘ropls’. The systems serology antibody titers, FcR binding and ADCD measurements were log10-transformed, and all measurements were z-scored. The PCA analyses was performed in R version 4.0.2.
Statistical analysis
Statistical analyses were performed with GraphPad Prism 9.0 software. Serum antibodies were plotted for individual animals and the geometric mean titer (GMT) and 95% confidence intervals plotted. Virus loads were plotted as the median value, interquartile range, and minimum and maximum values. Student’s t-test or two-way ANOVA was used to determine differences between paired groups as indicated in the figure legends. p ≤ 0.05 was considered significant. The AUCs and bootstrap confidence intervals were calculated using the R package ‘pROC’. For the case of AUC = 1 no confidence interval was provided. To visualize fold change of mutant humoral features with respect to the WT, a volcano plot was constructed. To calculate p-values, the R package ‘stats’ was used. The AUC and fold change analyses were performed in R version 4.0.2.